key: cord-0001023-w7uhqsio
authors: nan
title: ECR 2011 Book of Abstracts - B - Scientific Sessions
date: 2011-03-01
journal: Insights Imaging
DOI: 10.1007/s13244-011-0077-4
sha: 605fb055ce00385c008f0c27f53402824cf09db9
doc_id: 1023
cord_uid: w7uhqsio

nan

The application of diffusion tensor imaging in tumour of bone and soft tissue D.M. Deng 1 , Q.F. Meng 2 ; 1 Nangning/CN, 2 Guangzhou/ CN (ddemao@163.com) Purpose: To explore the value of DTI in diagnosis of bone and soft tissue tumours. Methods and Materials: Routine scanning of MR and DTI was performed prospectively on 42 cases with bone and soft tissue tumours betiding at calf, thigh or haunch. We analysed the statistical difference of DTI measurements between benign and malignant tumours' parenchyma, and between two groups' parenchyma 40 mAs; collimation 16 x 0.75. All the patients were clinically staged according to Durie and Salomon System. Diagnosis of osteolytic lesions was performed on the basis of axial and multiplanar reformatted images, whereas the assessment of spinal misalignment and fracture was done using multiplanar reformatted images Results: In all 99 patients, image resolution was diagnostic, enabling correct classification of multiple myeloma patients. WBLDCT showed a total of 216 pathologic bone findings in 64 patients. CT scanning resulted in complete evaluation of the bone lesions in these areas of the skeleton: skull (30), humerus (9), femur (16), ribs (3), scapulae (10), pelvis (30), clavicle (10), sternum (6), cervical (21), dorsal (40), lombar (31) and sacral rachis (10). In 30 patients the CT detection of bone involvement was the only criterion for the treatment. Furthermore, in 16 patients WBLDCT scanning demonstrated pulmonary or pleural lesions (9 cases due to infective source and 6 cases due to multiple myeloma localisations) and 1 case of renal neoplasia as a related pathology. The overall dose delivered to each patients was 3.2 mSv. Conclusion: Whole-body low-dose CT investigational protocol is a reliable imaging-based method for the direct management of patients with multiple myeloma. WBLDCT has a superior reproducibility and it is faster than conventional radiography, being, furthermore, able to demonstrate extraosseous findings.

A S153 C D E F G H B-078 14:54 Significant differentiation of focal breast lesions: raw data-based calculation of strain ratio M. Grigoryev, A. Thomas, T. Slowinski, S. Filimonow, T. Fischer; Berlin/DE (maria.grigoryev@charite.de) Purpose: Initial data suggest that elastography can improve the specificity of ultrasound in differentiating benign and malignant breast lesions. We compared elastography, B-mode ultrasound and mammography to determine whether raw data calculation of strain ratios (SRs) can further improve the differentiation of focal breast lesions. Methods and Materials: 201 women with histologically proven focal breast lesions (85 benign, 116 malignant) were included at two German breast centers. Patients underwent standardised ultrasound procedure using high-end ultrasound system (Aplio XG, Toshiba, Japan) with a 9-MHz broadband linear transducer. Two experienced readers analysed B-mode scans and mammograms using the BI-RADS criteria; elastograms were analysed using the Tsukuba score. SRs were calculated from a tumour-adjusted ROI; comparable ROI was placed in the lateral fatty tissue. Sensitivity, specificity, and cut-off values were calculated for SRs (ROC analysis). Results: Median age was 53 years [range . Lesion diameter was 17.4 ± 9.8 mm (mean ± SD). Sensitivity and specificity were 85%/60% for B-mode scanning, 85%/68% for elastography, 78%/62% for mammography, and 95%/74% for SRs. SR cut-off value of 2.27 (AUC 0.907) allowed significant differentiation (p < 0.001) of malignant and benign lesions. The quantitative SR calculation was superior to subjective interpretation of B-mode scans and sonoelastograms with a positive predictive value of 83% versus 78% and 74%, respectively, and was equal to that of mammograms. Conclusion: Calculation of strain ratios contributes to the standardisation of sonoelastography with high sensitivity and allows significant differentiation of benign and malignant breast lesions with a higher specificity compared to B-mode, subjective evaluation of elastography and mammography.

Qualitative and semi-quantitative evaluations of solid breast lesions by sonoelastography H. Yerli, T. Yilmaz, T. Kaskati, H. Gulay; Izmir/TR (hasanyerli@yahoo.com) Purpose: To determine whether the use of qualitative elasticity scoring method together with semi-quantitative strain index method by sonoelastography (SE) is useful to differentiate between benign and malignant breast masses. Methods and Materials: Some 78 lesions in 71 consecutive patients with solid breast masses (62 benign, 16 malignant) were prospectively included in this study. For each lesion, B-mode US and SE images were obtained. After elasticity scores had been determined with 5-point scoring method, strain indexes of the lesions were calculated using the same-level and normal-appearing breast region as an internal reference by means of the method of strain ratio measurement. The findings were compared with histopathology. Considering the receiver operating curves, the diagnostic performances for the elasticity scoring and the strain index methods were determined. Results: The mean scores on SE were 2.69 ± 0.59 for benign lesions and 3.75 ± 0.68 for malignant lesions. The mean stiffness index values were 2.03 ± 2.67 for benign lesions and 5.97 ± 4.45 for malignant lesions. The area under the curve value was 0.864 for 5-point scoring method and 0.840 for strain index method (P = 706). Sensitivity and specificity for 5-point scoring method were 80% and 95%, respectively; 87.5% and 72.6% for B-mode US; and 80% and 93% for strain index method when a cutoff point of 3.52 was used. A semi-quantitative evaluation using the strain index method did not contribute to qualitative evaluation by scoring. Conclusion: After 5-point scoring by SE, additional measurement of the strain index is not mandatory to differentiate between benign and malignant breast masses.

Purpose: Although tissue infarction is a well-known histologic feature of invasive fungal sinusitis, it has rarely been emphasised in radiologic literature in the head and neck. The purpose of this study was to investigate the prevalence and the characteristic MRI findings of cervicofacial tissue infarction (CFTI) in invasive fungal sinusitis. Methods and Materials: We retrospectively reviewed MR images in 33 patients with histologically proved invasive fungal sinusitis. CFTI was defined areas of necrosis. We divided CFTIs into those in the nasal cavity, and those extending to the extranasal areas. Attention was paid on the characteristics of CFTI on contrast-enhanced T1WI (CET1WI). Preservation of intervening fat on precontrast T1WI in cases of extrasinonasal CFTI and the presence of bone change on CT were recorded. Results: CFTI was found in 17 of 33 patients (52%). All of 17 patients had sinonasal CFTI, in whom 13 patients also had extranasal CFTI. 9 patients were infested with Mucor, and 6 with Aspergillus. 11 patients were immunocompromised and 6 patients had diabetes. On CET1WI, all lesions were characteristically ill-defined nonenhancing lesions. In cases of extrasinonasal CFTI, there was no respect for the intervening tissue planes, extending to the adjacent soft tissue, orbit, bony structure, or intracranial cavity. In 12 of 13 lesions with extrasinonasal CFTI, intervening fat was preserved on T1WI. 16 lesions with CFTI demonstrated no erosive or destructive bone changes on CT. Conclusion: On CET1WI, CFTIs were demonstrated in more than half of the patients with invasive fungal sinusitis. They may be a pathognomonic sign on MRI caused by this fulminant infection.

: 40 patients were included in our study after a radical mastoidectomy intervention ussing the following protocol for the MRI exam: sagittal 3DT2, axial T1WI fast SE, diffusion (b0, b300, b600 seg/mm), postgadolinium axial T1WI fast SE, delayed postgadolinium coronal GET1 WI. Hyperintensity on b600 values was interpreted as recurrent cholesteatoma. Tl report was done on two different moments by an experienced neuroradiologist. It was classificed as: CC (clearly cholesteatoma), DC (doubful cholesteatoma), CF (clearly fibrosis), DF (doubtful fibrosis). Results were compared with postsurgery reports. Results: Recurrent cholesteatoma was properly identified in 8 patients with PPV 100%. 6 of 7 patients were confirmed at surgery as fibrosis with only 1 false negative NPV 80%. Sensitivity and specificity were 89% and 100%, respectively. Conclusion: Results for specificity and PPV in our sample (100%) support the utility of IPD for detection of recurrent cholesteatoma. Our study shows the necessity of using 3DT2 sequences together with diffusion MRI to proper anatomical localization. Previous CT is also essential to determinate bony structures and compare with 3DT2 images. Results: Non-EPI DW MRI sequences showed restricted diffusion in 36 cases. Second-look surgery confirmed cholesteatoma in 24/26 cases; in 1/24 cases an empyema was diagnosed and in one case no cholesteatoma was found at surgery. In 2/36 cases restricted diffusion was accompanied by hyperintense signal on the T1-weighted images which was consistent with transplanted fat; without conventional imaging restricted diffusion in this two cases would be misinterpreted as cholesteatoma. The positive predictive value for detection of cholesteatoma with the full imaging protocol was 94% (34/36); whereas with non-EPI DWI as the only sequence this would drop to 88% (32/36). Conclusion: Residual or recurrent cholesteatomas after primary cholesteatoma surgery can be reliable detected by restricted diffusion on non-EPI DW MR imaging. DWI imaging without conventional sequences will increase the number of falsepositive findings in our patient setting, because transplanted fat within the cavity may show restricted diffusion.

Reducing the radiation dose for low-dose CT of the paranasal sinuses using iterative reconstruction: feasibility and image quality S. Bulla, F. Hassepass, P. Blanke, T. Kraus, M. Langer, G. Pache; Freiburg im Breisgau/DE (stefan.bulla@uniklinik-freiburg.de) Purpose: To retrospectively evaluate image quality of dose-reduced CT of the paranasal-sinus using the iterative reconstruction (IR) technique. Methods and Materials: In this retrospective study 80 patients (mean age: 46.9 ± 18 years) underwent low-dose CT of the paranasal-sinus (Siemens Definition, Forchheim, Germany), with either standard settings (A: 120 kV, 60 mAs) reconstructed with conventional filtered back projection (FBP) (n = 20) or with tube current-time product lowering of 20%, 40% and 60% (B: 48 mAs, C: 36 mAs and D: 24 mAs) using iterative reconstruction (n = 20 each). Subjective image quality was independently assessed by three blinded observers using a semiquantitative fivepoint grading scale (1 = poor, 5 = excellent). Effective radiation dose was calculated from the dose-length product. Mann-Whitney-U-test was used for statistical analysis. Results: Mean effective doses were 0.28 ± 0.03 mSv (A), 0.23 ± 0.02 mSv (B), 0.17 ± 0.02 mSv (C) and 0.11 ± 0.01 mSv (D) resulting in a maximum dose reduction of 61% with IR as compared to the standard low-dose CT. Best image quality was observed at 48 mAs (4.8; p < 0.05), whereas standard low-dose CT and maximum dose reduction (D) provided the same subjective image quality (4.3; p: n.s). Interobserver agreement was excellent (k values 0.77-0.93). Conclusion: As compared with filtered back projection, application of iterative reconstruction allows for significant dose reduction of up to 61% for paranasal-sinus low-dose CT without loss in diagnostic image quality.

3D PSIR gradient echo for imaging myocardial infarction A.M. Huber, K. Herrmann, M. Settles, M. Betz, A. Beer, E. Rummeny; Munich/ DE (armin.huber@roe.med.tu-muenchen.de) Purpose: The aim of the study was to assess the diagnostic accuracy in imaging viability of the myocardium with a 3D phase-sensitive gradient echo sequence. Methods and Materials: 19 patients with myocardial infarction were examined at a 1.5 Tesla MR System (Achieva, Philips) 10 minutes after application of contrast material with a 3D phase-sensitive inversion recovery (PSIR) multislice technique (IR gradient echo, TR/TE/bandwidth: 5.0 msec, 2.4 msec, voxel size 1.25 x 1.25 x 4.0 mm 3 ) that allows to image the entire short axis during three breath-holds, and with a segmented 2D single slice technique (IR gradient echo, TR/TE/bandwidth: 12.0 msec, 4.4 msec, 140 Hz/Px, 1.8 x 2.3 x 8 mm 3 ), which requires one breath-Results: Using a cut-off value of 20 ms, obtained from another study in normal subjects, T2* values were lower (ranging betwen 12 and 18 ms) in 9 patients. In these patients, T2 triple IR-FSE images were postive for haemorrhage in 5, while FGRET and or IR-FGRE were positive for microvascular obstruction in 7. All patients with haemorrhage showed also microvascular obstruction. Conclusion: T2* multiecho imaging is feasible for T2* measurements in patients with acute myocardial infarction. Furthermore, haemorrhage and severe microvascular obstruction represent the same type of complication as all patients with T2 triple IR FSE positive for haemorrhage showed microvascular obstrution, and T2* imaging was positive in the other patients with microvascular obstruction. The discrepancy is due to the superior sensitivity of T2* compared with T2 imaging. Purpose: To identify and characterise myocardial injury with magnetisation transfer (MT) balanced steady state free precession (SSFP) magnetic resonance (MR) imaging. Methods and Materials: Twenty four patients with suspected myocardial injury such as infarct (n = 13), myocarditis (n = 6), or other (n = 5) were subjected to cardiac MR (1.5 Tesla with phased array body coil) comprising precontrast SSFP, magnetisation transfer SSFP (MT-SSFP) and delayed contrast enhanced (0.1 mmol/ kg Gd-chelate) inversion recovery fast low angle shot (LE) images. Signal intensities (SI) were measured in blood, skeletal muscle, normal and injured myocardium. MT ratio was calculated as (SI SSFP -SIMT-SSFP)/SISSFP ± standard error of the mean. Results: MT ratios in blood, skeletal muscle, normal and injured myocardium were 0.03 ± 0.01, 0.3 ± 0.01, 0.31 ± 0.03, 0.21 ± 0.06, respectively. Moreover, injury that was visible on LE images was identified on MT images in 13 patients with a trend to lower MT ratios in subacute injury and higher MT ratios in chronic injury. Conclusion: Different MT ratios can be measured in blood, normal and injured myocardium with MT magnetic resonance imaging. Moreover, MT imaging may enable further characterisation of myocardial injury with respect to age and nature of the lesion.

The characterisation of anaplastic transformed WHO grade II gliomas by means of advanced magnetic resonance imaging G. Stasik-Pres, A. Hebda, B. Bobek-Billewicz; Gliwice/PL (gabastasik@poczta.onet.pl) Purpose: To characterise diffusion-weighted imaging (DWI), diffusion tensor imaging (DTI), perfusion weighted imaging (PWI) and proton magnetic resonance spectroscopy (1HMRS) parameters suggesting anaplastic transformation of WHO gradeII gliomas. Methods and Materials: Analysed group consisted of 47 (age 42±12 years) consequtive patients diagnosed in Maria Sklodowska -Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, between 2006 -2009 .4 months,range 2-37) WHO gradeIl gliomas did not transform to higher grade -nontransformers group. 8/47 (mean follow-up 15 months,range 6-24) WHO gradeII gliomas transformed to higher grade (Astrocytoma Anaplasticum WHOIII or Glioblastoma Multiforme WHOIV), which was histopathologically confirmed -transformers group. Selected parameters were analysed: rCBV max, rCBF max normalised to normal-appearing white matter, ADC min , Fa min , (Cho/Cr) max , (Cho/NAA) max ratio. Results: Higher normalised rCBF max ( 2.93±1.28 vs. 1.85±0.62;p < 0.05) and trend towards higher normalised rCBV max (2.57±0.99 vs. 1.84±0.55;p < 0.08) were in transformers group in comparison with nontransformers group. Transformers group characterised higher (Cho/NAA) max ratio (3.69.2.90-4.83 vs. 1.90.0.87-3.48 ; p < 0.001), whereas (Cho/Cr) max ratio (1.69.0.94-3.78 vs. 1.96.1.18-3.38;p=NS) was not different in comparsion with nontransformers group. No statistically significant difference between analysed groups was stated in terms of: ADC min (0.882±0.268*10 -3 mm 2 /s_ transformers vs.1.07±0.293*10 -3 mm 2 /s_nontransformers;p=NS), Fa min (0.122.0.093-0.145_transformers vs. 0.126.0.061-0.332 _nontransformers;p=NS). hold per slice (reference technique). The PSIR technique was used with a standard inversion time of 200 msec. The reference sequence was used after individual adaptation of the inversion time. The areas of hyperintense infarctions on selected slices and the entire volumes were compared for both sequence techniques. Results: The 3D PSIR gradient echo sequence provided sufficient image quality for evaluation in all patients without individual adaptation of the inversion time. The assessment of the area of infarction within one slice (r = 0.96, p < 0.002) and the volume of the entire infarction is possible with excellent correlation of both techniques (r = 0.95, p < 0.003). Conclusion: The 3D PSIR GE sequence allows stable imaging of myocardial infarction during a shorter acquisition time without individual adaptation of inversion time. Compared to the reference technique the infarct areas and volumes were determined accurately.

Purpose: Pseudoprogression (PsPD) following radiation and chemotherapy in patients with glioblastoma multiforme is difficult to distinguish from progressive disease. The aim of our study was to identify early signs of PsPD in brain MRI after standard therapy including surgery, radiation and temozolomide. Methods and Materials: Out of 74 patients with standard therapy, we retrospectively identified 47 with early progression in MRI within the first 6 months. Patients were divided in groups of "real" early progression versus pseudoprogression. We analysed DWI-ADC, also T2*-DSC-PWI changes in both groups. Results: Preceding studies have shown that median survival in patients with GBM was better in patients with PsPD than with PD. We identified 18 patients with median survival more than 12 months and early signs of progression in MRI within the first 6 months after therapy. In patients with PsPD, no signs of hypervascularisation were found (T2*-DSC-PWI), whereas in patients with early progressive disease restricted diffusion and hypervascularisation were found. Conclusion: Pseudoprogression is not an uncommon phenomenon after standard therapy of GBM. The amount of 30% given in the literature can be reaffirmed in our study. To our knowledge, this is the largest study population of patients with MRI signs of early progression. Our study strengthens the role of PWI and DWI as a diagnostic tool to identify PsPD, especially, since no definite examination procedures have been identified so far for reliably detecting PsPD and separating it from progressive disease.

Comparison of MRS, MR perfusion, and PET scan to formulate an appropriate protocol for evaluation of suspected recurrent gliomas G.S. Gujral, R. Deshpande, N.S. Saini; Delhi/ IN (girigujral@gmail.com) Purpose: Comparison of MRS, MR perfusion, and PET scan to formulate an appropriate protocol for evaluation of suspected recurrent gliomas. Methods and Materials: 30 patients with total resection of their brain tumours were followed up with MR spectroscopy, dynamic susceptibility contrast perfusion scan and FDG PET. Recurrence detected on imaging was confirmed by histopathology, the gold standard. Results: 14 of the 30 patients showed evidence of recurrence which was confirmed by histopathology in 12 patients. In one patient histopathology could not be done as the lesion was very small and deep seated and in the other a cerebellar lesion, histopathology would have resulted in deterioration of the patient's condition. However, all three imaging modalities showed evidence of recurrence in these two cases. Using a Cho/Cr cut-off of 2.4, MRS showed a sensitivity of 83% and a specificity of 100% to detect recurrence. Using a cut-off value for rCBV of 2.6, perfusion MRI showed a sensitivity of 91.7% and a specificity of 100%. PET exhibited a sensitivity of 83.3% and a specificity of 50%. Conclusion: All modalities detected recurrence but perfusion was the most sensitive as it detects early microvascular changes and increased blood flow within recurrent tumour. Advanced MR techniques such as perfusion are very robust in picking up recurrent tumour and should form an integral part of any protocol to detect recurrent tumours. 14:54 MGMT promoter methylation in glioblastoma can be predicted using diffusion weighted imaging A. Hiwatashi, T. Yoshiura, K. Yamashita, H. Kamano, H. Honda; Fukuoka/JP Purpose: Methylation of the O6-methylguanine-DNA methyltransferase (MGMT) promoter is related to the treatment effect of temozolomide, which is the only chemotherapeutic drug that prolongs survival in glioblastoma. The purpose of this retrospective study was to evaluate the value of magnetic resonance (MR) imaging including the apparent diffusion coefficient (ADC) on 3 T MR scanner to predict the status of MGMT promoter methylation. Methods and Materials: This study included 27 patients (15 females and 12 males) with glioblastomas. All patients obtained unenhanced T2-weighted and fluid attenuated inversion recovery images, contrast enhanced T1-weighted images and diffusion-weighted images on 3 T MR scanner before surgery. Minimum ADC ratio (mADCR) was calculated by the minimum ADC in tumour divided by that in the normal appearing white matter. Tumour volume on unenhanced and Purpose: [18 F]Fluoro-deoxy-glucose positron emission tomography and concomitant CT (PET/CT) is established for especially oncological imaging, and its increasing use yields more incidental image findings. Focal thyroid uptake of FDG is associated with a significant risk of thyroid malignancy but the reports of prevalence and significance are contradictory. In the present study, we assessed the Stockholm city regions prevalence of focal thyroid incidentalomas and their nature by means of pathological and clinical correlation. Methods and Materials: During 2006 and 2009, 3,637 patients in the Stockholm city region were examined by FDG-PET/CT. The anatomical position of focally increased thyroid FDG accumulations was registered and their standardised uptake value (SUV) was measured. Data on diagnostic and therapeutic follow-up were retrieved from the medical files. Results: 62 (1.7%) had increased thyroid FDG uptake which was focal in 35 (0.96%; 10 men and 25 women). Ultrasound and fine needle cytology was performed in 23/35 (66%) of the cases and 13/23 (61%) were malignant. Surgery was performed in 10 cases. Benign cytology was found in 9 patients (thyroiditis, colloid or hyperplastic nodules). The tumour size ranged from 8 to 40 mm. The maximum tumour SUV/normal thyroid SUV-ratio was significantly higher for malignant than benign nodules (8.9 ± 10.1 vs. 3.5 ± 1.1; p < 0.05). Conclusion: Incidental findings of focal thyroid FDG uptake are associated with a significant risk of a thyroid cancer. If a thyroid carcinoma may change the patient´s treatment plan or prognosis, a diagnostic work up with ultrasound and fine needle cytology is recommended.

PET/CT imaging of small objects with the clinically implemented 40-detector biograph mCT H. Kayed, M. Sadick, S.O. Schönberg; Mannheim/DE (hany.kayed@umm.de) Purpose: Small animal imaging with clinically implemented PET/CT (cPET/CT) scanners is inferior compared with animal PET/CT scanners. In this experiment we quantitatively analysed the applicability of the recently introduced Biograph mCT in imaging small objects. Methods and Materials: A single-spot phantom was constructed simulating a small mouse of 30-g body-weight bearing an 18 F-FDG PET positive tumor. The phantom was subjected to scanning with the Biograph mCT under clinical settings. The dataset was reconstructed and quantified using the TrueX algorithm and the TrueD software tool, respectively. Analysis of the SUV (standard uptake value) was performed using different doses (1, 10 and 20 MBq) of 18 F-FDG in different Gaussian postfilter reconstructions at 1 and 3 mm. Results: CT images showed phantom dimensions close to the manufactured dimensions in comparison with PET images. The 1-mm filter showed augmented SUVs compared with 3 mm filter irrespective of the dose. Low doses, such as 1 MBq, showed augmented SUVs compared with high doses (20 MBq). The average SUV showed a linear relationship and consistent data compared with the maximum SUV. Conclusion: Clinically implemented Biograph mCT with the TrueX reconstruction algorithms and TrueD quantification software are feasible for imaging small animals bearing very small tumours. These results show that application of low doses, determination of the volume of interest (VOI) in the reconstructed CT images, fused with the reconstructed PET images in 1 mm filter as well as measuring the SUV avg might be the optimal settings for evaluating small lesions.

A S161 C D E F G H

Evaluation of functional parameters in monitoring breast cancer treatment: 18 F-FDG-PET/CT mammography vs breast MRI L. Umutlu, T.A. Heusner, A. Koeninger, F. Otterbach, R. Kimmig, A. Stahl, G. Antoch, T.C. Lauenstein; Essen/ DE (Lale.Umutlu@uk-essen.de) Purpose: Aim of this study was the evaluation of functional parameters of 18-F-FDG-PET/CT mammography (SUV) with CAD-based analysis of breast MRI as monitoring tools of breast cancer under neoadjuvant chemotherapy (NCTX). Methods and Materials: 18 patients with breast cancer underwent both imaging modalities (a) prior and (b) after the third cycle of NCTX. Breast MRI was performed on a 1.5 Tesla scanner (Magnetom Espree, Siemens Healthcare). For dynamic imaging six dynamic T1-weighted gradient-echo sequences were collected (Gadovist, Bayer Schering Pharma). Contrast kinetics and pharmacokinetics were analysed quantitatively using a computer-aided detection system (CAD, iCAD). 18-F-FDG-PET/CT mammography examinations were performed within a time period of 48 hours on a Biograph PET/CT system (Siemens Molecular Imaging) and standardised uptake values (SUV) were assessed. Results were correlated to pathologic response evaluation. Functional changes were compared by Wilcoxon signed rank test. Results: According to histopathology all patients were pathologic responders. The SUV values showed a statistical significant under NCTX (p = 0.018). CAD-based analysis of breast MRI revealed a significant reduction of fast postinitial washout patterns (p = 0.07) and a significant increase of tumour tissue fibrotisation (p = 0.08). Pharmacokinetic parameters, by means of permeability and extracellular volume fraction, also revealed significant decrease. Conclusion: Both imaging modalities enable a valid evaluation of functional therapy response of malignant lesions under NCTX. Due to availability reasons, the omission of radiation and its superior (preoperative) assessment of residual tumour extent, breast MRI is recommended for therapy response evaluation in clinical routine.

Value of iodinated contrast administration in computed tomography exam for staging and restaging Hodgkin disease: comparison with 18 f-FDG PET/CT C. Capraro, D. Ippolito, L. Guerra, L. Monguzzi, C. Messa, S. Sironi; Milan/IT (letizia.monguzzi@gmail.com) Purpose: To assess the diagnostic value of administration of iodinated contrast in coregistered 18-fluorodeoxyglucose-positron-emission-tomographic (FDG/PET) with unenhanced low dose CT (ldCT) and contrast-enhanced CT (ceCT) for the staging and restaging of disease in patients with Hodgkin disease (HL). Methods and Materials: A total of 30 patients with histologically proven HL underwent a PET/ldCT+ceCT studies, performed in a single step procedure. Eighteen studies were performed for staging and 28 for the evaluation of residual disease after first line of chemotherapy (restaging). The two different datasets of images were evaluated by two expert readers blinded about the clinical data. Finally, the stage of disease was then defined according to PET/ldCT and PET/ceCT following the Ann Arbor classification. For restaging studies, images were interpreted considering the presence of residual mass with or without residual activity according to the International Harmonization Project. Results: Of 630 sites considered as potentially involved by the disease, PET/ldCT and PET/ceCT were concordant in 622/630 (98.7%) in staging studies and in 28/28 (100%)of the restaging studies. PET/ldCT and PET/ceCT were concordant with standard reference in 16/18 (88.9%) and in 17/18 (94.4%) patients, respectively, in staging studies. In comparison between restaging PET/ldCT and PET/ceCT results and standard reference both examinations were concordant with standard reference in 26/28 (92.8%)studies. Conclusion: In our study the addition of ceCT to PET/ldCT did not show significant added diagnostic value both in staging and, in particular, in restaging studies. If these results will be confirmed in a larger patients population, PET/ldCT could be considered the standard imaging technique for staging and restaging HL, sparing addicted radiation burden and iodinate contrast administration to the patient.

Investigation of dose minimisation protocol for 18 F-FDG PET-CT in the management of lymphoma post-chemotherapy follow-up L.I. Sonoda, B. Sanghera, T. Mills, W. Wong; London/UK (luke@sonoda.co.uk) Purpose: 18 F-FDG-PETCT plays an important role in the management of post-chemotherapy follow-up in lymphoma patients. Some centres perform prechemotherapy baseline CT and first post-chemotherapy PETCT from skull base to pubic symphysis. With a concern of radiation burden, especially in young patients, groups: 1) concordant: 1.1) suggestive of malignancy in both components, 1.2) suggestive of benign lesions; 2) discordant: 2.1) suggestive of malignancy on CT without FDG uptake, 2.2) suggestive of malignancy in PET without alterations in CT, 2.3) suggestive of benign lesions on CT with FDG uptake. Results: Of 1320 PET-CT studies, we found bone lesions in 243 patients. We review data from 124 patients, six of them were lost to clinical follow-up. They were 41 men, 77 women, mean age 61.4 (range 7-84). 73 were concordant, 70 suggesting malignancy, with 4 false positive (FP) results, and 3 suggestive of benign lesions (true negatives). We found 45 cases of disagreement: In 4 patients, CT identified blastic malignant lesions (subgroup 2.1), in 25, PET revealed the presence of malignant lesions (subgroup 2.2), 23 true positives (TP) and 2 FP. Benign lesions were noted on CT with FDG uptake of variable degree in 16 (subgroup 2.3), being CT very useful for diagnosis. Conclusion: In PET-CT, PET reduces FN of CT and CT reduces FP results of PET. The combined analysis is essential for a correct caracterisation of bone lesions.

in the supine and prone position using a low radiation dose and a faecal tagging protocol based on oral administration of iodinated contrast material. All datasets were wirelessly imported in DICOM format on a iPhone 4® 32GB (Apple Inc) running the OsiriX iPhone app version 1.1.3 (www.osirix-viewer.com) from a Macintosh desktop computer (iMac® 3.06GHz) connected to our hospital PACS and running OsiriX 3.7.1. Two experienced raters read CTC datasets independently on the iMac® and on the iPhone®. Lesions included 25 polyps sized less than 6 mm (37.3%), 29 between 6 and 9 mm (43.3%), 10 between 10 mm and 30 mm (14.9%), and 3 colonic masses (4.5%). Lesion detection rate and time needed to read each entire CTC study were recorded. Results: All lesions diagnosed on the iMac® were also detected on the iPhone®. However, the time required to complete reading of CTC datasets was significantly longer using the iPhone® (13.31±6.29 minutes) than with the iMac® (5.51±2.45 minutes; p < 0.01), predominantly due to the need to repeatedly pan and zoom over the various colonic segments on the small iPhone® screen.

The iPhone® can reliably be used for preliminary 2D reading of CTC datasets. However, it requires a substantially longer time for complete image analysis compared with a desktop workstation.

Automatic segmentation of the pancreas in contrast-enhanced CT data M. Erdt 1 , K. Drechsler 1 , M. Hammon 2 , M. Uder 2 , A. Cavallaro 2 ; 1 Darmstadt/DE, 2 Erlangen/ DE (marius.erdt@igd.fhg.de) Purpose: Pancreas segmentation in CT images is desirable for computer-aided diagnostics and operation planning, but so far considered unfeasible to automate. A method for fully automatic pancreas segmentation from contrast-enhanced CT data is introduced. Methods and Materials: Common portal phase CT scans of 40 patients were acquired. Liver, spleen and supporting vessels were automatically segmented. A subset of the data was used to build a pancreas tissue classifier using boosted Fourier and wavelet features. The built classifier was used for adapting a statistical geometric model in order to automatically segment the pancreas in the remaining images. Manual volume segmentation of the pancreas is taken as the gold standard. Threefold cross-validation was used with disjunct training and test folds. Results: Correlation with gold standard delineation was performed. Averaging resulted in 12% volume difference, 2.4 mm average surface distance, 4.5 mm RMS error, 23.4 mm max surface distance and 76% positive predictive value. Conclusion: Machine learning approaches combined with statistical shape modelguided segmentation can be used to segment the pancreas in contrast-enhanced CT data. Since this method is fully automatic and provides a sufficient accuracy, it enables new applications in computer-aided diagnosis and operation planning which were infeasible so far due to manual interaction needs. It is also a key factor towards automated full-body segmentation in CT.

Verification of feasibility and clinical role of computer-assisted analysis of functional parameters of the lung obtained from CE-MS-CT A. Malich, S. Mikulik, D. Hentrich; Nordhausen/ DE (sylvia.mikulik@shk-ndh.de) Purpose: This study aimed to analyse clinical role of CAD-based determination lung functional parameters of the lung obtained from MS-CT compared with whole body plethysmographic data. Methods and Materials: 88 patients were randomly selected who underwent MS-CT and whole-body plethysmography within < 3 days. Body plethysmography was performed using master-screen-body (Jaeger, Germany, version V4.67). CT was performed using 128-row CT (0.5 mm slice thickness, Siemens Definition AS+). All data were analysed using CAD (MeVis Pulmo3D,Fraunhofer MeviS, Germany). Vital capacity (VC), FEV1, intrathoracal gas volume (ITGV) were matched with CAD-based lung volume analysis (vol), mean low density (MLD), low-and high attenuation values (LAV/HAV), lung weight (w), P15, PK (maximum of histogramm) and FWHM (full width half maximum). Results: CAD-analysis was available at all cases. Analysis required 1 minute acquisition time. Correlation coefficients of VC versus Vol, MLD, LAV, HAV, weight, FWHM were 0.55 (<.01); -0.30 (<.05), -0.29 (<.05); 0.01 (n.s)., 0.55 (<.01), -0.29 (<.05). Correlation coefficients of FEV1 versus Vol, MLD, LAV, HAV and weight were.13 (n.s).,.02 (n.s)., -.25 (p <.05), -.04 (n.s). and -.37 (p <.01), respectively. Correlation coefficients of ITGV versus Vol, MLD, LAV, HAV, FWHM, P15, PK were.77 (<.001), -.60 (< 0.001), -.50 (<.01), -.52 (<.01), -.21 (n.s)., -.59 (p <.001), -.55 (p < 0.01). Degree of emphysema and fibrosis can be obtained from CT-data semiautomatically.

this study aimed to assess if PETCT acquisition area could be minimised, and how much radiation dose could be reduced. Methods and Materials: Retrospective data analysis of 100 consecutive lymphoma patients (55 male, mean 41.7 years-old, Hodgkin's lymphoma 50, non-Hodgkin's lymphoma 50) was performed to record sites of disease on pre-chemotherapy CT and post-chemotherapy PETCT. The potential reduction in radiation dose and timesaving achieved with PETCT scans limited to sites of known disease identified on pre-chemotherapy CT was calculated. Results: No FDG-uptake was seen in 72 of the 100 subjects. FDG-uptake at known disease sites was seen in 24 cases. Of the remaining 4 cases one had clinically significant pathology, a rectal adenocarcinoma. PETCT scans did not reveal any unexpected sites of lymphoma. Limiting PETCT to sites of known disease would have resulted in a mean radiation dose saving of 4 mSv (27.3 %), with a mean timesaving of 16 minutes per PETCT. Conclusion: Our study suggests young lymphoma patients may benefit from reduced radiation dose / scan time by limiting PETCT to sites of known disease with low risk of missing significant pathology. However in older patients, with increased incidence of asymptomatic synchronous malignancies, we recommend wholebody PETCT is advisable unless pre-chemotherapy PETCT has been performed.

Purpose: Low-cost non-invasive computer-aided diagnosis system was developed to automatically classify symptomatic and asymptomatic classes.

The classification is based on (i) Discrete Wavelet Transform (DWT), (ii) wavelet Package (WP) and combination of DWT and WP. The features selected were detailed and approximate coefficients for DWT method; energy, average horizontal and vertical WP coefficients for WP method; and a combination of these five (two of DWT and three of WP) features. These extracted features were fed to Variable-Increment (VIM) Perceptron functionality for automated decision making. We have also proposed a plaque malignancy index (PMI), a number to identify the symptomatic and asymptomatic carotid plaque class. This PMI can be used as an adjunct tool by the physicians during their screening to cross check their diagnosis. 274 carotid plaque ultrasound images (137 asymptomatic plaques and 137 symptomatic plaques) associated with retinal or hemispheric symptoms (33 stroke, 60 TIA, and 44 AF) were used for this work. Results: Results show that combination of DWT with WP showed the best results. To be specific, the classification accuracy, sensitivity and specificity of the DWTbased system was 89.1%, 90.7%, 84.4%, respectively, and 89.7%, 89.6%, 90.7%, respectively, for wavelet package based system, and the combination of DWT and WP achieved 90.3%, 90.2%, 90.6%, respectively. PMI clearly showed the separation of symptomatic and asymptomatic plaques. Conclusion: Results of this study showed that combination of DWT-and WP based features yield accuracy better than 90% with ability to classify symptomatic versus asymptomatic plaques.

Conclusion: CAD-analysis of volume, mean low density and lung weight calculation using MS-CT allow a reliable analysis of functional lung parameters with highly significant correlation to plethysmographic values and thus offer the opportunity to screen lung function at all patients undergoing MS-CT due to several reasons and perspectively to quantify degree of emphysema and fibrosis.

Separation of cortical and cancellous bone by 3D-texture-based analysis: a novel approach of threshold-independent calculation of cortical thickness A. Valentinitsch 1 , J. Patsch 1 , J. Deutschmann 1 , L. Fischer 1 , G. Langs 2 , F. Kainberger 1 ; 1 Vienna/AT, 2 Cambridge, MA/US (alexander.valentinitsch@meduniwien.ac.at) Purpose: To propose a fully automated threshold-independent method to segment cortical and cancellous bone in high-resolution peripheral quantitative computed tomography (HR-pQCT) images based on texture features Methods and Materials: We scanned 14 structure phantoms of radius sections (7 males and 7 females, isotropic resolution 41 μm and 82 μm) in a HR-pQCT. Expert segmentations on 41 μm data registered to 81 μm data served as grunt truth (GT). The method learns and selects texture features from a set of annotated training images. It learns a classifier that segments the bone segments with high accuracy. We evaluated segmentation error and compared the cortical thickness (Ct.Th) measurement with a state-of-the-art method. To test the agreement of Ct.Th between GT and other segmentation method, a two-tailed Student's t-test was then used to check whether linear regression slopes and intercepts were significantly different from the equality line.

The mean Dice similarity coefficient of GT and automatic segmentation is 0.904 (±0.045). There is no significant bias of the two tailed Student's t-test between measurement and GT with slope = 0.984 (p = 0.85) and intercept = 0.066 (p = 0.32) for our method, while both deviated significantly from GT for the standard method (slope = 1.377: p < 0.01, intercept = -0.319: p < 0.01). The mean absolute measurement error for the proposed method is 0.07 mm.

Conclusion: Experiments show that the proposed method measures the cortical thickness accurately. It outperforms a state of the art method in terms of bias and accuracy. This results in reliable thickness measurements, and provides an accurate basis for the study of the transitionary zone between cortical and cancellous bone.

The diagnostic value of direction information from diffusion tensor imaging of the prostate J.C. Apitzsch, F. Schoth, F. Goerg, N.A. Kraemer, C. Plumhans, M. Weibrecht, D. Pfister, T. Braunschweig, C.K. Kuhl; Aachen/ DE (apitzsch@rad.rwth-aachen.de) Purpose: Diffusion weighted imaging of the prostate is a common method of evaluating prostatic carcinoma: We assessed the diagnostic value of direction information resulting from diffusion tensor imaging of the prostate. Methods and Materials: MRI of the prostate was performed in 18 patients (range 54-74 years, mean age 67 years) at 3 Tesla (Archiva, Philips Medical Systems, Best, the Netherlands) combining an endorectal coil with a six channel surface coil. In additiona to multiplanar T2/TSE, 3D-Spectroscopy, DCE a DTI-sequence (Single-Shot SE-EPI, FOV 110 x 100 x 60 mm, Vox 1.7 x 1.7 x 3 mm, TE 64 ms, TR 3713 ms, B 1000) was performed for 32 directions. The results were compared with the histology either retrieved from prostatectomy or from prostate biopsy which had been performed in 10 of the patients. Further, the diagnostic value was compared to that of examining the pictures without DTI information.

The mean fractional anisotropy in the peripheral zone was 0.24 ± 0.08. A disturbance of the tangential orientation of the main diffusion direction in the peripheral zone was chosen as an indicator of carcinoma. The sensitivity/specificity of the reading of the direction information alone (superimposed to T1 weighted images) was 80%/75% of this cohort. Conclusion: DTI of the prostate provides information about prostatic microarchitecture in the peripheral zone and may help along with other criteria to increase the diagnostic accuracy in the diagnosis of prostate cancer. Clinical relevence/ application: The analysis of direction information from DTI images may improve noninvasive assessment of prostate carcinoma and provide useful information before prostate biopsy and/or operation.

Apparent diffusion coefficient as an imaging biomarker to predict the survival in patients with unresectable pancreatic cancer H. Nishiofuku, N. Marugami, T. Tanaka, H. Anai, S. Sueyoshi, M. Sho, Y. Nakajima, K. Kichikawa; Kashihara/JP Purpose: The purpose of this study is to investigate whether the apparent diffusion coefficient (ADC) on diffusion-weighted MR images at 1.5 T can predict the survival in patients with unresectable pancreatic cancer.

The institutional review board approved this study. The subjects were 31 patients (20 males, 11 females; mean age 65.7 years; range 53-79 years) with histologically confirmed unresectable pancreatic cancer who were treated with first-line chemotherapy (gemcitabine alone 17, TS-1 alone 7, gemcitabine plus TS-1 7) between July 2007 and December 2009. The minimum ADC (minADC) value of each primary tumour was determined by placing 3 regions of interest on ADC maps at pretreatment and 4 weeks after initiation of treatment, and the relative change of minADC value (%ADC) was determined. The effect of %ADC on overall survival (OS) was modelled by Cox proportional hazards regression. Results: Median time-to-progression was 4.0 months and median OS was 11.7 months. Patients with an increase of > 0% of the baseline minADC after 4 weeks of treatment (n = 19) had a significantly better median survival than patients with a decrease of <= 0% (n = 12) (18.9 vs 5.6 months; P = 0.001). The %ADC was the strongest independent predictor of OS in multivariate analysis (P = 0.009).

The relative change of ADC at 4 weeks from the initiation of treatment may serve as a useful imaging biomarker in unresectable pancreatic cancer.

Endocrine tumours of the pancreas: discordant patterns of enhancement at contrast-enhanced CT, contrast-enhanced MRI and CEUS G. Foti, N. Faccioli, R. Manfredi, R. Pozzi Mucelli; Verona/ IT (gfoti81@yahoo.it) Purpose: To compare contrast-enhanced computed tomography (CECT), contrast-enhanced magnetic resonance imaging (MRI) and contrast-enhanced ultra-sonography (CEUS) enhancement patterns of functioning (FPET) and nonfunctioning pancreatic endocrine tumours (NPET) Methods and Materials: In a 6-year-period, 125 consecutive patients preoperatively studied with at least two contrast enhanced exams (98 NPETs, 27 FPETs) were included in our retrospective study. A total of 94 CT, 83 MRI, 69 CEUS studies were performed; all CT and MRI exams were performed before and after i.v. administration of contrast material, including arterial pancreatic (45 sec), portal venous (80 sec) and late venous phase (120 sec); CEUS was performed with continuous observation of lesion enhancement from unenhanced (B-mode) to late venous phase. Student's t test was used to compare enhancement patterns (hyper-iso-hypovascularity in comparison with spared pancreas, homo-inhomogenous enhancement, peak enhancement phase). A value of p < 0.05 was considered significant. Results: Lesion hypervascularity was depicted in 76/94 (80.8%) CT, 70/83 (84.3%) MRI and 54/69 (78.2%) CEUS exams, without significant difference between NPET and FPET (p > 0.05). A non-significant difference was determined as cocers enhancement patterns (hypervascularity and homogeneity) between CT and MRI (p = 0.1), CT and CEUS (p = 0.1), MRI and CEUS (p > 0.05). Conversely, there was a significant difference regarding the peak enhancement phase between CT and MRI with respect to CEUS (p = 0.03). Conclusion: In case of doubtful findings, an additional enhanced study may be useful to confirm hypervascularity, representing the key for characterising NPET and for identifying FPET. Purpose: To investigate the prevalence of pancreatic duct variants in an ongoing epidemiological study using noninvasive secretin-enhanced magnetic resonance cholangiopancreaticography (sMRCP).

A total of 360 volunteers with a mean age of 57 (±13) years, 194 women and 166 men, and a mean body mass index of 27.8 (±4.3) kg/m 2 underwent prospective sMRCP. Navigator-triggered T2-weighted 3D turbo-spin-echo sMRCP with reconstruction of 3D-MIP was performed statically over 10 minutes after slow bolus injection of 1 U/kg secretin at 1.5 T (Magnetom Avanto, Siemens). Two readers evaluated the datasets for anatomic variants and pathologies of the pancreatic duct system. In cases of disagreement, a consensus reading was performed. Pancreatic duct variants were classified using a modified version of the guidelines of the Indiana University School of Medicine Office of Visual Media (IUSM 2005) . Results: All sMRCP examinations were of diagnostic quality. A nonvariant pancreatic duct was observed in only 61/360 cases (17%). Anatomic pancreatic duct variants were seen in 299/360 cases (83%), among them 67.5% with no accessory duct or an accessory duct not patent at the minor papilla; 7.5% with an complete or incomplete pancreas divisum. Clinical and morphologic signs of chronic pancreatitis were noted in 8 subjects (2.2%), none of which showed some form of pancreas divisum. Conclusion: Noninvasive secretin-enhanced MRCP allows screening for pancreas duct variant. Anatomic variants of the pancreatic duct system are more frequently seen in MRCP than described in anatomical data.

The volumetric measurement accuracy for ground-glass opacity nodules with low-dose high-definition CT: a chest phantom study E. Ning, W. He, Y. Xu, D. Ma, J. Li; Beijing/ CN (jianying.li@med.ge.com) Purpose: To study the volumetric measurement accuracy of a low-dose CT for ground-glass opacity (GGO) nodules on a high-definition Discovery CT750 HD (HDCT) using a chest phantom. Methods and Materials: An anthropomorphic chest phantom of 300 mm width containing a chest wall, heart, mediastinum and 16 artificial ground-glass opacity nodules with known volumes (~900 mm 3 ) was scanned on a HDCT with the following protocol: 100 kVp, automatic tube current modulation for noise index of 30, 1.25 mm collimation. Images were reconstructed with slice thickness of 0.625 mm and BONE algorithm with 30%ASiR (Adaptive Statistical Iterative Reconstruction). Nodule volumes were measured using a commercial volumetric software package. The GGO nodule volumes measured from the low-dose HDCT images were compared with the reference-standard volumes from the phantom. Effective dose was calculated on the basis of dose length product and volume CT dose index. Results: All artificial GGO nodules were successfully identified and measured. The radiation dose for the whole-chest scan was 0.94mSv. A good agreement was observed between the measured GGO nodule volumes and the reference-standard volumes from the phantom with average error of -5% (0 to -17%). Conclusion: High-resolution and low-dose GGO nodule detection and accurate volume measurement can be achieved on HDCT with Gemstone detector and ASIR algorithm.

First-pass perfusion CT on 320-detector row CT: comparison of differentiation capability of malignant from benign pulmonary nodules with first-pass dynamic MRI and PET/CT Y. Ohno 1 , M. Nishio 1 , Y. Onishi 1 , H. Koyama 1 , K. Matsumoto 2 , T. Yoshikawa 1 , S. Matsumoto 1 , D. Takenaka 1 , K. Sugimura 1 ; 1 Kobe/JP, 2 Kofu/JP (yosirad@kobe-u.ac.jp) B A S169 C D E F G H fication of the lung nodule a preoperative marking is necessary and reasonable. We report about 184 markings with a special lung marking wire, which is placed around the pulmonary nodule preoperatively using CT-guidance. Methods and Materials: In 184 patients (97m, 87f, mean age: 58.1 +/-13.7 years) with suspicious pulmonary nodules planned to undergo resection with VATS a special lung marking wire was placed preoperatively under CT-guidance. We evaluated technical success, safety, necessity of conversion to thoracotomy and histology in all patients. Results: Marking procedure was successful in 181 cases (98.4%). There were no major complications, minor adverse events such as pneumothorax (99 cases = 53.3%; maximum size 10.6+/-10.9 mm) or a perifocal bleeding (56 cases = 30.4%) did not necessitate a preoperative therapy. Complete resection of the marked pulmonary nodule was successful in 98.4% of patients. Conversion to thoracotomy was necessary in 29 patients (15.9%) due to bleedings, adhesions, malignancy or wire dislocation. Histology revealed a benign nodule in 96 cases (52.5%) and a malignant lesion in 87 cases (47.5%) out of which only 21 nodules (11.5%) turned out to present a primary pulmonary carcinoma. Conclusion: CT-guided marking of pulmonary nodules using a special marking wire followed by thoracoscopic resection is an efficient and safe method for diagnosing suspicious nodules in the periphery of the lung.

Purpose: To evaluate the effect of slice thickness on the detection of pulmonary nodule dimension and prove the efficiency of lung care software for the calculation of nodule volume and also to evaluate the diagnostic accuracy of pulmonary nodule at various locations clinically using dual-source CT. Methods and Materials: A DSCT scanner, an anthropomorphic phantom with two sets of tissue equivalent (30 HU) round nodules (diameters-2.3,4 and 5 mm); we produced images of thorax with pulmonary nodules. Four different protocols including dual energy and flash mode and four reconstruction methods with two different slice thicknesses were used for the scanning. We used Lin's concordance correlation coefficient statistical analysis with SPSS software as an appropriate statistical procedure. Results: Using manual ruler and RECIST method, the per cent deviation of ±13.06% and ±11.08% were noted for 2.0 mm nodule measured in 5.0 mm slices, and for 1.0 mm slices ±3.6% and ±2.71%, respectively. Further, volume determination of 2 mm nodule by lung care software showed the advantage of 3.57%,3.43%,0.49% and 0.57% compared with RECIST method for Th.5.0B31f, Th.5.0B80f, Th.1.0B31f, and Th.1.0B70f, respectively. The different positions of 5.0 mm diameter nodule measured by manual ruler shows the mean per cent difference of 2.77%,4.1%,1.79% and 4.45% for different reconstructions. Conclusion: Image slice thickness greatly affects the determination of pulmonary nodule dimension and thinner slice will dramatically improve accuracy. Lung care software showed highest accuracy for the calculation of volumetry compared with other methods. Further, the location of pulmonary nodules, whether it was medial or lateral did not affect its diagnostic accuracy.

Purpose: Computed tomography (CT) of the lung can be affected by respiratory artefacts especially in infants and incompliant patients. The objective was to analyse image quality of prospective respiratory-gated high-pitch (HP) CT in a ventilated ex vivo system. Methods and Materials: Four explanted porcine lungs were inflated inside a chest phantom (artiCHEST, PROdesign, Germany) and prepared with 20 artificial nodules simulating metastases (mean diameter: 11.3 mm±6). The lungs were ventilated by a computer-assisted diaphragm and scanned with a dual source CT scanner (Definition Flash, Siemens, Germany) . Prospective gated HP-CT (pitch=3.4, scan time 0.7 s) and conventional breath-hold CT scan were acquired in end inspiration phase. Non-gated HP-CT and conventional CT were also acquired. Semiautomatic volumetry of the nodules was performed (Oncology, Siemens, Germany). Results: The experiment provided a realistic respiration. The slow time resolution of conventional non-gated CT led to partial projection artefacts and inaccuracy of nodule delineation (mean nodule volumes: 772 mm3). All artefacts and nodule inaccuray predominated close to the diaphragm compared with the upper/middle parts of the lung (p < 0.05).With an inplane temporal resolution of 75 ms, the accuracy of prospective HP-CT was particularly improved. The mean volumes from the gated HP-CT, non-gated HP-CT and breathold CT scans were not significantly different (682, 691 and 672 mm3, respectively). Conclusion: Prospective HP-CT revealed excellent quality without relevant motion artefacts. Prospective HP-CT is promising and has the potential to increase significantly the quality of lung imaging in incompliant patient. HP-CT without gating is satisfactory for nodule detection and quantification. 11:42 CT-guided marking of pulmonary nodules with a special lung marking wire before video-assisted thoracoscopic surgery: review of 184 cases P. Stumpp, M.-K. Schulze, U. Eichfeld, T. Kahn; Leipzig/ DE (Patrick.Stumpp@medizin.uni-leipzig.de) Purpose: Minimally-invasive techniques such as video-assisted thoracoscopic surgery (VATS) nowadays are the method of choice for resection of small pulmonary nodules in the periphery of the lungs. For a quick and save intraoperative identi-S170 

Transarterial chemoembolisation of unresectable hepatocellular carcinoma (HCC): drug-eluting-bead embolisation (DEB-TACE) vs conventional TACE with cisplatin/lipiodol (cTACE) P. Wiggermann 1 , C. Brosche 2 , T. Brauer 2 , D. Sieron 2 , C. Stroszczynski 1 ; 1 Regensburg/DE, 2 Dresden/ DE (philippwiggermann@googlemail.com) Purpose: Outcome evaluation of DEB-TACE compared to cTACE in patients with unresectable HCC. Methods and Materials: Forty-four consecutive patients with an unresectable HCC (BCLC A/B) were analysed. All patients were either treated with one or more procedures of cTACE (n = 22) or with one or more procedures of DEB-TACE (n = 22). A total of 89 TACE procedures were performed. Two radiologists evaluated tumour response rates in accordance with the European Association for the Study of the Liver (EASL) response criteria by consensus reading. Results: Objective response (OR, complete response + partial response) for DEB-TACE was 22.7%; a further 68.2% was stable disease (SD). The respective response rates for the cTACE were OR 22.7 and SD 31.8%. Mean survival after DEB-TACE was significantly prolonged with 651 ± 76 days vs. 414 ± 43 days for cTACE (p = 0.01). Conclusion: This preliminary study indicates that DEB-TACE may have a higher efficacy than cTACE.

Intra-procedural C-arm dual-phase cone-beam CT imaging to predict response of hepatocellular carcinoma during drug-eluting bead transcatheter arterial chemoembolisation R. Loffroy 1 , M.D. Lin 2 , P. Rao 2 , N. Bhagat 2 , E. Liapi 2 , N. Noordhoek 3 , A. Radaelli 3 , J. Blijd 3 , J.F. Geschwind 2 ; 1 Dijon/FR, 2 Baltimore, MD/US, 3 Best/NL (romaric.loffroy@gmail.com) Purpose: To investigate whether intra-procedural dual-phase C-arm cone-beam CT (DPCBCT) can predict hepatocellular carcinoma (HCC) response treated using drug-eluting bead transarterial chemoembolisation (DEB-TACE) as compared to contrast-enhanced MRI, imaging gold-standard. Both modalities capture early-and delayed-phase tumour enhancement. Methods and Materials: Forty-seven lesions in 27 patients (15 males, age 61.9 ± 10.7 years) with unresectable HCC treated with DEB-TACE were retrospectively analysed. MRI was performed 1 month before and after DEB-TACE. Intra-procedural DPCBCT imaging was done immediately before and after DEB-TACE. Pre-and post-procedural DPCBCT tumour enhancement (TE) at early-and delayed-phases was assessed by two experienced radiologists blind to MRI findings. Tumour response (TR) was measured according to European Association

The 131 I-lipiodol group showed improved RFS (HR = 0·75 [95% CI 0·45 to 1·25, p = 0·27]) and, to a lesser degree, OS (HR = 0·90 [95% CI 0·51 to 1·58, p = 0·72]) compared to control, though neither reached statistical significance and the improvements were less than those seen in a previous similar trial. The combined results showed a non-significant improvement of RFS for the 131 I-lipiodol group (HR = 0·67 [95% CI 0·44 to 1·03, p = 0·07]), with a less striking trend in OS (HR = 0·69 [95% CI 0·44 to 1·11, p = 0·13]). This trial's data also showed a non-significant improved RFS, though not improved OS, in patients without vascular invasion. There were 2 serious adverse events reported, one patient with hypothyroidism caused by 131 I-lipiodol and one hepatic artery dissection during treatment. Purpose: Both loco-regional therapy with selective internal radiation therapy (SIRT) and systemic treatment with Sorafenib have shown to increase survival in patients with inoperable HCC. Combined treatment with SIRT and sorafenib may offer additional survival benefit. Methods and Materials: Prospective multicentre study of combination sequential SIRT-sorafenib therapy. Patients with inoperable HCC (Barcelona Clinic Liver Criteria (BCLC) stages B and C) were eligible. Enrolled patients were administered SIR-spheres (maximum dose 3 GBq) and subsequently administered sorafenib therapy (400 mg b.i.d). Monthly clinical and biochemical follow-up as well as 3-montly CT was performed.

Using an A'Hen's single stage design, we assumed target tumour response rate of 30% and no further interest response rate of 10% with type I error of 0.05 and power of 0.85. IT (irenebargellini@hotmail.com) Purpose: To retrospectively analyse the tumour response to transarterial chemoembolisation (TACE) in a series of consecutive HCC transplanted patients and evaluate the agreement between amended RECIST criteria and percentage of tumour necrosis, assessed at pathology. Methods and Materials: From January 1996 to December 2009, 384 cirrhotic patients (M/F= 336/48, mean age 55±7 years) with HCC underwent liver transplantation (LT). Among them, 200 (52%) patients (mean number of nodules 2.2±1.6, mean diameter 31.2±15 mm) were treated by TACE prior to LT. The CT examinations performed one month after TACE were retrospectively reviewed by two experienced radiologists in consensus, to assess tumour response to TACE according to the amended RECIST criteria. After LT, percentage of tumour necrosis was pathologically assessed on the explanted liver. Results: CT follow-up at 1 month was available in 165/200 (82.5%) patients; in the remaining cases, LT was performed within the first month after TACE. According to the amended RECIST criteria, the objective response rate was 87.3% (144/165), with 79 cases (47.9%) of complete response. On the explanted livers, mean percentage of tumour necrosis was 65±38% (range 0-100%), with 52% of patients with necrosis ≥90%. A significant correspondence (p <.0001) was found between amended RECIST criteria and percentage of tumour necrosis; mean rate of necrosis was 83% in patients with complete response, 75% in patients with partial response and 23% in subjects with stable disease. Conclusion: TACE is effective, inducing total /subtotal necrosis in over 50% of the patients. The amended RECIST criteria assessed by triphasic CT are reliable.

A S171 C D E F G H B-159 11:42 Differences in local tumour control of liver metastases from colorectal cancer between the treated liver lobes: treatment with transarterial chemoembolisation (TACE) T.J. Vogl, A. Salega, S. Zangos, N.N. Naguib; Frankfurt a. Main/DE (t.vogl@em.uni-frankfurt.de) Purpose: To evaluate differences in treatment response of palliative and symptomatic repeated transarterial chemoembolisation (TACE) of metastases from colorectal cancer between left and right liver lobes. Methods and Materials: Between January 2009 and March 2010, repeated TACE was performed in 28 patients (mean:54.61 years, range:27-79) with unresectable metastases of colorectal cancer in both liver lobes. Local chemotherapy protocols consisted of Mitomycin/Irinotecan/Cisplatin (n = 19), Mitomycin/Gemzar/Cisplatin (n = 5), Oxaliplatin (n = 1), and 3-bromopyruvate (3BP) (n = 1) (147 sessions, mean:5.3 sessions/patient). The overall response was assessed by separately measuring the volume in unenhanced MRI for target lesions in each liver lobe. The absolute change in tumour size was classified as progressive disease (PD), stable disease (SD) and partial response (PR). Results: Mean volume change in 147 treatments was a tumour growth of 5.18%, matching the criteria for stable disease. The mean response on metastases of the right liver lobe was a reduction in size by 7.59% versus an increase of 18.01% in the left lobe lesions. According to the volume change in the left liver lobe 14 patients were classified as SD, 9 as PD and 5 as PR. The right volume changes were classified as follows: 15 SD, 5 PD and 8 PR. Conclusion: Repeated TACE is a relevant palliative treatment option in liver metastases of colorectal cancer. However, the different vascularisation of the left liver lobe causes a more difficult access to the targeted tissues. Hence the used agents can be applied more effectively to the right liver lobe resulting in a better tumour control. 11:51 90Yttrium-radioembolisation of chemotherapy-refractory colorectal cancer liver metastases P.M. Paprottka, R.T. Hoffmann, F. Räßler, C. Trumm, G.P. Schmidt, M.F. Reiser, T.F. Jakobs; Munich/ DE (philipp.paprottka@med.uni-muenchen.de) Purpose: To evaluate safety and survival outcomes of radioembolisation with 90Yttrium-microspheres in patients with colorectal cancer liver metastases in whom currently available therapies had failed. Methods and Materials: One hundred patients with colorectal cancer liver metastases in whom chemotherapy had failed, underwent single-session, whole-liver 90Y radioembolisation. Response and toxicity were assessed by computed tomography/ magnetic resonance imaging utilising the response evaluation criteria in solid tumours (RECIST) and the National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE v3.0). Furthermore, tumour-marker levels, laboratory/ clinical toxicities and survival were evaluated. Results: The median activity delivered was 1.87 GBq. Imaging follow-up at 3 months demonstrated partial response, stable disease and progressive disease in 20.8%, 75% and 4.2%, respectively. Median overall survival was 11.2 months, with improved survival for patients with a decrease in carcinoembryonic antigen level (18.7 vs 5.2 months) and imaging response (28.7 vs 4.4 months; p < 0.001). Except for one instance of treatment-associated cholecystitis (grade 4 toxicity) there were no acute or delayed toxicities according to the CTCAE criteria higher than grade 2. No radiation-induced liver disease was noted. Conclusion: Single-session, whole-liver radioembolisation with 90Yttrium-microspheres is a safe and effective treatment option for patients in whom currently available therapies had failed. Anti-tumoural effect is supported by good local tumour control, decrease in tumour marker levels and significant improved survival for patients with imaging response. Further investigation is warranted to determine the optimal use of this emerging therapeutic modality.

for the Study of the Liver criteria (EASL). The change in TE compared between modalities. Results: A favourable (complete or partial) EASL TR was achieved in 74.5% and 76.6% of lesions at 1 month post DEB-TACE on MR early-and delayed-phases, respectively. Paired T-tests comparing pre-and post-TACE TE showed statistically significant average reduction in TE for both modalities by phase and lesion (p < 0.01). MRI TE decrease post-TACE linearly correlated with DPCBCT, Pearson correlation coefficient was excellent (0.80). A statistically significant relationship between post-TACE DPCBCT TE and favourable MRI TR was found: every 1 unit TE increment in DPCBCT resulted in decrease of TR by 5% (95%CI; range 0.91-0.99; p = 0.027) and 4% (95%CI; range 0.93-1.00; p = 0.034), for early and delayed-phases, respectively, on MRI. Conclusion: HCC 1-month post-DEB-TACE response as seen in MRI can be predicted by DPCBCT during the procedure.

Relationship of apparent diffusion coefficient to survival for patients with unresectable primary hepatocellular carcinoma after chemoembolisation X.-D. Ye 1 , Z. Yuan 2 , S. Dong 1 , X.-S. Xiao 1 ; 1 Shanghai/CN, 2 Nanjing/ CN (yuanyxd@163.com) Purpose: To evaluate the prognostic value of apparent diffusion coefficient (ADC) values from MR diffusion-weighted imaging of unresctable hepatocellular carcinoma after chemoembolisation. Methods and Materials: our study was proved by our institute and informed consent was obtained from all patients before commencement of the study. Twenty-three patients with unresctable hepatocellular carcinoma were scanned immediately before and after chemoembolisation within 24h using conventional anatomical MR imaging and diffusion-weighted imaging, from which ADC values in the lesions were measured. The changes in ADC values after chemoembolisation were calculated. The relationship between the lesion ADC and the survival time was analysed by correlation analysis. The overall cumulative survival was analysed by the Kaplan-Meier method, and survival curves were compared by the log-rank test. Results: The mean overall survival period was (25.0 ± 8.7) months. The pre-chemoembolisation lesion ADC value was (1.36 ± 0.249) ×10 -3 mm 2 /s; the change in ADC values post-chemoembolisation was (0.377 ± 0.332) ×10 -3 mm 2 /s. There were significant linear regression relation between the survival time and pre-chemoembolisation lesion ADC values (r = -0.698, P < 0.001) or the changes in ADC value post-chemoembolisation (r = 0.702, P < 0.001). And Log-rank test showed that pre-chemoembolisation ADC values (χ 2 = 7.339, P = 0.007) or the changes in ADC value post-chemoembolisation (χ 2 = 9.820, P = 0.002) significantly influenced the overall cumulative survival. Conclusion: Pre-treatment ADC values as well as changes in ADC values after treatment may provide useful information for predicting survival for patients with unresctable hepatocellular carcinoma. 11:33 Tumour response in liver metastases from the most common primary sites after repetitive transarterial chemoembolisation (TACE): a competitive evaluation P. Farshid, A. Davishi Nakhl Ebrahimi, A. Azizi, N. Naguib, E. Mbalisike, T.J. Vogl; Frankfurt a. Main/DE (parvizfarshid@gmail.com) Purpose: Competitive evaluation of results in the palliative treatment of patients with liver metastases from the most common primary sites using transarterial chemoembolisation (TACE). Methods and Materials: TACE was repeatedly performed in 105 patients with liver metastases from colon (31), breast (21), the uveal malignant melanoma (13), pancreas (31) and stomach (9). Radiologists retrospectively evaluated tumour response by MRI. The response evaluation criteria in solid tumours (RECIST) and specific growth rate (SGR) were used to assess tumour response. Survival parameters were calculated by Kaplan-Meier method. Results: Tumour response after TACE [mean, 4.18 sessions per patient (range: 2-12)] following RECIST and SGR evaluation showed no statistically significant difference between all groups [7.1015, (p = 0.5257) and 4.5121, (p = 0.3411)]. Mean of SGR (% per day), for liver metastases from colon, breast, uveal malignant melanoma, pancreas and stomach was -0.241, -0.208, -0.12, -0.099, and -0.074, respectively. Mean survival time and survival rate (1 and 3 years) from first TACE were for colon, 27, 89.5%, 51%; breast, 25.9, 88%, 49%; malignant melanoma, 20.83, 82.5%, 47.5%; pancreas, 16.72, 62%, 33%, and stomach 13, 55%, 28%. Median TTP and progression-free survival rate for 1 year were for liver metastases from colon 4.1, 89%; breast 3.5, 91%; uveal melanoma 4.3, 88%; Pancreas 2.91, 90%, and stomach 3.38, 53%. Conclusion: Despite lack of statistically significant difference in local tumour response after TACE using size-and volume-based evaluation, survival indexes were better in colon, breast, uveal melanoma, pancreas and stomach liver metastasis, respectively. S172 B C A D E F G H Conclusion: DW-MR imaging is a non-invasive imaging approach and a sensitive technique that allows the identification of spatially distinct regional responses to therapy within tumour tissues. For reader 2 these figures were 0.87, 84% and 74%. After addition of DWI, reader 1 achieved an AUC of 0.98, sensitivity 100% and specificity 91%. Compared with T2W-MRI, the difference in AUC was not significant (p = 0.78). For reader 2 these figures were 0.91, 89% and 83% (p = 0.09). Fusion images did not improve accuracy compared with T2W-MRI+DWI, but were helpful for a better anatomical correlation. Interobserver-agreement was κ0.69 for T2W-MRI, 0.82 for T2W-MRI+DWI and 0.84 for the fusion images. Conclusion: MRI has a high accuracy for the detection of locally recurrent rectal cancer. Addition of DWI does not significantly improve its performance. However, with DWI specificity and interobserver-agreement increase. Fusion images do not further improve accuracy.

Diffusion-weighted magnetic resonance imaging combined T2-weighted imaging evaluation in the pre-operative staging of primary rectal carcinoma R. Rong, X.Y. Wang, X.X. Jiang, X.W. Sun; Beijing/ CN (thornapplezixi@gmail.com) Purpose: To investigate the added value of diffusion-weighted (DW) MR imaging in the evaluation of pre-operative assessment of T staging in patients with primary rectal carcinoma. Methods and Materials: 27 patients have been selected in the study: 18 males and 9 females (mean age 60.7 ± 13.9 years; range 27-83 years). Retrospectively, one radiologist scored T2-weighted MR images and, in a second reading, T2-weighted images combined with DW MR images for tumour penetration through rectal wall and extension into mesorectal fascia. Spearman correlation was used to evaluate the pre-operative staging consistency of T2-weighted and combined T2-weighted plus DW MR images according to histopathological results. ROC curve was drawn to evaluate the diagnostic efficacy of the two. Results: For T staging of rectal cancer, addition of DW imaging correctly staged patients with rectal cancer more often than T2-weighted imaging (Spearman correlation: 0.695 vs. 0.456, P < 0.05). Addition of DW MR imaging was superior to T2-weighted imaging in the differentiation between T1\T2 tumour (positive group) and T3\T4 tumour ( Purpose: Recent publications show that MRI in rectal cancer is more accurate in selecting yN0-patients after chemoradiation (CRT) than in selecting the cN0-patients at primary-staging. This improved accuracy facilitates selection of patients for local excision after CRT. We aimed to evaluate whether changes in nodal distribution and size after CRT can explain this phenomenon.

We included 46 patients with locally advanced rectal cancer (LARC) who underwent preoperative CRT (group A) and 26 patients with non-LARC who underwent immediate surgery with/without preoperative 5x5Gy (group B). Group A underwent MRI before and 6-8 weeks after CRT and group B underwent pretreatment-MRI only. Nodes were analysed on 3DT1W-GRE with 1 mm 3 isotropic voxels. For each node, short axis diameter was measured.

For group A, nodes were measured and matched on pre-and post-CRT MRI. Results:In group A, 882 nodes were identified on pre-CRT-MRI, of which 380 (43%) disappeared after CRT. Mean number of nodes/patient decreased from 21 (±9) pre-CRT to 13 (±7) post-CRT (p < 0.0001). Mean nodal size (mm) decreased from 3.8 (±2) pre-CRT to 2.6 (±1.8) post-CRT (p < 0.0001). Mean size on pre-CRT-MRI of the 380 nodes that disappeared after CRT was smaller than that of the nodes that were still visible after CRT: 2.9 (±1.4) vs. 3.8 (±2.0) mm (p < 0.0001). In group B 558 mesorectal nodes were visualised (mean 22 (±11) nodes/patient). Conclusion:Pretreatment number of nodes does not differ between LARC and non-LARC. After CRT, mean number and size of nodes decrease significantly. Most small nodes disappear after CRT. These findings may contribute to the fact that size-criteria work better for restaging-MRI after CRT.

The added value of diffusion weighted imaging in combination with T2 weighted imaging for the evaluation of rectal cancer response to neoadjuvant treatments L. Monguzzi 1 , D. Ippolito 1 , C. Capraro 1 , M. Colombo 1 , C. Talei Franzesi 1 , F. Invernizzi 2 , E. De Ponti 1 , S. Sironi 1 ; 1 Monza/IT, 2 Desio/IT (letizia.monguzzi@gmail.com) Purpose: To assess the added value of DWI in combination with T2-weighted imaging in determining of tumour response to CRT, as compared to T2WI alone, correlated with tumour histopatologic downstaging as reference standard. Methods and Materials: A total of 29 patients with LARC, who underwent CRT and subsequent surgery, underwent pre-and post-CRT MRI examination (1.5 T scanner, Achieva, Philips) with following protocol: multiplanar T2 and T1 TSE images and in addition DWI sequences (b factor: 0 and 1000 mm 2 /sec).Two readers in consensus, blinded to the results, evaluated T2WI without and with DWI to detect the presence of complete response by a four-point scale as follows (1: definitely absent; 2: probably absent; 3: probably present; 4: definitely present). Mean value of ADC of pre-and post-CRT DWI were compared between responders and non-responders patients and correlated with histopatological response classified as tumour regression grade (Mandard's criteria; TRG1 = complete regression, TRG5 = no regression). Results: All of the patients underwent total mesorectal excision and Mandard's criteria were calculated (responders: TRG1-2; non-responders: TRG3-5). Twenty tumours (75%) showed complete (TRG1) or subtotal regression (TRG2) and were classified as responders; nine tumours (25%) were classified as non-responders (TRG3-5). Mean ADC value in MR1 examination was 0.86 ± 0.22 x 10-3 mm²/ sec, whereas mean values after CRT was 1.43 ± 0.25 x 10 -3 mm²/sec. On T2W images both readers correctly identified 12 of 18 patients with complete response (specificity 65%) while after addition of DWI the readers correctly identified 17of the complete responders (spec 95%) patients. The analysis showed that each of the two readers achieved more accurate results with T2 imaging combined with DW images than T2W images alone.

A S173 C D E F G H Results: CE-MRI indicated N0-status in 35/38 patients, who underwent follow-up. 3 patients had suspected N+status, of whom 2 had surgery, which confirmed the N+status in one patient, but refuted N+status in the other (false-positive). The third patient underwent chemoradiation which sterilised the nodes. These 3 patients are disease-free and alive. Median follow-up for all patients is 31 months (0-65). Three luminal and one nodal LR were found in the follow-up group. All recurrences were curatively treated. 2-year disease-free survival is 94%. Conclusion: CE-MRI with lymph-node-specific contrast-agents is useful for selecting N0-patients for LE. The low LR-rate and resectability of the local recurrences suggest that CE-MRI can play a role in safely selecting T1-2N0 tumours for LE, leading to lower treatment-related morbidity.

Usefulness of dynamic contrast-enhanced magnetic resonance imaging for evaluation of therapeutic response to neoadjuvant chemoradiation therapy in locally advanced rectal cancer S. Kim, J. Lee, J. Han, B. Choi; Seoul/ KR (radiresi@gmail.com) Purpose: To evaluate the efficacy of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) for assessing the therapeutic response to neoadjuvant chemoradiation therapy (CRT) for locally advanced rectal cancer. Methods and Materials: Between June 2009 and August 2010, 50 patients (32 males, 18 females; mean age, 64 years; range, 40-82 years) with locally advanced rectal cancer (≥ T3 or lymph node positive) who underwent CRT and subsequent surgery, were included in this study. All patients underwent pre-and post-CRT, 1.5-T DCE-MRI. A radiologist blinded to the final pathology drew a region of interest along the border of the tumour on the pre-and post-CRT MRI. Perfusion parameters, i.e. the volume transfer constant (K trans ), rate constant (K ep ), and volume of extracellular extravascular space (v e ) were calculated using the general kinetic model for the two compartments. The mean values of the perfusion parameters were compared not only between the T-downstaged group and the non-downstaged group, but also before and after CRT in each group. The histopathology results served as the reference standard. Results: The mean K trans (min -1 ) significantly decreased after CRT from 1.24± 0.53 to 0.76± 0.45 in the T-downstaged group (n = 24) (P = 0.0007), and it also tended to decrease from 1.02 ± 0.53 to 0.87 ± 0.48 in the non-downstaged group (n = 26) (P = 0.2358). None of the other parameters showed significant differences. Conclusion: A significant decrease in the mean K trans was a useful indicator for differentiating favourable responders from poor responders to CRT for locally advanced rectal cancer. (d.lambregts@mumc.nl) Purpose: Recently, we showed that MRI using a new lymph-node contrast (gadofosveset) has high accuracy for nodal staging in rectal cancer (results in press). The Aim of the current study was to determine the most accurate imaging criteria for nodal evaluation using gadofosveset-enhanced MRI. Methods and Materials: 33 patients underwent 3D T1-weighted MRI before and after administration of gadofosveset-contrast. For each visible lymph node the signal intensity (SI) was measured by placing an ROI on both the pre-and post-contrast MRI. Furthermore, the presence of a nodal relief effect was scored. Finally, nodes were subjectively scored (using a 5-point confidence score) by an MR-rectum expert, based on visual evaluation of SI and nodal relief. Nodes were drawn on an anatomical template for lesion-by-lesion histologic validation. Results: 289 nodes (55 malignant) were analysed. On pre-contrast MRI, the SI did not differ between benign and malignant nodes (p = 0.12). On gadofosveset-MRI, SI was significantly higher in the benign nodes (p < 0.0001), which resulted in an area under the ROC-curve (AUC) of 0.74 for prediction of metastatic nodes. The nodal relief rendered an AUC of 0.85. The combination of SI and nodal relief rendered an AUC of 0.88. When analysed visually, AUC for discrimination of metastatic nodes was 0.92 Conclusion: Only benign nodes show uptake of gadofosveset-contrast. The most optimal criteria for prediction of metastatic nodes are the nodal signal intensity on gadofosveset-MRI and the presence of a nodal relief. When analysed visually, these criteria provide high accuracy for discrimination of metastatic nodes.

All patients underwent MRI, including DWI (b-values 0.500.1000) before and 6-8 weeks after CRT. Mean ADC of the primary tumour was measured at both timepoints. When after CRT no clear tumour or scar tissue could be identified, ADC of the primary tumour bed was measured. Histology was the standard reference. The relative change (Δ) in ADC (%) after treatment was compared for good and poor-responding tumours. A good response was defined as a histological tumour regression grade (TRG, Mandard) of 0-2 and a poor response as TRG3-4. Results: At histologic evaluation 17 patients had a TRG0-2 and 12 patients a TRG3-4. After chemoradiation, a significant increase in ADC was observed in all patients: mean pre-treatment ADC was 1.11 ± 0.22 versus 1.47 ±0.25 *10 -3 mm 2 /s after treatment (p < 0.001). The relative increase in ADC (%) was larger in the good responders (43% ADC-increase) than in the poor-responders (31%) but this difference was not statistically significant (p = 0.2). Conclusion: ΔADC tends to be higher for the good responders as compared with the poor responders, although the difference was not significant. This could partly be attributed to the small number of patients. Larger studies will show whether relative changes in ADC will be useful to discriminate between good-and poorresponding tumours.

Spectral fat saturation inversion recovery MR sequence after gadolinium injection to differentiate fibrotic and neoplastic strands in the mesorectal fat in patients with restaging of rectal carcinoma after neoadjuvant chemoand radiation therapy E. Quaia, V. Ulcigrai, C. Federica, M. Cavallaro, M.A. Cova; Trieste/ IT (quaia@units.it) Purpose: To retrospectively assess the capabilities of contrast-enhanced spectral fat saturation inversion recovery (SPIR) MR sequence in the differentiation between fibrotic and neoplastic strands within the mesorectal fat in the restaging of patients with locally advanced rectal carcinoma after neoadjuvant radiochemotherapy. Methods and Materials: Forty consecutive patients (27 males, 13 females; mean age±SD, 65.1±8.5 years) with locally advanced rectal carcinoma underwent preoperative neoadjuvant radio-chemotherapy followed by surgery. MR imaging was performed 15-25 days after completion of neoadjuvant therapy by T2-weighted fast spin-echo (TR/TE, 2500/100 msec) and T1-weighted SPIR sequence (TR/ TE, 570/50 msec) after Gd HP-DO3 A injection (0.1 mmoL/kg). Two experienced radiologists (5, 10 years) visually analysed in consensus the MR images obtained before and after neoadjuvant therapy and evaluated the evidence of enhancing nodules or strands. Logistic regressions were conducted to test the significance of MR image findings with histologic analysis. Results: After neoadjuvant treatment, the disease was limited to the wall (n=18 patients), or revealed mesorectal fat infiltration (n=22) at histology. In 20 patients enhancing strands in the mesorectal fat were identified, and reticular enhancing strands reaching the mesorectal fascia presented the highest correlation with mesorectal fat tumoural infiltration (OR 28.52, 95%CIs: 1.34-604.1), while mesorectal enhancing nodules (OR.295, 95%CIs:.024-3.58) and separate enhacing strands reaching or not the mesorectal fascia (OR.07 or.12, 95%CIs:.007-.65 or.012-1.25) revealed the lowest correlation. Conclusion: The fibrotic or tumoural nature of the enhancing strands within the mesorectal fat evident after neoadjuvant radiochemotherapy may be characterized by contrast-enhanced SPIR MR sequence. Purpose: Patients with T1-2 rectal cancer have 5-20% risk for nodal metastases, which cannot be accurately evaluated with standard imaging. To minimise the risk for local recurrence (LR), total mesorectal excision (TME) is performed in these patients. Accurate prediction of N0-status could make local excision (LE) feasible. The aim of this study is to evaluate whether contrast-enhanced (CE)-MRI is useful in selecting patients with N0-status for LE. Methods and Materials: 38 patients who underwent LE for rectal cancer were included and underwent T2W&T1W-CE-MRI with ultrasmall superparamagnetic particles of iron-oxide (USPIO;n = 26) or gadofosveset (n = 12) to evaluate N-stage and residual tumour. A radiologist experienced in rectal MRI and the CE-images evaluated the images. In case of N0-status, patients underwent follow-up with regular visits, endoscopy and/or MRI as part of a prospective study. Patients with N+status underwent TME (±neoadjuvant therapy). Purpose: To compare 3 T MRI-based knee cartilage T2 measurements and focal knee lesions of normal, overweight and obese individuals over a 36 months period. Methods and Materials: 231 asymptomatic individuals (128 males, 103 females) aged 45-55 years with risk factors for knee osteoarthritis (OA) and no radiographic OA were randomly selected from the Osteoarthritis Initiative (OAI) incidence cohort. 78 individuals had normal weight, 84 individuals were overweight and 69 individuals obese. All individuals underwent 3 T MRI of the right knee at baseline and 36-month follow-up. Focal knee lesions were assessed using the whole organ MRI score (WORMS). T2 measurements (mean T2, texture and laminar analysis) were performed in five compartments (patella, medial/lateral femur/tibia). Groups were compared using multivariate regression models adjusted for gender, age, KL-score and OA risk factors other than overweight/obesity. Results: Compared to normal individuals, overweight and obese individuals had more and more severe meniscal and cartilage lesions at baseline and after 36 months (p < 0.05). Baseline prevalence of meniscal (cartilage) lesions was 50% vs. 68% vs. 73% (71% vs. 80% vs. 91%). Mean T2 values and all texture parameters were significantly different in all compartments for the three groups at baseline (p < 0.05). Obese individuals showed the highest mean T2 values and the most heterogeneous cartilage. In contrast to texture analysis parameters, mean T2 values were no longer significantly different between the groups after 36 months. Conclusion: Overweight and obesity are associated with more and more severe knee lesions and advanced cartilage degeneration according to T2 measurements.

New isotropic 3D PD-SPACE MR imaging in the comparison to isotropic 3D True-FISP and standard 2D MRI to depict the 3D MOCART score G.H. Welsch 1 , L. Zak 1 , A. Mauerer 2 , T.C. Mamisch 3 , S. Marlovits 1 , D. Paul 2 , L. Lauer 2 , S. Trattnig 1 ; 1 Vienna/AT, 2 Erlangen/DE, 3 Berne/CH (welsch@bwh.harvard.edu) Purpose: To use the capabilities of the new isovoxel sequences and their 3Dmultiplanar-reconstruction (MPR), an new isotropic 3D proton-density, turbo-spinecho (PD-SPACE) sequence was compared to an isotropic 3D true-fast-imaging with steady-state-precession (True-FISP) sequence and to a set of 2D standard MR sequences in their depiction of the 3D-magnetic resonance observation of cartilage repair tissue (MOCART) score. Methods and Materials: Sixty consecutive MR scans on 37 patients (32.8 ± 7.9 years) during standard follow-up intervals after matrix-associated autologouschondrocyte-transplantation (MACT) of the knee at 3 Tesla were prospectively included. The new 3D-MOCART score with eleven variables was assessed using the standard 2D sequences (~15 minutes) and the MPR of each isotropic sequence (3D-PD-SPACE ~7 minutes; 3D-True-FISP ~6 minutes).

The correlation between the 3D-MOCART scoring performed by the different sequences was highly significant for the variables 1) defect fill, 2)cartilage interface, 3) bone interface, 4) surface, 7) subchondral lamina, 8) chondral osteophyte, and 11) effusion (Pearson-coefficients 0.514 to 0.865 (p < 0.001)). The variables 5) structure, 6) signal intensity, 9) bone marrow oedema, and 10) subchondral bone showed lower correlations with best results with respect to the 3D-PD-SPACE sequence (0.307 to 0.633 (p = 0.016 to p < 0.001)). Grading of subjective quality revealed good results for all sequences (p ≥ 0.05). Artefacts were most often visible on the 3D-True-FISP sequence (p < 0.05). Conclusion: Different isotropic sequences can be used for the 3D evaluation of cartilage repair tissue with the benefits of isotropic 3D-MRI, MPR and a significantly reduced scan time.The 3D-PD-SPACE sequence reveals best results due to a better performance in the subchondral bone and because of the suppression of susceptibility artefacts produced by implantation and previous surgeries. Purpose: Meniscal extrusion is a risk factor for knee osteoarthritis (OA). Understanding risk factors for meniscal extrusion is important in order to develop targeted treatment. We aimed to assess the cross-sectional associations of meniscal tears, knee malalignment, cartilage damage, and body mass index (BMI) with meniscal extrusion.

Rectal adenocarcinoma: intratumoural oxygenation level assessed with BOLD MRI and correlates with prognostic factors -initial experience Y. Liu, R. Xu; Shenyang/ CN (liuli0016@yahoo.com.cn) Purpose: To establish the relationships between Intratumoral oxygenation level and immunohistologic markers of angiogenesis and metastasis in colorectal cancer. Methods and Materials: This study was approved by the local ethics committee, and written informed consent was obtained from all patients. BOLD MR imaging was performed in 51 patients with rectal adenocarcinoma to estimate R2* in tumour parenchyma, which is inversely related to tissue Po 2 . After surgery, resected specimens were sectioned and stained immunohistochemically to identify CD34 for quantification of microvessel density (MVD), and to identify matrix metalloproteinase-2 (MMP-2) for estimating of metastasis. The R2* value in tumour parenchyma was correlated with MVD and MMP-2 expression by using Pearson or Spearman rank correlation analysis, with significance assigned at the 5% level. Results: R2* value in tumour parenchyma is 26.6/sec±3.89; it was correlated positively with MVD (r = 0.333, P =0.017). However, there were no significant associations between R2* values in tumour parenchyma and MMP-2 score (rho <or= 0.158, P >or=0.27). Purpose: Bone marrow oedema-like lesions (BMLs) are predictors of cartilage loss in knee osteoarthritis (OA). Inversely, loss of cartilage may remove a tissue protecting underlying bone, increasing loading and possibly damage to bone. We assessed the association of prevalent cartilage damage and cartilage loss over time with incident BMLs in the same subregion of the knee. Methods and Materials: MOST is a longitudinal study of persons who have or are at risk for knee OA. Subjects with available baseline (BL) and 30 months follow-up (FU) 1.0 T MRIs were included. MRIs were read using the WORMS system. Tibiofemoral compartments were divided into 10 subregions. Cartilage morphology was scored from 0 to 6. BMLs were scored from 0 to 3. BL cartilage damage was defined as grade ≥ 2. Cartilage loss had two groups: incident (grades 0 at BL and ≥2 at FU) and progression (BL cartilage damage and increased grade at FU). Incident BMLs were defined as grades 0 at BL and ≥1 at FU. Logistic regression was applied, adjusting for potential confounders. 

A S175 C D E F G H Results: SNR was (20.9 ± 4.6) in cartilage, (18.0 ± 2.4) in muscle and (52.7 ± 7.0) in synovium. Mean ADC and FA values were, respectively (1.03 ± 0.06) × 10 -³ mm²/s, (0.44 ± 0.05) for cartilage, (1.18 ± 0.09) × 10 -³ mm²/s, (0.34 ± 0.04) for muscle and (2.59 ± 0.08) × 10 -³ mm²/s/(0.083 ± 0.005) for synovium. ADC decreased from the articular surface (1.52 ± 0.23) × 10 -³ mm²/s to the bone-cartilage interface (0.83 ± 0.19) × 10 -³ mm²/s. FA increased form the surface (0.23 ± 0.17) to the deep radial zone of the cartilage (0.73 ± 0.32). Mean reproducibility of ADC and FA were 6.5%/8.4% in cartilage, 7.3%/9.8% in muscle and 2.5%/5.3% in synovium. Conclusion: For a first time in vivo acquisition of cartilage DTI data become feasible. In vivo ADC and FA showed a good reproducibility and are consistent with published ex vivo results.

Cartilage quality assessment using gagCEST and sodium MRI at 7 Tesla B. Schmitt 1 , S. Zbyn 2 , D. Stelzeneder 2 , V. Jellus 3 , D. Paul 4 , P. Bachert 1 , S. Trattnig 2 ; 1 Heidelberg/DE, 2 Vienna/AT, 3 Erlangen/ DE, Purpose: gagCEST and 23Na MRI provide a contrast sensitive to glycosaminoglycan (GAG) content of articular cartilage. The aim of our study was to compare a newly developed gagCEST imaging method with 23Na imaging in patients after cartilage repair surgeries (matrix-associated autologous chondrocyte implantation and microfracture). Methods and Materials: One female (age: 67.3 years), and 11 male (mean age: 29.3 ± 6.04 years) patients were examined using a 7-T whole-body system with approval of the local ethics committee after written informed consent was obtained. A modified 3D-GRE sequence and a 28-channel knee coil were used for gagCEST imaging. 23Na imaging was performed with a circularly polarized knee coil using a modified GRE sequence. Statistical analysis of differences and Pearson correlation were applied. Results: The average asymmetry in gagCEST z-spectra summed over all offsets from 0 to 1.3 ppm was 8.01 ± 0.2 % (mean ± SEM) in native cartilage and 4.6 ± 0.2 % in repair tissue. A strong correlation (R = 0.885, p = 0.0001) was found between ratios of native cartilage to repair tissue obtained with gagCEST, and sodium imaging. The mean dimensionless ratios between native cartilage and repair tissue were 1.39 ± 0.3 (range: 1.05 -1.9) for gagCEST, and 1.34 ± 0.21 (range: 1.1 -1.78) for sodium MRI (mean ± SD). Conclusion: The high correlation between the new gagCEST method and 23Na imaging results demonstrates the GAG specificity and reliability of the technique. Both techniques suggest that cartilage repair tissue has lower GAG content than native tissue.

7T 23Na (Sodium) imaging after knee matrix-associated autologous chondrocyte transplantation: preliminary results S. Trattnig 1 , G.H. Welsch 2 , V. Juras 1 , M.E. Mayerhoefer 1 , D. Stelzeneder 1 , T.C. Mamisch 3 , O. Bieri 4 , K. Scheffler 4 , S. Zbyn 1 ; 1 Vienna/AT, 2 Erlangen/DE, 3 Berne/CH, 4 Basle/CH (siegfried.trattnig@meduniwien.ac.at) Purpose: To evaluate the feasibility of sodium imaging at 7T indicative of glycosaminoglycans in repair and native cartilage after matrix-associated autologous chondrocyte transplantation (MACT) and compare results with delayed gadoliniumenhanced MRI of cartilage (dGEMRIC) at 3 T. Methods and Materials: Six females and six males with mean age 32.8±8.2 [standard deviation] and 32.3±12.7 years were included, with mean follow-up 56±28 months after MACT. At 3 T, variable 3D-GRE dual flip-angle technique was used for T1-mapping before and after contrast agent administration. At 7T, sodium MRI of all patients using 3D-GRE sequence was performed with mean delay 70.5±80.1 days. 23Na-only transmit/receive knee coil was used. Statistical analysis of variance and Pearson correlation were applied. Results: Mean signal-to-noise ratio (SNR) in native cartilage was 24 and in transplants 16 (p < 0.001). Mean sodium intensities normalised by reference sample were 174±53 for repair tissue within cartilage transplant and 267±42 for healthy cartilage (p = 0.002). Post-contrast T1 values were 510±195 ms for repair tissue and 756±188 ms for healthy cartilage (p = 0.005). Mean score of magnetic resonance observation of cartilage repair tissue (MOCART) was 75±14. Association between post-contrast T1 and normalised sodium values show high Person correlation coefficient (R) of 0.706 (p = 0.001). High correlation R = 0.836 (p = 0.001) was found between ratios of normalised sodium values and ratios of T1 post-contrast values. Conclusion: Using modified 3D-GRE sequence at 7T, sufficiently high SNR in sodium images was achieved, allowing for the differentiation of repair tissue and native cartilage after MACT. Strong correlation was found between sodium imaging and dGEMRIC in patients after MACT.

MOST is a study of individuals who have or are at risk for knee OA. Subjects with available baseline (BL) knee radiographs and 1.0 T MRIs were included. MRIs were read using the WORMS system. Meniscal tears were graded from 0 to 4, where grades ≥ 1 exhibit tears. Meniscal extrusion was graded from 0 to 2, where grades ≥ 1 exhibit extrusion. Tibiofemoral cartilage morphology was scored from 0 to 6, where grades ≥ 2 exhibit cartilaginous defects. Associations were assessed using logistic regression. Multi-adjustments were performed when testing each predictor. Results: 1527 subjects were included. Medially, meniscal tears, varus malalignment, and cartilage damage were significantly associated with meniscal extrusion, with odds ratios (OR) of 6.5 [95%CI 5.1, 8.2 (j.vantiel@erasmusmc.nl) Purpose: Imaging techniques, e.g. delayed Gadolinium-enhanced MRI of cartilage (dGEMRIC), that accurately and quantitatively measure cartilage quality are increasingly important in osteoarthritis research. X-ray attenuation values in ex-vivo contrast-enhanced μCT have been used as a measure of glycosaminoglycans (GAG) distribution in the cartilage, and hence cartilage quality. The purpose of this study was to determine whether differences in GAG distribution can also be detected using a clinical application of CT-arthrography (CTa). Methods and Materials: On six human cadaveric knee joints a CTa was performed with a second-generation dual-source multidetector spiral CT scanner (Siemens, Germany) after intra-articular injection with a negatively charged contrast agent (Hexabrix 320, Guerbet). After CTa, the knees were dissected into five parts: both medial and lateral femoral condyles and tibial plateaus and the patella. Subsequently, all parts were scanned using ex-vivo contrast-enhanced μCT (Skyscan, Belgium) as a reference standard. X-ray attenuation values from both techniques were correlated between seven anatomically matching regions of interest (ROIs) in the cartilage using a Pearson's correlation coefficient. Results: X-ray attenuation of the different cartilage ROIs in the clinical CTa scans ranged from 144 -236 Hounsfield Units. Comparison with the ex-vivo contrastenhanced μCT resulted in a correlation coefficient of r=0.80 and a R2=0.65 (95% CI 0.58-0.92, p < 0.0001). Conclusion: The excellent correlation between clinical CTa and ex-vivo μCT shows that clinical CTa can be used to discriminate between GAG distributions within different cartilage regions and can thus be a cheap alternative for imaging techniques such as (dGE)MRI (C).

In vivo diffusion tensor imaging of the articular cartilage at 7 T: a feasibility study J.G. Raya 1 , A. Horng 2 , O. Dietrich 2 , L.S. Beltran 1 , S. Krasnokutsky 1 , M.F. Reiser 2 , M. Recht 1 , C. Glaser 1 ; 1 New York, NY/US, 2 Munich/ DE (Jose.Raya@nyumc.org) Purpose: To demonstrate the feasibility of in vivo diffusion tensor imaging (DTI) of articular cartilage with a line scan diffusion imaging (LSDI) sequence at 7 T using a dedicated knee coil. Methods and Materials: The right knee of 15 healthy volunteers ((29 ± 3) y) was two time scanned with knee reposition between the scans on a whole body 7 Tscanner using a birdcage transmit, 28-channel receive knee coil with a fat-saturated LSDI sequence (TE/TR/TReff = 45/180/2880 ms, matrix = 256 × 128, b-values = 10,450 mm²/s, 6 directions, FOV = 154 × 77 mm², thickness = 2 mm, acquisition time = 2:47 min/slice). Five axial slices over the patellar cartilage were acquired with a 3 mm gap to cover the complete patellar cartilage. ADC and FA maps and their mean values for cartilage, synovial fluid and muscle were calculated. SNR was calculated in the b0 image. Reproducibility was assessed by root-mean-square of coefficient of variation. S176 B C Purpose: To compare the performance of three high-end modalities for the ex-vivo assessment of bone microstructure. Methods and Materials: 3 bone samples from femoral heads of male patients with osteoporotic hip fractures (n = 1) and osteoarthritis of the hip (n = 2) were investigated with a HR-pQCT scanner (Scanco XtremeCT) at a resolution of 41 μm. Moreover, the cubes were scanned with two other dedicated high-resolution CT-scanners (Scanco μCT 40 and Viscom X8060 NDT) at a resolution of 20 μm, within the Research Network Osteology (RNO) Vienna which is an interdisciplinary platform for osteologic research focusing on qualitative and quantitative bone assessment in anatomical, anthropological and clinical bone samples. Parameters of bone microstructure were compared by mean absolute differences (%) of the measurements.

The comparison of HR-pQCT and Scanco μCT 40 showed an overestimation of trabecular bone volume (+15%) and trabecular thickness (+33.1%) by the HR-pQCT. Trabecular number was underestimated by the HR-pQCT scanner (-11.3%). The Viscom X8060 NDT μCT provides high image quality for 3D reconstructions, but still lacks a standardised evaluation software for quantitative microstructure analysis. Conclusion: HR-pQCT and micro-CT provide divergent results in ex-vivo human bone morphometry, which are most likely to be due to different resolutions and partial volume effects. This should be kept in mind when studies on bone microarchitecture are compared or cited. The Viscom system provides images of high quality, but still lacks standardised quantification options. 

A S177 C D E F G H degrees. Average glandular dose 0.7 mGy (range 0.3-1.4) per PhCT exposure. After training, ten accredited film readers read a set of 130 examinations, 64 with verified abnormalities and 66 with no findings, using two-view FFDM and two-view PhCT over 2 sessions. A further ten readers reviewed the same cases using two-view FFDM but single view (MLO) PhCT. The multireader-multicase (MRMC) ROC methodology was applied and the PROPROC fitted area under the ROC curve was used Figure of Merit. Results: No statistically significant difference was reached between detection accuracy with either FFDM and 2 views PhCT (readers' average AUC FFDM = 0.773, AUC PhCT = 0.810, p-value = 0.22), or FFDM and single MLO PhCT (readers' average AUC FFDM = 0.708, AUC PhCT = 0.680, p-value = 0.32). Average reading times were 125 s for two-view PhCT, 97 s for the single MLO PhCT, and 71 s for two-view FFDM. Conclusion: No differences were found in detection accuracy either between dualview PhCT and dual-view FFDM or between single-view PhCT and dual-view FFDM.

Digital breast tomosynthesis versus digital mammography in a clinical setting: side-by-side feature analysis of cancer conspicuity P. Skaane, R. Gullien, H. Bjorndal, U. Ekseth, U. Haakenaasen, G. Jahr; Oslo/NO (per.skaane@ulleval.no) Purpose: Compare cancer conspicuity of digital mammography and tomosynthesis in a clinical setting. Methods and Materials: 129 women underwent mammography (2D) and digital breast tomosynthesis (DBT). Indication included lump in 30 (23%), abnormal mammographic findings 54 (42%), and surveillance 45 (35%) women. 20 (16%) had a benign biopsy, 25 (19%) a cancer, and 84 (65%) no biopsy. BI-RADS density was 1 in 10 (8%), 2 in 70 (54%), 3 in 43 (33%) and 4 in 6 (5%) women. 2D exams were interpreted on-site and DBT retrospectively after women left the department. Sideby-side feature analysis of cancer conspicuity was performed by 3 external breast radiologists trained with cases not included in the study. The readers interpreted bilateral 2D and DBT of 50 women without information about prior imaging, indication, pathology, or cancer mix in the test set. The test set included 23 cancers. Results: DBT resulted in 2 false positive recalls and 2 additional cancers. Feature analysis of cancer conspicuity for the 3 readers (2D higher, equal, and DBT higher score): Reader A 1.14,8; reader B 1.13, 9; reader C 3.5,15, respectively. Conspicuity was higher for 2D in 5, equal in 32 and higher for DBT in 32 interpretations. Scores for 14 cancers presented as spiculated mass or distortion:1.10 and 31 respectively. For 5 DCIS, scores were 3, 12 and 0, respectively. The mean cancer scores was 3.42 for 2D and 3.98 for DBT. Conclusion: Tomosynthesis shows higher conspicuity for malignancy than mammography, especially for spiculated masses and distortions.

Role of mammography cranio-caudal view in adjunction to digital breast tomosynthesis: clinical performance study G. Gennaro 1 , E. Bezzon 1 , E. Baldan 1 , R. Chersevani 2 , C. di Maggio 1 , M. La Grassa 3 , L. Pescarini 1 , I. Polico 1 , A. Proietti 1 ; 1 Padua/IT, 2 Gorizia/IT, 3 Aviano/ IT (gisella.gennaro@ioveneto.it) Purpose: To evaluate the potential added value of mammography cranio-caudal (CC) view in adjunction to breast tomosynthesis (DBT) compared to standard twoviews digital mammography (FFDM).

The study included 250 consenting women with breast lesions classified as suspicious at mammography and/or ultrasound. They underwent bilateral digital mammography (CC, MLO) and tomosynthesis (MLO). Six readers reviewed the images per-breast in individual reading sessions, each including both FFDM and DBT images; after DBT evaluation, mammography CC-view of the same breast was added and evaluation refined using both modalities. Clinical assessments on 469 breasts from the six readers were compared with the truth and multiple-reader-multiple-case (MRMC) receiver operating characteristics (ROC) analysis performed. Clinical performance of mammography versus tomosynthesis and tomosynthesis plus mammography CC-view were compared using sensitivity, specificity, and area under the ROC curve (AUC). Results: Mean sensitivity was the same with FFDM and DBT (72.8% vs. 73.0%) while increased significantly adding FFDM (CC) to DBT (76.2%, p = 0.004), leading to 2.1 extra-cancers detected, on average. Mean specificity was significantly higher with DBT than FFDM (86.9% vs. 83.1%, p = 0.025), but the adjunction of FFDM (CC) to DBT reduced specificity, even not significantly. AUCs averaged across the six readers gave 0.831 (FFDM), 0.844 (DBT), and 0.852 (DBT+FFDM (CC)), respectively, but no difference between modality pairs was significant. Conclusion: Adjunction of mammography CC-view to tomosynthesis MLO increased mean DBT sensitivity, with no significant effect on specificity. The overall increase in clinical performance (AUC) due to the combination of DBT and FFDM was not statistically significant.

Second breast cancer size and survival following treatment for primary G.S. MacLennan 1 , G. Lawrence 2 , C. Boachie 1 , S.D. Heys 1 , F.J. Gilbert 1 ; 1 Aberdeen/UK, 2 Birmingham/UK (f.j.gilbert@abdn.ac.uk) Purpose: Currently there is uncertainty surrounding the role of surveillance mammography following primary breast cancer treatment. If in women with ipsilateral breast tumour recurrence (IBTR) or metachronous contralateral breast cancer (MCBC) smaller subsequent cancers are associated with improved survival then one could infer potential benefit from surveillance mammography. Methods and Materials: Survival in women with second cancers diagnosed between 1 st Jan 1990 and 3 1 st Jan 2007 from the West Midlands Cancer Intelligence Unit Breast Cancer Registry were analysed using Cox regression models. Outcomes were time from diagnosis of second cancer to all cause death and to breast cancer death. Estimates are hazard ratios (HR) and 95% confidence intervals. Risk factors included were prognostic factors of the primary tumour, size of the second tumour and time to second tumour. Results: There were 613/1174 all cause deaths after IBTR. Poorer prognosis was associated with second cancers > 2 cm compared with < 1 cm, HR 1.75 (1.29, 2.37), and second cancers within 60 months HR 1.66 (1.35, 2.04). After MCBC there were 358/975 all cause deaths and risk factors were similar; HR 2.14 (1.49, 3.06) for size and HR 1.29 (1.01, 1.63) within 60 months. Conclusion: In women with second cancers those with tumour size > 2 cm in diameter have a significantly greater risk of death. Also, time to the second cancer is important so frequency of surveillance mammography should be considered for maximum benefit. Caution is required interpreting these results as lead time bias from mode of detection may be a factor.

Density is in the eye of the beholder: visual versus semi-automated assessment of breast density on standard mammograms M. Lobbes, J. Cleutjens, V. Lima Passos, C. Frotscher, C. Boetes; Maastricht/NL (marclobbes@planet.nl) Purpose: BI-RADS guidelines advise assessment of breast density as part of the mammographic evaluation. In general, visual inspection is used to assess breast density in everyday clinical practice. Comparison of visual versus semi-automated analysis of breast density was never evaluated. Methods and Materials: Breast density was assessed by an experienced and an inexperienced reader in 200 mammograms. Breast density was also assessed by dedicated software using a thresholding technique. Breast density was scored according to the BI-RADS classification system. The correlation between breast density classification of both readers, as well as the correlation between their assessment versus the semi-automated analysis, was expressed as the weighted κ value. Results: Breast density measurements of the left and right breast and in both projections were excellent and highly significant (intraclass correlation coefficients (ICC) of > 0.9, p < 0.0001). There was frequent disagreement between the BI-RADS categorisation of the experienced and unexperienced reader (42% of the cases). The experienced reader correctly classified the BI-RADS classification in 58.5% of the cases. The classification was overestimated in 35.5% of the cases, and underestimated in 6.0% of the cases. The correlation between the classification of both readers versus the semi-automated analysis was considered only fair with weighted κ values of 0.367 (experienced reader) and 0.232 (unexperienced reader). Conclusion: Visual assessment of breast density on mammograms is inaccurate and observer-dependent. Semi-automated analysis of breast density can aid in the improvement of inter-observer variation of breast density classification on mammograms.

HighReX: a multi-reader, multi-centre observer study comparing diagnostic accuracy of photon counting tomosynthesis and full field digital mammography M.G. Wallis 1 , E. Moa 2 , F. Zanca 3 , K. Leifland 4 , M. Danielsson 2 ; 1 Cambridge/UK, 2 Solna/SE, 3 Leuven/BE, 4 Stockholm/ SE (matthew.wallis@addenbrookes.nhs.uk) Purpose: To compare, in an observer study, detection accuracy of full field digital mammography (FFDM) with that of two-view and single view (MLO) photon-counting tomosynthesis (PhCT). Methods and Materials: Sequential cases were collected from screening assessment and referral clinics in two European centres. In addition to standard management (Sectra MicroDose Mammography D40, GE Senograph DS and Essential) all women with breast composition BIRADS 2 or more underwent an examination on a prototype PhCT system, a modified Sectra MicroDose Mammography (L30), tomo angle 11 S178 the current presentation will focus on the radiologic technique.

A retrospective data analysis was undertaken for a total of 65 patients with a T1-2 oral cavity cancer, who were seen at a tertiary referral centre between 2004 and 2010. The correspondence between tumour thickness measured by ultrasonography and histopathology was assessed by Pearson's correlation coefficient, and also between tumour thickness and the development of neck metastasis. Results: In 11 cases,intra-oral measurement was not optimal due to limited mouth opening (n=2) or impossibility to depict the lesion (n=9). Tumour thickness measured by US correlated well with histopathology (n=23, R=0.93). Tumour thickness of ≤7 mm carries a risk of lymph node metastasis of 12%, whereas in tumours exceeding 7 mm this risk is 57% (p=0.001). Twenty-five per cent developed neck metastasis and 19% had local recurrence. Conclusion: Tumour thickness is an important predictive marker for lymph node metastases. As such, it can help in decision-making with regard to management of the primary tumour and neck. Intraoral ultrasound is feasible to measure tumor thickness, especially in tongue tumours. Based upon our findings, a wait-and-see policy is only warranted for superficial lesions with tumour thickness of less than 7 mm, but only if regular follow-up using US-guided aspiration of the neck is ensured. Purpose: To evaluate an innovative CAD application for automated cherry-picking of definitely normal cases and obvious cancers for computer-only reading.

A CAD algorithm assigns each suspicious finding a certainty level of malignancy, normalised from 0-100. Only candidates above a selected certainty are displayed. When operated at low certainty (many false CAD marks but extremely high sensitivity), almost all cases with no marks should be normal and bypass review by a radiologist. Operating at high certainty (almost no false marks but lower sensitivity) and searching for cases with matching CAD marks of the same type in both views, should identify obvious cancers for automatic recall. When run at high and low certainties, the algorithm functions as "extreme-CAD", enabling automatic cherry-picking of both definitely normal cases and obvious cancers. 15892 mammograms (280 malignant) were run at various certainty levels with a prototype CAD algorithm (Siemens) to investigate the feasibility of developing "extreme-CAD". Results: Operating at "10" certainty, yielded 1411 normal cases and 1 malignant case with no CAD marks, allowing "Extreme-CAD" to automatically bypass review of 9% of the normal cases with only 0.36% missed cancers. When subsequently operated at "100" certainty, 17.1% of the malignant cases had matching CAD marks allowing automatic recall with only 0.06% recall of normal cases. Conclusion: "Extreme-CAD" could be used to cherry-pick for computer-only reading 9.3% of the cases, including definitely normal cases and obvious cancers, with a negligible number of missed cancers. The remaining cases are interpreted by a radiologist.

Does the application of CAD prior to planned stereotactic biopsy prevent unneccesary biopsies of B2 lesions? An analysis of screening-selected cases A. Malich, A. Kott, M. Klima, C. Roth; Nordhausen/DE (ansgar.malich@shk-ndh.de) Purpose: Despite independent double reading and preinterventional consensus conference, the number of lesions recommended for stereotactic biopsy due to suspicious microcalcification-clusters finally proven as B2 is still high. This study aimed to verify whether preinterventional application of CAD allows a reduction of unnecessary biopsies. Methods and Materials: All screening-selected stereotactic biopsies (9G) performed at our hospital in the previous year were retrospectively analysed by CAD B A S179 C D E F G H endoscopy, followed by MRI. The MRI images were evaluated for invasion of paraglottic, preepiglotic space, cartilage invasion, extension to extralaryngeal tissue and lymph nodes, and were compared with surgical findings. Results: In 15 patients (44%) the tumour was glottic, in 10 patients (29%) supraglottic and in 9 (15%) patients transglottic. Ten patients (29%) had subglottic extension. No statististical difference was seen in number of patients with positive MRI findings in comparisson with surgery for extension to subglottic subside (29% vs 38%), growth to the anterior commisure (48% vs 48%), preepiglottic space (33% vs 29%) and cartilage infiltration (19% vs 14%), respectively. However, infiltration of the paraglottic spaces was observed more frequently on MRI (81% vs 52%), p < 0.05. Twelve tumor specimens were classified as T2 (36%), 18 as T3 (53%) and 4 as T4 (11%). According to our MR findings, 4 (12%) patients had T2, 26 (76%) had T3 and 4 (12%) patients had T4 staging. The MRI staging was correctly classified in 77% patients and the overall correlation was statistically significant (ρ = 0.59, < 0.05). Fifteen patients had metastases in regional lymph nodes and only two were detected on MRI. Conclusion: MRI has been shown to be a reliable method for presurgical assessment of patients with tumour of the larynx.

MRI with surface microcoils in the follow-up of laser resections for glottic tumours: initial experience D. Farina, P. Rizzardi, M. Ravanelli, R. Maroldi; Brescia/IT (jam.81@virgilio.it) Purpose: To assess the effectiveness of MRI with surface micro-coils in the follow-up of glottic tumours treated by endoscopic laser surgery. Methods and Materials: Nineteen patients treated for primary glottic tumour (5 T1, 9 T2, 5 T3) by endoscopic laser resection underwent MRI with surface coils (two 4 cm loop coils). TSE T2 sequences were applied on axial and coronal planes. Signal/noise (SNR) was calculated for lateral crico-arythenoid muscle (LCAM), scars and recurrences. Contrast/noise (CNR) between LCAM and scar and between LCAM and recurrence was also calculated. Results: Submucosal recurrences were proven in 9/19 patients, 3/9 cases completely negative on endoscopy. MRI detected recurrence in 8/9 cases. In one case imaging findings were considered doubtful and the patient was addressed to biopsy. No false negative and false positive were seen. Mean SNR of submucosal recurrences was significantly higer than those of LCAMs and scars (10.23; 5.39 and 3.77, respectively, on axial TSE-T2 images; 22.33; 7.69 and 3.26, respectively, on coronal TSE-T2 images). Mean CNR between LCAMs and scars was significantly higer than CNR calculated between LCAMs and recurrences (2.43 vs -4.32 on axial images; 2.37 vs -8.26 on coronal images). Conclusion: MRI with surface microcoils is effective in the detection of submucosal recurrences after laser larynx surgery even if endoscopy is completely negative. TSE-T2 sequences enable differential diagnosis between scar tissue and recurrent tumours on both axial and coronal planes.

Evaluation of role of ultrasonography and Doppler in assessment of cervical lymphadenopthy H.V. Vadvala, N. Patel, A. Arya, S. Shah, H. Rajpura; Ahmedabad/IN (harshna_21984@yahoo.co.in) Purpose: To determine the most reliable ultrasonographic parameters for differentiation of benign and malignant cervical lymph nodes with emphasis on intranodal vascularity and spectral indices. Methods and Materials: A prospective study was conducted on 94 patients using high-resolution 7.5 MHz linear transducer. Size, L/T Ratio, echogenic hilum, nodal cortex, nodal margin, necrosis and calcification were evaluated on grey-scale ultrasound. On colour Doppler, type of intranodal vascularity was characterised as central, peripheral, mixed or abnormal flow such as displacement, aberrant vessel and focal absence of perfusion. RI and PI values < 0.6 and < 1.2, and > 0.8 and > 1.5 were considered as specific for benign and malignant lymph nodes, respectively. Diagnosis on each mode was categorised and histopathologically correlated. Results: L/T ratio of < 2, partial or complete absence of hilum with eccentric cortex and heterogeneous echotexture showed a sensitivity of 71.4%, 68.3% and 65.3%, respectively, for malignant nodes. Grey scale, colour Doppler, spectral Doppler and combined grey scale and Doppler showed sensitivity and specificity of (86.2&74.6%), (82.1&90%), (82.2&94.2%) and (93&88.3%), respectively, for diagnosis of malignant nodes. Sensitivity and specificity for cut off values of RI > 0.676 and PI > 1.2 for malignant nodes were 81.5% and 88.75%, respectively. Conclusion: The L/T ratio, hilum and echogenecity were the critical indicators for diagnosing malignant nodes on grey-scale sonography. The distribution of intranodal vascularity was found more accurate than spectral indices in differentiating cervical lymphadenopathy. Addition of colour Doppler to conventional ultrasonography improved the diagnostic accuracy for characterisation of lymph nodes.

In sane volunteers, single anatomical layers were differentiated in all cases: cheek mucosal, submucosal and muscular layers; tongue mucosal and muscular layers; hard palate mucosa and periosteum. In neoplastic patients, tumour sizes were correctly evaluated in all cases, while depth of infiltration was correctly determined in 8 out of 9 cases, with sensitivity and diagnostic accuracy values of 89%. Conclusion: Sonography examination seems to be a promising tool for the study of oral cavity tumours because it allows to demonstrate the tumour depth of infiltration.

Dual energy CT of head and neck cancer: average weighting of low and high voltage raw data improves lesion delineation and image quality J.M. Kerl, A. Tawfik, R.W. Bauer, T.J. Vogl, M.G. Mack; Frankfurt a. Main/DE (matthias.kerl@kgu.de) Purpose: To compare between five dual energy image data sets pure 80 kv, 140 kv and average images with three different weighting factors (WF) for the detectability of head and neck squamous cell carcinoma (HNSCC). Methods and Materials: 35 patients with HNSCC who underwent a contrast-enhanced dual-source scan in dual energy mode (Siemens Flash) included in this study. Images were reconstructed as pure 80 kv, 140 kv, and average weighted image data sets in 3 different WF (0.3, 0.6, 0.8) incorporating (30%, 60%, 80% from the 80 kv tube, respectively). Vascular enhancement, objective noise, lesion and muscle attenuation and CNR were recorded in the 5 image data sets. Three radiologists in consensus reviewed all images and rated the image quality and lesion detectability on a 5-point scale. Results: Tumour enhancement and muscle attenuation increased with decreasing tube voltage from 140 kVp through 80 kVp, but the total increase in muscle attenuation from 140 kVp through 80 kVp (21 HU, 36%) was small compared to that of the tumour (74 HU, 97%), resulting in an increase in tumour-muscle difference. CNR was significantly higher at WF 0.6 than at 0.3, and 0.3 than at 140 kV (p < 0.0001). In addition, WF 0.6 showed the best image quality, followed by 0.8, 0.3 and 80 kV dataset. Conclusion: Mixing the dual energy raw data from the 80 kv and 140 kv tubes in WF 0.6 results in a high CNR and the best subjective overall image quality and lesion detectability.

Clinical impact of FDG PET-CT on management decisions in patients with head and neck tumours P. Bhatnagar 1 , J. Martin 2 , A. Goldstone 1 , F.U. Chowdhury 1 , K. Dyker 1 , M. Sen 1 , A.F. Scarsbrook 1 ; 1 Leeds/UK, 2 Bradford/UK (drpriyabhatnagar@gmail.com) Purpose: To review the impact of FDG PET-CT in the management pathway of patients with head and neck tumours. Methods and Materials: Consecutive patients with primary head and neck tumours undergoing FDG PET-CT between April 2007 and December 2009 were retrospectively analysed. Scan indication, findings on PET-CT compared to conventional imaging and impact on subsequent patient management were evaluated. Clinical impact was divided into: major: detection of occult disease or characterisation of indeterminate lesion (s) on conventional imaging; minor: confirmation of suspected metastases seen on conventional imaging; no impact. Results: 77 patients underwent 92 PET-CT scans during the study period (mean age 57, range 14 to 83 years) including 38 patients with oro-pharyngeal carcinoma, 5 with naso-pharyngeal carcinoma, 25 with squamous carcinoma in cervical nodes but unknown primary and 9 with other primary sites in the head and neck. 56 scans were performed for staging, 36 scans were for restaging or assessment of suspected disease recurrence. In 31 cases (33%) PET-CT had a major impact on subsequent patient management. 6 patients had unsuspected nodal involvement, 5 patients had occult primary tumours and 2 had unsuspected metastatic disease. In a further 18 patients, PET-CT clarified indeterminate lesions as either malignant (11 patients) or benign (7 patients). PET-CT had a minor impact on 41 patients (46%) and no impact on 20 cases (21%). Conclusion: FDG PET-CT has a valuable role in the management of patients with head and neck tumours by improving staging accuracy and helping characterise indeterminate lesions. Purpose: To assess regional iron distribution by MRI-R2* within the heart of patients with β-thalassemia major ( TM ) and other iron overload diseases. Methods and Materials: Breathhold ECG gated MRI (1.5 T) of the heart was used for the measurement of transverse relaxation rates R2* in 32 patients (11-79 y). In a mid-papillary short-axis slice divided into septal, anterior, lateral, and posterior quadrants, R2* was analysed from ROI-based signal intensities from 12 echo times (TE = 1.3-26 ms). Typical boundary effects were evaluated in detail.

The segmentation of the cardiac wall resulted in highly significant correlations of R2* between septal and all other quadrants. In the patient group with R2* < 50 s-1 (normal), all quadrants show significantly higher median rates (126 ± 25%, 150 ± 24%, 174 ± 51%) than the septum (p < 10-4). Typical boundary effects on segmental R2* from blood, lung tissue, epicardial fat and hepatic iron could not be easily separated from segmental iron distribution.

The measurement of MRI-R2* in the interventricular septum is representative for the cardiac wall and the least affected method by boundary effects to detect patients with iron overload at risk of developing heart failure. AT (julia.furtner@meduniwien.ac.at) Purpose: To determine whether there is an accordance between the region of the most malignant part of the tumour related to metabolic changes as demonstrated by multivoxel MR-spectroscopy and the region of the highest tumour perfusion using arterial spin labelling (ASL). Methods and Materials: 45 patients with primary brain tumours, histologically classified as low-grade (n = 16) or high-grade (n = 29) gliomas according to the WHO brain tumour classification were included. A 3 Tesla MR scanner was used to perform multivoxel MR-spectroscopy and pulsed-ASL sequences at an inversion time of 614 ms. The two different sequences were coregistrated and the cerebral lesions were assessed for location of the Cho/Naa ratio peak on MR-spectroscopy and highest signal intensity on the ASL sequence. Results: The accordance of the Cho/Naa ratio peak in multivoxel MR-spectroscopy and the region of highest signal intensity in the ASL sequence was 85% in highgrade gliomas and 72.5% in low-grade gliomas. Conclusion: Our findings may suggest that ASL allows to detect the region of the highest tumour vessel density, which shows a high correlation with the metabolically most malignant region of the brain tumour. The combination of these two non-invasive MR sequences is able to define the tumour hot spot without the use of any contrast media. In the clinical setting this would mainly be an advantage for patients with elevated serum creatinine values or known previous allergic reaction associated with contrast media, in which case contrast media application would be a contra indication.

Estimation of water diffusion and tumour micro-circulation in human glioblastomas: repeatability of the appliedtechniques, effect of the b-values, and relationship between the functional parameters S. Bisdas, A. Dimostheni, R. Ritz, T. Nägele, U. Ernemann; Tübingen/ DE (sotirios.bisdas@med.uni-tuebingen.de) Purpose: To assess the repeatability of dynamic contrast-enhanced (DCE)-MRIderived microcirculatory parameters and apparent diffusion coefficient (ADC) calculations in glioblastomas and to examine the association between them taking account the effect of b-value. Methods and Materials: 25 patients with glioblastomas underwent multi-shot echoplanar DWI and DCE-MRI. ADC images sets were generated using 0, 150, 500, 1000, and 2000 s/mm 2 as b-values. Commercially available software using an arterial input function (AIF) was applied for the calculation of the transfer constant between the intravascular and extravascular space (K trans ), extravascular blood volume (v e ), transfer constant from the extracellular extravascular space to plasma (k ep ), and initial area under the signal intensity-time curve (iAUC). The qualitative and quantitative DCE-MRI and ADC assessments were analysed using parametrical and non-parametrical tests. Measurement repeatability was determined by the Bland-Altman method. Results: K trans , v e , and iAUC values in tumour tissue were higher than in peritumoral area (P = 0.02). In the tumour tissue, significant correlations were observed between K trans and v e , K trans and iAUC, iAUC and v e (r = 0.64, P = 0.0014). Further significant correlations were found in the tumour tissue between v e and ADC 0-500-1000 (r = 0.56, P = 0.007), in the peritumoral region between K trans and ADC 150-500-1000-2000 (r = 0.55, P = 0.05) as well as between v e and ADC 0-500-1000 (r = 0.56, P = 0.05). The most satisfactory repeatability was shown for K trans , v e, and for ADC calculated by 4 b-values. Conclusion: DCE-MRI and ADC using a 4-point calculation showed satisfactory repeatability in glioblastomas. The demonstrated correlations imply that the choice of the b-values is crucial for the association of ADC and DCE-MRI parameters.

Orbital lymphoma: imaging findings at diagnosis and follow-up G. Priego, C. Majos, F. Climent, A. Muntane, N. Romero, S. Mercadal; Barcelona/ES (gemika@gmail.com) Purpose: Imaging findings in orbital lymphoma, at diagnosis and follow-up, allow a proper radiological assessment, and hence an appropriate staging. Methods and Materials: We included 44 patients (mean age 61 years, 20 men and 24 women) with orbital lymphoma studied by imaging (CT or MRI) between 1996 and 2010. Twenty three cases were collected with radiological exams at diagnosis. Radiological data were analysed including orbital distribution, quadrants involved and infiltration of orbital structures. In the follow-up, 39 cases were collected and radiological findings were reviewed regarding treatment response. P <.0001). Mean NADC ratio in the malignant group (0.93 ± 0.18) was also significantly lower than the benign group (1.29 ± 0.12; P <.0001). The mean maximal rCBV values of benign and malignant meningiomas were 7.18±4.12 and 6.01±3.89, in the parenchyma, and 1.07±0.95 and 3.85±1.41 in the peritumoral edema. The mean rMTE values were 1.14±0.23 and 1.24±0.27, in the parenchyma, and 0.88±0.24 and 1.21±0.31, in the peritumoral oedema. The defferences in rCBV and rMTE values between benign and malignant meningiomas were not significant (P > 0.05) in the parenchyma, but both were significant (p < 0. 05) in the peritumoral oedema. Conclusion: Diffusion and perfusion MR imaging are useful in the preoperative differentiation between benign and malignant meningiomas.

Contribution of diffusion and perfusion weighted magnetic resonance imaging in the differentiation of meningiomas and other extra-axial tumours A. Zimny, J. Bladowska, P. Szewczyk, J. Wagel, M. Sasiadek; Wroclaw/PL (abernac@wp.pl) Purpose: Different extraaxial tumours often show very similar appearance in standard MRI; there are many tumours that can mimic meningiomas, e.g. dural metastases or schwannomas. The aim of our study was to evaluate usefulness of diffusion (DWI) and perfusion weighted MR imaging (PWI) in the differential diagnosis of extraaxial tumours. Methods and Material: Forty-six extraaxial tumours (32 meningiomas including one located intraventricularly, 11 schwannomas, 2 dural metastases, and 1 choroid plexus papilloma) underwent standard MR imaging as well as DWI and PWI using 1.5 Tesla MR unit. In each tumour min. ADC and max. rCBV measurements were performed as well as analyses of perfusion curves with evaluation of the rate of their return to the baseline (return coefficient -RC). Results: The values of min. ADC, max. rCBV and RC for different tumour types were as follows: meningiomas (0.79; 8.61; 0.21), schwannomas (1.2; 0.7; 1.8), dural metastases (0.8; 1.2; 1.0), choroid plexus papilloma (1.1; 2.0; -0.6), respectively. There were significant differences between meningiomas and all other tumour types in perfusion parameters and between meningiomas and schwannomas also in min. ADC values. There were no significant differences in min. ADC between meningiomas and dural metastases or between intraventricular meningioma (ADC = 0.9) and choroid plexus papilloma. Conclusion: Though several of the evaluated extraaxial tumours showed identical appearance in standard MRI, they differed significantly in PWI. PWI and less significantly DWI seem to be very useful in distinguishing different dural tumours, neoplasms located in the cerebello-pontine angle or intraventricularly.

Comparison of contrast-enhanced susceptibility-weighted imaging with contrast-enhanced T1-weighted imaging at 3 Tesla in patients with newly diagnosed Glioblastoma A. Radbruch, P. Bäumer, B. Wiestler, B. Stieltjes, L. Kramp, C. Hartmann, W. Wick, S. Heiland, M. Bendszus; Heidelberg/DE (alexander-radbruch@gmx.de) Purpose: Susceptibility-weighted imaging (SWI) is a promising tool for evaluating brain tumours by assessing microbleeds/intratumoural susceptibility signals (ITSS). The purpose of this study was to determine the added value of contrast-enhanced susceptibility-weighted imaging (CE-SWI) to visualise enhancement in CE-SWI in addition to ITSSs, and to compare both parameters with contrast-enhanced T1 Images (CE-T1). Methods and Materials: CE-SWI (TR 26, TE 19.2; flip-angle 15°) and CE-T1 (MPRAGE TI 1100, TR 1710; TE 4.0 flip-angle 15°) images were acquired from 40 patients with newly diagnosed and histologically confirmed glioblastoma, before surgery at a 3-Tesla MRI. The images were overlaid on the slice demonstrating the largest tumour area. Two experienced neuroradiologists a) measured the area of the enhancement and calculated percentage-wise overlap of the enhancement in both sequences and b) quantified the number of ITSSs and assessed their overlap in both sequences. Results: All 40 patients presented a nearly identical enhancement in CE-SWI and CE-T1 sequences with a percentage-wise overlap of the enhancement of 92.2 ±4.6 %. ITSSs were detected in 36/40 patients on CE-SWI and in 15/40 on CE-T1. There was a significant correlation between tumour area and number of ITSSs (R2 = 0.80). Only in 15.0 ±12.4% ITSS that were detected on CE-SWI could also be detected on CE-T1. Conclusion: CE-SWI images provide additional information regarding the appearance of ITSSs that often cannot be visualised in CE-T1 images, whereas the enhancement in both sequences is nearly equivalent. Therefore, CE-SWI could be a useful sequence, e.g. for biopsy-planning, to avoid taking the biopsy from an ITSS. Purpose: To investigate CT perfusion paremeter changes in a rat brain C6 glioma model using CT perfusion imaging, and analyse correlations between CT perfusion paremeter and CD105. Methods and Materials: A total 20 male SD rats were divided into glioma and control groups (normal rats) at random. Rats in glioma group were implanted C6 glioma cells at right caudate nucleus of the brain. After 3 weeks, perfusion CT was performed in glioma and control groups. Perfusion CT values (CBV, CBF, PS and MTT) were measured at the right nucleus caudatus in rat brain. The samples at nucleus caudatus in rat brain were examined histologically using HE and immunohistochemical staining for CD105. The data were analysed using SPSS11.5 statistical software. Spearman correlation coefficients were used to investigate relationships between perfusion CT parameters (CBV, CBF, and PS) and immunohistochemical staining for CD105. Analysing the first control post-treatment, the following findings were observed: complete response in 36% (14/39), major response in 59% (23/39) and minor response (including growth) in 5% (2/39). Controls every 12 months in average showed stability of the images (or even decrease) for 2 years (range 1-5 years). Conclusion: Lymphoma can involve the orbit, being the most common the superiorexternal quadrant and the orbital structures described. In follow-up there may be residual images that remain stable over time. However, recurrence is a possibility and follow-up by imaging is required.

Volumetric increase in Schwannoma on surveillance brain MRIs following stereotactic radiosurgery: showed no elevation of Cho or Lip/Lac compared with the normal brain tissue, but reduction of NAA was observed in 86% of cases. Conclusion: Choline is frequently elevated in relatively high proportion of patients with intracranial tuberculomas; knowledge of possible choline elevation can avoid the misinterpretation of neoplasms. Awareness of the fact that a significant proportion of tuberculomas show Choline (Cho) elevation on MRS is important in clinical imaging interpretation of brain masses and mass-like lesions.

A S185 C D E F G H was measured up to 480 min after bolus injection of 25 μM/kg at 1.5 Tesla using a T1-weighted GRE-flash-3D sequence. Pharmacokinetics in serum and renal and faecal excretion of gadoxetate were quantified using an LC-MS/MS-assay. Results: Peak signal intensity was reached significantly earlier (T max ) in subjects with variant OATP1B1 alleles (*1a/*1a, 66 ± 30 min; *1b/*1b, 43 ± 21 min; *15/*15-*1a/*5, 35 ± 9 min, p < 0.05). Hepatic enhancement (AUC 3-¥ in arbitrary units) was significantly lower in carriers of *15 or *5 (28 ± 8.8, p < 0.05) compared with *1a/*1a wild-type (42 ± 8.8) and *1b/*1b (40 ± 15). None of the genetic polymorphisms of OATP1B1, OATP1B3, or ABCC2 had any significant effect on systemic exposure to gadoxetate. Conclusion: OATP1B1 is a major factor influencing hepatic uptake of gadoxetate in humans. Hepatic enhancement is reduced in subjects with deficient OATP1B1 function. OATP1B3 and ABCC2 polymorphisms have no clinical relevance in gadoxetate-enhanced liver-MRI.

Gd-EOB-DTPA-enhanced MR imaging (MRI) in the early dynamic phase: is bolus timing necessary to obtain an optimal arterial phase for liver MRI? C. Schmid-Tannwald, K.A. Herrmann, M.F. Reiser, C.J. Zech; Munich/ DE (christine.schmid-tannwald@med.uni-muenchen.de) Purpose: To evaluate the necessity to perform bolus timing for the early dynamic phase in Gd-EOB-DTPA-enhanced liver MRI. Methods and Materials: Three groups of patients which were referred to a Gd-EOB-DTPA-enhanced liver MRI were retrospectively evaluated. In group 1 a CareBolus technique (n = 20), in group 2 a test bolus technique (n = 18), in group 3 no bolus timing with a fixed delay of 20 seconds (n = 18) was used to determine the start of the arterial-dominant phase after bolus injection. Two readers independently evaluated the arterial phase with five criteria (proper enhancement in aorta; in liver arteries, slight enhancement in portal vein, missing enhancement in hepatic veins, cortical enhancement of kidney). If a criterion was fulfilled, one point was assigned. Sums per patient were calculated. Significance was tested with Wilcoxon-test. Results: The mean of the sum scores for the CareBolus technique and the test bolus technique were similar with 9.1 and 9.0, respectively, and significantly higher than the mean for the fixed delay with 6.2 (p = 0.0013 and 0.0028, respectively). The most critical parameter was the enhancement in the portal vein, which was correctly hit in 12/20 cases with CareBolus, 11/18 cases with test bolus and only in 3/18 cases for the fixed delay.

The results indicate that for Gd-EOB-DTPA-enhanced liver MRI a bolus-timing technique is mandatory. The differences between CareBolus and test bolus technique are negligible.

Clinical and pathological effects of gadodiamide on a new, adeninesensitised, rat model of nephrogenic systemic fibrosis N. Bouzian 1 , N. Fretellier 1 , P. Bruneval 2 , F. Daubiné 3 , J.-M. Idée 1 , C. Corot 1 ; 1 Roissy/FR, 2 Paris/FR, 3 Nantes/FR (ideej@guerbet-group.com) Purpose: To set up a new rat model of nephrogenic systemic fibrosis (NSF) and investigate the role of baseline renal function. Methods and Materials: Wistar rats received a 0.75% adenine-enriched diet for either 4 weeks (characterisation study vs. control rats), 16 days (Study 1) or 8 days (Study 2). Rats received daily injections of gadodiamide (Study 1: 5 x 2.5 mmol/ kg/day, Day 14 to Day 18, Study 2: 5 x 1.0 or 2.5 mmol/kg/day, Day 7 to Day 11). The model was characterised from a clinical, pathological (femur and kidneys) and biochemical standpoint up to 3 weeks after completion of the adenine treatment. In studies 1 and 2, adenine-treated rats were sacrificed 3 weeks after the last injection of gadodiamide or saline. Skin biopsies were performed on 1 st and 9 th days of treatment and at sacrifice. All histopathological studies were blinded. Results: Adenine induced major tubular lesions with lithiasis and interstitial fibrosis. Fibrous osteitis-like lesions occurred in both femoral epiphyses and diaphyses. Early decrease in creatinine clearance (Day 8: 0.12 ± 0.03 vs. 1.2 ± 0.3 mL/mn/100 g, p < 0.01), hyperphosphoraemia (peak at Day 22: 4.1 ± 1.0 vs. 2.4 ± 0.2 mmol/L), hypocalcaemia (2.0 ± 0.2 vs. 2.3 ± 0.1 mmol/L) and gradual mortality were found. Gadodiamide-induced mortality (Study 1: 10/11, Study 2: 1 death at 2.5 mmol/kg/ day) and skin lesions (dermal fibrosis, calcification, inflammation) were related to the baseline renal function. Conclusion: Gadodiamide-induced cutaneous and systemic toxicity depends on baseline renal function in an adenine-sensitised model of NSF.

Twenty-eight liver transplant patients with ischaemic changes of the biliary tree underwent MRI at 1.5 T-device. After acquisition of T1w/T2w images, conventional MRC was performed through thin-slab 3D FRFSE T2w and thick-slab SSFSE T2w sequences. In each patient a 3D fat-suppressed breath-hold T1w sequence (TR/TE, 3.6 ms/1.6 ms; 2.4 mm thk/-1.8 mmsp; matrix, 224 x 192; one NEX) was performed before and 15, 20, 25, 30, 35, 40 minutes after intravenous administration of gadolinium-EOB-DTPA (Primovist®, Bayer Schering Pharma; 10 ml) at a flow-rate of 2 ml/s, followed by injection of isotonic saline (20 ml). T1w sequence was also obtained after 90 minutes in 9/28 cases. Depiction of intra-/extra-hepatic bile ducts, dilation of biliary system, presence of biliary stricture and stone/sludge were evaluated on images by consensus reading of two observers using a 3-point-scale: 0, absent; 1, visible; 2, excellent. Imaging results were correlated with direct cholangiography and/or follow-up. Results: The grading of visualisation and depiction of dilation of the intra-hepatic bile ducts was significantly superior on T2w MRC compared with Gd-EOB-DTPAenhanced T1w MRC; on the contrary, this latter significantly tended to out-perform conventional T2w MRC in delineating strictures involving hepatic bifurcation and extrahepatic donor bile duct (p < 0.0001). Sludge/stone formation was correctly identified with both techniques. Biliary enhancement was recognised within 20 minutes post-Gd-EOB-DTPA injection in 6 subjects, between 25 and 40 minutes in 16, and at 90 minutes in 6 cases. Conclusion: Gd-EOB-DTPA-enhanced MRC may provide both anatomical and functional information of ITBLs in liver transplants.

Multi-center, randomised comparison study to eVALUatE outcomes and resource needs of imaging and treatment following Primovist-enhanced MRI of the liver in comparison to extracellular contrast media-enhanced MRI and contrast-enhanced CT in patients with colorectal cancer and suspected metachronous liver metastases: an interim analysis C.J. Zech 1 , S. Gschwend 2 , V. STUDY Group 2 ; 1 Munich/DE, 2 Berlin/ DE (Christoph.Zech@med.uni-muenchen.de) Purpose: To perform an economic evaluation of Gd-EOB-DTPA-enhanced MRI (PV-MRI) compared to MRI with extracellular-contrast-media (ECCM-MRI) and contrast-enhanced MDCT (CE-MDCT) as initial diagnostic modalities in the workup of patients with colorectal liver metastases. Methods and Materials: 35 study centres included 296 patients from 10/2008 to 07/2010, which were referred for staging of the liver in case of suspected liver metastases. The initial imaging study was randomised. A consensus between the investigating surgeon and radiologist was reached whether further imaging was required to obtain a therapeutic decision. The primary efficacy variable was the proportion of patients for whom further imaging was required after initial imaging with PV-MRI as compared to ECCM-MRI and CE-MDCT. The results were compared using a one-sided approximate test for equality at a significance level of 2.5%. The diagnostic confidence was rated by the investigator on a four-point-scale. Results: 98 patients received PV-MRI, 95 ECCM-MRI and 91 CE-MDCT. Further imaging was required in 0 (0%), 15 (15.8%) and 33 cases (36.3%), respectively. These results are statistically significant. Diagnostic confidence was high/very high in 96 (98%), 84 (88.4%) and 51 cases (67.1%), respectively.

The results show superiority of PV-MRI over CE-MDCT and ECCM-MRI with regard to the requirement for further imaging. With PV-MRI as the initial staging strategy, no further tests to assess operability were needed, which has implications for work-flow and costs. Further evaluation of the collected data with regard to diagnostic efficacy of each method compared to surgery and an cost calculation is planned.

Hepatic uptake and pharmacokinetics of the MRI contrast agent gadoxetate in healthy subjects: role of OATP1B1, OATP1B3 and ABCC2 J.P. Kühn, A. Nassif, S. Oswald, W. Weitschies, W. Siegmund, R. Puls, N. Hosten; Greifswald/DE (kuehn@uni-greifswald.de) Purpose: Liver-MRI with the specific contrast agent gadoxetate improves detection of focal lesions. There is evidence that gadoxetate is a substrate of hepatic organ ion transport proteins (OATPs) and ABCC2. To confirm their role in hepatic uptake and elimination of gadoxetate in humans, we measured hepatic enhancement and pharmacokinetics of the agent in volunteers with functional polymorphisms of OATP1B1, OATP1B3 and ABCC2.

The study was performed in 41 white subjects, among them 8 with wild-type of OATP1B1, 8 carriers of the *1b/*1b haplotype and 12 of the *15/*15 and *1a/*5 haplotype. The remaining subjects were carriers of OATP1B3 G699 A or ABCC2 C-24T and G1249 A. Gadoxetate-induced hepatic enhancement S186 DE (muenzel@roe.med.tum.de) Purpose: To evaluate the effect of heart rate variability and heart rate on intra-scan and inter-scan "stairstep" artefacts in step-and-shoot coronary CT angiography (CCTA) with the use of a wide detector. Methods and Materials: 70 patients underwent CCTA using a step-and-shoot technique with prospective ECG-gating at a 256-slice CT scanner (Brilliance iCT, Philips). Heart rate during the CT scan was registered. Variability of heart rate was estimated by the difference of maximum and minimum heart rate during the scan. Motion artefacts were quantified using a 5-point scale with 1 point indicating no motion artefacts whereas coronary segments rated with 5 points were considered as nondiagnostic. Stairstep artefacts between single CT scan volumes per step were defined as misalignment of continuous cardiac structures quantified by measurement of the off-set on coronal and sagittal CT multiplanar reformations (MPR) of transverse single slices. Results: Mean heart rate during the CT scans was 66 bpm (standard deviation SD: 16.7). Mean variability of the heart rate was 10.7 bpm (SD 17.5) . A significant correlation between heart rate and stairstep artefacts (r = 0.43, p < 0.001) and between heart rate and motion artefacts (r = 0.67, p < 0.001) was found. There was no significant effect of heart rate variability on stairstep artefacts (r = 0.5, p < 0.001) and motion artefacts (r = 0.18, p = 0.154). Conclusion: In contrast to CCTA using narrow CT detectors with low z-coverage heart rate variability has no significant effect on motion and stairstep artefacts using a wide CT detector with high z-coverage. However, a higher heart rate still increases stairstep and motion artefacts.

Dual-source 128-section CT coronary angiography at variable heart rates: comparison of image quality and radiation dose L. Zhang, H.H. Pan, N. Xu, Y.S. Du; Shanghai/CN (zhanglei4302@hotmail.com) Purpose: To determine the effect of heart rate frequency (HRF) and heart rate variability (HRV) on radiation exposure and image quality in patients undergoing DS CT coronary angiography with adaptive ECG pulsing. Methods and Materials: 253 consecutive patients were categorised in three groups: group A: low HRF (≤65 beats/min) with normal to minor HRV (mean interbeat difference (IBD), 0-3), coronary CTA proceeding with prospective ECG-triggered spiral model; group B: intermediate to high HRF (≥66 beats/min) with normal to minor HRV (IBD, 0-3), with prospective ECG-triggered sequence scan; group C: intermediate to high HRF (≥66 beats/min) with moderate to severe HRV (IBD > 4), with retrospective ECG-gated spiral model. Optimal ECG pulsing windows for low (≤65 beats/min), intermediate (66-79 beats/min), and high HR (≥80 beats/min) were at 60%-76%, 30%-77%, and 31%-47% of the R-R interval, respectively. Outside the ECG-pulsing window, the tube current was reduced to 20% of the full current. Results: Coronary CTA yielded good image quality in 87.7% of patients (222/253). No significant differences among three groups with different HRF and HRV in image quality and diagnostic performance were found. Significant difference (P < 0.001) was found among the effective doses of group A (0.602 ± 0.363 mSv), group B (1.253 ± 1.804 mSv) and group C (9.039 ± 5.657 mSv). Conclusion: The higher temporal resolution of dual-source 128 sections CT coronary angiography performed with adaptive ECG pulsing results in preserved diagnostic image quality and performance independent of HRF or HRV at the cost of limited dose reduction in arrhythmic patients.

Incremental value and safety of oral ivabradine for heart rate reduction in computed tomography coronary angiography E. Maffei 1 , C. Martini 1 , T. Arcadi 1 , S. Seitun 1 , A. Clemente 1 , N. Brunetti 2 , F. Cademartiri 1 , A. Guaricci 2 ; 1 Parma/IT, 2 Foggia/IT (filippocademartiri@hotmail.com) Purpose: Heart rate (HR) reduction is essential to achieve optimal image quality and diagnostic accuracy with computed tomography coronary angiography (CTCA). Administration of ivabradine could be an attractive alternative to beta-blockade to reduce HR. Methods and Materials: One-hundred twenty-three patients referred for CTCA were prospectively enrolled. Patients were divided into two groups depending on the absence or presence of chronic beta-blockade treatment. Within the two groups were blindly analysed in central core-laboratories. Significant CAD was defined as ≥50% stenosis. The assessment of MDCT accuracy was performed according to a per-patient, per-vessel and per-segment approach analysis. Results: Two hundred and six patients were excluded from analysis (protocol deviations, MDCT or CA not conducted). The final study population comprised 1048 patients. On a per-patient analysis, sensitivity, specificity, positive and negative predictive value were 83%, 74%, 78% and 79%, respectively. On a per-vessel analysis, they were 62%, 90%, 66% and 88%, respectively, and on a per-segment analysis, they were 38%, 94%, 43% and 93%, respectively. The best sensitivity and worst specificity were observed in heavily calcified segments (Agatston score ≥ 600), as well as in the population at high risk of CAD when compared to patients with low or intermediate probability of CAD. Conclusion: In this large cohort of stable patients in whom an elective CA was scheduled, diagnostic accuracy of MDCT may help establish clinical application guidelines in patients with suspected CAD.

Long-term predictive value of assessment of coronary atherosclerosis by contrast-enhanced coronary computed tomography angiography: metaanalysis and systematic review F. Bamberg 1 , W. Sommer 1 , V. Hoffmann 1 , K. Nikolaou 1 , M.F. Reiser 1 , U. Hoffmann 2 , C. Becker 1 ; 1 Munich/DE, 2 Boston, MA/US (fabian.bamberg@med.uni-muenchen.de) Purpose: We conducted a systematic review and meta-analysis to determine the predictive value of findings of coronary CT angiography for incident cardiovascular events. Methods and Materials: We searched PubMed, EMBASE and the Cochrane Library through January 2010 for studies that followed ≥ 100 individuals for ≥ 1 year and reported at ≥ 1 hazard ratio (HR) of interest. Risk estimates for the presence of significant coronary stenosis (primary endpoint; ≥ 50% diameter stenosis), left main stenosis, each coronary stenosis, 3-vessel disease, any plaque, per coronary segment containing plaque, and non-calcified plaque were derived in random effect regression analysis and causes of heterogeneity were determined in meta-regression analysis. Results: We identified 11 eligible articles including 7,335 participants (59.1±2.6 years, 62.8% males) with suspected CAD. The presence of ≥ 1 significant coronary stenosis (9 studies, 3,670 participants and 252 outcome events [6.8%] with 62% revascularisations) was associated with a tenfold higher risk (HR: 10.74, 95%-CI: 6.37-18.11) with an 11.9% annualised event rate; 6.4% in studies excluding revascularisation (HR: 6.15 [95%-CI: 3.22-11.74]). Adjustment for coronary calcification did not attenuate the prognostic significance (p = 0.79). The estimated HR (95%-CI) for left main stenosis, presence of plaque and each coronary segment containing plaque were 6.64 (2.6-17.3), 4.51 (2.2-9.4), and 1.23 (1.17-1.29), respectively. Conclusion: Presence and extent of CAD on coronary CT angiography are strong, independent predictors of cardiovascular events despite heterogeneity in endpoints, categorisation of CT findings and study population.

Study repetition of nonsatisfactory prospective high pitch acquisitions in cardiac imaging with 128 DSCT: frequency, radiation dose, and meaning G. Tardaguila de la Fuente, C. Delgado Sanchez-Gracián, C. Martínez, D. Castellón Plaza, F. Tardáguila Montero; Vigo/ES (gonzatar@gmail.com) Purpose: Prospective high pitch acquisitions in cardiac imaging allow us to obtain the whole heart in one beat. Low dose studies are performed but when artefact appears we cannot edit the ECG; the solution is to repeat the study. The aim of this study is to determinate how often studies have to be repeated, why they have to be done again and radiation dose. Methods and Materials: 229 patients underwent high pitch cardiac 128-DSCT. High pitch technique was used when heart rate < 65 bpm spontaneously or after the administration of betablockers. Imaging protocol: rotation speed: 0.28 s, table speed: 460 mm/s, pitch: 3.4, collimation: 2 x 128 x 0.6 mm and tube current 320 mAs/ rotation. Tube potential was 100 kV (patients < 80 kg), 120 kV (patients > 80 kg) and 140 kV (stent studies). We analysed dose of the study, number of cases that need to be repeated, cause of the repetition, type of acquisition of the repetition (high pitch, prospectively or retrospectively triggered) and total dose of repeated studies. Results: 40 studies needed to be repeated (17.5%). Causes: 26 studies had to be repeated because of rhythm alterations, 9 due to isolated extrasystoles and 5 for unclear reasons. The second acquisition was another high pitch in 27 patients (68%), in 4 (10%) prospectively triggered scan was obtained and a retrospectively triggered CT was performed in 9 (22%). Radiation dose of non-repeated studies was 1.23 ± 0.49 mSv. Total dose of repeated studies (including the initial acquisition) was 3.98 mSv ± 2.62. Conclusion: Although a relevant number of high pitch cardiac studies (17.5%) had to be repeated, the low dose of these studies makes it worth to choose this type of acquisition when possible. S188 B C A D E F G H patients was reduced from 2 to zero for the inexperienced readers but remained unchanged for experienced readers. Conclusion: Computer-aided detection as a second reader has an overall beneficial impact on the performance of observers of all experience levels for diagnosing significant coronary artery stenosis during cCTA interpretation. In experienced readers benefit the most from the addition of computer-aided detection to their interpretation, while the added value is only modest for more experienced observers.

Correlation between myocardial bridge and myocardial ischemia by dual energy CT S. Ou, Y. Qi, G. Peng; Guangzhou/CN (lhq2008xxh@163.com)

Purpose: The aim of this investigation is to evaluate the detective rate, morphology of myocardial bridge-mural coronary artery (MB-MCA), and the relationship between MB-MCA and myocardial ischaemia by dual energy CT (DECT). Methods and Materials: 252 patients suspected or confirmed ischaemic heart disease (IHD) underwent DECT were prospectively analysed. 

Bronchial artery evaluation in cardiac computed tomography (CCT) patients: is there an association between coronary artery disease (CAD) and bronchial arteries hypertrophy? S. Tresoldi 1 , F. Villa 1 , G. Di Leo 2 , N. Flor 1 , R. Bisogno 1 , G. Cornalba 1 ; 1 Milan/IT, 2 San Donato Milanese/ IT (silvia.3soldi@gmail.com) Purpose: To estimate the correlation between the diameter of bronchial arteries at cardiac computed tomography (CCT) and the degree of coronary artery disease (CAD). Methods and Materials: Two radiologists in consensus retrospectively evaluated CCT images of 100 patients (77 males; mean age 66 ± 11 years; range 34-88 years) with 4 degrees of CAD [gr1 = patent coronary arteries (n = 25), gr2 = significant disease (n = 25), gr3 = stent (n = 25), gr4 = by-pass (n = 25)] and measured the number and diameter of bronchial arteries. Examinations were performed either on a 64-MDCT or on a DSCT after administration of 85-100 ml of contrast media (370-400 mgI/ml) at 5 ml/s followed by a 50-ml saline flush at the same rate. Technical parameters were: 100-120 kVp; 200-700 mAs; retrospective ECG gating with mAs modulation; 64 x 0.6 mm collimation. Lung-windowed images were reviewed to search for diseases possibly related with bronchial artery hypertrophy. Kruskal-Wallis and Chi-square test were used. Results: Patient demographics were similar in the 4 groups (mean age, respectively, 60 ± 12, 68 ± 9, 65 ± 8, 69 ± 9 years). Mean bronchial artery diameter in the 4 groups was, respectively, 1.5 mm, 1.7 mm, 1.6 mm, 1.9 mm with a statistically significant difference (p = 0.008). A significant difference (p = 0.023) was found also in the detection of hypertrophic (> 1.5 mm) bronchial arteries, more represented in group 2 (58%) and 4 (63%) than in 1 (35%) and 3 (50%). There were no significant differences in pulmonary pathology between the four patients' groups. Conclusion: Diameter of bronchial arteries significantly increases with CAD and particularly with surgical revascularisation, suggesting the existence of a physiological shunt between the two systems that can compensate each-other in case of need.

patients were randomised to either no additional premedication or oral ivabradine for 5 days prior to CTCA. In the presence of chronic beta-blockade therapy it was shifted to atenolol 50 mg twice a day for 5 days prior to CTCA. HR and blood pressure were assessed at admission (T0), immediately before CTCA (T1) and during CTCA (T2). The target HR was < 65 bpm. Results: Ivabradine significantly reduced HR during CTCA. Mean relative HR reduction was 15% for controls, 12% for chronic beta-blockade, 19% for ivabradine and 24% for both chronic beta-blockade and ivabradine at T2 (p for trend < 0.001).

The rate of patients who reached the target HR at T2 was 83% in controls, 71% with chronic beta-blockade, 97% with ivabradine and 97% with both (p for trend < 0.05). The percentage of patients that needed additional IV beta-blockade at T1 decreased from 69% to 40% with ivabradine and 30% with both (p for trend < 0.05). Conclusion: Ivabradine is safe and effective in increasing the rate of patients at target HR and in reducing the need for additional IV beta-blockade in patients referred for CTCA.

Individual optimisation of contrast enhancement in coronary CT angiography: a multivariate analysis J.-F. Paul 1 , L. Grandin 2 , J.-P. Daurès 2 ; 1 Le Plessis-Robinson/FR, 2 Montpellier/FR (pauljf@ccml.fr)

Purpose: To optimise contrast enhancement in coronary CT angiography for each patient. Methods and Materials: A first cohort of 58 patients underwent coronary CT angiography using a fixed protocol (73 mL iobitridol 350 mgI/mL, 5.6 mL/s). Influence of various individual parameters (weight, height, body mass index (BMI), body surface area (BSA), cardiac frequency) were studied with regard to contrast enhancement in aortic root. The Pearson's correlation coefficient was used to assess the strengths of associations. Multivariate linear regression analysis with a backward selection process was performed to evaluate the impact of independent factors on aortic enhancement. Piecewise linear regression analysis was used for the factors where the relationship with enhancement was not linear. Then, a second cohort of 61 patients underwent an individually tailored injection protocol based on this model with various injection rates in order to achieve the range of target enhancement within 400-500 HU. Results: From univariate analysis, a significant negative correlation was found with weight, height and BSA. No significant correlation was found with cardiac frequency, but enhancement tends to decrease for heart rate above 75 bpm. From multivariate analysis, the best model for explaining variability included height, weight and cardiac frequency (R 2 = 0.63). In the second cohort of patients using tailored protocol, 85% achieved the target range. Conclusion: Simple physiologic/morphologic individual parameters may explain for more than 60% of aortic enhancement variability in coronary CT angiography. Individually tailored injection protocols may reduce variability of enhancement, and achieve a predefined target enhancement in most cases.

Computer-aided detection of coronary artery stenosis at computed tomography angiography: effect on performance of readers with different experience levels E.M. Arnoldi 1 , U. Schoepf 2 , M. Rosenblum 2 , P.L. Zwerner 2 , R. Goldenberg 3 , M.F. Reiser 1 , C. 

shown that high-resolution black-blood carotid MRI (hr-bb-cMRI) is able to identify complicated atherosclerotic AHA lesion type VI (AHA-LT6) with haemorrhage, thrombus or rupture of the fibrous cap with good correlation to histopathology. The purpose of our study was to evaluate the prevalence of AHA-LT6 in carotid arteries of subjects with cryptogenic stroke. Methods and Materials: 30 consecutive patients (24 men, mean age 69.9±11.9 years) with cryptogenic stroke and eccentric plaque by ultrasound were recruited from our stroke unit. All patients underwent extensive clinical workup (lab, brain MRI, ultrasound, 24-hour ECG, transesophageal echocardiography) to exclude other causes of stroke. All patients received a hr-bb-cMRI at 3.0-Tesla with fat-saturated pre-and post-contrast T1w-, PDw-, T2w-and TOF images using surface coils and parallel imaging techniques. Prevalence of AHA-LT6 was determined in both carotid arteries based on previously published MRI criteria by two experienced reviewers who were blinded to the clinical information. Results: AHA-LT6 with haemorrhage, rupture of the fibrous cap and / or thrombus was found in 14 out of 30 arteries (46.7%) ipsilateral and in 1 out of 30 arteries (3.3%) contralateral to the ischaemic, "cryptogenic" stroke (P < 0.001).

This study suggests that arterio-arterial embolism from non-stenotic carotid atherosclerotic plaques play a role in the pathogenesis in a subset of subjects previously diagnosed with a cryptogenic stroke.

Association between carotid artery plaque volume, composition and ulceration L. Saba, S. Sannia, R. Montisci, R. Sanfilippo, R. Bura, G. Mallarini; Cagliari/IT (lucasaba@tiscali.it) Purpose: Ulcerated carotid plaques are associated with a higher risk of embolism and occlusion, producing ischaemic neurological events. The purpose of this study was to evaluate the relationship between carotid artery plaque volume, composition and presence of ulceration. Methods and Materials: 70 consecutive patients (46 males, 24 females; mean age 63 years; age range 42-81 years), studied by using a multi-detector row CT were retrospectively analysed. Component types of the carotid plaque were defined according the following HU ranges: lipid < 60 HU; fibrous tissue from 60-130 HU; calcification > 130 HU and plaque volumes of each component was calculated.

Each carotid artery was analysed by 2 observers. Statistical analysis, by using ROC analysis, was performed to evaluate association between presence of ulceration and specific components of plaque (and its volume). Results: 14 ulcerated plaque were detected. No significant association was observed between presence of ulceration and total carotid plaque volume (Az = 0.641; p value = 0.39), calcified plaque volume (Az = 0.545; p value = 0.67) and fibrous plaque volume (Az = 0.612; p value = 0.18) whereas a significant statistical association was detected with lipid plaque (Az = 0.784; p value = 0.021). Conclusion: Results of our study suggest that the amount of lipid plaque volume is associated with presence of ulceration; this finding suggests that the lipid component, identified by MDCTA, should be considered a significant determinant of carotid plaque vulnerability

Association between carotid plaque enhancement and micro-vessel density: preliminary results by using MDCTA and histological validation L. Saba, L. Lai, R. Sanfilippo, R. Montisci, G. Faa, G. Mallarini; Cagliari/ IT (lucasaba@tiscali.it) Purpose: Previous studies demonstrated that carotid plaques analysed by MDCTA can show contrast enhancement. The purpose of this preliminary work was to explore the association between carotid plaque enhancement (CPE) and micro-vessel density. Methods and Materials: 21 consecutive (males 15; median age 63) patients studied using a 16-detector row CT scanner were prospectively analysed. Exams were performed before and after contrast medium administration and analysis of plaque enhancement was performed. Patients underwent carotid endarterectomy "en bloc"; histological sections were prepared and the presence of micro-vessel was quantified. Logistic regression analysis as well as ROC curve and area under the curve was calculated. Results: A statistically significant association between the degree of CPE and micro-vessel density (p = 0.009; R2 = 0.32) was observed. The ROC curve analysis confirmed this association with an area under the curve of 0.735 (standard deviation = 0.112) and a p value of 0.0355. Conclusion: Results of this preliminary study indicated that the CPE is associated with the micro-vessel density. Histological analysis demonstrated that the degree of intra-plaque neo-vascularisation is statistically associated with the entity of CPE. 

Association between carotid plaque type and cerebral microbleeds L. Saba 1 , R. Sanfilippo 1 , M. Anzidei 2 , R. Bura 1 , R. Montisci 1 , G. Mallarini 1 ; 1 Cagliari/IT, 2 Rome/IT (lucasaba@tiscali.it)

Purpose: Cerebral microbleeds (CMBs), are being increasingly recognised with the widespread use of MRI techniques that are sensitive to iron deposits The purpose of this work was to explore the association between carotid plaque type and CMBs. Methods and Materials: 43 consecutive (males 31; median age 64) patients were prospectively analysed. Carotid arteries were studied by using a 16-detector row CT scanner, whereas brain was explored with a 1.5 Tesla system. CMBs were studied using a T2*-weighted gradient-recalled echo (GRE) sequence. Microbleeds were classified as absent (grade 1), mild (grade 2; total number of microbleeds, 1-2), moderate (grade 3; total number of micro-bleeds, 3-10) and severe (grade 4; total number of microbleeds, > 10). Chi square, multiple logistic regression analysis as well as ROC was calculated. Results: The prevalence of CMBs was 29.9%. A statistically significant difference was observed between symptomatic (43%) and asymptomatic (15%) patients (p value = 0.0425; OR = 4.36). Liner regression analysis demonstrated an association between the number of CMB and the symptoms (p = 0.022). A statistically significant association was observed between the presence of fatty plaque and CMBs (p = 0.015). Conclusion: Results of this study confirm an increased number of CMBs in symptomatic patients and indicate an association between the presence of carotid artery fatty plaque and CMBs.

Prevalence of complicated carotid atherosclerotic plaques ispilateral to ischaemic cryptogenic stroke using high-resolution MRI Purpose: To evaluate the use of contrast-enhanced ultrasound in the neovascularisation within carotid atherosclerotic plaques. Methods and Materials: 35 patients with known atherosclerotic plaques in the carotid artery were examined with contrast-enhanced ultrasound to role out the features of neovascularisation within this plaque. In addition, these plaques were analysed and correlated with plaque size and echogenicity. For contrast-enhanced ultrasound we injected 2.4 cc of SonoVue (Bracco, Italy) i.v. The examinations were performed using the S 2000 or Sequoia 512 (Siemens/Acuson, Mountain View) with an 9.15 MHz or 17 MHz probe using the CPS-software. Results: There were 41 atherosclerotic plaques, 27 of which (19 soft and 8 mixed) enhanced after injection of SonoVue. The enhancement occurred from the carotid wall to the center of the plaque with a short-line pattern in 15 plaques, whereas 12 plaques enhanced from both the carotid wall and the carotid lumen, with just a little spot pattern. The arrival time of contrast was later in the plaques than in the carotid artery and the time to peak was longer in the plaques than in the carotid lumen. Among the 14 unenhanced plaques, 4 were hard, 3 were calcified, 2 were soft, and 5 were mixed. The unenhanced plaques had a thickness of < 2.7 mm. Conclusion: In our small patient population, contrast-enhanced ultrasound allows the dynamic evaluation of neovascularisation within carotid plaques and neovascularisation may correlate with plaque morphology. Purpose: To evaluate a non-enhanced time-resolved 4D SSFP MRA for dynamic visualisation of intracranial collateral blood flow. Methods and Materials: 22 patients (59.0 ± 11.8 years) with steno-occlusive disease of brain-supplying arteries were included in this study, 4D SSFP MRA of the intracranial arteries was acquired with 15 temporal phases and a temporal resolution of 115 ms on a 1.5 T MR scanner. ECG-gated image acquisition was performed with increasing trigger times following spatially selective and non-selective inversion (FAIR technique) to obtain time-resolved images. The presence of collateral flow B A S191 C D E F G H Conclusion: On the basis of the small patient population, this feasibility study shows that quantitative analysis of perfusion could provide an in vivo early biomarker for prediciting treatment response in patients with HCC lesions.

CT-perfusion in HCC patients: evaluation of quantitative map as early predictor of tumour response to radiofrequency ablation and transarterial chemoembolisation treatment P.A. Bonaffini, D. Ippolito, F. Meloni, D. Leni, R. Corso, S. Sironi; Monza/IT (pa.bonaffini@gmail.com) Purpose: To prospectively investigate the role of CT-perfusion (CT-p)technique in detection of blood flow changes related to therapeutic effects of transarterialchemoembolisation and radiofrequency ablation in patients with HCC. Methods and Materials: Fifty-four patients with known cirrhotic liver disease and biopsy proven diagnosis of HCC lesion, that underwent to TACE or RFA treatment, were prospectively enrolled in our study. Perfusion study of hepatic parenchyma and of treated lesion was performed about 1 month after treatment on a multidetector 16-slice CT (Philips Brilliance,16p,NL). Dynamic-CT study was performed with static table position: 8 dynamic slice/scan; total of 40 scans; intravenous injection of 50 ml of contrast medium. The treated lesion and surrounding parenchyma were evaluated using a dedicated perfusion software which generated a quantitative map of perfusion by means of colour scale. The following perfusion parameters for whole liver and HCC lesions were considered: hepatic perfusion (HP), arterial perfusion (AP), blood volume (BV), hepatic perfusion index (HPI) and time to peak (TTP). Results: Perfusion parameters of treated lesions could be quantitatively assessed using CTp analysis. 16/54 patients had a residual disease and values of perfusion parameters measured within tumour tissue were: HP = 38.5 ± 14.6 ml/sec/100 gr; AP = 41.5 ± 8.9 ml/min; TTP = 19.1 ± 4.3 sec; BV = 18.5 ± 6.3 ml/100 mg; HPI = 57.8 ± 32.3%.The corresponding values obtained in remaining 26 cases in whom a complete necrosis were achieved were: HP = 12.8 ± 6.1 ml/100 gr/ sec; AP = 13.1 ± 7 ml/min; BV = 7.4 ± 9.8 ml/100 gr; HPI = 13.1 ± 8.7% and TTP = 36.2 ± 12.3 sec. A significant difference (p < 0.001) was observed in the mean value of all parameters calculated between treated lesions with residual tumour and those successfully treated. CT-p technique has demonstrated highly perfused areas related to the presence of residual arterial vessels within the viable portion of treated lesions.

Conclusion: This feasibility study shows that quantitative analysis of perfusion could provide an in vivo early biomarker for prediciting the treatment response in patients with HCC lesions.

Role of perfusion CT in evaluation of early acute pancreatitis and correlation with clinical outcome M. Kang, A. Sagar, D. Bhasin, R. Gupta, A. Bhalla, S.S. Rana, N. Khandelwal; Chandigarh/IN (mandykang2005@yahoo.com) Purpose: To evaluate efficacy of CT perfusion for detection of early ischaemic changes in patients with acute pancreatitis and to assess the correlation of CT perfusion measurements with the clinical outcome. Methods and Materials: 39 consecutive patients with clinical diagnosis of acute pancreatitis presenting within 72 hours of symptom onset were enrolled in the study. A control group of 15 patients was also enrolled. Eleven perfusion parameters were evaluated for both the groups: EFP (extraction fraction product, FP (blood flow), VP (blood volume), PS (permeability surface product), KEP (outflow rate), VE (leakage rate into extravascular space), E (extraction measure), MTT (mean transit time), DT (delay time), PEI (peak enhancement intensity) and TTP (time to peak). The severity of disease was assessed using CTSI (Balthazar) and modified CTSI (Mortele) scores. The following parameters of clinical outcome were recorded: duration of hospital stay, need for intervention, development of infection, organ failure and mortality. Results: EFP, FP, VP, KEP and PEI were significantly lower in the study group as compared with the control group. VE, TTP and E were higher in the study group. EFP, FP, VP, KEP and PEI correlated well with regard to need for percutaneous intervention. FP, VP, KEP and PEI correlated well with duration of hospital stay. EFP correlated well with occurrence of infection. Between mild and severe pancreatitis groups the parameters EFP,FP,VP, KEP,VE,E,PEI and TTP were statistically significant. Conclusion: Our data suggest that pancreatic perfusion measurement using MDCT could help in early assessment of severity and predict clinical outcome of acute pancreatitis. Purpose: To evaluate the feasibility of CT-perfusion for the prediction of early response to chemotherapy in advanced gastric or colon cancer patients with liver metastasis using FDG-PET as a standard of reference. Methods and Materials: Colonic (n = 18) or gastric (n = 2) cancers patients with liver metastasis underwent perfusion-CT and FDG-PET before and after 1 cycles of chemotherapy. CT-perfusion parameters (MIP), blood flow (BF), blood volume, permeability (PMB), arterial liver perfusion (ALP), portal liver perfusion, hepatic perfusion index for the largest hepatic metastasis were obtained. PET were analysed to determine maximum SUV (SUVmax) for the corresponding liver metastasis. Reduction rate was calculated for CT and PET. Patients who showed more than 25% reduction in PET-SUVmax after chemotherapy were designated as a responder group. Mann-Whitney-U test was used to determine the differences of CT-perfusion parameters between responders and non-responoders. Spearman correlation test was used to find correlation between perfusion parameters and SU-Vmax. Performance of CT-perfusion parameters in discriminating non-responders from responders was evaluated using ROC analysis. Results: On PET, there were responder (n = 8) and non-responder groups (n = 12). Among CT-perfusion parameters, BF, PMB, and ALP were significantly different between the groups. The reduction rate of BF (correlation coefficient = 0.414) and PMB (0.394) was significantly correlated with that of SUVmax. ROC analysis revealed area under the curve was highest in BF (0.781) for differentiating responders from non-responders. In the prediction of non-responders, BF showed 87.5% sensitivity and 75% specificity at a cut-off value of 15% reduction rate. Conclusion: CT perfusion parameters, particularly BF, can be used as an alternative tool to PET in predicting early response after chemotherapy. Purpose: To evaluate the effect of iterative reconstruction on perfusion values and perfusion pixel noise in dynamic volume CT. Methods and Materials: 18 patients underwent dynamic volume 320-slice CT of the kidneys for pre-operative assessment of focal lesions. The study had ethics approval. Following intravenous injection of 60 ml of iodined contrast material at a flow rate of 10 ml/s, the imaging protocol consisted of 17 sequential low-dose scans at every 2-8 s (decreasing frequency) covering both kidneys. Scanning parameters were 100 kV, 20-40 mAs (depending on body habitus), 0.5 s gantry revolving time, 16 cm detector coverage, and 0.5 mm slice thickness for all dynamic scans. Raw data were reconstructed using conventional filtered back projection (FBP) with coneXact+ for full-field coverage and true iterative reconstruction (IR). ROIs were placed in the aorta as the reference and within 10 levels in each kidney cortex. Two postprocessing techniques were applied for calculation of perfusion maps and regional perfusion, and perfusion pixel noise values were statistically compared between both reconstruction methods. Results: We found no difference in regional perfusion values between FBP and IR (2.61 and 2.79 in normal cortex, 1.92 and 1.71 in renal neoplasia (n.s)., respectively), independent of the postprocessing method used. Standard deviation of perfusion pixel noise was 43 and 21 (p=0.01) in normal tissue, and 46 and 23 in renal tumours (p=0.001) for FBP and IR, respectively. Conclusion: Iterative reconstruction leads to significant reduction in perfusion pixel noise while maintaining regional perfusion values, therefore potentially improving reproducibility and allowing for radiation dose reduction.

Iterative reconstruction algorithm for hepatic enhanced 256-slice CT: image quality of the liver tumours on routine and low-dose CT S. Xu, Y. Hou, Q.Y. Guo; Shenyang/CN Purpose: To evaluate 256-slice CT images of the liver tumours reconstructed using iterative reconstruction (IR) on routine-dose (RD) and low-dose (LD) CT.

This prospective study was approved by the institutional review board. Informed patient consent was obtained. Forty eight patients (24 men, 24 women; average age: 58 yrs) with known or suspected liver tumours underwent hepatic enhanced 256-slice CT scans (Brilliance iCT, Philips). RD and LD CT scans were acquired sequentially during the hepatic portal venous phase of contrast enhancement. The CT dose index of LD acquisition was 50% lower than RD acquisition. RD and LD data were reconstructed using IR algorithm (iDose,Philips) at eight different levels (0%, 20%, 30%, 40%, 50%, 60%, 70%, 80%) separately. Quantitative noise measurements were performed. Two radiologists blinded to scanning technique evaluated the four iDose level (0%, 20%, 50%, 80%) reconstructions. Sharpness of tumours, contrast between tumours and normal liver tissue, and image quality were graded [scale: 1 (worse) -4 (best)]. Data were analysed using randomised block design analysis of variance and Friedman rank test. Results: On RD and LD CT, the quantitative noise significantly improved with increasing level of IR (P < 0.05).As the level increased, however, sharpness of tumours tend to be obscure. The LD data reconstructed at 50% iDose level got identical observer image quality scores with RD data reconstructed at 0% iDose level.

Quantitative analysis of Iodine concentration in hepatic tumours using dual energy multiphasic CT as objective index for assessment of tumour vascularity: initial experience S. Kobayashi 1 , O. Matsui 1 , K. Otani 2 , T. Gabata 1 , W. Koda 1 , T. Minami 1 , Y. Ryu 1 , K. Kawai 1 , K. Kozaka 1 ; 1 Kanazawa/JP, 2 Tokyo/JP (satoshik@staff.kanazawa-u.ac.jp) Purpose: To investigate whether quantitative analysis of iodine concentration in hepatic tumours using multiphasic dual energy (DE) CT could be useful as objective index for assessment of tumour vascularity. Methods and Materials: Eight patients (14 hepatic tumours) who underwent abdominal multiphasic dual source DE CT (non-enhanced single energy CT data followed by arterial, portal and equilibrium phase DE CT) and subsequent angiography-assisted CT for closer examination were included in this study of 176 consecutive abdominal multiphasic DE-CT cases. Virtual-non-contrast images and iodine maps of the livers were generated from the DE CT data using commercial software. Iodine concentrations in the tumours were measured on the iodine maps. Average iodine concentration in hypervascular hepatocellular carcinoma (HCC), (n = 10) and hypovascular hepatocellular tumours (n = 4) were compared with t-test. Additionally, average iodine concentration in background liver parenchyma and abdominal aorta were measured. Results: Average iodine concentrations in each contrast phase (arterial, portal, and equilibrium phase) were as follows: hypervascular HCC = 2.5±0.8, 2.3±0.4, 1.7±0.4 (mg/ml); hypovascular hepatocellular tumour = 1.1±0.8, 1.8±1.2, 1.4±0.8 (mg/ml); background liver parenchyma = 1.0±0.5, 2.3±0.5, 1.4±0.3 (mg/ml); abdominal aorta = 12.5±2.8, 5.3±1.0, 4.0±0.6 (mg/ml). On arterial phase, average iodine concentration of hypervascular HCC was significantly higher than that of hypovascular hepatocellular tumour (P = 0.01). Conclusion: Quantification of iodine concentration using dual energy CT might be useful as objective index to assess the degree of vascularity in hepatic tumours.

Spectral CT with metal artefacts reduction software (MARS) for improvement of tumour visibility in the vicinity of gold fiducial markers O.R. Brook 1 , S. Gourtsoyianni 1 , A. Brook 2 , C. Wilcox 1 , V. Raptopoulous 1 ; 1 Boston, MA/US, 2 Haifa/IL Purpose: Metal artefacts reduction software (MARS) utilising spectral CT has been recently introduced for reduction of beam-hardening artefacts. We demonstrate its use for visualisation of lesions in the vicinity of implanted gold fiducials for imaging guided focused radiotherapy.

The study was approved by institutional IRB with a waiver of informed consent. Consecutive patients with implanted fiducial seeds for radiotherapy of various abdominal lesions scanned at one month post-treatment were included in the study. Spectral imaging was utilised for arterial phase CT performed on Discovery CT750 HD (GE Healthcare, Waukesha, WI). Images were reconstructed with and without MARS. Two observers independently evaluated images with and without MARS for single best image, visibility of the tumour and amount of artefacts in three planes. Mann Whitney test was used to test statistical significance. Results: Seven patients (23 fiducials) were included in the study. Fiducials were placed in pancreas (n = 11), liver (n = 8) and periportal lymph nodes (n = 4). In 65% (95% CI, 51-77%) of the cases MARS images were graded better than regular reconstruction. Median score of visibility of the tumour for regular images was 2 and for MARS was 4 (possible scores: 1-worst to 5-best). A significant reduction of blooming artefact was noted on MARS images (p = 0.01). Quadratic weighted kappa values showed moderate to perfect agreement between observers for the various measurements. Conclusion: Metal artefacts reduction software utilising spectral CT improved visibility of the tumour in the vicinity of gold fiducial seeds used for focused radiotherapy planning. 

B A S193 C D E F G H

Vertebral fracture assessment: between experts and young physicians A. Bazzocchi 1 , P. Spinnato 1 , F. Fuzzi 1 , C. Sassi 1 , E. Salizzoni 1 , G. Battista 1 , G. Guglielmi 2 , R. Canini 1 ; 1 Bologna/IT, 2 Foggia/ IT (paolospinnato@email.it) Purpose: The aim of our study was to determine the diagnostic accuracy of new dual-energy x-ray absorptiometry technologies in the detection of vertebral fractures and to analyse intra-and inter-observer variability and time consumption of densitometric methods (MXA) compared with radiographic ones (MRX) in two groups of readers: experts and residents. Methods and Materials: Forty-five patients were submitted to both MRX and MXA (Lunar-iDXA TM , GE-Healthcare, USA) on the same day. Lateral images of the spine were independently evaluated by 3 expert radiologists (group-1) and by 3 radiology residents (group-2) in two sessions with 7 days between evaluations of the same anonymous images. "True" fractures were defined by an independent expert radiologist on MRX evaluation. Cohen test was used to analyse interpretation agreement. Results: Forty-eight "true" fractures were detected (48/585 vertebrae, 8.2% -28 mild, 16 moderate, 4 severe) in 45 analysed patients (61.2±11.1years, BMI=25.3±3.0Kg/m 2 ). MXA accuracy was 97% and 80% on a lesion-based and patient-based analysis, respectively. Inter-observer agreement on presence/absence of vertebral fracture (lesion-based) was 67% for MXA versus 65% for MRX in group-1, 65% versus 58% in group-2. The average intra-observer agreement in the two groups was equally 98% for both methods. The mean time spent for a single examination was 1.35 (min.sec) for MXA versus 2.10 for MRX in group-1 and 2.11 for MXA versus 3.36 for MRX in group-2. Conclusion: Nowadays technical improvements make MXA comparable with traditional radiographic gold standard, providing consistent advantages and attractions even for less-experienced physicians. Purpose: To obtain BMD values of the lumbar spine from sagittal reformations of routine abdominal contrast-enhanced MDCT and assess their performance in differentiating patients with and without prevalent and incidental osteoporotic fractures of the spine. Methods and Materials: Eight postmenopausal women (65±5years) underwent standard lumbar QCT (L1-L3) and afterwards routine abdominal contrast-enhanced MDCT. The sagittal reformations were used for corresponding lumbar BMD measurements. The calculated MDCT-to-QCT conversion equation for BMD was applied to baseline and follow-up routine abdominal contrast-enhanced MDCT scans of 91 postmenopausal women (68±8years). Their vertebral fracture status (no, prevalent or incidental osteoporotic fracture) was assessed in the sagittal reformations. Results: A correlation coefficient of r=0.94 (p < 0.05) was calculated for the BMD values of MDCT and standard QCT with the conversion equation BMD QCT =0.69xBMD MDCT -11 mg/ ml. Mean follow-up time of the 91 patients was 10±3 months. Ten patients (11.0%) had a prevalent osteoporotic vertebral fracture at baseline. Incidental osteoporotic vertebral fractures were diagnosed in 12 patients (13.2%). At baseline, patients with prevalent and incidental fractures showed significantly lower BMD values (averaged over L1-L3) than patients without fracture (66.33 mg/ml vs. 73.46 mg/ml vs. 93.55 mg/ml; p < 0.05). BMD values of patients with incidental and prevalent fractures were not different at baseline (p> 0.05). Short-and long-term reproducibility errors for BMD measurements in the sagittal reformations amounted 2.09% and 7.70%, respectively. Conclusion: In this longitudinal study, converted BMD values of the lumbar spine derived from sagittal reformations of routine abdominal contrast-enhanced MDCT could predict incidental osteoporotic vertebral fractures.

Cortical porosity identifies fragility fractures in type-2 diabetic postmenopausal women S.P. Yap, T. Baum, A.J. Burghardt, T.M. Link; San Francisco, CA/US (thomas.baum@ucsf.edu) Purpose: Subjects with type-2 diabetes mellitus (T2DM) have an increased risk of fragility fractures. Currently, fracture risk is assessed via BMD measurements from dual x-ray absorptiometry (DXA). However, patients with T2DM have normal to elevated BMD measurements with respect to normal controls. The purpose of this study is to determine whether T2DM patients with fragility fractures have higher cortical porosity using high-resolution peripheral quantitative computed tomography (HR-pQCT).

The iDose IR improves the image quality by decreasing image noise. When the radiation dose was reduced by 50%, Images maintaining the diagnostic quality of RD CT (0% IR) could be acquired with 50% iDose level correspondingly. However, excessive IR may obscure the sharpness of tumours.

Is intravenous morphine comedication effective in improving bile duct visualisation in dual-energy CT-cholangiography? C.M. Sommer, C.B. Schwarzwaelder, W. Stiller, U. Stampfl, N. Bellemann, T. Heye, L. Grenacher, H.U. Kauczor, B.A. Radeleff; Heidelberg/DE (cmsommer@gmx.com) Purpose: To prospectively evaluate whether intravenous morphine comedication is effective in improving bile duct visualisation in dual-energy CT-cholangiography in potential donors for living-related liver transplantation. Methods and Materials: This prospective study was approved by our local ethics committee. Forty subjects underwent CT-cholangiography with infusion of a hepatobiliary contrast over 40 minutes. Twenty minutes after beginning of the contrast infusion, either normal saline (n = 20 patients; group 1) or morphine sulphate (n = 20 patients; group 2) was injected. Another 25 minutes later, a dual-energy CT scan of the liver was performed (tube currents of 140 kV and 80 kV). Applying a dualenergy post-processing technique, pure iodine images were generated. Primary study goals were determination of bile duct visualisation scores (on a scale of 0 to 3: 0 -not visualized; 3 -excellent visualization) and diameters. The secondary study goal was determination of the bile duct volume. Results: Bile duct visualisation scores for second-and third-order branch ducts were significantly higher in group 2 compared with group 1 (2.9±0. Purpose: Juxtafacet cysts (JFC) are often found in degeneration of the facet joints. The method of choice to diagnose JFC is MRI. However, some JFC are missed and only found intraoperative. The present study addresses the incidence of JFC in a symptomatic population, positional MRI features and factors leading to variation in size of JFC. Methods and Materials: 67 distinct intraspinal or intraneuroforaminal JFC in 50 patients were investigated using positional MRI. The signal intensity, size of the JFC, the presence of a slip, the variation of the slip and the angular movement of the affected segment were assessed in supine, neutral sitting, flexion (sitting) and extension (standing) positions. The overall movement of the spine and the lordosis angle were measured. Results: JFC varied in size in segments with an unstable slip and with an increased angular movement. JFC with bright signal tend to vary in size compared with JFC with intermediate or low signal. Joint effusion and displacement of the effusion leads to formation of the cysts and is responsible for variations in size. In our study, the sensitivity for JFC detection was 100% for extension (standing), 89% for supine and 78% for neutral sitting position. JFC were the most prominent in extension (standing).

The present study proves that JFC are related to degenerative disease of the facet joints. Detection rate of JFC rises with increasing lordosis of the spine and under weight-bearing condition. Unstable slipping or increased angular movement affect the size of JFC. Results: Of all measured angles, LSA is significantly smaller in patients. Lumbar Stability Index is strongly changing to instability in both groups (patients and controls) with increasing BMI from 30 to 35 and also with changing position from lying to standing. The Oswestry number has a positive relation to BMI which is stronger in control group than in patients.

In reporting lateral lumbosacral x-ray, jumbar stability index should be noticed in people with high BMI, especially measured in standing position, even if the person has no low back pain. Lumbosacral angle should be noticed especially in people with chronic nonspecific low back pain.

The diagnostic advancement of axial loaded lumbar spine MR in patients with clinically suspected central canal stenosis Y. Kim 1 , J. Axial and sagittal T2 weighted images were obtained in preloaded phase and axial loaded phase. Axial images were evaluated with concerns of 1) gross change of central canal stenosis, 2) findings of facet joint change, including arthrosis, effusion, effacement of effusion, and 3) formation of ventral synovial cyst after axial loading. In addition, dural sac cross-sectional area (DSCSA) was measured in L3/4, L4/5 and L5/S1 levels to quantify the change of stenosis. The significant change in DSCSA was defined as 10% decrease. Changes of neural foraminal stenosis, curvature and spondylolisthesis were evaluated with sagittal images.

Methods and Materials: 49 postmenopausal females (age = 60.5 ± 5.3 years) were recruited into four groups: controls (n = 19), controls with osteoporotic fractures (n = 10), T2DM (n = 17), T2DM with osteoporotic fractures (n = 6). Fractures were assessed from radiographs and QCT of hip and spine. DXA of the spine and hip was performed for BMD measurements. HR-pQCT of distal and ultra-distal (UD) regions of radius and tibia were performed and cortical porosity was assessed. Purpose: Baastrup's disease is characterised by the development of abnormal contact between adjacent spinous processes. The clinical significance remains unresolved, with a few studies reporting Baastrup's disease as a cause of back pain. The aim of this study was to perform a large population analysis of the prevalence of Baastrup's disease in an unselected population and to determine if it is part of the normal ageing spine.

The scans of 1,008 patients, divided equally between seven age groups, who had undergone abdominopelvic CT were retrospectively reviewed. Baastrup's disease was judged to be present, by two independent observers, if there was close contact between adjacent spinous processes, and if the apposing ends were sclerotic. The presence of degenerative changes, i.e. disk degeneration, spondylolisthesis and facet OA was also noted. Results: 413 patients (41.0%) had evidence of Baastrup's disease. A decade-ondecade increase in prevalence was seen, with a peak of 81.3% in the over 80's. Up to 5 affected levels were seen in some patients (4.1% of positive cases), but most commonly, one level was affected (35.4%). Baastrup's disease was most common at L4/5. There were associated degenerative changes at almost all affected levels (899/901). Conclusion: Baastrup's disease has a very high prevalence amongst the elderly. Our data demonstrates that it develops with increasing age, and is part of the normal ageing process in the spine. Further, given the near universal association with degenerative changes, we urge caution before diagnosing Baastrup's disease as the cause of a patient's back pain.

A S195 C D E F G H

Prognostic HRCT patterns in fibrotic pulmonary sarcoidosis S.L.F. Walsh 1 , N. Sverzellati 2 , A. Devaraj 1 , A.U. Wells 1 , D.M. Hansell 1 ; 1 London/UK, 2 Parma/ IT (slfwalsh@gmail.com) Purpose: The aim of this study was to identify prognostic HRCT patterns in a group of patients with a diagnosis of fibrotic pulmonary sarcoidosis. Methods and Materials: The need for patient consent was waived by the institutional ethics committee. HRCTs of 96 patients (mean age 42.1 ± 13.3 years, 48 male:48 female) with a clinico-radiologic diagnosis fibrotic pulmonary sarcoidosis were studied. Visual estimates were made of the extent of abnormal lung and the proportional contribution of fine and coarse reticulation, microcystic (cysts ≤ 4 mm) and macrocystic honeycombing. In addition, a score for severity of traction bronchiectasis was assigned. Using death as the primary outcome measure, variables were analysed by Cox proportional hazards model. Results: CT features predictive of a worse outcome on bivariate analysis were proportion of coarse reticulation (HR = 1.06, CI = 1.01-1.08, p < 0.004) and main pulmonary artery size (HR = 2.45, CI = 1.18-5.08, p < 0.016). Patterns of honeycombing and traction bronchiectasis were not significantly associated with increased mortality. Conclusion: In fibrotic sarcoidosis the HRCT pattern of coarse reticulation is significantly associated with increased mortality. Unlike, in the setting of fibrotic interstitial pneumonias, patterns of honeycombing and traction bronchiectasis have no significant predictive value. In addition, main pulmonary artery size is a strong predictor of mortality in patients with fibrotic pulmonary sarcoidosis.

Acute exacerbation of usual interstitial pneumonia after resection of lung cancer: CT manifestations of the preoperative lung H. Sugiura 1 , M. Jinzaki 1 , M. Inoue 1 , T. Hosokawa 1 , K. Hosoda 1 , S. Koga 1 , T. Hoshi 2 , Y. Kawabata 2 , S. Kuribayashi 1 ; 1 Tokyo/JP, 2 Saitama/JP (hsugiura@momo.so-net.ne.jp) Purpose: To investigate whether graphical changes in preoperative chest CT were correlated with the occurrence of acute exacerbation (AE) of usual interstitial pneumonia (UIP) after resection of lung cancer. Methods and Materials: This study included 205 cases (182 males and 23 females) of lung cancer with histologically proven UIP pattern in the specimen of the resected lungs. The preoperative chest CT scans were retrospectively evaluated, and assessed for the presence of interstitial pneumonia (IP) findings. The IP pattern was classified into three groups on the basis of IP findings as follows: group 1, with typical honeycombing; group 2, with atypical honeycombing; group 3, neither typical nor atypical honeycombing (only ground glass opacities and/ or reticulation). Atypical honeycombing indicated the finding of cysts with some fibrosis whose appearance corresponds neither to the definition of honeycombing nor pulmonary emphysema. Results: Acute exacerbation of UIP developed in 11 cases after thoracotomy. The incidence of AE occurred in the groups 1, 2, and 3 of IP pattern was 13.6% (6/44), 6.4% (5/78), and 0.0% (0/83), respectively. Comparison between the IP pattern and the incidence of AE indicated significant difference (P < 0.01). Conclusion: Postoperative AE of UIP developed in cases with atypical honeycombing as well as in cases with typical honeycombing, while AE was not observed in cases with no findings of typical nor atypical honeycombing. Therefore, atypical honeycombing may also be a risk factor of postoperative AE of UIP, and is important to be recognised before surgery for lung cancer. Purpose: To provide morphological and quantitative CT information on emphysema in healthy smokers in correlation with pulmonary function tests (PFT).

The study population included 75 smokers (current smokers: n = 39; ex-smokers: n = 36) and 25 non-smokers who underwent volumetric high-resolution CT of the chest and PFTs with automated quantification of emphysema. Results: The frequency (29/75; 39% vs 1/25; 4%; p = 0.001) and extent (0.13% vs 0.0004%; p = 0.01) of emphysema was significantly higher in smokers compared to nonsmokers. Among smokers: (a) emphysema was an isolated finding in 8 subjects whereas it was found in association with CT features of airway disease in 21 smokers; (b) the severity of emphysema did not differ according to the GOLD Results: With a use of axial loading device, the additional diagnosis of severe central canal stenosis was made in 13 patients (25%) in both gross interpretation and quantitative study (DSCSA < 75 mm 2 ). The significant decrease of DSCSA was demonstrated in 22 patients (42%). The significant decrease was related to facet joint effusion and effacement of effusion. Conclusion: Measurable advancement in diagnosis of severe central canal stenosis was possible with axial loaded MRI. Patients with facet joint instability had tendency to show significant changes in central canal area. Purpose: To assess the rotation of the spinal cord and vertabra in axial planner using 3D-MRI in scoliosis patients. Methods and Materials: Ten patients with a thoracic or thoracolumbar scoliosis underwent 3D-Space T2-weighted MRI. The rotation of the spinal cord and vertebra in axial section were measured for T4, the apical vertebra and L1 through multiplanar reconstruction.The rotational data were compared with Cobb angle; also their relationship was anlysed using non-parametric tests. Results: There was different rotation of spinal cord at T4, the apical vertebra, L1 and rotation angle of spinal cord (RAc) was different from vertebra rotation angle (RAsag), especially in patients with large Cobb angle; most of them had smaller vertebra rotation than spinal cord. There were three rotation types : overrotation (O) type, under-rotation (U) type and reverse-rotation (R) type. 80%(8 of 10)patients were of O type rotation. Statistical analysis showed a significant correlation between spinal cord rotation and Cobb angle at measured level. The position of apex was statistically correlated with the Cobb angle. Conclusion: To quantify the degree and rotation type of the spinal cord using 3D-MRI in scoliosis patients are important to make surgical planning.That may decrease the operation dangers and prevent the complication of nervous dysfunction caused by spinal cord injury due to the over derotation. Purpose: The aim of this study was to identify HRCT features that predict mortality in patients with chronic fibrotic hypersensitivity pneumonitis (CHP).

A cohort of 95 patients with a clinico-radiologic diagnosis of CHP was identified (mean age 55.1 ± 12.6 years, 40 male:55 female, biopsy proven = 27). HRCT scans at baseline were scored by two observers for the extent of abnormal lung and the proportional contribution of fine and coarse reticulation, microcystic (cysts ≤ 4 mm) and macrocystic honeycombing. Main pulmonary artery and segmental pulmonary artery measurements were also assessed. Finally, a score for severity of traction bronchiectasis was assigned. Using death as the primary outcome measure, variables were analysed by Cox proportional hazards model. Results: On multivariate analysis, percentage of coarse reticulation (HR = 1.02, CI = 1.01-1.04, p = 0.014), microcystic honeycombing (HR = 1.09, CI = 1.01-1.17, p = 0.019) and macrocystic honeycombing (HR = 1.06, CI = 1.01-1.10, p = 0.005) were independently associated with an increased mortality. Macrocystic honeycombing was a stronger prognostic index than microcystic honeycombing. The degree of traction bronchiectasis was also a strong predictor of poor prognosis (HR = 1.11, CI = 1.05-1.17, p < 0.001). Pulmonary vasculature measurements were not predictive of mortality either on univariate or multivariate analysis. Conclusion: Patterns of coarse fibrosis such as coarse reticulation and honeycombing are associated with increased mortality in patients with a diagnosis of CHP. In addition, degree of traction bronchiectasis is strongly predictive of mortality. S196 Purpose: To evaluate the incidence of gemcitabine-induced lung toxicity and the role of HRCT in detecting pulmonary drug-related events in asymptomatic patients. Methods and Materials: 140 patients with proven neoplasia (NSCLC, pancreas, breast and urotelial cancer) were treated with gemcitabine (as single or combined therapy) for 3-18 months. All of them underwent at least two HRCTs before and during treatment. Respiratory symptoms were assessed the same day of the exam, using the WHO toxicity grading. Three patients treated also with radiation therapy were not included in the study. Results: During treatment, 15 patients (10.9%) presented with HRCT findings compatible with drug-related toxicity. The HRCT showed four major radiological appearances: sub-lobar or segmental opacities (n = 6) (40%), interstitial pneumonitis with ground glass opacities (n = 6) (40%), bronchiolitis (n = 2) (13.3%) and diffuse alveolar damage (DAD) (n = 1) (6.7%). Respiratoty symptoms, however, were present in only 8 patients (5.8%): Grade 1 (n = 4); grade 2 (n = 5); grade 4 (n = 1). There was no correlation between incidence of lung disease and gemcitabine dosage. All symptomatic patients were successfully treated with steroids and gemcitabine dose reduction or suspension. Asymptomatic patients underwent steroid treatment with close clinical and HRCT follow-up.

The incidence of gemcitabine-induced lung toxicity in our series was similar to those reported in the literature. HRCT allowed for identification of lung alteration in an asymptomatic phase and can therefore be considered a valid instrument to support clinical management of oncologic patients treated with gemcitabine. Purpose: The hypothesis of this study was that distinct morphological emphysema phenotypes assessed by CT show characteristic perfusion defect pattern. Methods and Materials: Forty-one patients with severe (GOLD III or IV) emphysema underwent inspiratory 3D-HRCT and contrast-enhanced MR-perfusion (1.5 T; spatial resolution 3.5 mm x 1.9 mm x 4 mm). 3D-HRCT data were visually analysed in consensus by 3 experts in chest radiology, for phenotyping and quantification of emphysema. The predominant phenotype per segment was categorised as normal, centrilobular, panlobular or paraseptal. Segmental lung perfusion was visually analysed using 6 patterns of pulmonary perfusion (1 -normal, 2 -mild homogenous reduction in perfusion, 3 -inhomogenous perfusion without focal defects, 4 -inhomogenous perfusion with focal defects, 5 -inhomogenous absense of perfusion, 6 -homogenous absence of perfusion), the extent of the defect given as a percentage. Results: A total of 730 segments were evaluated. CT phenotyping categorised 566 (78%) as centrilobular, 159 (22%) as panlobular and 5 (< 1%) as paraseptal with no normals. Scores with regard to MR perfusion patterns: 1 -0; 2 -0; 3 -28 (4%); 4 -425 (58%); 5 -169 (23%); 6 -108 (15%). The predominant perfusion pattern matched as follows: 70% centrilobular emphysema -inhomogenous perfusion with focal defects (score 4); 42% panlobular -homogenous absense of perfusion (score 6) and 43% panlobular -inhomogenous absense of perfusion (score 5 Purpose: To assess the role of MDCT with automated measurement of airways in the quantification of airflow obstruction in inspiratory and expiratory scans in patients with COPD. Methods and Materials: 20 patients (age range 48-85 years) with clinical diagnosis of COPD prospectively underwent pulmonary functional tests (PFTs) and chest CT scan. All CT exams were performed with a 64-rows scanner (slice thickness/ interval 0.625 mm). Two consecutive acquisitions were obtained, one standard dose full inspiration scan followed by one low-dose (20 mAs) scan at the end of forced stage (p = 0.77). Smokers with emphysema had significantly higher mean values of FRC (p = 0.0012), RV (0.5). Correlations were found between the percentage of emphysema and: (a) tobacco consumption of current (r = 0.34215; p = 0.0330) and exsmokers (r = 0.44104; p = 0.0071); (b) alterations of TLC, FRC, RV, FEV1/ VC and DLCO of smokers. Although there was no statistically significant difference in the overall percentage of emphysema between current smokers and exsmokers, current smokers had a higher percentage of emphysema in the right lung (p = 0.041) and in the right upper lobe (p = 0.037). Conclusion: Quantitative CT allows recognition of regional specificities and subclinical functional alterations in smokers with emphysema.

The Purpose: To evaluate the usefulness of quantitative lung parenchymal assessment in bronchiolitis obliterans (BO) following hemopoietic stem cell transplantation (HSCT). Methods and Materials: Ten patients with BO after HSCT and 20 volunteers were included. They underwent MDCT including inspiration and expiration scans. The patients performed pulmonary function tests (PFTs). The source images were reconstructed two ways for visual (VA) and qunatitative assessment (QA). In VA, radiologic abnormalities were reviewed. In QA, mean lung density (MLD), relative areas of lung (RA) with attenuation coefficients below specific thresholds and nine percentiles of the distribution of attenuation coefficients were computed. The relationships between these parameters and PFT results of patients were correlated. Results: All patients showed abnormal CT findings including expiratory air trapping on VA. In QA, MLD of inspiration scan (MLDin), MLD of expiration scan (MLDex), difference between MLDin and MLDex (ΔMLD) were -862.6, -813.2 and 49.4 for patients with BO, and -815, -690.25 and 124.75 for volunteers. There was significant difference between all these parameters for two groups (p = 0.001). Relative area with attenuation coefficient less than -800 HU (RA 800 ) and -750 HU (RA 750 ) and the 80 th percentile on expiration were most highly correlated with patients' residual FEV1 (rho = -0.927, p < 0.001). MLDin had a slight tendency to be lower in patients whose disease was more severe. But this was not statistically significant (r = 0.46, p = 0.07). Conclusion: RA 800 , RA 750 and the 80 th percentile on expiratory scan are valid index to quantify BO on MDCT. Quantitative lung analysis is useful tool for the assessment of BO after HSCT.

Pulmonary findings in ankylosing spondylitis by HRCT: correlation with disease indexes Z.M. Metafratzi 1 , E. Alexiou 1 , M. Fanariotis 1 , I. Saougou 2 , K. Vlahos 2 , A. Drosos 2 , K. Tsampoulas 2 ; 1 Larisa/GR, 2 Ioannina/GR Purpose: To identify the lung abnormalities in ankylosing spondylitis (AS) on HRCT and to correlate them with the disease severity indexes. Methods and Materials: Thirty-two patients with AS (age 48±8ys, and disease duration 17.5±7 ys) free of respiratory symptoms, underwent rheumatological examination and lung HRCT. A control group of fifteen age-and sex-matched subjects were investigated with the same technique. Results: HRCT revealed at least three different coexisting abnormalities in twentythree (74%) patients, predominantly pleural tags (19, 59.4%), air-trapping (17, 53.1%), pleural thickening (14, 43.8%), parenchymal bands (13, 40.6%), irregular interfaces (10, 31.3%), interlobular thickened septa (9, 28.1%), GGOs (8, 25%), bronchiectasis (7, 21.9%), interstitial lung disease (5, 16.6%), bullae (3, 9.4%), honeycombing (2, 6.3%) and tracheal dilatation (8, 25%). None of the control group had more than two coexisting HRCT findings including mild air-trapping (5, 35.7%) and mild bronchiectasis (4, 26.7%). No correlation was noted between HRCT findings and the smoking history or drug administration of the patient group. The only HRCT abnormality observed to be statistically significant in the patients compared with the controls was the parenchymal bands (p = 0.002). Furthermore, CRP was significantly correlated to the presence of parenchymal bands (p < 0.001). No other correlation was detected between HRCT findings and the disease severity indexes or duration of the disease. Conclusion: A spectrum of HRCT lung abnormalities is often detected in asymptomatic patients with AS. The most common findings demonstrate evidence of mild airway and interstitial disease and tracheal dilatation.

A S197 C D E F G H tions in the same time) or goup 2 (single transforaminal injection). Follow-up was conducted within 7 days and one month (short-term) and one year (long-term). Short-and long-term outcomes were measured using a visual analog scale (VAS) and the Oswestry Disability Index. Univariate analysis (using Fisher exact test and λ2 tests) and multiple logistic regression analysis were performed to evaluate the relationship between possible outcome predictors and the therapeutic effect. Results: Significant pain relief (> 50%) was seen at one month in 49% of patients in group 1 and 43% in group 2. Significant differences were noted in pain relief characterictics at one month between groups 1 and 2 in patients presenting a disc extrusion or a posterolateral herniated disc. The only significant outcome predictor at short-term follow-up was previous history of lumbar surgery. No major complication was noted at long-term follow-up. Conclusion: Association of transforaminal and interlaminar steroïd lumbar injection under CT guidance is more effective than single transforaminal injection. Purpose: Biopsies of metastatic sites are usually not recommended to evaluate changes in biological features in women with ABC. We assessed discordant expression of estrogen receptor (ER), progesterone receptor (PgR) and epidermal growth factor receptor 2 (HER2), between primary tumour and bone metastases and its clinical impact on patients' management. Methods and Materials: ABC patients who underwent CT-guided bone biopsy from 1997 to 2009 were included. Data on receptors status of primary tumour and bone metastases were collected and analysed. Results: 109 patients were enrolled. Median time from primary diagnosis to bone biopsy was 4.2 years. Overall discordance rate (ODR) in ER expression was 20.5%: 3/7 pts (42.9%) shifted from negative to positive; 19/100 pts (19%) from positive to negative (p = 0.0006). ODR in PgR expression was 43.9%: 4/21 pts (19%) changed from PgR negative to positive and 43/86 pts (50%) from positive to negative (p < 0.0001). ODR in HER2 status was 6.9%: 4/78 pts (5.1%) changed from HER2 negative to positive and 2/8 pts (25%) from positive to negative (p = 0.41). According to the ER status, 13/21 pts (61.9%) changed endocrine therapy when discordant and 23/77 pts (29.7%) when concordant (p = 0.01). According to the HER2 status, 4/6 pts (66.6%) changed immunologic therapy in discordant status and 6/72 pts (8.3%) in concordant status (p = 0.002). All patients well tolerated the bone biopsy. Conclusion: Receptors profile may significantly change between primary breast cancer and bone metastases, with relevant impact on therapy. CT-guided bone biopsy may therefore be considered a safe procedure to optimise therapeutic choices.

Complications of CT-guided transgluteal pelvic abscess drainage with catheter placement L. Zheng, M. Kataoka, A. Roycjowdhury, J. Ferrucci; Worcester, MA/US (larry.zheng@umassmemorial.org) Purpose: To assess complications of CT-guided percutaneous drainage catheter of deep pelvic abscess by transgluteal approach in a teaching medical centre. Methods and Materials: 101 cases of CT-guided pelvic abscess drainage with catheter placement by the transgluteal approach performed in our institution from 1, 2006 up to 3, 2010 were reviewed. All the records were reviewed for clinical diagnosis, catheter size and follow-up. The CT images were evaluated for signs of inflammation at the procedure site, angle/location of the approach, technique of catheter placement (Trocar vs Seldinger) and numbers of attempts of catheter placement. Approaches were classified as parasacral or non-parasacral route. Results: Patient age ranged from 4 to 94 years (average 50 years; 69 females and 32 males). The majority of our patients had inflammation or vascular congestion (76%) in the pelvis. The catheter was placed using the parasacral in 87 cases and non-parasacral approach in 14 cases. Seldinger technique was used in 66 cases and Trocar technique in 35 cases. The size of the catheter used varied from 6 to 14 Fr. The number of the punctures varied from 1 to 5 (average 1.4). Complications were seen in 8 cases (8%) including 3 pseudoaneurysm (all done with nonparasacral approach), 3 haematomas (2 < 3 cm and 1 > 5 cm) and 2 fistulas. All complications occurred with theTrocar technique. Conclusion: Complications of CT-guided transgluteal approach for pelvic abscess drainage are more common with non-parasacral approach and Trocar technique in our experience. expiration. Datasets were analysed using an automated commercial software for airways analysis (Thoracic VCAR, GE Healthcare). One chest radiologist reported morphologic airway parameters, automatically calculated from lobar (second generation) to sub-subsegmental bronchi (fifth generation: diameter < 4 mm) on both inspiratory and expiratory scans: lumen diameter (LD), lumen area (LA), wall thickness (WT), wall area (WA), wall area ratio (WA%: wall area/total bronchial area%). Ratio between mean values of all automatic measurements obtained in inspiratory and expiratory scans were correlated with PFTs for each patient and each anatomic level (Pearson correlation coefficient). Results: The best correlation (p = 0.04; p = 0.08) was observed between functional parameters of airflow obstruction (FEV 1; FEV 1 /FVC) and the ratio between mean values of LA in expiration and mean values of LA in inspiration (LA exp /LA insp ), at the level of the fifth bronchial generation. Conclusion: Dynamic modifications of distal airways lumen area correlate with functional parameters indicative of airflow obstruction. MDCT with automatic measurement of the airway parameters may have a role in quantification of airflow obstruction in COPD patients.

CT quantitative measurements correlated with pulmonary function tests in chronic obstructive pulmonary disease: a meta-analysis X. Xie, R. Vliegenthart Proença, Y. Wang, G.H. de Bock, M. Oudkerk; Groningen/NL (x.xueqian@rad.umcg.nl) Purpose: To determine the relationship between CT quantitative emphysema or small airway measurements and pulmonary function test (PFT) in chronic obstructive pulmonary disease (COPD). Methods and Materials: Web of Science, PUBMED, EMBASE and Cochrane Central were searched for studies before September 2010. Out of 1386 citations, 18 articles were included in the meta-analysis according to pre-defined inclusion criteria. Attenuation area percentage < -950 HU (%LAA-950), mean lung density (MLD) and wall area percentage (WA%) were considered as CT quantitative parameters. The predicted forced expiratory volume in the first second (FEV1, %p), and FEV1 divided by the forced volume vital capacity (FEV1/FVC) were considered as reference PFT parameters. Data were independently evaluated and abstracted with a standardized protocol and abstraction form. Pooled correlation coefficients as well as 95% confidence intervals (95%CI) were calculated in a meta-analysis using a random effect model. 

Ultrasound-guided versus computed tomography-controlled pararadicular injections in the lumbar spine: a prospective randomised clinical trial A. Loizides, S. Peer, S. Ostermann, J. Obernauer, K. Galiano, H. Gruber; Innsbruck/AT (alexander.loizides@i-med.ac.at) Purpose: Pararadicular injections are widely used for alleviation of back pain.

Injections are preferentially performed as fluoroscopy or computed tomography (CT)-controlled interventions. Ultrasound provides real-time monitoring, does not produce ionising radiation, and is broadly available. This study was performed to evaluate the overall effect of ultrasound-guided pararadicular injections.

We studied feasibility, accuracy, time-savings, radiation doses, and pain relief of ultrasound-guided pararadicular injections versus CT-controlled interventions in a prospective randomised clinical trial. 22 adult patients with chronic low back pain were consecutively enrolled and assigned to an ultrasound or a CT-group. Results: 12 subjects from the group randomised to ultrasound were judged (exclusion criteria: BMI> 35 + postsurgical state) to be feasible for this type of approach.

In 12 patients of the ultrasound group the intertransverse ligament was clearly visible and all of the associated pararadicular injections were performed correctly. The mean duration of procedure and mean radiation dose was 4.82 minutes and 26.73 mGy.cm in the ultrasound group, and 6.9 minutes and 63.57 mGy.cm in the CT group. Both groups showed a benefit from pararadicular injections. Conclusion: These initial results which will be evaluated in a larger study show that the ultrasound approach to the pararadicular compartment in the lumbar spine is feasible with minimal risks in a large majority of patients and results in a significant reduction of procedure duration and radiation dose.

Image-guided biopsy of pleuric and peripheral lung lesions: comparison between ultrasound (US) and computed tomography (CT) guidance C. Murolo 1 , G. Serafini 2 , E. Savarino 1 , L.M. Sconfienza 3 ; 1 Genoa/IT, 2 Pietra Ligure/IT, 3 Milan/IT (io@lucasconfienza.it)

Purpose: Image-guided biopsy represents the reference standard in the characterisation of pleuric and peripheral lung lesions that can be sampled under CT or US guidance. Our purpose was to compare the outcome of CT or US guidance when sampling peripheral lung or pleuric lesions. Methods and Materials: From 1/ 2000 to 8/2010, 711 thoracic biopsies were performed at our institution. Among these, 284 lesions had pleuric origin or had a peripheral location in the lung with at least a small contact with the pleura. These lesions were biopsied either under CT (179 lesions; 170 patients, 71/99 males/ females, mean age 64 ± 12.5 years) or US guidance (105 lesions; 103 patients, 44/59, 67 ± 9.9 years), according to location of the lesion and patients' general conditions, using a 23G modified-Menghini needle. For each biopsy, duration of the procedure, occurrence of post-procedural pneumothorax, and sample adequateness were recorded. Chi-square and U-Mann-Whitney statistics were used. Purpose: To assess CT scan as a guiding tool for percutaneous placement of internal fiducial markers for stereotactic radio-surgery of inoperable pancreatic cancers. Methods and Materials: Consecutive 31 patients with inoperable pancreatic cancer were treated with stereotactic radio-surgery at our institute in the last 12 months. CT scan was used as a guiding tool for placement of internal gold fiducial markers to track the translational and rotational movements of the organ during treatment. A total of 148 fiducials were placed. At least 4 fiducials each (not more than 5) were placed in and around the tumour in all patients in non-co-linear manner with at least 2 cm distance between any two fiducials. Results: Stomach, small bowel, colon and liver were traversed in 22, 1, 1 and 2 patients, respectively. No major complications such as pancreatitis, significant haemorrhage or sepsis occured in any patient. Small self limiting focal haematomas occurred in 4 patients. A minority of fiducials (17 i.e. 12%) were not usable of tracking due to either migration or improper placement. Conclusion: CT scan guidance is accurate and safe for percutaneous fiducial placement in and around the pancreatic tumours. At least 4 fiducials (three are necessary for tracking) are necessary for each lesion, as migration of fiducials is unpredictable, sometimes rendering them useless for tracking. Thorough knowledge of principles of fiducial placement is necessary. Access routes and methods are similar to CT-guided pancreatic biopsies but may need more punctures. Antibiotic coverage is recommended when traversing colon is inevitable.

Ultrasound-guided aspiration and ethanol sclerotherapy for the treatment of endometriotic cysts G. Gatta, V. Parlato, G. Di Grezia, A. Porto, R. Grassi, A. Rotondo; Naples/IT (graziella.digrezia@yahoo.it)

Purpose: To check the effectiveness of ultrasound-guided aspiration and ethanol sclerotheraphy in treating endometriotic cysts and adopt this procedure as an alternative to surgery. Methods and Materials: 50 consecutive patients with an average age of 25.2 years (SD 6.5; range 16-14 years) diagnosed with endometriotic cyst through ultrasonography, who were pregnant or presenting with highly anaesthesiologic risk, adherences, refusal of surgery, underwent ultrasound-guided aspiration of 54 endometriotic cysts (40 cysts with a transabdominal approach and the remaining 14 with a transvaginal one) and ethanol sclerotherapy.

The procedure was successful in all patients. Imaging check-ups were carried out after a span period of: 12/ 24 hours, 15 days and 3/6/12 months. After 12 months, 4 of our patients (8%) showed recurrence. 3 of them opted for a second session of ultrasound-guided aspiration and ethanol sclerotherapy. Conclusion: Ultrasound-guided aspiration and sclerotherapy with 95° ethanol provide a valid alternative to surgery in treating endometriotic cysts. 

A S199 C D E F G H before CTA (group 1), 20 in reversed order (group 2). Mean density (Houndsfield Units, HU) was determined at baseline and peak enhancement of CTP, as well as on source-images of CTA in distinct regions (basilar/ middle cerebral artery, transverse and superior sagittal sinus, grey/ white matter (GM/ WM)). Contrast of extra-/intracranial arteries and veins was rated according to a 5-point-scale (1 = excellent, 5 = poor). CTP maps were assessed by determining mean transit time (MTT), cerebral blood flow (CBF) and blood volume (CBV) in identical regions. Results: HU between groups 1 and 2 were not significantly different for CTA and CTP at peak enhancement. At CTP baseline, HU between group both groups were different for all points except for GM and WM. There was no significant difference between both groups for the mean delta (difference between baseline and peak enhancement), as well as for MTT, CBV and CBF. Quality of CTA was rated to be superior in group 2 concerning extracranial venous contrast; however, not for all other criteria. Conclusion: Reversal of CT stroke protocol had no significant influence on quantitative parameters of the CTP. Quality of CTA concerning extracranial venous contrast was superior when CTA was performed before CTP.

Whole brain CT perfusion maps with paradoxical low mean transit time to predict infarct core S. Chakraborty, M.E. Ahmad, R. Glikstein, M. Hogan, D. Dowlatshahi, G. Stotts; Ottawa, ON/CA (santanoo@gmail.com) Purpose: CT perfusion (CTP) is based upon central volume technique (CBF = CBV/MTT) and uses complex deconvulotional method to generate various perfusion maps. Traditionally, low CBV values are indicative of infarct core and low CBF with high MTT/ TTP is used to describe ischaemic penumbra. The paradoxical low MTT in the region of ischaemia has not described in the current CTP literature to best of our knowledge. Methods and Materials: We retrospectively reviewed the clinical and whole brain perfusion data of 134 patients presenting with acute stroke in less than 6 hours from onset of symptoms over 18 months. 22 patients with low MTT in the region of ischaemia were further evaluated for clinical and imaging outcome. Results: Out of 134 patients, 41 were excluded due to inadequate studies (n = 14) and no follow-up scans (n = 27). 43 of remaining 93 patients had obvious abnormality on MTT perfusion maps. 22 of these 43 patients showed low MTT (colour coded as blue on RGB scale) in the centre of otherwise high MTT (red on RGB scale) area. Overall, MTT map has lower sensitivity (71%) for the detection of ischaemia and paradoxical low MTT sign has even lower sensitivity, seen only in 36% of cases. However, this has very high specificity and all 22 patients with low MTT developed infarct in the same region on follow-up imaging. However, the final volumes of infarct were approximately 20% bigger. Conclusion: Low MTT in the region of ischaemia is an interesting finding that accurately predicts the infarct core.

Perfusion CT in stroke imaging: influence of post-processing M.M. Lell 1 , B. Abels 1 , E. Klotz 2 , B. Tomandl 3 , M. Uder 1 ; 1 Erlangen/DE, 2 Forchheim/DE, 3 Bremen/ DE (michael.lell@uk-erlangen.de) Purpose: There is an ongoing discussion about the appropriateness slope (MS) or deconvolution (DC) approach for post-processing of perfusion CT data in stroke imaging. We performed a qualitative and quantitative comparison of DC and MS using identical source data sets and pre-processing parameters. Methods and Materials: 50 data sets were used to calculate colour maps of cerebral blood flow (CBF), cerebral blood volume (CBV), and various temporal parameters with software implementing DC and MS algorithms. Colour maps were qualitatively rated/categorised. Quantitative region-of-interest-based measurements were made in non-ischaemic grey and white matter, suspected penumbra and suspected infarction core. Qualitative results, quantitative results as well as PCT lesion sizes from DC and MS were statistically compared. Results: CBF and CBV colour maps based on DC and MS were of comparably high quality. Quantitative CBF and CBV values calculated by DC and MS were within the same range in non-ischaemic regions. In the suspected penumbra, average CBF (DC) values were lower than CBF (MS). In the suspected infarction core, average CBV (DC) was similar to CBV (MS). Using adapted tissue-at-risk/non-viable-tissue thresholds we found excellent correlation of DC and MS lesion sizes. Conclusion: DC and MS yielded comparable qualitative and quantitative results. Lesion sizes indicated by DC and MS showed excellent agreement when using adapted thresholds.

: 224 patients were included in the current prospective study. TACE (mean 3.4 sessions/patient, SD 1.33, range 1-10) was performed as a downstaging treatment to meet the previously described LITT requirements (number of metastases < 5, diameter of each metastasis < 5 cm). The intraarterial protocol consisted of Irinotecan and Mitomycin (77 patients), Gemcitabine and Mitomycin (49 patients) or Mitomycin alone (98 patients) in addition to Lipiodol and Embocept in all patients. Post TACE, all patients included underwent MR-guided LITT (mean 2.2 sessions/patient). Results: Overall, TACE resulted in a mean reduction in diameter of the target lesions of 21.4%. The mean time to progression was 5 months, the mean local tumor control rate was 10.7 months, both calculated as of therapy completion. Calculated from the beginning of the TACE therapy, mean survival rate in the patients was 26.1 months (SD 16.01, range 4-110 months), in patients treated with Irinotecan and Mitomycin 24.17 months, in patients treated with Gemcitabine and Mitomycin 26.9 months and in patients treated with Mitomycin only 27.24 months with a statistically significant difference between the groups (p < 0.01). Conclusion: Repeated TACE offers adequate downstaging of liver metastases of colorectal carcinoma to allow MR-guided LITT. The combined treatment illustrates substantial survival rates and high local tumour control with statistically significant differences between the three different chemotherapeutic drug combinations used.

Can real-time US-CT/MRI fusion imaging guidance enable ablation of liver malignancies that are undetectable with conventional US? L. Solbiati 1 , T. Ierace 1 , L. Cova 1 , T. Tondolo 2 , N. Goldberg 3 ; 1 Busto Arsizio/IT, 2 Segrate/IT, 3 Jerusalem/IL (lsolbiati@aobusto.it) Purpose: To assess the ability of real-time US-CT/MRI fusion imaging for guiding percutaneous ablation of liver malignancies undetectable with US. Methods and Materials: From 2003 to 2009, 334 patients with 884 focal liver tumours (544 HCCs and 340 metastases) underwent percutaneous radiofrequency or microwave ablation guided by a image fusion system that combines real-time US with fusion to CT/MRI images based upon magnetic field tracking (Esaote, Genoa, Italy). Of these, 85 HCCs and 68 metastases (N = 153 [17.3%]) in 67 patients were detectable only with contrast-enhanced CT or MRI, but undetectable with US due to small size (n = 66), isoechogenicity with liver parenchyma (n = 46), obscuration due to location (n = 41). 138/153 (90.2%) tumours were smaller than 2.0 cm and 58/153 (37.9%) smaller than 1 cm (mean 1.3 cm). CT or MRI was performed at 24 hr after ablation to assess the technical efficacy (completeness of ablation including a 5 mm ablative margin) and at 8 months to search for possible local tumour progression (clinical efficacy). Results: At 24-hr follow-up imaging, 140/153 (91.5%) malignancies were completely ablated; 4 (2.6%) were partially ablated and 9 targets (5.9%) were completely missed. The four partially ablated nodules were successfully re-treated at 28-40 days. At 8-month follow-up, local tumour progression was detected in 27/144 (18.7%) tumours (11 HCCs and 16 metastases). Thus, overall efficacy was 91.5% per tumour and 86.6% per patient. Conclusion: A real-time fusion imaging system enables sufficiently precise targeting of many tumours undetectable with US alone to achieve complete ablation in the majority of cases. Purpose: To evaluate if the order of a CT stroke protocol (CT-perfusion before or after CT-angiography) has an impact on image quality. Methods and Material: 40 consecutive patients with clinical suspicion of brain infarction underwent CT stroke protocol including non-enhanced CT, CT angiography (CTA) and CT perfusion (CTP). Twenty examinations were performed with CTP S200 B C A D E F G H rial (iomeprol, 400 mgI/ml, 80 ml, 4 ml/s). The simultaneous exposition of 140 kV and 80 kV was used with the collimation of 2 x 32 x 0.6 mm of both system. The virtual nonenhanced images were compared with the "real" nonenhanced images previously acquired by conventional way. The radiologist blinded to the findings of the real images describe findings of the CT angiography including the evaluation of virtual nonenhanced images of 6 mm thickness were created with the threematerial decomposition algorithm called brain haemorrhage.

The following findings were found on "real" noncontrast images: no hypodense area (112x) cases, territorial hypodense area (73x), intracranial bleeding (7x), space occupying masses (6x), extensively calcified tissue (2x). All findings were found on virtual nonenhanced images even if the image quality was decreased according higher noise level. The additional value of the vascular structures and masses enhancement was found in improved depiction of the hypodense area and the enhancement pattern of expansive processes.

In the nonselected population of the patients with clinically suspected stroke, the conventional nonenhanced CT images could be omitted without the risk of missed important finding. This approach could reasonably decrease the radiation dose burden to the patient Purpose: To compare susceptibility weighted imaging (SWI) with dynamic susceptibility contrast (DSC) MR imaging for evaluation of acute cerebral ischaemia at 3.0 T.

A high-resolution SWI and DSC were both performed with a 3.0 T MRI on 34 patients with acute ischzemic stroke within 3 days of symptom onset. For qualitative analysis, MR findings of DSC were grouped into four categories (hyperperfusion, normal perfusion, delayed perfusion, or absent perfusion), while SWI was evaluated according to the changes of signal intensity. Three regions of interest (ROI) were defined by diffusion-weighted imaging abnormalities and mirror regions in the contralateral normal brain tissues, respectively. Signal intensity ratios (infracted region versus the normal contralateral region) were calculated and compared using a paired t test.

Results: SWI showed vein dilatation in affected area in 26 of 34 patients, which were confirmed by DSC with hypoperfusion. In 8 cases, SWI showed normal whlie DSC showed delayed perfusion (n = 2), hyperfusion (n = 3) or perfusion deficit (n = 3). The signal intensity ratio on DSC was 0.82 ± 0.23, while SWI was 0.91 ± 0.26. Both qualitative and quantitative analyses demonstrated that there were no significant difference between the two techniques (P > 0.05). In addition, SWI revealed haemorrhage within ischaemia in 8 cases which were not showed by routine MRI and DSC.

The high-resolution SWI was sensitive in detecting cerebral ischaemic lesions and evaluating cerebral haemodynamics of decreased diffusion, and can be combined with plain MRI to provide meaningful information in the diagnosis of acute cerebral ischaemia.

Isolated cerebellar infarct: imaging findings, aetiopathogenesis and clinical outcome A. Zourla, A. Vemmou, S. Papakostas, M. Tsalikis, P. Stavrakas, A. Chalazonitis;

Purpose: To describe computed tomography (CT) and/or magnetic resonance imaging (MRI) findings in patients with isolated cerebellar infarct (ICI) correlated to aetiopathogenetic mechanisms and clinical outcome. Methods and Materials: From a consecutive series of 568 patients with posterior circulation infarction evaluated prospectively we identified by imaging studies 140 (24.6%) patients with cerebellum involvement. After extensive study, cases were (a) classified according to the aetiopathogenetic mechanisms: atherosclerosis, cardioembolism, infarcts unknown cause (IUC) and other atiologies, (b) categorised based on vessel cerebellum territories and (c) followed up to 5 years regarding mortality and stroke recurrence. Results: We identified 74/140 (53%) cases with ICI. Infarcts distribution was: posterior inferior cerebellar artery (PICA) 23 (31%), anterior inferior cerebellar artery (AICA) 8 (11%), superior cerebellar artery (SCA) 37 (50%) and multiple territories in 6 (8%) cases. Haemorrhagic transformation was more common in SCA infarcts (40.5%) compared to PICA (8.7%) and AICA (11.1%), (p = 0.012), while brain oedema with mass effect was similar among vascular territories (11.9%, 8.7%, and 11.1%, respectively, p = 0.923). The most common causes were different (p = 0.004) among vascular territories: cardioembolism in SCA (60%), atherosclerosis in AICA (67%) and in PICA equal cardioembolism and atherosclerosis (34% and

Evaluation of a novel post-processing filter of native CT images for detection of early ischemic stroke: improved differentiation of early infarct core A. Kemmling, L. Feyen, J. Minnerup, V. Hesselmann, T. Niederstadt, W. Heindel, H. Seifarth; Münster/ DE (akemmling@web.de) Purpose: In non-enhanced CT identification of early infarct in acute ischaemic stroke is limited. This study evaluates the use of a novel post-processing filter for native CT images to detect early ischaemic infarct. Methods and Materials: Non-enhanced CT (NE-CT) was performed in 25 patients with acute (< 6 hours) middle cerebral arterial (MCA) stroke. Standard NE-CT images were processed using an novel edge-contrast enhancement filter (Neuro Best, Siemens). Standard and Neuro Best enhanced scans were evaluated for parenchymal hypoattenuation and loss of gray white-matter distinction (1-5 scale; definite absence to definite presence). Infarct lesions were segmented slice-byslice (fixed window, 10 to 80 HU). Infarct core in admission cerebral blood flow CT perfusion maps and follow-up infarct in NE-CT were segmented. All CT scans and lesion maps were co-registered (Analyze 9.0, AnalyzeDirect). Voxels in standard and Neuro Best CT images were defined as "true positive" (TP), "true negative" (TN), "false positive" (FP) or "false-negative" (FN) by comparison with reference lesion maps (follow-up CT and CTP). Sensitivity (TP/TP + FN) and specificity (TN/ TN + FP) for detection of infarct core was calculated.

Results: With standard NE-CT, sensitivity and specificity for stroke detection were 67% and 98%. Sensitivity increased to 75% using Neuro Best CT images with slight loss of specificity (96%). Receiver operating characteristic analysis of overall image rating revealed an improvement in detecting infarcted tissue using Neuro Best CT post-processing (p = 0.02).

In acute ischaemic stroke, detection of early infarct in NE-CT is facilitated using a novel edge-contrast enhancement filter (Neuro Best). Purpose: Early identification of patients developing malignant ischaemic stroke is crucial. The purpose of this study was to evaluate rater-independent parameters based on precise anatomical volume measurements of infarct, cerebrospinal fluid reserve volume and midline-shift to quantify the risk of malignant infarction.

In 25 consecutive cases with acute proximal middle cerebral arterial occlusion, CT was used to determine cerebrospinal fluid (CSF) reserve volume in the healthy hemisphere normalised to intracranial volume (IV): after automated skull-stripping and thresholding (HU range -10 to 20) of CT images, CSF volume was calculated and normalised to IV (nCSF = volume of CSF/IV). Volume of infarct at admission (reduced CBV in CT perfusion) and stroke volumes (SV) at maximum infarction prior to decompression or death were segmented (Analyze 9.0) and normalised to IV (nSV = SV/IV). Herniation was quantified by volume of midline shift (MSV) normalised to IV (nMSV = MSV/IV). The ratio of nSV/nCSF was correlated with nMSV. A voxel-wise probability map of midline-shift, strokevolume, and CSF in MNI-152 space was created. The cut-off value of nSV/nCSF for malignant infarction was determined by ROC curve analysis.

The ratio of nSV/nCSF significantly correlated with nMSV (r = 0.836, p < 0.001). The cut-off value for malignant infarction was 1.7 (sensitivity 92.3%, specificity 90.9%). The voxel-wise probability map in MNI-152 space allows visual estimation of nMSV with a priori knowledge of nCSF and nSV.

The risk of malignant infarction highly correlates with rater independent parameters of volume of midline-shift and ratio of normalised infarct-volume and CSF reserve volume.

Dual energy computed tomography: the value of the virtual nonenhanced brain images in the assessment of acute stroke J. Purpose: To prospectively determine the accuracy of diffusion-weighted (DW) magnetic resonance (MR) imaging for identifying cancer in the prostate peripheral zone (PZ) in association with T2-weighted and DCEMR imaging before biopsy.

The institutional review board approved this study and all the patient subscribed an informed consent. Twenty-seven patients underwent endorectal MR at 3 T magnet (Discovery M750, GE Healthcare) equipped with surface phased array and endorectal coil. Scan protocol included morphologic imaging with TSE T2-weighted sequences on the axial, sagittal and coronal planes, DWI sequences at different b value (500, 1000, 3000) and dynamic contrast enhanced imaging using GRE 3D T1-weigthed sequence. Image cluster analysis was performed on voxels within the suspected tumour regions. Two readers in consensus recorded the presence of prostate cancer at magnetic resonance imaging and rated the imaging quality of DWI. Results: For all the patients, the DWI sequence images were suitable for the evaluation of the zonal anatomy of the prostate gland and the tumour localisation. In the prediction of prostate tumour foci, we noticed an improvement for tumour detection with a b value of 3000 in comparison with other b value (500 and 1000) with 89% and 87%, respectively, for sensitivity and specificity, and was crucial especially in cases with negative or borderline pattern at DCEMR. Conclusion: DWI is a feasible technique that can be used for the differentiation of malignant and benign tissues in the prostate gland tissue. We obtained a significant improvement using high value of b (1000-3000). Purpose: To evaluate the incremental value of DW-MRI over T2-weighted MRI (T2WI) at 3 T for prostate cancer detection; and to investigate the use of quantitative diffusion parameters to characterise aggressiveness, using whole-mount step-section pathology as the reference standard. Methods and Materials: 51 patients (median age 58 y) underwent prostate T2WI and DW-MRI at 3 T for the assessment of prostate cancer; and were subsequently treated by prostatectomy. Two readers independently evaluated all the studies, and provided a score for the qualitative analysis according to the level of suspicion for the presence of cancer on a 1-5 index scale, first using T2WI alone and then using T2WI and DW-MRI in conjunction. Areas under the curve (AUCs) were estimated to evaluate performance. Quantitative analysis was performed using generalised estimating equations to test the ADC difference between benign and malignant prostate regions, as well as the associations between the ADC and tumour Gleason scores. Results: AUCs for readers 1 and 2 were 0.79 and 0.76 for T2WI and 0.79 and 0.78 for DW-MRI, respectively. There was a significant difference in the mean ADC between malignant and benign prostate regions [1.09 and 1.08 x 10 3 mm 2 /s for readers 1 and 2, respectively, vs 1.81 x 10 3 mm 2 /s for both readers]. A higher Gleason score was significantly associated with a lower mean ADC (p = 0.017) Conclusion: The addition of DW-MRI to T2WI did not significantly improve the readers' performance in detecting prostate cancer. However, quantitative analysis showed that ADC may be a useful tool to detect and establish the aggressiveness of prostate cancer. 31%). After a median follow-up of 30 months (interquartile range 20-60), mortality was 47%, 16%, 11% for SCA, PICA and AICA respectively, while for the same period recurrence rate was 65%, 29% and 62%, respectively. Conclusion: Cerebellar infarct vascular distribution is related to different stroke mechanism and radiological features and is associated with different long-term stroke outcome. The relationship between corticospinal tract injury and motor function outcome was also analysed.

The rFA values significantly decreased with time, while the rADC values significantly increased (p < 0.05). However, there was no significant correlation between the rFA, rADC values and Fugl-Meyer scores. DTT maps showed that if the corticospinal tract was not compressed, motor function was almost intact; if the corticospinal tract was compressed, curved or slightly disrupted, motor function was impaired but improved in a short time; and if the corticospinal tract was mostly or completely disrupted, then motor function was severely impaired and recovered very slowly. Conclusion: rFA and rADC values in brainstem infarction areas evolve dynamically and in regular patterns. There are relationships between the extent of corticospinal tract injury and clinical manifestations and motor function recovery, which can greatly help the evaluation of clinical prognosis.

Abciximab does not prevent ischaemic lesions related to cerebral angiography: a randomised placebo-controlled trial C. Brockmann; Mannheim/DE Purpose: To assess the efficiency of GB IIb/IIIa platelet receptor inhibition by abciximab in the prevention of silent embolism during DSA by diffusion weighted magnetic resonance imaging (DWI).

This randomised, double-blind, prospective study assessed 184 participants with a need for DSA. The participants were monitored by MRI before and after the angiography. DWI was evaluated by two independent neuroradiologists for the occurrence of silent embolism. Results: Using DWI a total of 105 lesions consistent with silent embolism were counted in 31 patients. Fifteen of 90 (16.7%) patients receiving abciximab and 16 of 94 (17.0%) patients in the placebo group were affected by silent embolism. No significant relationship was found between the two groups and the post-angiographic appearance of silent emboli (Wilcoxon two-sample test: p-value 0.9).

Conclusion: The present study shows that IIb/IIIa receptor inhibition by abciximab does not diminish the occurrence of silent infarct during diagnostic catheter angiography. These results challenge the current concept of thrombembolism and air embolism equally contributing to post-angiographic ischaemic cerebral lesions.

Our findings indicate that solid blood clots are not the origin of hyperintense lesions observed on DWI and enhance the role of alternative mechanisms such as air embolism. Silent embolism as a surrogate parameter for ischaemic stroke therefore must be questioned. Purpose: We assessed the ability of magnetic resonance imaging (MRI), proton magnetic resonance spectroscopy (1H MRS) and carbon-11-acetate PET/CT to image localised prostate cancer and detect its aggressiveness. Methods and Materials: Twenty-one patients with untreated prostate cancer, diagnosed using transrectal ultrasound guided biopsy, underwent pelvic carbon-11-acetate PET/ CT and 1.5 T MRI using surface coils. Cancer laterality was based on the percentage of cancer and the highest Gleason score determined from biopsies. 3D 1H MRS of the entire prostate volume was performed. Cramer-Rao minimum variance bound was automatically calculated for each individual metabolite. For each lobe choline + creatine + polyaminesto-citrate (CCP/C) ratios and standardised uptake values (SUV) were obtained from 1H MRS and carbon-11-acetate PET/CT, respectively. The visual and quantitative findings on MRI and PET/CT were compared with cancer laterality and aggressiveness based on the Gleason score, and with PSA velocity and international risk group classification. Results: Using visual analysis, the sensitivity, specificity and accuracy of contrastenhanced MRI were 89%, 29%, 79% and using carbon-11-acetate PET/CT were 80%, 29%, 71%, respectively, on a lobar level. The agreement of visual interpretation, with prostate cancer laterality based on biopsy findings, of contrast-enhanced MRI and carbon-11-acetate PET/CT and was 71%. The dominant lesion CCP/C and SUV values were neither associated with prostate cancer aggressiveness nor did PSA velocity correlate with the SUV or CCP/C values from the entire gland. Conclusion: MRI, 1H MRS and carbon-11-acetate PET/CT enables detection of localised prostate cancer with comparable and limited accuracy but fail to provide information on cancer aggressiveness. 

Purpose: To compare T2-weighted MRI and functional MRI techniques in repeat prostate biopsies. Methods and Materials: Sixty-eight patients with a history of negative biopsies, negative digital rectal examination and elevated PSA were imaged on a 1.5 T MR system using an integrated endorectal-pelvic coil before repeat biopsies. T2-w MRI, diffusion-weighted imaging (DWI), 3D-MR spectroscopy imaging (MRSI) and dynamic-contrast-enhanced MRI (DCE-MRI) were successively acquired in all patients. Interpretation was based on robust literature-adapted dichotomous criteria with cutoffs for DWI and MRSI. Systematic sextant biopsies and additional biopsies were targeted towards areas whose features raised the suspicion of tumour. Results: Prostate cancer was found in 28/68 (41.2%) patients corresponding to 45 segments. In 62 segments suspicious areas were evidenced in locations not sampled by sextant biopsies of which 20 (32%) were found to be positive. The other 25 positive segments were found in the routine eight spots. The probability of a segment being malignant was calculated from logistic regression modelling from the results of T2w-MRI and of functional technique. T2-w-MRI + DWI exhibited the highest positive predictive value (0.52), compared with T2-w MRI alone (0.29), +DCE-MRI (0.33) and +3D-MRSI (0.25). The probability of cancer in a segment varied 12-fold from 63.4% when T2-weighted and diffusion -weighted imaging MRI were both suspicious compared with 5.2% when both were non-suspicious. Conclusion:

The proposed system of analysis and reporting was proven of clinical relevance in the decision process to repeat -or not to repeat -targeted biopsies.

Biexponential apparent diffusion coefficients in different prostate tissue with echo planar imaging (EPI) sequences and high b value X. Liu, L. Zhou, W. Peng; Shanghai/CN (liuxiang_1940@163.com)

Purpose: To investigate the feasibility of biexponential signal decay modelling for diffusion signal decays of different prostate tissues with EPI and high b value. Methods and Materials: Twenty-six patients (11 with prostate cancer, 15 benign) underwent EPI diffusion-weighted MR imaging in which b-factors from 0 to 3000 s/mm 2 in 10 steps were sampled. The acquired signal decay curves were fit with biexponential signal decay functions, the fast ADC (ADCf), fraction of fast component (f) and the slow ADC (ADCs) were measured for the peripheral zone (PZ), central gland (CG), cancer and benign lesions. The data of different tissue were compared with a oneway bonferroni test.

The prostate cancer has the lowest ADCf (1.67 ± 0.18) and ADCs (0.42 ± 0.09) in all the tissue (P < 0.05, Bonferroni), and smallest f (43.3 ± 3.7%)(P < 0.05, Bonferroni). The PZ tissue shows a higher f (66.8 ± 10.4% vs 52.7 ± 4.4%)and ADCs (1.40 ± 0.32 vs 0.75 ± 0.18) than CG (P < 0.05); however, there is no significant difference between the ADCf of the two tissue (3.87 ± 0.42 vs 4.01 ± 0.70, P > 0.05). The benign lesions exhit a lower ADCf (3.2 ± 0.52) than both the normal tissue (P < 0.05), and a lower ADCs (0.75 ± 0.25) and f (57.8 ± 2.6%) than PZ; however, no significant different between the CG and benign in the ADCs and f.

The biexponential apparent diffusion coefficients with EPI at high b value provide additional, unique tissue characterisation parameters for prostate tissue and could be practical in clinical use.

Whole-torso MRI including DWI for patients with recurring prostate cancer: technical feasability and assessment of lesion conspicuity in DWI M. Eiber, K. Holzapfel, E.J. Rummeny, A.J. Beer; Munich/DE Purpose: In a technical development study the principal methodological aspects of whole-torso magnetic resonance imaging (MRI) including diffusion-weighted imaging (DWI) with background-suppression using a time optimised protocol for restaging of prostate cancer patients were evaluated. Methods and Materials: Twenty-two patients underwent MRI at 1.5 T from the base of the skull to the proximal thigh using axial T1w, T2w STIR and DWI (b-values: 50 and 500 s/mm 2 ) and sagittal T1w and T2w STIR of the spine. ADC-values of liver, spleen, kidney, muscle and bone were measured. Image quality in DWI was assessed using a scale from 0 to 9. Contrast-to-noise ratios (CNR) of lymph node and bone metastases were determined in T1w, T2w STIR and DWI. Bone metastases were further subclassified according to their Hounsfield units (HU) in CT. Results: Mean acquisition and mean room times were 66:52 min and 77:34 min, respectively. ADC values of normal organs showed good concordance with reported data. Good to excellent image quality was observed for DWI (mean scores B A S203 C D E F G H experienced breast radiologists reviewed the MR images according to the MR-BI-RADS lexicon. Surgical excision was recommended regardless of the imaging findings. MR findings ("non suspicious lesions": BI-RADS 1 and 2 categories, versus "suspicious lesions": BI-RADS 3, 4 and 5 categories) were correlated with histologic results on surgical specimen. Overall and by-lesion sensitivities, specificities, positive (PPV) and negative (NPV) predictive values of MR in predicting presence of malignancy were calculated. Results: Lesions were categorized as follows: 63 BI-RADS 1.13 BI-RADS 2.42 BI-RADS 3.54 BI-RADS 4 and 3 BI-RADS 5. Surgical excision demonstrated 26/175 malignancies, corresponding to an overall CNB underestimation rate of 14.8% (10% for papillomas; 4% for RSL, 28% for ADH, 32% for LN). MR provided an overall sensitivity of 81%, specificity of 48%, PPV of 21% and NPV of 93%. By-lesion analysis demonstrated the following specificities, sensitivities, PPV and NPV: 100%,29%,14%,100%, respectively, for papillomas; 50%,55%,4%,97%, respectively, for RSL; 60%,46%,30%,75%, respectively, for ADH; 83%,80%,67%,91%, respectively, for LN. Conclusion: MR demonstrated high NPV for malignancy in case of B3 lesions associated with low underestimation rates at CNB (papillomas and RSL). In cases of non-suspicious MR findings, follow-up rather than surgical excision might be recommended for those lesions.

Purpose: To assess whether MRI-BIRADS lesions features can predict the probability of malignancy for high-risk breast lesions diagnosed at core-needle-biopsy (CNB). Methods and Materials: Retrospective review of MRI findings in 138 high-risk breast lesions -71 benign papillomas (PAP), 30 radial sclerosing lesions (RSL), 23 lobular neoplasias (LN), 14 atypical ductal hyperplasias (ADH) -diagnosed at imaging-guided CNB was performed. MRI-BIRADS morphologic and kinetic features were compared with histology at surgical excision (SE) in 117 (85%) and with imaging follow-up (> 24 months) in 21 (15%) lesions. Chi-square test was used for statistical analysis (statistically significant: p-value < 0.05). Results: 23 (17%) lesions were upgraded to malignancy at SE (seven papillomas, one RSL, eleven LN, four ADH): 9 (39%) were invasive carcinomas and 14 (61%) were ductal carcinomas in situ. Lesion-specific underestimation rates were: benign papillomas 10% (7/71); RSL 3% (1/30); LN 48% (11/23); ADH 29% (4/14). Lesions classified as nonmasslike enhancement (NMLE) were significantly more likely to be malignant (36%) than those described as masses (13%) [p = 0.01]. Lesions size was not a significant predictor of malignancy[p = 0.19]. For papillomas, RSL and ADH presenting as masses, BI-RADS margins, internal enhancement pattern and kinetic characteristics did not help to predict presence of malignancy. For LN, risk of underestimation was significantly higher for lesions classified as 'suspicious masses' according to Fischer score (7/8 [88%]) rather than "non-suspicious" (1/7 [14%]) [p = 0.01] and for lesions > 10 mm (5/5 [100%]) rather than ≤10 mm (3/10 [30%]) [p = 0.04]. Conclusion: Among MRI features, NMLE in any B3 lesion and 'suspicious mass' in the case of LN are highly predictive of malignancy. MRI finding of focus or foci (< 5 mm) does not exclude presence of malignancy.

N.K. Singh, P. Crystal, S. Kulkarni; Toronto, ON/CA (drnitinsingh@hotmail.com) Purpose: To evaluate outcome of otherwise occult likely benign breast lesions detected on screening MRI in asymptomatic high-risk women. Methods and Materials: Retrospective review of breast MRI database performed to identify all nonpalpable, mammographically & sonographically occult, MRIdetected likely benign lesions in asymptomatic women without prior history of breast carcinoma. These BI-RADS 3 "MRI-only" lesions were followed from 18 to 77 months (mean 28 months) and rate of malignancy was calculated. Characteristics of probably benign lesions on screening breast MRI were analyzed for predictors of malignancy. Results: 131 lesions were reported as likely benign in 114 patients out of 695 screening breast MRI studies. These included 42 (37%) foci, 48 (42%) non-mass like enhancement and 41 (36%) masses. Eighteen lesions were biopsied during follow-up period due to lesion progression on follow-up MRI or for other reasons. Malignancies were found in four women, constituting 22% of the lesions that underwent biopsy and 3 % of probably benign lesions. Tumor histology was invasive Methods and Materials: A total of 143 consecutive patients with a first negative random TRUS-guided prostate biopsy with elevated PSA levels (total PSA between ≥4 ng/ml and < 10 ng/ml) underwent MR examination on 3 T magnet (Verio, Siemens, Germany) equipped with surface phased array and endorectal coil. Scan protocol included morphologic imaging with TSE T2-weighted sequences on three planes, spectroscopic imaging with 3D CSI sequences and dynamic contrast enhanced imaging using 3D FLASH T1-weigthed sequence. The second prostate biopsy was performed no later than 2 weeks from PCA3 test and MR examination. Results: For PCA3 test, the total number of urinary sediments that could be analysed successfully was 95.3% (141 on 143 specimens) as for MR, the examination and the following comparison with biopsy results were possible in all patients (143/143, 100%). The overall sensitivity and specificity of PCA3 in this cohort were 76.9% and 66.6%, respectively, with a PPV of 80% and a NPV of 62.5%. Sensitivity and specificity, for combining MRI, MRS and DCEMR, were respectively, 92.8% and 86.6% with a positive predictive value (PPV) of 92.8% and a negative predictive value (NPV) of 86.6%. Conclusion: The association MRSI/DCEMR in comparison with PCA3 results shows a better detection of prostate cancer foci in patients with prior negative TRUS-guided biopsy and alterated PSA serum levels.

TRUS-MRI image fusion prostate biopsy in patients with no previous biopsy and a PSA level between 4 and 10 ng/ml F. Cornud, D. Eiss, N. Barry de Longchamps; Paris/FR (francois.cornud@imagerie-tourville.com) Purpose: To show the feasibility of TRUS-MRI image fusion prostate biopsy and to compare the positive biopsy rate and the prognosis value of TRUS-MRI fusion biopsy with those of a 12 cores sextant protocol. Methods and Materials: 72 consecutive patients (mean age 64 ± 6) with a PSA level between 4 and 10 (mean 6.7 ± 6) and no palpable abnormality at digital rectal examination underwent an endorectal multiparametric MRI combining T2-weighted, diffusion-weighted and dynamic contrast enhanced MRI. During TRUS, a dedicated electromagnetic device (virtual navigator, ESAOTE, Italy) permitted a dynamic overlay of TRUS and MR images when a suspicious area, defined by a multiparametric MRI score, was detected, allowing for stereotaxic biopsies. Four TRUS-MRI fusion biopsies were performed in the target, followed by the sextant protocol. Results: TRUS-MRI image fusion was possible in every case. Both targeted and sextant biopsies were positive in 28 patients (28/72, 39%). Fusion biopsies were solely positive in 12 patients (12/72, 17%), concerning tumours originating in the anterior part of the transition zone in 10/12 cases (83%). No patient had solely positive sextant biopsies. Increased cancer detection rate by fusion biopsies (40/72, 55.5%), approached significance (p = 0.057). No difference in Gleason score was observed between fusion and sextant biopsies, but the mean cancer length per core was significantly higher on fusion biopsies (5.9 mm ± 2.5) than that of sextant biopsies (4.3 mm ± 2.7, p = 0.02). Conclusion: TRUS-MRI image fusion is a feasible technique which significantly increases the assessment of tumour burden. It questions the pertinence of the indication of MRI prior to biopsy. lesions (evident on MR images, non-palpable, first-look ultrasound and mammographically occult) were recommended for second-look ultrasound. The retrospective evaluation included second-look ultrasound identification according to lesion's location (ipsilateral -in the same/different quadrant -controlateral), distance from the primary lesions (< 4 cm/> 4 cm). Results were correlated with pathological findings. Results: Of 50 lesions depicted at second-look ultrasound, 26 lesions were ipsilateral to the known cancer and 24 were controlateral. Overall, 16/26 (61%) ipsilateral lesions were malignant. Of the 26 ipsilateral lesions, 12 (24%) were in the same quadrant as the index cancer, of which 8 (66%) were malignant (4-ILC, 3-IDC, 1-DCIS) and 4 were benign. The remaining 14/26 (28%) ipsilateral lesions were in different quadrant from the known cancer; 8 (57%) were malignant (5-ILC, 3-DCIS) and 6 were benign. Of the 24 controlateral lesions to known cancer, 4/24 (16%) were malignant (2-ILC, 2-IDC).

Overall, of the 20/50 (40%) malignant lesions identified by second-look ultrasound, 13/20 (65%) were within 4 cm distance from the primary (p < 0.05). Conclusion: Second look ultrasound identified 71% of incidental breast MRI lesions in newly diagnosed breast cancer patients, of which 40% proved malignant. The majority (65%) of additional tumour foci are in the same quadrant and within 4 cm of the index cancer. At a minimum, when performing ultrasound in a patient known to have cancer, thorough evaluation of at least that quadrant of the breast is recommended.

Breast MRI contribution to the study of papillary lesions M. Vaño, J. Camps Herrero, M. Forment Navarro, V. Ricart Selma, J. Cano Gimeno, C. Martínez Rubio; Alzira/ES (mavanyo@hospital-ribera.com)

Purpose: To evaluate the diagnostic performance of breast MRI in the evaluation of papillary lesions and to assess the influence of this technique on the therapeutic approach.

In a retrospective study of breast MRI exams performed from January 2003 to July 2010, three groups of patients were included: patients with nipple discharge (n = 40), patients with a radiological suspicion of a papillary lesion (n = 28) and patients with a retrospective pathological diagnosis of a papillary lesion (n = 16). Enhancing lesions associated with a dilated duct or cyst, lesions located in the nipple-areolar complex or with a segmental distribution were considered suspicious. Lesions were classified according to BI-RADS categories and the gold standard was pathological results after surgery. Results: Of the 84 patients studied with MRI, 30 had a negative MRI study for a papillary lesion (with a mean follow-up of 20 months) and in 54 cases findings consistent with papillary lesions were diagnosed. They were pathologically confirmed: 17 intracystic papillomas (31'5%), 15 papillary carcinomas (27'7%), 7 papillary carcinomas in situ (13%), 9 diffuse papillomatosis (16'6%) and six false positives. In 27 cases (26'2%), MRI provided new information that changed therapeutic approach. Sensitivity was 87.9%, especificity 77%, NPV was 90.9% and PPV was 70%. Conclusion: MRI helps to characterise not only papillary lesions but also allows us to identify other lesions not seen with conventional techniques and has a high NPV. New findings identified with MRI are important because they may change the therapeutic approach in 1 out of every 4 patients.

Kinetic analysis of non-mass-like enhancing lesions on breast MRI using manual and computer-assisted methods Purpose: To analyse the kinetic characteristics of non-mass-like enhancing lesions in dynamic breast MRI using manual and computer assisted methods.

The enhancement pattern of 82 histopathologically verified non-mass-like enhancing lesions (36 malignant, 46 benign) was evaluated on breast MRI using manual placement of a region of interest (ROI) and commercially available computer analysis software. The latter can automatically assess volume enhancement characteristics of a lesion voxelwise. Kinetic features evaluated included classification of the signal-intensity time curve as washout, plateau or persistent enhancement. A washout enhancement pattern is regarded as suggestive for malignancy.

ductal carcinoma in two patients and ductal carcinoma in situ in two patients. All four malignancies were diagnosed in enhancing masses (10% of enhancing masses).

No cancers were diagnosed in subgroup of enhancement foci. Conclusion: A probably benign interpretation was given to 16% of high-risk women at their breast MRI screening examination. In 3% of women with likely benign lesions, breast carcinoma was subsequently diagnosed in an area initially judged to be probably benign. Additional studies are required to refine criteria for probably benign lesions on breast MRI. Purpose: To provide patient-specific estimates of doses and theoretical risks of radiation-induced cancer for patients undergoing 256-slice CT colonography (CTC) screening. Methods and Materials: Patient-specific voxelized phantoms were generated using 256-slice CTC images of 8 males and 13 females, while corresponding dose-images were produced by employing a novel Monte Carlo CT dosimetry software. A low-dose screening, CTC 120 kVp protocol was simulated with prone and supine imaging at 30 and 50 mAs, respectively. Tube load was adapted depending on patient body-size. Normalised (to free-in-air CTDI) organ dose data were derived for all primarily irradiated radiosensitive organs, i.e. 11 main organs and 11 remainder organs. Organ doses were determined for each patient for the specific exposure parameters. Effective dose was estimated for each patient using 2007 ICRP recommendations. Risks of radiation-induced cancer were estimated using appropriate sex-, age-and organ-specific cancer risk factors. Individualised total cancer risk was estimated by adding respective organ risks.

The higher doses were received by colon, liver, stomach and bladder. Mean effective dose was estimated to be 1.9 and 2.1 mSv for male and female patients, respectively. The mean total theoretical life-attributable risk factor for cancer induction following screening 256-slice CTC was estimated to be 20 x 10 -5 and 17 x 10 -5 , for female and male individuals, respectively. Conclusion: For a typical patient cohort undergoing low-dose screening CTC on a 256-slice CT scanner the total theoretical risk for radiation-induced cancer is as low as 1 in 5000 and 1 in 5900 for females and males, respectively.

Low dose CT colonography: feasibility of using iterative reconstruction algorithms for improved 3D interpretation of polyps D. Mehta, S. Virmani; Highland Heights, OH/US (dhruv.mehta@philips.com)

Purpose: 3D visualisation is used as primary interpretation method for CT colonography. Low-dose acquisition techniques are prone to significant artefacts particularly in the pelvis region. Previously published literature indicates high prevalence of polpys in the rectum. The purpose of this study was to assess the image quality impact of iterative reconstruction techniques on 3D interpretation of polpys. Methods and Materials: Randomised selection of 30 low-dose CT colonography datasets was performed from the database. Selection criteria included scanner type (64-slice CT), acquisition protocol (120 kV, 15 mAs), and BMI > 25 < 30.

Reconstructions were performed using conventional filtered-backprojection (FBP) and an iterative reconstruction technique (iDose, Philips). Qualitative image quality assessments were performed relative to FBP [scale: 1 (worse than FBP) -5 (superior to FBP)] by two blinded readers on a per-segment basis (rectum, sigmoid, descending, transverse, ascending and caecum) using a 3D workstation (EBW, Philips). Quantitative noise measurements were performed in a ROI placed in the air-pockets of each segment. Statistical differences were tested using wilcoxon signed-rank test and t-test. Results: Unlike manual ROI placement, computer-aided analysis demonstrated a significant difference in enhancement pattern between benign (washout: 32.6%, plateau: 32.6%, persistent: 34.8%) and malignant lesions (77.1%, 8.6%, 14.3% respectively, P < 0.01, two-sided Chi-squared test) following initial rapid signal increase. Mean percentage of washout voxel volumes within a lesion was significantly higher in malignant lesions than in benign lesions (11.9% ± 12.7 (SD) vs. 6.9% ± 11.3 (SD), P < 0.01, Mann-Whitney U test). Conversely, the mean percentage of persistent voxel volumes was significantly lower in malignant lesions than in benign lesions (60.1% ± 21.1 (SD) vs. 79% ± 23 (SD), P < 0.01, Mann-Whitney U test). Conclusion: Computer-assisted enhancement pattern analysis might have diagnostic benefit in the evaluation of non-mass-like enhancing lesions. Average time interval for follow-up was comparable between CTC and OC (9.00 ± 4.24 vs 9.00 ± 4.39 months, respectively; p = 0.7188). No patients chose to undergo a 3-year follow-up with either CTC or OC. Most patients elected to have follow-up with either CTC or OC before 18 months rather than later (Chi 2 = 12.66, p = 0.0004, and Chi 2 = 12.53, p = 0.0004, respectively). Between the two techniques, no statistically significant difference was found between the ratios of patients undergoing follow-up before 18 months and from 18 months onwards (Chi 2 = 0.01, p = 0.9203). Conclusion: A substantial fraction of C-RADS2 patients prefer to undergo immediate OC and polyp removal rather than follow-up (either with CTC or OC), and the majority of those accepting follow-up are willing to wait for less than 18 months. Such findings may suggest a revision of the proposed C-RADS2 category.

Digital colonic cleansing of CT colonography images: influence on quality of 2D and 3D reconstructions P. Vagli, L. Faggioni, E. Picano, R. Scandiffio, A. Mantarro, P. Bemi, L. Cini, E. Neri, C. Bartolozzi; Pisa/IT (paolavagli@yahoo.it)

Purpose: To evaluate the performance of digital colonic cleansing of CT Colonography (CTC) datasets with faecal tagging and its impact on quality of 2D and 3D post-processing images.

We reviewed 26 cases of CTC positive for small polyps confirmed by conventional colonoscopy, in which patients underwent a 3-day-noncathartic preparation and a same day faecal tagging protocol. Digital cleansing was performed using a dedicated plug-in embedded in the ColonVCAR software (GE Medical Systems, Milwaukee, WI). Two radiologists experienced in CTC, who were unaware of conventional colonoscopic findings, analysed CTC datasets in 2D mode, as well as in 3D mode using Virtual Endoscopy, with and without digital bowel cleansing. CTC image quality in terms of severity of digital subtraction artefacts in 2D mode before and after cleansing was estimated visually in consensus on a per-segment basis using a five-point scale (1=poor through 5=excellent), while 3D images with and without cleansing were classified as either diagnostic or not diagnostic. 

Computer-aided detection in CT colonography (CTC): which CAD paradigm is best in a screening population? G. Iussich, L. Correale, C. Senore, D. Campanella, G. Galatola, C. Laudi, P. Falco, N. Segnan, D. Regge; Turin/IT Purpose: To prospectively compare the diagnostic performance and time efficiency of primary and second reader CAD paradigms. Methods and Materials: Individuals participating to a CRC screening program and with a positive FOBT test were recruited for same day CTC and CC. Two experienced radiologists independently analysed the CTC studies following randomisation, using CAD either as a first (CAD1) or as a second reader (CAD2); levels of confidence were assigned to positive findings. Reporting time, per-patient sensitivity for patients with adenomas or cancer of at least 6 mm in size, specificity, along with the relative 95% confidence intervals, and areas under ROC curves (AUC) were calculated for both reading paradigms. Colonoscopy (CC) and histology were the reference standard. Results: Seventeen of the 199 enrolled individuals (8.5%) were excluded for: refusal to perform CC (3), unavailable histological data (4), protocol violation (3) and non-diagnostic image quality (7). The remaining 182 individuals included 93 patients (51%) with at least one adenoma 6 mm or more. Sensitivity with CAD2 and CAD1 was 86% (80/93) (95% CI: 77-92%) and 89% (83/93) (95% CI: 81-95%), respectively (P = 0.5). There was no difference in specificity between CAD2 and CAD1 [90% (80/89) (95% CI: 81-95%) versus 91% (81/89) (95% CI: 83-96%)]. The mean AUCs for CAD2 and CAD1 were similar (P = 0.09). CAD1 reading took 2 minutes less than CAD2 (6 m: 2 s ± 55 s versus 8 m: 7 s ± 30 s, P = 0.008). Conclusion: CAD1 is more time efficient and has similar diagnostic performances to CAD2 and should be considered for future mass screening programs, where cost-effectiveness may represent a key issue.

Diagnostic value of CT-colonography as compared to colonoscopy in an asymptomatic screening population: meta-analysis M.C. de Haan 1 , R.E. van Gelder 1 , A. Graser 2 , S. Bipat 1 , J. Stoker 1 ; 1 Amsterdam/NL, 2 Munich/ DE (m.c.dehaan@amc.uva.nl) Purpose: Previous meta-analyses on computed tomography (CT)-colonography included both average risk and high risk individuals, which may overestimate the diagnostic value in screening. A meta-analysis was performed to obtain the diagnostic value of CT-colonography for screening. Methods: A comprehensive search (until July 2010) was performed using the MEDLINE, EMBASE and Cochrane databases. Initial selection of articles and critical appraisal was done by two reviewers using QUADAS. Inclusion criteria were: prospective, randomised trials or cohort studies comparing CT-colonography with colonoscopy (≥50 participants), ≥95% asymptomatic average risk participants ≥50 years. Study-design, participant characteristics, CT-colonography techniques and 2 x 2 contingency tables were recorded. Sensitivity and specificity estimates were calculated on per patient basis and summary sensitivity on per polyp basis, using bivariate and univariate analyses. Conclusion: Compared to colonoscopy, CT-colonography is highly sensitive for colorectal cancer. CT-colonography is less sensitive for clinically relevant (advanced) adenomas ≥6 mm; however, its detection rate is acceptable for screening purposes especially when the attendance is higher than for colonoscopy.

Clinical results of a five-year CT colonography screening programme E. Picano, E. Neri, L. Faggioni, P. Vagli, C. Bartolozzi; Pisa/IT Purpose: To analyse the diagnostic results of the CT Colonography (CTC) screening programme at our Institution over a five-year period.

We retrospectively reviewed 1894 screening CTC examinations performed at our institution from January 2003 to March 2010 on asymptomatic patients. Out of them, 810 had been obtained on a 4-row CT scanner and the remaining 1084 on a 64-row CT scanner using a low-radiation dose protocol in the supine and prone positions and a same-day preparation with iodine faecal tagging (500 mL of water and 50 mL of diatrizoate dimeglumine). Lesions were classified according to the C-RADS system (C0 = not diagnostic, C1 = unremarkable, C2 = polyps sized 6-9 mm, C3 = polyps over 10 mm, C4 = colonic masses over 3 cm). Results: Overall, 249/1894 examinations were classified as C0 (13.1%), 1338 as C1 (70.7%), 142 as C2 (7.5%), 95 as C3 (5.0%) and 70 as C4 (3.7%). C0 cases were not diagnostic mainly due to the persistence of marked faecal residues. The greater axis of polyps was 7.6±1. 

Endoscopically validated CTC datasets, containing 8 lesions reported by experienced readers in a prior study were selected (mean polyp size = 10 mm; range 5-25 mm), including true positive and false positive lesions. Video clips (mean = 26s) incorporating these lesions were interpreted by faculty members (n = 10) or participants (n = 19) selected from 3 CT colonography workshops. Case order and polyp appearance time were randomised. Visual search was recorded using a Tobii X120 eye tracker and observers indicated potential polyps with a mouse click. Results: Visual search patterns were recorded satisfactorily. Of 5 true positive lesions, novice readers indicated a median of 3 lesions (range 1-5) and experienced readers indicated a median of 4 lesions (range 2-5), giving good accuracy calibration data for the development of new 3D vision measures. Both novices and experts identified true positive lesions at a median of 3.3 seconds after the lesion was first visible, although times varied by lesion. Conclusion: Visual tracking during 3D interpretation is technically feasible. Experienced readers on average identified more lesions, although both novices and experts took on average similar times to indicate polyps. Further studies will assess the effect of training on visual search pattern. 

We developed a dose adaptation strategy modulated to the patients AP chest diameter (APD) with the goal to minimise variation in image quality while ensuring the highest possible vascular attenuation. Methods and Materials: 92 patients (mean age 56 year, BMI 27.7 ± 4.4 kg/sqm) in two institutions underwent prospectively triggered 320-slice single-beat cardiac CT. Tube voltage was 100 kV in patients with APD up to 25 cm, 120 kV between 26 and 30 cm, and 135 kV with APD greater than 30 cm; mA was gradually adjusted to the patient's APD within the kV settings. All other imaging parameters were kept constant including scan-field-of-view, slice thickness, reconstruction method and kernels, and contrast material injection volume and flow. Image noise was defined as standard deviation of attenuation and measured using a ROI in the descending aorta. Variation in image noise, aortic attenuation, signal-to-noise ratios were calculated and statistically compared between the three kV settings. Results: Average pixel noise (SD) for 100, 120, 135 kV scans were 29.35 (4.23), 27.7 (3.89), 27.9 (3.98) HU, respectively, without statistical difference. Vascular attenuation (SD) were 488.7 (99.5), 400.0 (67.8), 361.2 (45.9) HU, respectively, and signal-to-noise ratios (SD) were 16.73 (3.07), 14.70 (3.01), 13.24 (3.16), respectively, all significantly different. Conclusion: Tube output modulation with stepwise increase of kV and gradual adaptation of mAs based on patients' APD is an effective means to reduce variation in image quality in CT coronary angiography whilst assuring the highest possible vascular attenuation in patients with varying body habitus.

Economic analysis of the use of coronary calcium scoring as an alternative to stress ECG in the non-invasive diagnosis of coronary artery disease V. Raman 1 , E.T.M. McWilliams 2 , S.R.M. Holmberg 1 , K. Miles 1 ; 1 Brighton/UK, 2 Hastings/UK (ramanvivek@doctors.org.uk)

Purpose: A publication by National Institute of Health and Clinical Excellence (NICE) recommends that in low pre-test probability population suspected of coronary artery disease (CAD), a zero coronary calcium score (CCS) excludes CAD. This study evaluates this recommendation by a) undertaking a meta-analysis of studies involving CCS and CAD and b) using the meta-analysis to inform an economic evaluation of a range of diagnostic pathways that use CCS as an alternative to stress-ECG (sECG). Purpose: To compare the detection performance of a context-sensitive computeraided detection (CS-CAD) scheme with that of conventional polyp-detection CAD in detecting flat lesions in CT colonography (CTC). Methods and Materials: 105 colonoscopy-confirmed lesions ≥6 mm categorised as "flat" in cathartic or non-cathartic multi-center CTC studies were measured retrospectively for their maximum height in 2-D multi-planar views. The lesions with maximum height of < 3 mm were considered as flat lesions for this study. A fully automated CS-CAD scheme was developed to detect colorectal lesions based on context-sensitive algorithms and hyper-shape features. The detection performance of the CS-CAD and conventional polyp-detection CAD were compared with the flat-lesion patients and randomly sampled normal patients. Per-lesion sensitivity, false-positive (FP) rate, and normalised area under free-response operating characteristic curve (AUC) were assessed by the use of leave-one-patient-out evaluation and McNemar's test. Results: There were 127 patients (male: 52%; female: 48%; mean age: 56 years) from 20 institutions: 109 were normal and 18 had 20 flat lesions (≥10 mm: 10; 6-9 mm: 10; 3 cancers, 13 adenomas). The use of CS-CAD increased AUC from 0.28 to 0.92. For flat adenomas/carcinomas, CS-CAD detected 100% (100%) of lesions ≥10 mm (6-9 mm) with 0.8 (3.1) FPs per patient on average, whereas conventional polyp-detection CAD detected only 63% (13%) of lesions ≥10 mm (6-9 mm) with 1.4 (4.8) FPs per patient. The improvement in detection sensitivity by CS-CAD was statistically significant (p < 0.001). Conclusion: CS-CAD yields superior performance over conventional CAD in detecting flat lesions < 3 mm in maximum height. Purpose: Despite colonic cleansing prior to CT colonography, prone and supine acquisitions remain necessary to differentiate pathology from luminal content. Observers must therefore match endoluminal locations between prone and supine studies, a process that is complicated by considerable colorectal deformation. We developed and validated a novel algorithm that establishes endoluminal spatial correspondence between prone and supine CT acquisitions. Methods and Materials: Ethical approval was obtained to use anonymised CT colonography data from 24 patients randomised into 12 development (13 polyps) and 12 validation sets (13 polyps). A radiologist (experience > 500 validated datasets) indicated the polyp volumes in both prone and supine CTC scans using separate multiplanar reformats. These polyps were masked from the algorithm so as not to influence registration and used as reference points for validation. To perform registration, a triangulated mesh of the endoluminal colonic surface was first extracted from segmented colonic data. Surface parameterisation was then achieved using the Ricci flow method followed by non-rigid B-spline registration in the 2 dimensions of colon length and circumference to recover the deformations. Results: Using the development set, 8 of the 13 polyps were registered successfully (62%) with a mean error of 6.6 mm (+-4.8). In the validation set, 10 of the 13 polyps were registered successfully (77 %) with mean error of 6.7 mm (+-4.6) Conclusion: This novel algorithm allows accurate registration of prone and supine datasets, and should considerably simplify the interpretative task in CT colonography. It additionally has the potential to improve computer-aided detection. Purpose: Diagnostic accuracy can vary between observers because of differences in ability to detect and characterise lesions. Eye tracking technology has been used to explore observers' visual search during plain radiographic interpretation but has not been applied previously to complex 3D tasks. We compared visual search strategies of novice and experienced readers interpreting CT colonography.

Using DECT as the first-line imaging modality for myocardial perfusion assessment for known or suspected CAD has the potential to increase QALYs at lower cost compared with SPECT. Prospective studies are needed to verify these findings.

The impact of NICE guidelines for the investigation of chest pain on outpatient radiology services in the UK C.M. Patterson, E.D. Nicol, L. Bryan, T. Woodcock, D. Bell, S.P. Padley; London/UK (cmpatterson@doctors.net.uk)

Purpose: The National Institute for Health and Clinical Excellence (NICE) have released guidelines for the investigation of chest pain of recent onset. There is concern that the guidelines will increase the burden on imaging services, requiring service reconfiguration and investment (2, 3). This study was performed to assess the impact of the guidelines on outpatient radiology services in the UK. Methods and Materials: 595 consecutive patients attending chest pain clinics at two hospitals over six months preceding release of the NICE guidelines (51% male; median age 55 yrs (range 22-94 yrs)) were risk stratified using NICE criteria. Preliminary cardiac investigations recommended by NICE were compared with existing clinical practice and the relative costs calculated. Results: NICE would have recommended 443 patients (74%) for discharge without cardiac investigation, 10 (2%) for cardiac computed tomography (CCT), 69 (12%) for functional cardiac imaging and 73 (12%) for invasive coronary angiography (ICA). Relative to existing practice there would have been a trend towards reduced functional cardiac imaging (-24%; p = 0.06) and increased CCT (+43%; p = 0.436) but a significant increase in ICA (+508%; p < 0.001). The cost of investigations recommended by NICE would have been £15,881 greater than existing practice.

This study suggests implementation of the NICE guidelines will require investment in imaging services, particularly ICA. It will be necessary to establish and maintain CCT for relatively few patients; also to establish and maintain functional cardiac imaging even though referrals are likely to decline. Individual hospitals should assess their local cardiac populations prior to service reconfiguration. .5 and 27.5 (n = 53) and > 27.5 kg/m 2 (n = 13). IQ was assessed per-segment by two observers independently using a 4-point scale (1 -excellent, 4 -undiagnosable). Image noise and signal-to-noise ratio (SNR) at aorta root were measured. The effective dose was calculated using conversion factor of 0.017. Results: Diagnostic IQ was obtained in 503 of 507, 687 of 668, 164 of 167 coronary segments in 80, 100, 120 kV groups without significant difference (p = 0.482). The proportions of segments with score 1 to 4 were not significantly different among three kV groups (all p > 0.05). Image noises were significantly higher in 80 kV group than 100 and 120 groups (P < 0.001), while SNR was not (p = 0.097). The effective dose of 80 kV group (0.36 ± 0.03 mSv) was significantly lower than that of 100 kV group (0.86 ± 0.08 mSv) and 120 kV group (1.77 ± 0.18 mSv). Conclusion: MDCT-CA is the reference method for the non-invasive exclusion of critical coronary stenosis. Up to a low-medium CAD risk, the MDCT-CA diagnostic workup is the most cost-effective protocol, being superior to the traditional exercise ECG-based protocol.

Cost-effectiveness and lifetime benefits of substituting dual-energy CT for SPECT in the assessment of myocardial perfusion in patients with suspected or known coronary artery disease M. Meyer 1 , M. Weininger 2 , U.J. Schoepf 2 , G. Rowe 2 , J.W. Nance 2 , D. Kang 2 , S.A. Chiaramida 2 , C. Fink 1 , T.

Purpose:To compare the cost-effectiveness and potential lifetime benefits of replacing SPECT with dual-energy computed tomography (DECT) for myocardial perfusion assessment in coronary artery disease (CAD).

A decision and micro-simulation model was developed to estimate the health effects of using DECT myocardial perfusion imaging instead of SPECT for identifying patients in need of revascularisation. The model was based on the performance of stress/rest DECT and stress/rest SPECT for detecting myocardial perfusion deficits in 50 patients (13 female; 63±11 years) with known or suspected CAD. Stress/rest perfusion cardiac MRI with delayed enhancement and decision to revascularise served as the reference standards. For DECT, a reimbursement of $1700 was assumed. All other actual healthcare costs in these patients were derived from our hospital billing system using Cranware software. Results: Compared with MRI, DECT (versus SPECT) had 90% (85%) sensitivity, 71% (58%) specificity and 87% (79%) accuracy for identifying patients with obstructive CAD in need of revascularisation. Compared with no-imaging and no-treatment strategies, routine SPECT gained 13.49 quality-adjusted life-years Purpose: Cardiac MRI (CMR) is often used to distinguish physiologic cardiac adaptation in athletes from cardiomyopathies when screening remains inconclusive. Earlier research, both with echocardiography and CMR, is ambivalent on the nature of cardiac changes in strength athletes. In an attempt to clarify the extent of ventricular adaptations in strength athletes we investigated strength athletes either admitting or denying the use of anabolic androgenic steroids (AAS) and compared them with non-athletes and athletes practising combined strength-endurance sports. Methods and Materials: 150 men aged 18-62 years underwent CMR (mean age 28 ± 8 years): 52 non-athletes (0-3 hrs/wk training), 52 strength-endurance athletes and 46 strength athletes (> 6 hrs/wk training) of which 20 admitted and 26 denied the use of AAS. Blinded observers experienced in cardiac MRI used a reproducible contour tracing protocol for the assessment of volumes, function, and wall mass. Results: Non-using strength athletes show little difference in cardiac dimensions as compared to non-athletes (LV volume 103.5 vs. 100.9 ml/m 2 , LV wall mass 48.2 vs 47.0 g/m 2 ). Steroid-using strength athletes show an increased LV volume (118.5 ml/m 2 ) and LV wall mass (60.8 g/m 2 ), not exceeding dimensions found in strength-endurance athletes (127.5 ml/m 2 and 65.9 g/m 2, respectively). Ventricular and septal diameters show similar results, but maximum septal wall thickness is higher in steroid-using strength athletes than in strength-endurance athletes (4 chamber/short axis view 12.5/12.0 vs. 11.0/11.1 mm). Conclusion: Left ventricular dimensions in steroid-using strength athletes are increased as compared to non-using strength athletes and non-athletes, but do not exceed those of strength-endurance athletes. Purpose: To evaluate non-enhanced-MRA (NE-MRA) acquired with a 3D Turbo Spin Echo sequence with ECG triggering versus DSA as reference standard in symptomatic PAOD patients. Methods and Materials: 23 PAOD patients with clinical indication for DSA-guided intravascular intervention underwent NE-MRA 1day after intervention (native SPACE, Magnetom Espree 1.5 T, Siemens). Two blinded readers (r1, r2) scored 66 preselected stenoses-grades (StG) from 1 (< 10%) to 4 (> 90%).The intervetionalist scored the same preselected stenosis-grades from DSA as goldstandard (DSAStG). Segmentwise comparison consisted of comparison between MRA and DSA in 95 vessel segments from 1 (good quality) to 4 (inadequate). Results: 3 datasets were excluded due to insufficient image data. Stenosis were graded higher in NE-MRA than in DSA (mean StG r1:3.0, r2:3.1, κ 0.59; DSAStG 2.7; p (r1), p (r2)0.001). Segmentwise comparison revealed mean r1:2.7 and r2:2.9 (κ 0.34).

Conclusion: This study can be seen in context of a NE-MRA renaissance due to NSF and in second line increasing cost pressure. The new NE-MRA technique revealed only limited visualisation quality in symptomatic patients with multiple-and high-grade stenosis. We assume that especially flow disturbance due to multiple stenosis may have hampered visualisation quality of the flow-dependent technique in this collective. Even though we assume restricted clinical potential as "fall back" after failed CE-MRA, anatomic pre-imaging for future MRI-guided interventions and maybe in subjects with less flow disturbance (i.e. screening). A further advance of NE-MRA techniques towards a reliable diagnostic tool for patients with renal impairment and PAOD remains desirable.

anonymised datasets were analysed by three readers. The lowest level of tube current at which CCS remained accurate was correlated with six parameters A: weight, B: body mass index, C and D: AP and lateral thoracic diameters, E: the arithmetic mean of (C) and (D) and F: Scout attenuation coefficient (SAC) measured from AP and lateral CT scout projections. Results: SAC had the closest correlation with the lowest tube current for accurate CCS. The 95% confidence intervals were established for SAC and lateral thoracic diameter. The lowest value in this range was used to predict the lowest possible tube current that could have been used. These data were used to create an algorithm for dose stratification. Using SAC or lateral diameter, mean effective radiation dose for CCS could be reduced from 1.83mSv to 0.80mSv or 0.87mSv, respectively. Conclusion: CCS can be performed using at least 50% reduction in tube current. Radiation dose can be easily tailored to the patient by utilising data from CT scout projections. Purpose: Radiation exposure is a major healthcare concern. We aimed to optimise scanning protocols in order to reduce radiation dose, while maintaining image quality.

In a single centre cohort study, consecutive patients underwent contrast-enhanced prospective, electrocardiogram-gated, CTCA using a 320-multidetector scanner. We assessed patients before (group 1, n = 31) and after (group 2, n = 32) the implementation of a dose reduction protocol. Dose reduction techniques included reducing detector range, tube current and voltage, limiting acquisition to one heart-beat with a 70-80% phase window, reducing heart rate to 60 rather than 65 beats per minute (bpm) and using delayed, intermittent and low dose bolus tracking. We present effective dose calculated with k-factors of 0.014 (and 0.023) mSv/(mGy*cm). Results: Patients in both groups were of similar age (58 ± 14, 55 ± 11), BMI (28 ± 6, 27 ± 5 kg/m 2 ) and heart rate (59 ± 13, 56 ± 8 bpm). The median DLP was lower in group 2 (171 (interquartile range 137-259) vs 308.60 (182-393), p < 0.001). The median DLP for bolus tracking was lower in group 2 (20 (10-23.4) vs 31.7 (19.5-51.4), p < 0.001). Thus, the median effective dose fell from 4.3 (7.1) to 2.1 (3.9) mSv for CTCA and 0.72 (0.44) to 0.46 (0.28) mSv for bolus tracking. Conclusion: Radiation dose for 320-multidetector CTCA can be reduced by optimising protocols, without compromising image quality. It is important to consider which k-factor is used when assessing CTCA radiation dose. The radiation dose from bolus tracking and scout images should not be ignored when considering patient exposure, and these can also be markedly reduced.

The influence of chest wall tissue composition in determining image noise during cardiac computed tomography N.S. Paul, H. Kashani, D. Odedra, P. Rogalla; Toronto, ON/CA (narinder.paul@uhn.on.ca)

To determine the influence of chest wall composition on image quality during cardiac CT compared with other biometric parameters.

A prospective study recruiting 50 consecutive patients (35M), mean age 55.2 years, were referred for coronary artery calcium quantification. All scans used 120 kVp and 150 mAs. Images were analysed on a 3d workstation (Aquarius Intuition, Terrarecon, CA). Patient weight and body mass index were recorded. Following parameters were measured:anteroposterior and lateral diameters, total (CS Total), fat (CS fat) and muscle and bone cross-sectional area (CSMB). Image noise was measured by prescribing an ROI in descending aorta. Correlation coefficient (R) was calculated to assess the association between image noise and biometric parameters (MedCal, v11.2, Belgium). Data were stratified based on gender and BMI. For BMI groups, data were stratified based on fat to muscle ratio. Results: For total study population, image noise was highly correlated with weight (R = 0.72) and CS total (R = 0.68). The strongest association with image noise in males is observed for weight and BMI (R = 0.68) followed by CS total (R = 0.59). For females it is seen for CS total (R = 0.76) with a CS fat R value of 0.72. For BMI, in groups with a fat: muscle ratio > 1, CS total and tat had the strongest correlations with image noise. Conclusion: Image noise in cardiac CT correlates more strongly with chest wall component than other patient parameters including BMI and weight. Incorporating this parameter into dose reduction strategies will potentially result in improving protocols.

A D E F G H Methods and Materials: Ten patients with proven PAD (Fontaine II, TASC-II type-A lesion of the superficial femoral artery) and 10 healthy control subjects were included. Patients underwent DCE perfusion imaging of the calf musculature once, whereas controls were examined twice on different days to determine normative values and the reproducibility of the technique. Hyperaemic fractional microvascular blood volume (V p , unit: %) of the anterior tibial, gastrocnemius and soleus muscle were determined during reactive hyperaemic conditions as provoked by a standardised cuff paradigm. Results: Fractional microvascular blood volume V p was significantly lower for all muscle groups in PAD patients compared to controls (range differences 3.2-4.8 %, p < 0.01). Differences in V p between the muscle groups were not significant (p = 0.49 and p = 0.15 in patients and controls, respectively). Interscan and interreader coefficient of variation in controls ranged from 10 to 13% and 3 to 4%, respectively, for the different muscle groups. Conclusion: DCE MRI using a blood pool contrast agent is able to reproducibly determine the fractional hyperaemic microvascular blood volume of the calf musculature and to discriminate between PAD patients and healthy control subjects. Blood pool agents are therefore a valuable addition to DCE MRI in PAD. Purpose: To assess the influence of x-ray tube voltage (100 vs 120 kVp) on various efficacy parameters and dose reduction in abdominal 64-slice MDCT angiography.

The present research is an ancillary study of a previously published prospective multicentre multivendor trial. Patients (n=310) were randomised to undergo MDCT angiography with administration of iobitridol (350 mgI/ mL) or iomeprol (400 mgI/mL). Each centre applied its own mAs/kVp for abdominal MDCT. Types of contrast medium (CM), body mass index (BMI) and presence or absence of endoprosthesis (EP) were studied as confounding factors. Dose estimates were retrospectively collected. The primary endpoint was to assess the influence of x-ray tube voltage on the diagnostic efficacy. Secondary objectives were to assess the image quality and dose savings obtained at 100 vs120 kVp. Results: Among the 307 analysed patients, 18% were explored at 100 kVp. Baseline characteristics including confounding factors were balanced between both kV groups. The diagnostic contribution was equivalent between both kV groups with no impact of BMI, EP or CM iodine concentration. Effective dose for MDCT angiography was 5.7±1.9 mSv versus 15.5±8.4 mSv, respectively, in 100-and 120 kVp patients. Image quality, including arterial enhancement and vessel wall visualisation, was deemed significantly superior at 120 kVp. Conclusion: Although setting the x-ray tube at 100 kVp does not impact the diagnostic efficacy of abdominal 64-slice MDCT angiography, it negatively impacts the image quality. Scanning patients at 100 kVp is compatible with a two-third reduction of the radiation dose. Moreover, using thin slice (1 mm) for 3D reconstruction we reduce artefacts from bone-vessel closeness. Thus, diagnostic accuracy was improved and no significant difference in lesion detection was reported between above or below the knee. Purpose: To compare a standard (0.1 mmol/kg BW) and a low (0.05 mmol/kg BW) gadobenate dimeglumine dose, respectively, at 1.5 T and 3 T in terms of image quality and diagnostic accuracy for contrast-enhanced MR-angiography (CE-MRA) of the peripheral vessels relative to CT-angiography (CTA). Methods and Materials: 60 patients were randomly assigned to 2 groups of 30 subjects. Group 1 underwent CE-MRA at 1.5 T using a 0.1 mmol/kg dose while group 2 underwent CE-MRA at 3 T using a 0.05 mmol/kg dose of gadobenate dimeglumine. All patients underwent CTA. Three blinded observers reviewed MRA datasets while one independent observer evaluated CTA images for the assessment of stenosis degree (> 50%), extent and location. Signal-to-noise (SNR) and contrast-to-noise (CNR) ratios were calculated for each MRA dataset. Interobserver agreement for CE-MRA was analysed using Cohen-k statistics. Sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) were calculated for each group. Differences between groups were assessed using McNemar's test. Results: Interobserver agreement was substantial for image quality and stenosis assessment. Similar values for SNR and CNR were obtained for both groups (p > 0.05). Sensitivity, specificity accuracy, PPV and NPV values of Group1/Group2 for stenosis assessment compared to CTA were 95/96%, 95/95% 97/97%, 96/95% and 97/97%, respectively. The diagnostic performance of groups 1 and 2 was not significantly different (p > 0.05).

Conclusion: Image quality and diagnostic performance of CE-MRA with a low 0.05 mmol/kg BW dose of gadobenate dimeglumine at 3 T is similar to that achieved with a standard 0.1 mmol/kg BW dose at 1.5 T.

Dynamic contrast-enhanced MRI with gadofosveset in peripheral arterial disease: initial findings of hyperemic fractional microvascular blood volume assessment B. Versluis 1 , T. Leiner 2 , P. Purpose:

The aim of the current study was to describe a method to assess the fractional hyperaemic microvascular blood volume of the calf musculature in patients with peripheral arterial disease (PAD) and healthy control subjects, using a blood pool contrast agent (Gadofosveset) in dynamic contrast-enhanced (DCE) MRI. 15:12 "Fail to plan and you plan to fail". Patient management and therapy decision according to transatlantic inter-society consensus (TASC) guidelines for peripheral artery obstructive disease (PAOD): a prospective 64-CT angiography study using DSA as reference A. Napoli, F. Zaccagna, M. Anzidei, G. Cartocci, L. Molisso, L. di Mare, C. Catalano, R. Passariello; Rome/ IT (alessandro.napoli@uniroma1.it) Purpose: To evaluate the diagnostic accuracy of 64-slice CTA for the detection of peripheral artery stenosis and to compare provisional therapeutic approach suggested after 64-CTA using DSA as reference standard. Methods and Materials: 212 consecutive patients with disabling intermittent claudication scheduled for DSA underwent CTA (64x0.6 mm; 110 ml of Iomeprol-400 at 4 ml/sec). Image evaluation was performed with the arterial vascular system divided into 35 segments, detected stenoses were graded using a 4-point scale (0-4: normal-occlusion). Lesions were classified according to TASC guidelines and a provisional indication was made as to whether treatment should most likely be conservative, endovascular, surgical or combined. Diagnostic performance (accuracy, sensitivity, specificity, PPV and NPV) was obtained considering also uninterpretable CTA results. Chi-square, McNemar and Wilcoxon's rank tests were used as appropriate. Results: A total of 7420 arterial segments pooled into 1060 vascular regions were assessed. McNemar's test showed no significant difference between CTA and DSA (0.62<p0.05). After DSA, 50/210 (28%) patients were referred for conservative treatment, 94/210 (44%) underwent endovascular procedures, 39/210 (19%) underwent surgery and 18/210 (9%) underwent an intra-operative hybrid (endovascular and surgical) treatment. Therapy recommendation based on CTA findings alone was identical to that of DSA for all but one patient who had a type D infra-inguinal lesion. Conclusion: 64-CTA yields very high diagnostic accuracy without influence of plaque composition; these data translate into reliable therapy decision-making and patient management as compared to DSA. Purpose: The ankle-brachial index (ABI) measurement is an important functional measure in the diagnosis and follow-up of peripheral arterial disease (PAD). ABI measurements, however, are hampered by large interobserver variability and inconclusive results in patients with uncompressible arteries. Alternatively, MRI can be used to determine arterial peak flow (APF) as a substitute functional measure. In this study, APF of the popliteal artery and the ABI in PAD patients is compared. Methods and Materials: 124 patients with PAD, varying from intermittent claudication (n = 67) to critical ischaemia (n = 57), were included. All subjects underwent vector cardiography triggered quantitative 2D cine MR phase-contrast angiography of the popliteal artery to determine APF during a routine contrast-enhanced MR angiography. All APF measurements were analysed by one radiologist with 3 years of experience. ABI measurements were performed by a small team of Doppleroperators with > 10 years of experience. Results: APF was successfully measured in 87% of the patients, whereas the ABI could only be determined in 70% of the patients. Pearson correlation between APF and ABI was 0.72 in intermittent claudication and 0.47 in critical ischaemia, both were significant at the 0.01 level. Both APF and ABI were significantly lower in critical ischaemia as compared to intermittent claudication (p < 0.01). Conclusion: The APF is a functional vascular measure that can be determined in more patients than the ABI, is not influenced by uncompressible arteries, and correlates significantly with the ABI and is therefore a valuable measure in PAD patients, both for diagnostic and follow-up purposes.

Conclusion: DE CTA has diagnostic accuracy higher than other non-invasive imaging technique and very similar to DSA that still represent the gold standard in vascular imaging. Moreover, automatic bone removal application allow fast post-processing time generating in few seconds 3D MIP images of vascular tree (luminogram) similar to DSA images. This improve confidence for non-vascular radiologist and other specialists and also allow a perfect therapeutic planning for endovascular treatment.

Dose reduction in lower extremity arterial imaging using a reduced tube voltage and automatic tube current modulation technique with 64-slice computed tomography Y. Ge 1 , Y. Guo 1 , Z. Wen 1 , J. Li 2 , X. Su 1 ; 1 Zhengzhou/CN, 2 Beijing/CN (cjr.geyinghui@vip.163.com) Purpose: To evaluate the use of automatic tube current modulation (ATCM) and 80 kVp tube voltage for radiation dose reduction for lower extremity arterial imaging in multi-slice CT. Methods and Materials: Eighty-two consecutive diabetic outpatients with lower extremity occlusive disease were randomly divided into three groups with different scanning techniques on a 64-slices CT scanner. The first group used standard 120 kVp and fixed tube current of 180 mAs (group1, n = 26), and the second (group2, n = 26) and third groups (group3, n = 30) used 120 kVp and 80 kVp with ATCM, respectively. The pelvic, knee, calf and foot were four observing levels for noise measurement and image quality assessment with a 3-point scale. Paired analyses were performed on radiation dose, image quality and image noise using t-test. Results: The scan lengths for all three groups were statistically the same (p > 0.05). The CTDI vol and DLP values for the three groups were (14.48±3.60, 5.56±2.16 and 2.99±0.73 mGy) and (1703.54±347.7, 679.6±267.3 and 360.95±91.47 mGy·cm). Radiation dose was significantly reduced by 60% and 79% for group2 and group3, respectively, compared with group 1 (p < 0.05). Group3 and group1 had the highest and lowest image noises, respectively. However, there was no significant difference in image quality scores (2.57±0.50, 2.26±0.48 and 2.56±0.56) between the three groups (p > 0.05). Conclusion: Using ATCM for lower extremity arterial scanning can reduce radiation dose. Radiation dose can further be reduced with equivalent image quality by using 80 kVp.

Evaluation of an iterative reconstruction algorithm (iDose) to reduce image noise in a preclinical model of low dose CT angiography A.A. Fingerle, P.B. Noël, A. Huber, E.J. Rummeny, M. Dobritz; Munich/ DE (alexander.fingerle@gmx.de) Purpose: In this work we present the evaluation of a hybrid-iterative reconstructionalgorithm (iDose, PhilipsHealthcare) for the reduction of image noise and improvement of image-quality in a preclinical model of low-dose CTA. Methods and Materials: A swine intestinal specimen consisting of the bowel and the mesentery was used as a model. After injection of a contrast solution in the SMA, the intestinal specimen was scanned in a water bath on a 256-slice multi-detector CT (Brilliance iCT, PhilipsHealthcare) using 15 low-dose combinations of tube-voltage and tube-current settings (80-120 kVp; 15-130 mAs). In additiona, a high-dose scan (140 kV/650 mAs) was performed as reference to the low-dose acquisitions. Low-dose images were reconstructed with a conventional reconstruction-algorithm and in three different iDose-levels. Altogether 60 different image sets were generated. Image-noise was measured in vessels and the intestinal wall. Subjective image-quality and delineation of vessels were addressed. Statistical analysis was performed. Results: CTA performed with the low-dose protocol showed an average dose reduction by a factor of 2 compared to the standard-dose protocol. Image-noise was higher by a factor of 2 in low-dose images reconstructed with the standard reconstruction. Reconstructing low-dose data with the iDose reconstruction-algorithm, image-noise was lowered to a comparable level of standard-dose CT angiography (p < 0.01). On this note, all image information was retained in low-dose images reconstructed with iDose. Conclusion: iDose allows dose reduction in MDCT angiography. Further iDose provides a significantly reduction of image-noise and thereby retains the imagequality equivalent of standard-dose CT. This may allow a routine dose reduction in the clinic with high impact on patient health. Purpose: Aim of the study is to describe the main angiographic patterns of the cocaine-induced vasculitis changes on cerebral angiograms in SAH/stroke patients. Methods and Materials: From April 2007 to September 2010, we retrospectively identified thirteen patients (mean age 43.7 years), referred for acute SAH/stroke, after use of cocaine, with positive urine toxicology or its metabolites. The subjects underwent computed tomography angiography (CTA), magnetic resonance angiography (MRA) and digital subtraction angiography (DSA) valuation shortly after.

The main symptoms were headache, hemiparesis, agitation, and aphasia. Seven patients showed on CT mild SAH and a large putaminal haemorrhage, while at DSA multiple mild-and small-vessel occlusions were found. In four patients, CT showed a large SAH in the basal cisternal, while MRA and DSA demonstrated segmental narrowing/dilatations of medium-sized intracranial arteries. In two patients, a wide basal cisternal and interemispherical SAH were found on CT; in one of the above patients DSA revealed multiple infundibulomas of the main intracranial bifurcations, with a 23 mm diameter size, anterior communicating aneurism that were subsequently embolised. Data were analysed with a Fischer exact test (p < 0.05).

The rate of strokes among cocaine abusers is broadly increasing in the last few years. The exact mechanism of cocaine-induced stroke remains still unclear, but is probably a consequence of alteration in the cerebral flow autoregulation, due to several factors including vasospasm, cerebral vasculitis, enhanced platelet aggregation, and hypertension. Although histopathologic examination is required for a definitive diagnosis, CTA, MRA, and DSA can support the neuroradiologic evaluations.

Manganese-induced toxic-metabolic encephalopathy: clinical and radiological findings F.I. Todua, D. Miminoshvili, M. Okujava; Tbilisi/GE (dmiminoshvili@yahoo.com)

Purpose: To review neurological abnormalities (clinical and neuropsychological changes) and results of radiological (Multi-slice CT, MRI) examinations in methcathinone users. This psychoactive substance is synthesised from drugs which contain ephedrine and potassium permanganate.

We have investigated 32 patients (28 males and 4 females) with chronic intravenous use of methcathinone. Patient age ranged from 23 to 54 years. All patients underwent MRI and MSCT of the brain. Results: On MRI examination the brain abnormalities were revealed in 28 cases: there was a hyperintense signal on T1-weighted images in globus pallidum (n=24), substantia nigra (n=22), nuclei dentatus of cerebellum (n=18), anterior pituitary lobe (n=11), and white matter of the great hemispheres (n=3). The signal intensity was higher after recent and frequent use of the Methcathinone. The signal abnormalities on MR were caused by the accumulation of the manganese, which is byproduct of the synthesised street drug methcathinone, in the basal ganglia and other structures. On CT investigation leukoaraiosis was revealed in 15 cases, basal gangia calcifications in five and cerebellar atrophy in four cases. Most patients (85 %) had extrapyramidal disorders: bradykinesia, postural disturbances, ataxia, tremor, muscular dystonia (mostly of the face and lower extremities). Pseudobulbar syndrome (dysarthria, dysphonia) were revealed in 10 % cases. Cognitive abnormalities (personality changes, stupor) were noted in 5% cases. Conclusion: Thus, in cases of methcathinone abuse MRI and MSCT examinations are strongly recommended to find out the cerebral anomalies and determine treatment tactics.

Supratentorial superficial siderosis: localisation, clinical symptoms and differential diagnoses J. Linn 1 , K. Bochmann 1 , T. Pfefferkorn 1 , P. Cerebral microbleeds can be visualised in up to 40% of patients with MRI scanners at standard field strengths. However, these scanners have a limited sensitivity in detecting more subtle cerebral damage associated with iron deposition. The aim of this study was to quantify focal and diffuse iron deposition in CADASIL, using 7 Tesla MRI. Methods and Materials: Twenty-five NOTCH3 mutation carriers and 15 healthy controls were examined using high-resolution T2*-weighted imaging on a 7 Tesla whole body MRI scanner. MRI scans were analysed for focal and diffuse areas of decreased signal intensity. Mean signal intensity measurements in affected brain regions were compared between mutation carriers and controls. Results: Diffuse areas of decreased signal intensity were found in mutation carriers. Compared with healthy controls, mutation carriers had significantly lower signal intensity in the putamen (p = 0.005) and caudate nucleus (p = 0.0007). Focal areas of decreased signal intensity were found in 36% of mutation carriers and in none of the healthy controls. These areas were predominantly (66%) located in the thalamus. Other locations included the subcortical and deep white matter. Conclusion: 7 Tesla MRI reveals increased areas of hypointensity in the putamen and caudate nucleus of CADASIL patients that is likely caused by increased diffuse iron accumulation. Focal areas of decreased signal intensity were also found in 36% of CADASIL patients, in a pattern consistent with microbleeds.

Diagnostic accuracy of thin slice angiographic reconstructions of volume perfusion CT: comparison with CT angiography A.M.J. Frölich 1 , E. Klotz 2 , R. Schramm 1 , M. Knauth 1 , P. Schramm 1 ; 1 Göttingen/DE, 2 Forchheim/ DE (andreas.froelich@med.uni-goettingen.de) Purpose: Volume perfusion CT (VPCT) is increasingly used for diagnostic imaging in acute stroke. The goal of this study was to compare the diagnostic accuracy of thin slice angiographic reconstructions obtained from VPCT datasets (4D-CTA) with conventional CT angiography (CTA) data. Methods and Materials: 59 consecutive patients were examined with a comprehensive CT protocol including non-enhanced CT (NECT), CTA and VPCT (Siemens Definition AS+). Contrast protocol for CTA was 60 ml at 6 ml/sec, for VPCT 45 ml at 6 ml/sec. VPCT used a periodic spiral technique: 30 spirals within 45s, z-coverage 96 mm, 4D-CTA recons every 1 mm with a slice width of 1.5 mm. Axial and coronal maximum-intensity projections (10 mm every 3 mm) were generated from CTA and 4D-CTA at peak arterial time. Two experienced neuroradiologists blinded to diagnosis and patient data evaluated all images for presence of vessel occlusion or stenosis. Results: 34% of patients had an acute intracranial occlusion which was detected on both 4D-CTA and CTA. 54% showed no acute vascular pathology on CTA or 4D-CTA. In 24% arterial stenosis was detected in agreement with CTA and 4D-CTA. In 12% of cases, a relevant pathology of the extracranial arteries was not detected because they were not included in the VPCT scan volume. Conclusion: For the evaluation of acute stroke, obtaining NECT and VPCT alone (with 4D-CTA reconstructions instead of traditional CTA) may be a time-and radiation dose-efficient approach to detect acute intracranial vessel occlusion and intracranial stenosis. CTA remains mandatory in cases of extracranial vessel pathology outside of the VPCT scan range.

Evaluation of middle cerebral artery stenosis: comparison of highresolution MR imaging and digital subtraction angiography M. Wang 1 , D. Shi 1 , T. Li 1 , Y. Han 1 , Y. Dai 2 ; 1 Zhengzhou/CN, 2 Shanghai/ CN (meiyun9999@gmail.com) Purpose: To explore the accuracy of high-resolution MR imaging (HR-MRI) in quantifying the severity of middle cerebral artery (MCA) stenosis by comparing with gold standard, conventional digital subtraction angiography (DSA). Methods and Materials: A prospective study of 26 patients with symptomatic MCA stenosis was enrolled in this study. HR-MRI was performed on a 3.0 T MRI system (TrioTim, Siemens Medical Solutions) and compared to DSA. High-resolution T1-, T2-and PD-weighted images were acquired along MCA short axis, and T1-and PD-weighted postcontrast images were obtained 5 minutes after 15 ml Gd-DTPA injection. A quantification of wall thickness area, lumen area, and degree of stenosis was performed on T2-weighted cross-sectional images at the level of maximal MCA stenosis by three independent blinded readers, the degree of stenosis was calculated and compared with that defined by DSA. Results: Spearman correlation demonstrated that HR-MRI correlated well with DSA in evaluating the severity of MCA stenosis (r = 0.81, p < 0.05). The inter-observer agreement was good (Kappa = 0.72). The sensitivity and specificity of HR-MRI in the detection of severe stenosis was 83% and 80%, respectively. Conclusion: HR-MRI could potentially be a useful tool in providing information about severity of MCA stenosis. haemorrhage, head trauma or neurosurgical intervention. T2*w was analysed by two neuroradiologists in consensus regarding the localisation of SupS and with regard to the presence of microbleeds (MBs) or macrohaemorrhages (ICHs). Clinical records were assessed to determine the clinical symptoms. The most probable causes underlying the SupS were identified based on all clinical, laboratory, imaging and histological information. Results: 52 patients (73 ± 4 years) matched the inclusion criteria. In most cases SupS involved the central sulcus. Additional ICHs were detected in 29 patients, 17 patients had MBs. Most common symptoms were headaches, focal seizures or transitory ischaemic attacks. The following causes were identified: probable CAA (39 cases), definite CAA (3 cases), reversible cerebral vasoconstriction (2 cases). In 7 cases the underlying diagnosis could not be determined. Conclusion: SupS presents with typical clinical symptoms, which are most probably caused by the common involvement of the central sulcus. In older patients, CAA is the most important cause of SupS.

Carotid artery plaque histology and brain MRI L. Saba, L. Lai, R. Sanfilippo, R. Montisci, G. Faa, G. Mallarini; Cagliari/IT (lucasaba@tiscali.it)

Purpose: There are several evidences that carotid plaque structure could be an independent risk factor for ischaemic stroke. The purpose of this work was to explore the relationship between carotid plaque histological composition and brain MRI findings. Methods and Materials: 18 consecutive (males 14; median age 64) patients were prospectively analysed. Brain MRI was performed using a 1.5 Tesla system and lesion's pertinence to the anterior circulation were recorded. Patients underwent carotid endarterectomy "en bloc"; histological sections were prepared and the presence and integrity of fibrous cap, haemorrhage	hrombus and plaque composition (calcification, lipid rich necrotic core) were analysed. Results: The prevalence of haemorrhage	hrombus was 56% (10\18) and the prevalence of ruptured fibrous cap was 72% (13\18). Liner regression analysis demonstrated an association between the number of brain infarct (lacunar and non-lacunar) and the presence of haemorrhage	hrombus (p = 0.003), whereas no statistically significant association was observed between the presence of fatty plaque, ruptured fibrous cap and the type of the plaque. Conclusion: Results of this preliminary study indicate that haemorrhage	hrombus play a crucial role in the pathogenesis of stroke, and it supports the notion that cerebrovascular lesions are probably caused by emboli arising from thrombotic lesions.

Vertebral artery variations at the C1-2 level diagnosed by MR angiography A. Uchino, N. Saito, T. Watadani, Y. Okada, E. Kozawa, N. Nishi, W. Mizukoshi, R. Nakajima, M. Takahashi; Hidaka/JP (auchino@saitama-med.ac.jp) Purpose: The craniovertebral junction is clinically important as the location of many kinds of congenital or acquired lesions requires surgical interventions. The vertebral artery (VA) in its several variations runs within this area. We report the incidence of these VA variations on magnetic resonance angiography. Methods and Materials: We retrospectively reviewed MRA images, obtained using two 1.5-tesla imagers, of 2,739 patients, and paid special attention to the course and branching of the VA at the level of the C1-2 vertebral bodies. Results: There were 3 types of VA variation at the C1-2 level−(1) persistent first intersegmental artery (FIA), (2) VA fenestration, and (3) posterior inferior cerebellar artery (PICA) originating from the C1/2 level. The overall incidence of these 3 variations was 5.0%. There was no laterality in frequency, but we found female predominance (P < 0.05). We most frequently observed the persistent FIA (3.2%), which was sometimes bilateral. We found VA fenestration (0.9%) and PICA of C1/2 origin (1.1%) with almost equal frequency. Two PICAs of C1/2 origin had no normal VA branch. Conclusion: We frequently observed VA variations at the C1-2 level and with female predominance. The persistent FIA was most prevalent and sometimes seen bilaterally. Preoperative identification of these variations in VA is necessary to avoid complications during craniovertebral junction surgery.

Purpose: The application of MBIR to clinical CT applications is still a relatively new area, and a number of challenges have to be overcome for clinical application of MBIR. One of them is the elimination of pinwheel artefacts, which are sometimes also observed in FBP reconstructions from helical scans, in the presence of strong longitudinal gradients. We proposed an improved MBIR method to reconstruction images with fewer artefacts. Methods and Materials: Pinwheel artefact is caused by high-z gradient in object and non-perfect sampling of detector along z direction. Here, we present a method to prevent such kind of artefacts using MBIR by improving its forward model. Instead of a uniform voxel as used in conventional methods, we incorporate a longitudinal slope in each voxel to better model longitudinal gradients, which can better represent edges within a voxel. This eliminates possible model mismatches and therefore reduces artefacts. Results: We validated the proposed method to a simple post-processing method to minimise pinwheel artefacts, which is commonly used for FBP reconstructions. Both methods have been implemented and tested using patient data sets acquired from GE Lightspeed CT systems. Both result in significant artefact reduction. The improved voxel model method outperformed the post-processing method, especially for the region close to bone structures. Conclusion: The proposed method can efficiently suppress pinwheel artefacts without significant computational penalty.

Advanced interpolation for normalised metal artefact reduction (NMAR) in CT E. Meyer 1 , R. Raupach 2 , M. Lell 1 , B. Schmidt 2 , M. Kachelrieß 1 ; 1 Erlangen/DE, 2 Forchheim/ DE (esther.meyer@imp.uni-erlangen.de) Purpose: To correct metal artefacts in clinical circular and spiral cone-beam CT. Methods and Materials: Metal artefact reduction (MAR) is typically performed by replacing corrupt projection data by interpolation. We have recently proposed NMAR that consists of a normalisation step prior to interpolation and that turned out to significantly outperform all other known MAR techniques and thereby appears to be the most promising MAR approach available today [MedPhys 37 (10):inpress]. Nevertheless, even with NMAR slight artefacts remain. We therefore enhanced NMAR and replaced its simple linear interpolation by better interpolation and inpainting methods: directional interpolation and anisotropic diffusion inpainting. Directional interpolation aims at properly connecting traces of high-contrast objects though the metal parts in the projections. Inpainting with diffusion type equations ensures a smooth transition between measured and surrogate projection data. For the evaluation, ten patient data sets with hip prostheses and dental fillings were processed with NMAR and the image quality was evaluated. Purpose: We present an iterative reconstruction scheme in image space based on a non-isotropic noise model for purpose of reducing radiation dose and image noise. Methods and Materials: Most recently, iterative reconstruction in image space was introduced providing a huge potential of noise and/or dose reduction, while maintaining spatial resolution. In all currently published methods, non-linear regularisation is based on an isotropic, local noise model. We propose an improvement of this noise model, which takes into account the higher noise of projection data in directions of large attenuation. This is done by applying non-isotropic attenuation maps derived from parallel forward projections of axial images, and can be efficiently implemented using standard Fourier techniques. The strength of non-linear regularisation becomes non-isotropic adjusted by the non-isotropic local noise values, derived from the attenuation maps.

Results: Based on simulation data and clinical data, respectively, it is shown that for non-symmetric objects non-isotropic image noise can be substantially reduced, revealing low contrast details otherwise obscured by noise. Clinical data of thoracicabdominal scans demonstrate a potential radiation dose reduction by 70%, while maintaining image noise and visually perceived sharpness for high contrast objects. Conclusion: Iterative reconstruction in image space based on the presented nonisotropic noise model is an efficient iterative reconstruction scheme providing significant reduction of image noise and/or radiation dose by up to 70%. For each image slice, a single-slice forward projection has to be computed prior to the iteration loop.

Translation of statistical iterative reconstruction into non-linear image processing H. Bruder, R. Raupach, J. Sunnegardh, K. Stierstorfer, T. Flohr; Forchheim/DE (rainer.raupach@siemens.com)

Purpose: We demonstrate that statistical iterative reconstruction can be translated to non-linear image processing in case of data-dependent Gaussian noise. Methods and Materials: Statistical iterative reconstruction is known to produce images with better signal-to-noise ratio compared to conventional FBP-type reconstruction. Originally based on the Poisson noise model, it can be simplified to a data-dependent Gaussian noise model for large numbers of quanta, manifesting as a signal weighting of sinogram data according to their statistical reliability. Based on the update equation of Iterative Filtered Backprojection reconstruction (IFBP), we introduce the signal weighting operator (method A) and show that, even in case of non-linear regularisation, under certain conditions the update equation can be formulated as an iterative reconstruction in image space. We also derive an equivalent approximation of iterative reconstruction in image space based on a non-isotropic noise model (method B). Results: In case of non-linear regularisation it is demonstrated that methods A and B are equivalent regarding noise characteristic and spatial resolution. For non-symmetric objects, non-isotropic image noise can be substantially reduced, revealing previously invisible low contrast details. Image sharpness of objects with contrast beyond the noise level is maintained. Conclusion: We prove that statistical iterative reconstruction with data-dependent Gaussian noise and noise weighted IFBP is equivalent. Under certain conditions, it can be translated into a non-linear image processing. If iterative reconstruction is to be used for the reduction of artefacts due to non-exact reconstruction, a few IFBP iterations can be applied prior to image space processing.

A S215 C D E F G H a dual-source system. In this work, we pursue a different approach, namely employing an iterative algorithm to reconstruct images from less than 180° of projections and using a histogram constraint to prevent the occurrence of limited-angle artifacts. The method is fundamentally different from previously published approaches using prior images and TV minimisation. Furthermore, motion detection is used to enhance dose usage in those parts of the image where temporal resolution is not critical. Results: The proposed method yields good results, both with simulated and with clinical data. Our evaluations also show that an enhancement of temporal resolution to a value equivalent to about 120°-135° of projections is viable. Furthermore, by employing motion detection, a substantial noise reduction can be achieved in those parts of the image where no motion occurs. Conclusion: Using our proposed method, an enhancement of temporal resolution in CT images by about 30% is possible. Furthermore, an optimisation of dose usage can be achieved by use of a motion detection system.

CT image reconstruction from limited data L. Ritschl 1 , C. Fleischmann 2 , M. Kachelrieß 1 ; 1 Erlangen/DE, 2 Nürnberg/ DE (marc.kachelriess@imp.uni-erlangen.de) Purpose: To obtain clinically relevant and artefact-free images from incomplete CT data (limited angle, metal implants, few projections, projection truncation). Methods and Materials: Image reconstruction based on compressed sensing (CS) is a promising technique to reconstruct from incomplete raw data. Existing algorithms (TV-Min, PICCS, ASD-POCS), however, do not always yield clinical image quality and require to adjust several object-dependent parameters. We propose an improved total variation (iTV) reconstruction algorithm to overcome these difficulties. It controls the reduction of the TV while improving raw data agreement throughout each iteration step. Thereby, iTV does not need to be tuned for a specific patient. For example, the same set of parameters for metal artefact reduction, for sparse projection sampling, for limited angle artifact reduction and for projection detruncation is used for both clinical CT and flat detector CT. To evaluate the performance of iTV clinical patient cases (from a dual source clinical CT scanner and from a mobile C-arm CT scanner) with limited data were processed and compared with today's gold standard cone-beam filtered backprojection.

In all examples, the algorithm was able to suppress most of the image artefacts. Especially in the case of metal implants and dental fillings and in the case of limited angle tomograhy (160°) the resulting images were of surprisingly high image quality. Conclusion: With iTV a highly robust and highly versatile algorithm that guarantees high-fidelity images for actual patient data has become available.

Assessment of CT spectral imaging with metal artefact reduction sequence for metal artefact reduction: in vitro and clinical studies X. Zheng 1 , Y. Hua 1 , G. Shen 1 , J. Li 2 , Y. Shen 2 , G. Zhang 1 ; 1 Shanghai/CN, 2 Beijing/ CN (jianying.li@med.ge.com) Purpose: To assess CT spectral imaging with metal artefact reduction sequence (MARS) in reducing metal artefacts in vitro and in patients.

A phantom consisting of two rows of 16 needles (7-22G), two 10-mm diameter aluminum tubes and four tibial bone nails in a homogenous agar gel block within a water bath was scanned with the regular CT at 120 kVp and the single-tube, instantaneous kV switching spectral imaging on GE HDCT for monochromatic images. Twenty orthopaedic patients with intramedullary nails received CT for evaluation. Multiple regions of interest were placed around implants for CT value and standard deviation measurement and comparison. In patient study, images were also scored for implant margin visualisation, surrounding tissue details, and artefact severity. The surface profile and profile plot through implants were also analysed. Results: The monochromatic images with MARS in phantom study reduced the maximal CT number variation in areas surrounding implants from 32 HU in regular CT to 11 HU (p < 0.05). In patient study, metal artefacts significantly compromised the delineation of adjacent structures and implant margins in the regular kVp images. These artefacts were dramatically reduced in the 2D and 3D monochromatic images with MARS, which made the implant margins, intramedullar content, trabecular integrity, periosteal reaction and soft tissue abnormalities readily visualised and evaluated (score of 4.2 vs. 2.8, p < 0.01). The plots from the HDCT monochromatic images also reflected the actual structural layers. Conclusion: HDCT spectral imaging with MARS reduced metal artefact from metal implants and improved the depiction of surrounding focal structures.

Results: Subtle artefacts remaining with NMAR, such as in the presence of multiple dental fillings, can be significantly reduced when using the more advanced interpolation techniques. For example tiny dark regions close to the dental fillings are removed. For the cases studied, the directional interpolation is slightly more efficient than the anisotropic diffusion inpainting. Conclusion: NMAR's image quality can be further enhanced using sophisticated interpolation methods. Considerable improvements are found in particular for structures in the utmost vicinity of implants and dental fillings.

Intrinsic beam hardening correction (IBHC) for cone-beam CT R. Grimmer, M. Kachelrieß; Erlangen/ DE (rainer.grimmer@imp.uni-erlangen.de) Purpose: To remove first and higher order non-linear beam hardening and scatter artefacts from patient CT images. Methods and Materials: X-ray polychromacy and scatter are non-linear effects appearing as cupping and banding artefacts in CT images. A first order water correction (e.g. [MedPhys 33 (5):1269-1274]) followed by higher order BHC (e.g. [MedPhys 37 (10):5179-5187]) are typically used to remove those artefacts. To avoid this two-step process and to avoid any calibration steps depending on the scanner or the spectrum, we propose IBHC which performs a simultaneous correction of the first and higher order beam hardening artefacts. It separates the initial image into different materials, and forward projects those to obtain various rawdata sets. A combination of these is reconstructed and used as basis images to correct the initial image such that the final image is nearly artefact-free. A wavelet decomposition is used to determine those frequency components that contain artefacts to find the best combination of the basis images. Patient data from clinical and flat panel CT scanners are used to evaluate IBHC. Results: The The IBHC can almost completely eliminate cupping and streak artifacts. Typically, more than 90% artefact reduction is observed. For example, dark streaks were reduced from -246 HU to -8 HU in one of our head scans. Conclusion: IBHC is an efficient and stable new approach to reduce cupping and beam hardening. It is scanner-independent and therefore can be applied also offline to any CT image.

Dose-efficient deformation-constrained 4D cone-beam CT image reconstruction M. Brehm, F. Bergner, M. Kachelriess; Erlangen/ DE (marcus.brehm@imp.uni-erlangen.de) Purpose: To reconstruct high-quality respiratory-correlated 4D volumes from flat panel detector cone-beam CT scans and to obtain 100% dose usage. Methods and Materials: Respiratory-correlated cone-beam CT suffers from streak artefacts and from reduced dose usage because only a subset of the acquired projection data contributes to each time frame. We developed a new approach that iteratively reconstructs the volumes and simultaneously estimates the respiratory deformation fields. This estimate is used to generate a non-moving reference volume from which the other motion phases are derived by deformation. A spatiotemporal regularisation is used to suppress the effect of artefacts from the iterative 4D CBCT reconstruction. The method is verified using simulated rawdata obtained by deforming a clinical patient dataset using realistic deformation fields and by processing patient data acquired with the flat panel cone-beam CT scanner. Results: The streak artefacts in the reconstructions from both, simulated data and patient data, are highly reduced. Compared to the conventional phase-correlated method with 10 reconstructed phases, the noise level is reduced by more than 60% with our method in all cases. This corresponds to a dose reduction of more than 80%. Streak artefacts can almost completely be suppressed. Conclusion: Motion-compensated, spatiotemporal regularisation combined with iterative 4DCBCT reconstruction is capable of removing the streak artefacts to a large degree while maintain a good spatial and good temporal resolution. The dose usage is significantly increased because all data now contribute to each time frame.

Using a histogram constraint to enhance temporal resolution in CT H. Schöndube, T. Allmendinger, K. Stierstorfer, H. Bruder, T. Flohr; Forchheim/DE Purpose: To present and evaluate a new algorithm to enhance temporal resolution in CT beyond the short-scan limit by making use of a histogram constraint. Methods and Materials: A minimum scan angle of 180° plus fan-angle is needed to acquire complete data for reconstructing an image. Conventionally, this means that a temporal resolution of half the gantry rotation time is achievable and that an enhancement of temporal resolution can only be accomplished by a faster gantry rotation or using

Five-year ultrasound survey of functional overload injuries of patellar and Achilles tendons in elite rugby players C. Martini 1 , G. Ferrero 1 , G. Cicio 1 , L. Scofienza 2 , E. Silvestri 1 ; 1 Genoa/IT, 2 San Donato Milanese/IT (chiarapio@libero.it)

Purpose: Tendon overload syndromes affecting patellar (PT)and Achilles tendons (AT) represent a frequently disabling painful condition in rugby athletes. We used US to re-evaluate AT and PT in elite rugby athletes five years after a previous report, comparing the results with a group of healthy non-athletes volunteers in order to highlight any morpho-structural differences related to functional overload. Methods and Materials: US evaluation of each player tendons was performed at baseline (July 2005) and five years later (July 2010) using two different US systems (1 st :HDI5000,ATL;2 nd :MyLab70XvG,Esaote) equipped with high-resolution probes (8-18 MHz). Weight bearing and at rest bilateral PT evaluation was performed and proximal (PPT), mean (MPT) and distal (DPT) longitudinal measures were recorded. Weight bearing and at rest bilateral AT evaluation was performed and latero-lateral (LL)/antero-posterior (AP) diameter was measured and noted.Tendons were also scanned to highlight fibrillar structure alterations, focal hipoechoic areas and calcifications.The examination was repeated on 26 healthy non-athletes volunteers matched for age and sex. Results: Athletes' mean tendon thickness was significantly higher than that of controls (mean value PT atlete/controls:PPT 5.3 mm/3.8 mm;MPT 5.1 mm/3.6 mm;DPT 5.0 mm/3.6 mm; AT:AP 6.4 mm/5.7 mm;LL 16.1 mm/12.8 mm). Hypoechoic areas were found in 27% of athletes and 5% of controls, calcifications in 15% and 3%, respectively. All athletes who presented focal hypoechoic areas and/or calcifications at baseline were found to have them increased in number and size at five years. Two athletes healthy at baseline were found to have one focal hypoechoic area and one calcification at five years. Conclusion: AT and PT US evaluation in rugby players allows for detecting and highlighting overload degenerative alterations without clinical symptoms. US seems to have a great potential in prediction and prevention of damage progression in high-level athletes.

Tendon and ligament injuries of the ankle and foot: the role of diffusion weighted MR imaging in the diagnosis H. Aydin; Ankara/TR (dr.hasanaydin@hotmail.com)

Purpose: Our aim was to determine the diagnostic performance and utility of diffusion weighted imaging (DWI) against the routine magnetic resonance imaging (MRI) in evaluation of the patients with ankle and foot injuries. Methods and Materials: After institutional review board approval, ankle-foot MR imaging and DWI-apparent diffusion coefficient (ADC) mapping was performed to the 106 injured tendons and ligaments of 60 patients. All the injuries of tendons and ligaments were named as Rupture (R)-Partial tear (PT) and Tenosynovitis (T Purpose: The objective of this study was to apply magnetic resonance imaging (MRI) using displacement encoding with stimulated echoes (DENSE) to investigate the dynamic displacements and strain fields in the medial gastrocnemius muscle. Methods and Materials: Five volunteers were imaged using a 1.5 T Siemens MRI scanner (MAGNETOM Symphony, Siemens, Erlangen). Informed consent was obtained and all studies were approved by our institutional review board. The volunteers were positioned feet-first and right-side down into the MRI bore with their right ankle resting on a platform with a straight right knee and a flexi-coil around the right calf. The sequence was gated using a plantarflexion trigger mechanism and a cine series of echo-planar DENSE images was acquired over several successive plantar flexions. Displacement encoding was applied in three orthogonal directions. First and second principle strain measurements were derived from the displacement fields after spatio-temporal phase unwrapping. Results: Regional displacements were largest in the medial portion of the medial gasctrocnemius adjacent to the interface with the soleus muscle. The directions of the first principle strain vectors agreed well with the pennation angle of the myofibres. The magnitude of the first principle strain was largest in the middle of the muscle belly and the mean first principle strain along this midline at maximum plantar flexion for all five volunteers was 74.5±38.4%. Conclusion: This study shows how DENSE MRI can be used to quantitatively determine regional displacement and strain in the gastrocnemius muscle. These strain measurements compare favourably with previous studies using ultrasound.

Sonoelastography of plantar fascia: comparison among healthy subjects and patients affected by plantar fasciitis D. Orlandi 1 , L. Sconfienza 2 , C. Martini 1 , G. Ferrero 1 , E. Silvestri 1 ; 1 Genoa/IT, 2 San Donato Milanese/IT (theabo@libero.it)

Purpose: Plantar fasciitis is a common condition in athletes and general population. In early stages, the diagnosis of such affection can be challenging. Our aim was to evaluate real-time sonoelastography in the assessment of the plantar fascia in healthy volunteers and in patients with clinical suspicion of plantar fasciitis, comparing these findings with those obtained at conventional ultrasound. Purpose: Femoroacetabular impingement is a pre-arthritic condition with two main types; cam and pincer. Each type has been described at surgery to lead to specific patterns of articular cartilage wear. We sought to evaluate if the dGEMRIC-index would produce any identifiable patterns of cartilage degeneration similar to these. Methods and Materials: We compared the dGEMRIC-indices in a group of cam (n = 6; mean age 33 years) and pincer patients (n = 7; mean age 36.3 years) to a control group of asymptomatic controls (n = 12; mean age 25.2 years) with no MRI findings of osteoarthritis. T1maps and dGERMIC indices were obtained from a dual-flip angle 3D GRE-sequence at 1.5 Tesla. The superior portion of the hip joint was divided into seven regions from posterior anterior. These regions were then subdivided into peripheral and central regions and at each ROI, a dGERMIC index was recorded. Differences between study groups were determined using Student's t-test. Purpose: The aim of this study was to describe a novel semiquantitative scoring system for MRI of hip osteoarthritis, to test its reliability and analyse associations of intraarticular damage with radiographic OA severity. Methods and Materials: 52 patients (> 50 years of age) with hip pain were examined with 1.5 T MRI. Of these, 44 (85%) had weight-bearing x-rays and 37 (71%) received i.v. gadolinium administration. All MRIs were read by one expert radiologist. Intra-and interobserver reliability exercise was performed on a subset of 15 examinations. Cartilage, bone marrow lesions (BMLs), cysts, osteophytes, labral damage, synovitis and effusion were assessed. X-rays were scored in consensus according to the Kellgren-Lawrence (KL) grading scheme. Reliability was calculated using weighted (w) kappa statistics. Associations between presence of MRI features (large BMLs, severe cartilage damage, labral damage and synovitis) and radiographic OA severity were calculated using Cochran-Armitage test for trend. Results: Distribution of radiographic grading was: KL0 12 (27.3%), KL1 11 (25%), KL2 14 (31.8%), KL3 5 (11.4%) and KL4 2 (4.6%). Intra-reader reliability for the different features ranged between 0.69 (synovitis) and 0.85 (cartilage). Inter-reader reliability ranged between 0.48 (labrum) and 0.85 (BMLs). Scoring of subchondral cysts was not reliable with a w-kappa of 0.18 and 0.15. There was a strong association between presence of MRI-detected lesions and radiographic KL grade (p = 0.002) apart from labral damage (p = 0.26). Conclusion: MRI-based semiquantitative assessment of the hip joint appears to be reliable apart form scoring of subchondral cysts. Presence of more severe MRI-detected intraarticular pathology shows a strong association with radiographic OA apart from labral damage. acetabular roof, rotating from anterior to superior to posterior: (1) acetabular diameter (width of the cup), (2) acetabular-extension (depth of the cup), (3) acetabular depth as defined by Pfirrmann et al (depth correlated to the centre of the femoral head) and (4) the beta-angle, as distance to the head-neck offset. We related these parameters to age, status of the epiphyseal scar and location in the joint. Results: Looking at the influence of age, we found that the acetabular diameter increases highly significant (p < 0.001). The acetabular extension, also growth by aging highly significant (p = 0.05). The beta angle decreases with growing older, but not significant (p > 0.05). Dividing the population into two groups, one with open epiphysis and the other with a closed one, we could not find any significant changes between these both groups (p < 0.05). Significant changes could mostly be found in the superior, ventral superior and dorsal superior parts. Conclusion: Within our normal cohort no asymmetric changes could be observed in coverage during maturation using MRI. These data can be compared to hips with pathological coverage (hip dysplasia, pincer-type FAI) to better understand potential influences during growth, genetic preposition and to develop preserving possibilities.

The femoral epiphyseal torsion angle in MR arthrography of the hip: diagnostic utility in patients with femoroacetabular impingement syndrome C. Schaeffeler, K. Holzapfel, M. Eiber, H. Gollwitzer, E.J. Rummeny, K. Woertler; Munich/ DE (schaeffeler@roe.med.tum.de) Purpose: To evaluate if the epiphyseal torsion angle (ET-angle) shows pathologic values in patients with clinically suspected femoroacetabular impingement syndrome (FAI) referred to MR arthrography (MR-A) of the hip and to analyse whether patients with pathologic values show higher incidence of hip pathology. Methods and Materials: 68 consecutive MR-A of the hip in patients with positive clinical impingement tests were included to the study group. The control group was composed by 30 hips of healthy volunteers without clinical signs of FAI, who underwent standard MRI. The angle between a line perpendicular to the proximal epiphyseal scar of the femur and the long axis of the femoral neck (ET-angle) was measured independently in both groups on oblique axial T1w images by three radiologists. Furthermore, all MR-A were retrospectively evaluated for internal derangement and clinical data were reviewed. Results: Mean values of the ET-angle in the FAI-group were 12.1° (range: -3.1°-29.4°) for reader 1, 12.2° (0.8°-28.9°) for reader 2 and 12.8° (0.5°-29.9°) for reader 3. In the control group the mean values were 7.5° (2.4°-17.8°) for reader 1, 6.9° (0.7°-18.5°) for reader 2 and 9.6° (2.1°-19.8°) for reader 3. T-test analysis showed highly significant lower ET-angles in healthy volunteers compared to the FAI-group for all three readers (p < 0.010). The calculated intraclass correlation coefficient of 0.75 represented good interrater agreement. Patients with labral lesions detected by MR-A did not show significantly increased ET-angles. Conclusion: Patients with clinically suspected FAI show significant higher ETangles than healthy volunteers. Values > 20° should be considered pathologic.

Superior acetabular notch (pseudocartilage defect): prevalence in MRarthrography and correlation with surgery T.J. Dietrich, A. Suter, C.W.A. Pfirrmann, C. Dora, M. Zanetti; Zurich/CH (tdiet@gmx.net) Purpose: In literature, based on a few dried skeletal specimens, the supracetabular notch has been described as an anatomic variant. The purpose of this study was to evaluate the prevalence of such superior acetabular notches (SAN) (pseudocartilage defects) in MR-arthrographies of the hip and to correlate with surgery. Methods and Materials: 838 consecutive MR-arthrograms were reviewed for the presence of superior acetabular notches (SANs). SANs were classified into two types: SAN type-1 with an accessory fossa in the apex of the acetabulum and a distinct contrast media inflow into the fossa, and SAN type-2 with a fossa without contrast media inflow. MR-arthrographies with SANs were correlated with surgery if available. Results: Twenty-three (2.7%) patients (10 females, 13 males; mean age 29.3 years) presented with SAN type-1: depth, 2.6 mm; width 5.0 x 4.5 mm; localisation lateral to the acetabular notch, 7.7 mm. Seventy-seven (9.2%) patients had a SAN type-2 (37 females, 40 males; average age 38.6 years): depth 2.4 mm, width 4.8 x 4.6 mm, localisation lateral to the acetabular notch 8.1 mm. Four hip joints with SAN type-1 and 11 joints with SAN type-2 underwent surgery. No cartilage defect was found in the area of the SAN; however, in one case the surgeon described an accessory fossa in the dome of the acetabulum as an anatomic variant. Purpose: The purpose of this study was to assess hepatic steatosis in morbidly obese patients undergoing gastric bypass surgery using open magnet 1.0 T 1H-MRS with histopathological control. Methods and Materials: Patients underwent 1H-MRS for the assessment of steatosis before and three months after surgery. Liver biopsy was performed during surgery. Steatosis was expressed as a ratio of the fat peak area over the cumulative water and fat peak areas. Histopathological percentages of steatosis were graded as follows: none (0-5%), mild (5-33%), moderate (33-66%) and severe (> 66%). Spearman's correlation and accuracy was assessed. Differences between groups were assessed with the Wilcoxon signed rank and Mann-Whitney U analysis. Results: We included 38 patients, with mean age and BMI of 45.6 years and 47.9 kg/m 2 , respectively. Before surgery mean steatosis measured with 1H-MRS was 10.3%. 1H-MRS showed an accuracy of 32/36 (89%) and significantly correlated with histopathological assessment of steatosis (r = 0.85, p < 0.001). 1H-MRS was able to discriminate none from mild steatosis (p = 0.011), mild from moderate steatosis (p < 0.001) and moderate from severe steatosis (p = 0.037). Three months after surgery steatosis decreased to 5.2% (p < 0.001). Purpose: To evaluate the feasibility and accuracy of dual energy CT (DECT) for ex-vivo quantification of liver fat content (LFC) in the presence of iron and contrast media (CM).

Sixteen phantoms with a defined LFC (0-50%) and with varying liver iron content were scanned with a second-generation DECT-system. Phantoms were scanned unenhanced and contrast-enhanced using an iron-specific dual energy 3-material decomposition algorithm providing virtual non-iron images (VNI) to subtract iron and CM from the data. Single-energy (SECT) and DECT measurements (HU) of unenhanced and contrast-enhanced phantoms were compared to the known LFC using Pearson's analysis and Student's t-test for related samples. 

The mean ADC value decreased as the severity of liver fibrosis increased (P < 0.01), and was significantly different between all groups of fibrosis (P < 0.05). The mean choline/lipid ratio increased as the severity of liver fibrosis progressed (P < 0.05), and was statistically different between severe fibrosis/cirrhosis and the other groups of fibrosis (P < 0.05). Conclusion: There is a decrease of ADC value and increase of choline/lipid ratio with increasing degrees of liver fibrosis. The preliminary results demonstrate the potential usefulness of diffusion-weighted imaging and proton MR spectroscopy for quantification of liver fibrosis.

Tissue echo patterns analysis with acoustic structure quantification (ASQ): preliminary experience of new non-invasive method for staging liver fibrosis in chronic hepatitis V. Cantisani, P. Ricci, A. Marcantonio, C. Marigliano, V. Maldur, A. Porfiri, R. Passariello; Rome/IT Purpose: The acoustic structure quantification technology is a novel ultrasound method that provides analysis of the fibrosis level in the liver using the statistical information in the raw data echo signals. The aim of this study was to evaluate ASQ in the assessment of liver fibrosis in patients with chronic autoimmune, criptogenetic and viral hepatitis B and C. Methods and Materials: 57 patients (26 females, 31 males, mean age 57 years) with chronic viral hepatitis underwent ASQ imaging and consecutively liver biopsy. The results were compared with the histological fibrosis degree (F), which served as the reference. 20 healthy volunteers received ASQ quantification of different segments of the liver. Results: The best ASQ assessments with the lowest rate of invalid measurements were carried out by an intercostal approach to segment VII/VIII of the liver. The ASQ values of the healthy group had a mean of 108 (range 88-122), the means of the patient group ranged from 100 to 150. ASQ quantification correlated significantly with the histological fibrosis stage: no fibrosis or mild fibrosis (< or = F2), severe fibrosis (F4 or F5) and cirrhosis (= F6), respectively. Conclusion: This study shows the ASQ is a noninvasive, quick and suitable method for assessing liver fibrosis or cirrhosis in patients with HBV, HCV or autoimmune disorder. ASQ measurements of the liver should be performed via a subcostal access.

Increasing ASQ values velocities correlate with higher degree of hepatic fibrosis. Purpose: Comparison of MR elastography (MRE) and fibro-C index for quantification of liver fibrosis in chronic hepatitis B. Methods and Materials: A retrospective comparison of quantification of liver fibrosis with MRE and fibro-C index was performed in 30 patients with chronic hepatitis B. Stiffness of the liver was estimated by placing regions of interest on the stiffness maps MRE and a mean value in kilopascals (kPa) obtained. The liver biopsy samples were processed with standard histology technique and an experienced hepatopathologist graded for fibrosis using METAVIR score. On the same biopsy sample, quantification of the amount of collagen with fibro-C index was performed with Gaussian mixture model segmentation method and quantification index representing total aggregated collagen amount was obtained. The stiffness values, METAVIR scores and quantification index scores were correlated. Receiver operator curve (ROC) analysis was performed for accuracy of MRE and fibro-C index for detection of fibrosis. Results: Liver stiffness values with MRE and quantification by fibro-C index correlated well (R2 = 0.64, p < 0.01). There was no significant difference between the accuracies of MRE and fibro-C index for detection of fibrosis (0.94 vs. 0.90, P = 0.5), significant fibrosis (0.95 vs. 0.90, P = 0.41) and cirrhosis (0.99 vs. 0.94, P = 0.2). However, MRE performed better than fibro-C index in differentiating the different grades of fibrosis from normal liver. Conclusion: MRE derived stiffness of liver correlates well with fibrous content of liver and differentiates fibrotic liver and normal liver, therefore MRE may be suitable alternative for non-invasive assessment of liver fibrosis burden in chronic hepatitis B patients.

MRI was performed on a 1.5-Tesla scanner using a 3D-DE-SPGR sequence with automatic chemical-shift selective reconstruction of in-/ out-of-phase (IP/OP) and fat-signal-only images. Absolute LFC was estimated by means of MR fat-fractions (FF-MRI) as derived from IP/OP (FF-IP/OP) and from fat-only images (FF-FAT). In vitro, seven phantoms containing homogenised liver/ fat concentrations from 0 to 50% (FF-Phantom) were examined. In vivo, 43 biopsies in 22 patients were retrospectively analysed for the percentage of hepatocytes containing lipid droplets (cell fraction; CF-HISTO) and for the absolute fat-fraction per field-of-view (FF-Phantom). Results: In vitro significant linear correlation was observed between FF-MRI and FF-Phantom (r = 0.993; p < 0.0001). Mean measurement bias of absolute values was significantly smaller for FF-IP/OP compared to FF-FAT (+2.6/+4.5). In vivo histopathological measurements ranged from 5 to 90% for CF-HISTO and from 0 to 20% for FF-HISTO. Correlation of FF-MRI with FF-HISTO was significantly better when compared to CF-HISTO (p < 0.0001). Mean difference of FF-MRI to CF-HISTO was 21.1% and to FF-HISTO 4.8%. Mean measurement bias was significantly smaller for FF-IP/OP compared to FF-FAT (p < 0.0001). Conclusion: FF-MRI derived from 3D-DE-SPGR reflects the absolute LFC per area or volume and not the percentage of fatty hepatocytes as determined with traditional histopathological evaluation. FF-IP/OP is superior to FF-FAT for estimation of LFC showing significant less overestimation especially in a clinically relevant range from 0 to 5%. Purpose: Few longitudinal imaging studies of liver-engrafted-islets are available for ITA/IAK patients. The link between islet-function and the appearance of islet-induced-LFFC-US is controversial. The aims of this study were to assess prospectively LFFC-US and their relationship with islet-function. Methods and Materials: LFFC-US timing of detection, prevalence and duration were assessed in 30 IAK, in 5 ITA and, retrospectively, in full-, partial-, and no-function-groups, according to islet-function evaluated one year after islet-transplantation. Patients with persistent LFFC-US underwent liver biopsy. LFFC-US-positive and LFFC-US-negative patients with functioning-islet were compared for islet-function and C-peptide-levels during the follow-up. Results: LFFC-US were found in 12 patients (10/30 IAK, 2/5 ITA). First detection was at 6 months in 8 cases, at 12 months in 4 cases. LFFC-US last longer than 1 year in 8 cases. Steatosis was found histologically in 8/8 patients. At 12 months, LFFC-US were detected at a higher extent in patients with partial-islet-function (10/12, 8 IAK, 2 ITA) compared with patients with full-islet-function. C-peptide-levels were significantly lower in LFFC-US-positive than in LFFC-US-negative patients. At 18 months, LFFC-US-positive were more prone to worsen their function (9/12) compared with LFFC-US-negative-patients (3/18). Conclusion: US represents a sensitive tool to diagnose/monitor LFFC-US, resulting more related to partial-than full-islet-function, in both IAK and ITA patients. We could suppose that in case of full-function the local insulin secretion is not enough to induce steatosis. Differently, in some cases of partial function, steatosis could be the expression of remaining vital islets stressed in insulin overproduction due to lost function of other islets.

Quantification of liver fibrosis using diffusion-weighted MR imaging and proton MR spectroscopy in experimental rabbit models Q. Wang, C. Liang, H. Liu; Guangzhou/ CN (wangqscmu@gmail.com) Purpose: To investigate whether the degree of liver fibrosis can be quantified non-invasively with diffusion-weighted MR imaging and proton MR spectroscopy using an experimental rabbit liver fibrosis model. Methods and Materials: 5 male New Zealand White rabbits of control group and 40 rabbits of liver fibrosis group were examined with diffusion-weighted MR imaging and 1 H MR spectroscopy. Rabbit liver fibrosis models were induced by transperitoneal injection of CCl 4 and were divided into no fibrosis (stage 0), mild/ moderate fibrosis (stages I and II) and severe fibrosis/cirrhosis (stages III and IV) groups based on histological examinations from resected liver. Diffusion-weighted SE EPI single shot sequence (b1 = 0 sec/mm 2 , b2 = 600 sec/mm 2 ) and 1 H-MRS PRESS technique (TR = 1500 msec, TE = 35 msec) were used to measure ADC values and choline-to-lipid ratios. We compared the mean ADC value and choline/ lipid ratio among different groups of liver fibrosis. Purpose: Evaluation of a CTPA study for the detection of pulmonary embolism (PE) represents a time consuming and tiring reading process. Computer-aided detection (CAD), when used as second reader, was found to improve reader performance especially for the detection of small emboli but inevitably at the expense of a substantial increase of reading time. We hypothesised that CAD may be used more advantageously when used as concurrent reading tool.

B C A D E F G H

In this institutional review board-approved retrospective study, six observers of varying experience evaluated 157 negative and 39 positive 64-slice consecutively acquired CTPA. With a time interval of 6 weeks, all cases were read twice in different order using CAD as second and as concurrent reader, respectively. Per patient, observers were asked to determine the presence of PE using a 5 point confidence scale and to document their reading time with and without CAD. Sensitivity and specificity were calculated by comparing the reader data with an independent consensus standard. Results: Baseline performance without CAD was high with a mean sensitivity of 91% and increased further to 95% with CAD as second reader and 94% with CAD as concurrent reader. Mean specificity was not adversely affected (96%, 94% and 96%, respectively). The mean reading time with CAD as concurrent reader significantly decreased (108 s, 136 s and 91s, respectively, p < 0.001). Conclusion: CAD used as concurrent reader has the potential to achieve the same high sensitivities without loss of specificity at the advantage of a significantly shorter reading time. Purpose: To determine the correlation between increasing thrombus load as a predictor of right ventricular failure as determined by computerised tomography pulmonary angiography (CTPA) findings to form a basis for correlation with clinical outcome.

Intraobserver and interobserver variability (ie-iaOVar). in the estimation of liver fibrosis with shear wave elastography (SWE) in chronic liver disease: a proposition for standardisation through a modified protocol P.S. Zoumpoulis, I. Theotokas, A. Plagou, K. Mpouhra, E. Vafeiadis; Athens/GR (p.zoumpoulis@echomed.gr)

Purpose: To estimate and correct ie-iaOVar. of measurements of liver fibrosis using SWE.

We studied 60 cirrhotic patients using SWE. All patients had liver biopsy performed in a period not exceeding ±4 months from US examination. During the first part of the study, all patients had a SWE examination of the liver performed by two independent examiners. Both examiners obtained two SWE measurements representing the mean tissue elasticity in KPa. During the second part of the study, each examiner performed a third SWE examination which was processed by a prototype analysis software to quantify the degree of liver fibrosis. The findings were compared for fibrous liver staging (F stage) using liver biopsy.

We observed large ie-iaOVar. differences in the mean elasticity measurements during the first part of the study. The SWE images collected during the second part of the study were analysed by calculating statistical features (mean, standard deviation, kurtosis, skewness and entropy) of the elasticity values within multiple measurement ROIs as well as relative ratios between different ROIs. Multiple regression analysis was then performed to quantify the index of liver fibrosis. We found good correlation of the SWE-based quantification of liver fibrosis and the histologically derived liver fibrosis staging results. Conclusion: SWE of the liver is an effective approach for estimating liver fibrosis, which exhibits good correlation with histologically proven F stage results. A modified analysis, adapted to the liver anatomic features and relying on multiple statistical features was found to reduce the ie-iaOVar. of measurements. angiography, DE perfusion maps were visually scored for perfusion defects using a novel scoring system (P-Score). Vascular obstruction score was determined using an established scoring system (Mastora) based on the CT pulmonary angiography (CTPA) data. Both scores were correlated with short-axis diameters of the left heart ventricle, right ventricular to left ventricular diameter ratio (RV/LV ratio), and width of the pulmonary trunk, as well as clinical data related to PE severity (oxygen saturation, plasma troponin I, arterial partial pressure of oxygen (PaO 2 ), heart rate, blood pressure, shock index, 30-days-survival). Reading times were measured. Results: P-Score correlated significantly (p < 0.05) with vascular obstruction score (O-score) (r = 0.65), RV/LV ratio (r = 0.47), left ventricular diameter (r = -0.36), width of the pulmonary trunk (r = 0.26), and with the clinical parameters troponin I (r = 0.43), and PaO 2 (r = -0.50). O-Score correlated significantly with RV/LV ratio (r = 0.36), width of the pulmonary trunk (r = 0.27), PaO 2 (r = -0.41), and troponin I (r = 0.37). Mean time for generation of P-score and O-score was 140 ± 72 and 252 ± 108 seconds. Conclusion: A newly introduced, DECT-based pulmonary perfusion defect scoring system correlates well with clinical and CT parameters of PE severity. It is an easy-to-perform method and represents a valuable clinical tool for diagnosis and prognosis of acute PE.

A S223 C D E F G H Conclusion: CTA identified PE in 10.3% of ED patients. In addition, CTA provided an alternate diagnosis which impacted acute treatment in 17% of patients. Pneumonia was the most common non-PE diagnosis and was radiographically occult in nearly 50% of patients. Seventeen per cent of patients had indeterminate nodule (s).

Do incidentally discovered pulmonary emboli on contrast-enhanced abdominal CT scan warrant further evaluation? Y. Yagil, A. Engel, L. Guralnik; Haifa/IL (y_yagil@rambam.health.gov.il) Purpose: Unsuspected pulmonary embolism (PE) is a well-recognised entity; however, there are no publications regarding incidentally discovered PE on contrastenhanced (CE) abdominal CT scans. The aim was to determine the added value of a CE-chest CT scan to abdominal CT scan for the confirmation of PE. Methods and Materials: CE-chest CT and abdominal CT examinations of 35 patients in whom PE was detected on the abdominal CT examination were retrospectively studied. Data pertaining to patient characteristics and CT examinations were obtained from computerised medical records. PE was evaluated by the observation of occlusive pulmonary arterial filling defects on contiguous images. Verification of PE and quantification of arterial clot load using the Miller scoring system was assessed on both examinations. Results: The study population included 15 (43%) men and 20 (57%) women, with a mean age of 63.5 years (range 22-89 years). Seventeen (49%) patients had a history of active malignancy and 7 (19%) of previous thromboembolic episode. All patients diagnosed with PE by the abdominal CT scan had positive PE findings on the chest CT scan as well. The average difference in arterial clot load score between the two scans was 22%. In 46% of the cases no difference in score was observed. Conclusion: Positive PE findings on CE-abdominal CT scan can provide the necessary information required for decision on patient management. Therefore, incidental discovery of PE on CE-abdominal CT scan does not warrant further evaluation by a CE-chest CT scan.

Clinicians' response to a report of an incidental pulmonary embolism detected on contrast-enhanced CT J.M. Franklin, F.V. Gleeson; Oxford/UK (jamiemfranklin@hotmail.com) Purpose: Incidental pulmonary emboli (iPE) may be detected on contrast-enhanced CT (ceCT) of the chest performed for other reasons; the prevalence of this finding is approximately 2.5%. The purpose of this study was to report on both clinical response related to the finding of an iPE on ceCT in the UK, and the reasons for the responses. Methods and Materials: Patients with an iPE detected on a ceCT incorporating the chest were retrospectively identified for a 12-month period. Case note review was performed to assess clinical response to this finding. Patient and iPE characteristics, subsequent treatment, and outcome related to treatment and nontreatment were recorded. Results: There were 68 patients with iPE. 37% of iPE were in the pulmonary trunk or main pulmonary arteries, 35% were lobar and 28% were in the segmental or subsegmental arteries. 62 of 68 (91.2%) patients with iPE were treated with therapeutic anticoagulation. There were 2 (3.2%) serious adverse events related to treatment. Six (8.8%) patients were not treated. Two (33%) developed recurrent venous thromboembolism (VTE); the remainder suffered no adverse sequelae. Clinicians' decisions to treat iPE were not significantly associated with patient age, PE risk factors or PE size. A smaller proportion of single iPE were treated than multiple iPE (19/24 vs 43/44; p=0.033). Conclusion: There are adverse outcomes associated with both treatment and non-treatment of iPE. However, despite the uncertainty about the natural history and clinical significance of iPE, the majority of patients at our institution with iPE receive prompt anticoagulation.

Pulmonary arterial hypertension associated with systemic sclerosis: prevalence in the absence of interstitial lung disease V. Wilkinson, K. Karunasaagarar, J. Hurdman, C. Elliot, D. Kiely, C. Hill; Sheffield/UK (Victoria.Wilkinson@sth.nhs.uk) Purpose: Pulmonary arterial hypertension (PAH) may occur independenly of interstitial lung disease (ILD) in systemic sclerosis (SSc). We aimed to assess the proportion of patients in our SSc population with PAH, in the absence of ILD. Methods and Materials: Retrospective analysis of 153 SSc patients investigated for possible PAH. 102 patients included who had high resolution computed tomography (HRCT) and right heart catheterisation (RHC) within 3 months. PAH diagnosed by Methods and Materials: Retrospective cohort analysis of positive CTPA versus negative CTPA scans was performed over 40 months. Right ventricular:left ventricular ratio (RV:LV ratio), presence of septal shift, pulmonary artery and aortic size were analysed with respect to thrombus load as determined by the modified Miller score (MMS) using non-parametric analysis (two-tailed or Chi-squared test) and Pearsons rank correlation. Results: 504 positive CTPA scans and a representative cohort of 100 negative CTPA scans were included. Increasing MMS was found to correlate with higher RV:LV ratio with statistically significant difference between the RV:LV ratio between the negative cohort versus the positive cohort with MMS 12 or more (mean RV:LV ratio 1.323 vs 0.930, p < 0.0001) with no significant difference with MMS < 12. Septal shift was also more likely with a MMS 12 or more compared to MMS < 12 (63.8% vs 18.6%, p < 0.001) with an increase in pulmonary artery size with increasing MMS (Pearson rank coefficient r = 0.221, p < 0.001). Conclusion: This is the first study to accurately describe the relationship between thrombus load and an increase in RV:LV ratio beyond a MMS score of 12 using CT pulmonary angiography. This suggests a 'tipping point' beyond which right ventricular decompensation occurs and may add to further risk stratification in patients with acute pulmonary embolism on CTPA. Purpose: To evaluate current MRI performance for PE diagnosis, by reference to 64-detector multislice CT angiography (CTA) findings. Methods and Materials: IRB approval and informed consent were obtained. Three hundred patients out of 1796 patients referred for PE suspicion in our institution between June 2007 and June 2009 were included. MRI was performed within 24 hours from CTA on a 1.5 Tesla unit and included 3 different sequences: 1-Unenhanced steady state free precession (SSFP) (2D FIESTA) axial transverse sequences, with ECG gating, with and without breath-holding, 2-fast SPGR coronal perfusion sequence (TR: 2, TE: 0.8) after injection of 0.1 mmol/kg DOTA-Gd at 5 ml/s, 3-fast SPGR axial transverse angiographic sequence (TR: 3.4, TE:1.2) after injection of 0.1 mmol/kg DOTA-Gd at 3 ml/s. Two independent readers analysed all sequences together (global reading) and each sequence separately in a random order. Results: MRI protocol was completed in 274 patients who had a conclusive CTA. The rate of indeterminate result was 28% for reader 1 and 30% for reader 2. Excluding indeterminate results, MRI sensitivity and specificity on global readings were 84 and 99% for reader 1 and 79 and 100% for reader 2. The kappa value for interobserver agreement was 0.59 (95% CI: 0.50 -0.67). Unenhanced breath-holded SSFP sequence sensitivity and specificity were 76 and 96% for reader 1 and 69 and 99% for reader 2. Conclusion: MRI shows high specificity, even on unenhanced sequences and acceptable sensitivity for PE diagnosis at the cost of a high rate of indeterminate results. Purpose: Pulmonary embolism (PE) is potentially life-threatening. Its clinical findings can mimic other conditions. CT angiography (CTA) is the cornerstone of evaluating patients with suspected PE presenting to the Emergency Department (ED). The purpose of this study was to 1. determine the incidence acute PE among adult ED patients undergoing CTA and 2. determine the frequency of alternate diagnoses which affect patient management and its concordance with chest radiography. Methods and Materials: Two board-certified radiologists retrospectively reviewed the clinical information and radiologic findings of 496 patients suspected of PE who presented to the ED in 2009. We documented the presence of PE, alternate diagnoses and correlation to radiography. Non-PE CT findings were categorized into three groups: A = acute change in clinical management; B= follow-up, but no change in clinical management; C = normal or no clinical significance. Results: CTA was nondiagnostic in nine patients. The frequency of PE among 487 patients was 10.3%. The prevalence of the categorized groups was A = 17%, B = 24%, and C = 69%. The predominate category A finding was pneumonia (88%). Fifty-two per cent of these pneumonia patients had concordant findings on chest radiograph. The most common group B finding was an indeterminate pulmonary nodule (70%) requiring follow-up in 17% of total patients.

limbs 40 were primarily treated by endovascular-, 45 by surgical-and 5 by combined techniques. Endovascular treatment was applied to significantly lower TASC stages (p = 0.001). Limb salvage rates were 95% (1 year), 95% (2 years), 95% (3 years) and 87% (4 years), respectively. Survival rates were 86% (1 year), 81% (2 years), 70% (3 years), and 66% (4 years), respectively. There were no significant differences between the treatment groups for limb salvage (n = 0.20) and survival (n = 0.58), respectively. Costs for initial surgical treatment were 6.298 Euro, and 3.869 Euro (for endovascular treatment) per patient. Total treatment costs (including 5-year follow-up) were 27.429 Euro for the surgery first approach and 17.443 Euro for the endovascular first approach per patient. Conclusion: Interdisciplinary co-operation achieves good long-term results for limb salvage and survival. An endovascular first approach has lower costs than a surgery first approach and should be preferred when applicable.

Randomised clinical study of the Zilver ® PTX ® self-expanding Nitinol stent with polymer-free paclitaxel coating shows improved 12-month effectiveness over angioplasty and bare metal stents for the superficial femoral artery

Purpose: Zilver ® PTX ® is the first drug-eluting stent with CE approval for the superficial femoral artery (SFA). This multicenter, randomised study compared the safety and effectiveness of the Zilver PTX to balloon angioplasty (PTA) and bare metal stenting (BMS). Methods and Materials: Patients with de novo or restenotic SFA lesions were randomised to PTA or Zilver PTX stent placement. PTA patients experiencing procedural failure (> 30% residual stenosis) underwent secondary randomisation to provisional stenting with Zilver BMS or Zilver PTX. Endpoints included event-free survival (EFS) and primary patency by Duplex ultrasound (peak systolic velocity ratio < 2.0). Results: 479 patients were enrolled, with 241 randomised to the Zilver PTX group and 238 to the PTA group. Demographics and lesion characteristics were similar for the groups. Approximately half the PTA group experienced procedural failure and underwent secondary randomisation, assigning 59 and 61 patients to provisional stenting with Zilver BMS and Zilver PTX, respectively. 12-month results included non-inferiority of EFS rate and superiority of patency rate for the Zilver PTX groups compared to the PTA group. The stent fracture rate was less than 1%. There was significant clinical improvement with Zilver PTX over both PTA and BMS. Furthermore, provisional Zilver PTX showed significantly improved primary patency over provisional BMS, demonstrating the benefit of the polymer-free paclitaxel coating.

The Zilver PTX results support safety and show significantly improved effectiveness over PTA and BMS for treatment of the SFA.

Subintimal angioplasty for superficial femoral artery tasc II d lesions in critical limb ischaemia: outcomes with and without stenting and value of stent position for secondary patency after subintimal channel occlusion R. 

In a prospective trial consecutive patients with stenosis/ occlusion of the femoro-popliteal vasculature with a length ≤ 90 mm are treated by atherectomy with the SilverHawk/TurboHawk catheter. The intra-arterial pressure gradient over the target lesion, measured by Radi PressureWire, is controlled before and after each pass with the atherectomy device. Haemodynamically sufficient treatment is defined as a complete reduction of the pressure gradient to 1.0. The endpoint of atherectomy is defined angiographically by the interventionalist who is blinded to the results of pressure measurements. Results: 20 patients were enrolled. Mean lesion length was 5 cm, mean pressure gradient 0.72. Complete reduction of the pressure gradient was achieved with 4 ± 2 passes in mean. For a satisfactory angiographic result in mean 11 ± 4 passes were necessary (p < 0.001). Conclusion: Our results demonstrate that directional atherectomy can achieve haemodynamically sufficient stenosis treatment with significantly less catheter passes than angiography indicates, allowing for a shortening of the procedure and reduced complication rate.

Alternative techniques for below-the-knee recanalisation: procedural and clinical results from a single center experience R. Gandini, L. Boi, C. Del Giudice, A. Spinelli, C. Reale, G. Simonetti; Rome/IT (boiluca.md@gmail.com) Purpose: Percutaneous arterial recanalisation with PTA is becoming cornerstone of treatment of below-the-knee vascular peripheral disease. Nonetheless, successful recanalisation is not always obtained with traditional antegrade approach. Recently, new alternatives techniques have been introduced to recanalise these cases not treatable with antegrade approach. We report our results focusing on anatomical indications technical aspects and rationale of these new approaches.

In a single-center retrospective clinical analysis, from January 2005 to July 2010, we collected 1602 patients with critical limb ischaemia for complex lesions of the popliteal and infrapopliteal vascular territory. In 224 patients (14%), traditional antegrade revascularisation failed and alternative techniques were adopted. Of these subgroups, 85 limbs were treated with transpedal retrograde approach, in 46 patients a pedal to plantar and 36 patients a plantar to pedal retrograde revascularisation was performed. A transcollateral angioplasty and retrograde revascularisation was suitable in 57 patients. Results and Conclusion: Successful recanalisation was obtained in 219 patients (98%) with alternative techniques. Adjunctive popliteal stenting was performed only in 13 patients.

Use of a percutaneous suture-mediated closure system to achieve haemostasis after 6-8 F transfemoral approaches P. Kamusella 1 , C. Wissgott 1 , H.J. Steinkamp 2 , R. Andresen 1 ; 1 Heide/DE, 2 Berlin/DE (pkamusella@wkk-hei.de) Purpose: The safety and effectiveness of a percutaneous suture-mediated closure system for achieving haemostasis after a transarterial femoral approach were investigated. Methods and Materials: A percutaneous suture closure system (Perclose/ Per-closeProglide, Abbott Vascular) was used in 2200 patients (average age 67.5 ± 12.5 years) who had undergone angiographic intervention. Vascular access sites with a sheath size of between 6 F and 8 F were closed. Platelet aggregation was inhibited with 100 mg/d ASA and all of the patients additionally received 5000 IU heparin periinterventionally. After application of the suture-mediated closure system and achievement of sufficient haemostasis, a light compression bandage was applied in all patients and 6 hours' bed rest was recommended. Post-interventionally (following day and after 6 weeks), the puncture site of all patients was checked using colour-coded duplex sonography. Results: Immediate haemostasis was achieved in 2103/2200 patients (95.6 %). In the remaining 97/2200 cases, a correct development of the suture was not possible, because of calcifications. In such cases, haemostasis was achieved by manual compression. Major complications involved three infections that required vascular surgical debridement with the use of an interposition graft. In addition, there were five cases of secondary bleeding requiring transfusion. Minor complications involved 3 (0.14 %) pseudoaneurysms, 256 (11.6 %) groin haematomas (up to max. 3 cm) and 85 (3.9 %) palpable suture granulomas. Conclusion: At puncture sizes of 6-8 F, safe and effective haemostasis is possible with the percutaneous suture-mediated closure system. Insufficient suture closure is to be expected in the case of strongly calcified vessels.

Transfemoral and transtibial combined approach in subintimal recanalisation of SFA obstructions extending on popliteal and distal vessels origin R. Gandini, E. Pampana, V. Da Ros, L. Boi, M. Stefanini, G. Simonetti; Rome/IT (boiluca.md@gmail.com) Purpose: To propose our technique, to obtain the optimal recanalisation to the foot in case of long SFA occlusions involving the popliteal trifurcation, using a combined antegrade and retrograde subintimal approach in patients presenting critical limb ischemia (CLI). Methods and Materials: On 664 diabetic patients treated for limb salvage, in 75 with SFA long occlusions involving the leg vessel tree, subintimal antegrade and retrograde approaches were performed (posterior tibial artery in 43 and anterior tibial artery in 32 cases). The patent portion of the runoff vessel was previously assessed by magnetic resonance angiography and directly punctured under ultrasound Doppler guidance. A subintimal channel rendezvous was performed to allow snaring of the guidewires. Subsequent balloon dilatation was performed. Results: We achieved 98.2% technical success. At Doppler-US mean follow-up of 24.5 months, the patent vessels were 71.2%, but we had a 92.3% in limb salvage with complete healing of limb lesions and rest pain resolution. The oximetry value showed an increase from mean original value of 17.8 mmHg to 45.1 mmHg at 6 months follow-up. Conclusion: In patients with SFA occlusion involving the popliteal trifurcation, secure candidate in amputation, combined antegrade and retrograde subintimal recanalisation approaches, is probably the most suitable and efficacy endovascular option to obtain a direct flow to the foot and so, an high percentage of limb salvage.

Directional atherectomy of heavily calcified stenotic lesions of the lower limb within joint segments P. Minko, S. Jaeger, A. Buecker, M. Katoh; Homburg a.d. Saar/DE (peterminko@yahoo.com) Purpose: To evaluate the mid-term outcomes (2 year) of patients with peripheral occlusive disease (POD) and heavily calcified stenotic lesions in segments with high biomechanical stress after atherectomy. Method and Materials: Patients suffering from POD (Rutherford 3 to 6) were treated with the Silverhawk atherectomy device (ev3 Endovascular, MN, USA) if heavily calcified lesions in the most proximal or distal 3 cm of the superficial femoral artery and/or popliteal artery were present. Overall 42 patients with 46 lesions (15 females, 27 males; mean age: 69 ± 8.8) were included into this prospective study. Patients were followed up after 6, 12, 18 and 24 months for clinical re-evaluation including the measurement of the maximum walking distance and the ankle brachial index (ABI) and to perform duplex sonography. Results: The primary success rate of the procedure was 89%. In five cases additional PTA and/or stenting was necessary. Procedure-related complications namely peripheral embolism occurred in 3 cases (6%). The mean Rutherford score decreased significantly from 4 to 0.8 (p < 0.001), while the mean ABI increased from 0.62 to 0.94 (p = 0.088) after 24 months. The primary and assisted primary patency rate was 66% and 79%, respectively. Conclusion: Directional atherectomy should be considered as an alternative for treatment of heavily calcified stenotic lesions in segments with high biomechanical stress as in most cases stenting can be avoided. Results after two year demonstrated a significant decrease of the Rutherford score and an increase of the ABI and a reasonable patency rate.

Haemodynamic endpoint definition of directional atherectomy in the superficial femoral and popliteal artery by direct intraarterial pressurewire measurments J. Becker-Lienau, M.F. Reiser, M. Treitl; Munich/ DE (mtreitl@med.uni-muenchen.de) Purpose: A major drawback of directional atherectomy is potential over-excision with the risk of vessel rupture or aneurysm formation. Since there is no study concerning the endpoint of atherectomy, the interventionalist decides by angiogram and experience when to stop atherectomy. The typically irregularly vessel silhouette induces the interventionalist to do more passes then maybe necessary. Our study compares the least necessary count of passes of directional atherectomy for sufficient stenosis treatment based on (1) intra-arterial pressure measurements and (2) 

A low correlation between the automatic method and MRI results was found for all workstations. Correlation and Bland and Altman plots between DS-CT and MRI, obtained with the semiautomatic analysis on all WS were good (r2 = 0.81, 0.86 and 0.83, respectively). A constant error was found for all WS (+5%, -3% and +4%) for the automatic evaluation of EF. Mean time was significantly shorter for Leonardo WS (8'35") compared to Vitrea WS (10'20") and not significantly longer than for Aquarius WS (7' 32"). Intra-reader agreement was good for all WS (wK = 0.83, 0.79 and 0.81). Inter-reader agreement was significantly better for Leonardo and Aquarius than for Vitrea (wK = 0.85 and 0.91 vs wK = 0.75). Conclusion: A good correlation with MRI was obtained for all workstations using a semiautomatic analysis. The LV function is a predictive marker for morbidity and mortality in patients with coronary artery disease. This study investigates the variability of LV function assessment on three commercially available WS.

Gender differences in response to cold pressor test assessed with velocity-encoded cine-MR imaging of the coronary sinus A. Jacquier, A. Flavian, F. Kober, P. Cozzone, M. Bernard; Marseille/FR (alexis.jacquier@ap-hm.fr) Purpose: Gender-specific differences in cardiovascular risk are well known, and current evidence supports an existing role of endothelium in these differences. The purpose of this study was to assess non-invasively coronary endothelial function in male and female young volunteers by myocardial blood flow (MBF) measurement using coronary sinus (CS) flow quantification by velocity encoded cine MRI at rest and during cold pressor test (CPT). Methods and Materials: Twenty-four healthy volunteers (12 men, 12 women) underwent MRI in a 3 tesla scanner. Coronary sinus flow was measured at rest and during CPT using non-breath-hold velocity-encoded phase contrast cine-MRI. Myocardial function and morphology were acquired using a cine steady-state free precession sequence. Results: At baseline, MBF was 0.63 ± 0.23 mL.g -1 .min -1 in men and 0.79 ± 0.21 mL ·g -1 ·min -1 in women. During CPT, the rate pressure product in men significantly increased by 49 ± 36% (p < 0.0001) and in women by 52 ± 22% (p < 0.0001). MBF increased significantly in both men and women by 0.22 ± 0.19 mL·g -1 ·min -1 (p = 0.0022) and by 0.73 ± 0.43 mL·g -1 ·min -1 (p = 0.0001), respectively. The increase in MBF was significantly higher in women than in men (p = 0.0012). Conclusion: MRI coronary sinus flow quantification for measuring myocardial blood flow revealed a higher response of MBF to CPT in women than in men. This finding may reflect gender differences in endothelial-dependent vasodilatation. This non-invasive rest/stress protocol may become helpful to study endothelial function in normal physiology and in physiopathology. (mhatake@radiol.med.kyushu-u.ac.jp) Purpose: To determine whether chemoradiotherapy for oesophageal cancer affects left ventricular function (LVF).

The LVF of 27 patients (21 males, 47-85 years) with oesophageal cancer (CE 6, UT 3, MT 13, LT 5, Stage I-IVa) who received concurrent chemoradiotherapy (total dose to primary lesion of 40-71Gy, 5-FU and CDDP) was evaluated with 1.5-T MRI before (n = 27), at about 40Gy (n = 23), and after chemoradiotherapy (≥ 61.4Gy, n = 18). Several parameters related to LVF (end diastolic volume index (EDVI), end systolic volume index (ESVI), stroke volume index (STVI), ejection fraction (EF), cardiac output index (COI), heart rate (HR), wall motion (WM) and wall thickening (WT)) were compared among three time points using Wilcoxon single rank test. The dose of the left ventricular segments was also evaluated. Results: In thoracic oesophageal cancer, significant change (p < 0.05) was observed between before and after chemoradiotherapy in STVI (38.7±1.8 mL/m 2 vs. 33.5±2.8), EF (59.4±2.1 % vs. 56.3±2.0), HR (67.8±3.0 min -1 vs. 82.3±4.6), and WM of segment 9 (7.6±0.49 mm vs. 6.3±0.88) and segment 10 (9.8±0.54 vs. 8.9±0.7), and between before and at about 40Gy in HR (67.8±3.0 min -1 vs. 72.2±3.8) and WM of segment 8 (7.3±0.43 mm vs. 6.1±0.58) and segment 9 (7.6±0.49 vs. 6.2±0.63). No significant difference was observed in cervical oesophageal cancer except EF between before and at about 40Gy (57.6±3.8 % vs. 57.4±5.0, p < 0.05). Irradiated dose of segments 8, 9 and 10 was more than 40Gy in the patients with thoracic oesophageal cancer. Conclusion: These results indicate that chemoradiotherapy for thoracic oesophageal cancer impairs LVF.

Patient-specific left ventricular function and biomechanical parameters from cardiac CT imaging A. Baali 1 , R. Moreno 1 , M. Chau 2 , H. Rousseau 1 , F. Nicoud 3 ; 1 Toulouse/FR, 2 Castelnau le Lez/FR, 3 Montpellier/FR (baali.adil@gmail.com) Purpose: Oriented towards a more complete diagnosis for clinicians, we propose a new Computational Fluid Dynamics (CFD) that uses cardiac computed tomography (CT) explorations to automatically generate functional results including computational haemodynamics and segmental contractile parameters. This is done within the same numerical domain all over the cardiac cycle. This new method provides us relevant patient-specific functional and biomechanical parameters since they are known to be related to pathological processes. Methods and Materials: We performed a dynamic multi-slice computed tomography coronary angiography on fifteen patients. Left ventricle was extracted by segmentation and non-structured numerical domain was generated in order to represent the cavity as a grid-based on tetrahedral finite volumes with an inlet (mitral valve) and an outlet (aortic valve). The grid volume, qualified as the native mesh, is then transformed over the cardiac cycle through realistic morphological. Results: Ejection fraction, time-resolved spherical index and segmental strain rate indexes were performed for each study from deformation process. Functional imaging with typical biomechanical fields: velocity, pressure, wall shear stress and vorticity were obtained from CFD by means of arbitrary lagrangian eulerian (ALE) formulation into the AVBP code (Cerfacs-Toulouse). All results were available after 3h, including geometrical extraction (semi-automatic), deformation process and CFD results. Purpose: Diastolic dysfunction is defined as an impaired left ventricular filling. The aim of this study was to establish the ability of multislice spiral computed tomography (MSCT) in the assessment of left ventricular diastolic function. Methods and Materials: Left ventricular filling was assessed in 38 subjects (12 without any cardiac abnormality, 26 with various cardiac diseases) by transthoracic echocardiography (TTE) and MSCT. The time separating the two modalities was less than ten days. The early left ventricular filling percentage (ELVFP) defined as the sum of LV volumes during the two first diastolic phases on the ejection volume and left atrial volume indexed to body surface area (LAVi) were measured by MSCT. Mitral inflow velocity and myocardial tissue velocities were measured by TTE. A dysfunction was diagnosed from the echocardiography abnormalities. The MSCT parameters were evaluated to classify the patients into normal or abnormal diastolic function. Results: Diastolic dysfunction was found in seventeen patients by TTE. There was a significant difference of ELVFR measured by MSCT between normal diastolic function group and abnormal diastolic function group (p < 0.0001). ELVFR correlated well with E/Ea used in the estimation of the left ventricular filling pressures (Rhô = -0.52; p = 0.002). Enlarged LAVi was found in 36% of patients with diastolic dysfunction, whereas it was enlarged in only 20% with normal diastolic function. Conclusion: MSCT is a feasible method to assess left ventricular diastolic function. It can separate patients with diastolic dysfunction from patients with normal diastolic function. Purpose: Right ventricular (RV) function is an important prognostic factor in pulmonary arterial hypertension (PH). The RV fractional area change (RVFAC) is a widely used echocardiographic index of RV function but has been poorly documented in cardiac magnetic resonance imaging (CMR) studies, and this was the aim of our prospective study. Methods and Materials: 43 patients (22 F/21M; aged 60 ± 12 yrs) with precapillary PH underwent right heart catheterisation and CMR on a 1.5 T scanner (Siemens) with eletrocardiographic gating (delay ±48h). In the four chamber cine, the tricuspid annular plane systolic excursion (TAPSE) was measured and RV end-diastolic and end-systolic areas (RVEDA, RVESA) were manually delineated. RVFAC was calculated (100 x (RVEDA-RVESA)/RVEDA). RVEF was calculated using contiguous 6 mm RV short axis cines (balanced SSFP). Results: Patients had increased mean pulmonary arterial pressure (45 ± 10 mmHg) and pulmonary vascular resistance (16 ± 7 wu.m 2 ). CMR results were as follows: RVEF = 31 ± 17% (range 3-75), RVFAC = 26 ± 13% (4-65) and TAPSE = 13.6 ± 8 mm (6-26). The RVEF was related to RVFAC (r 2 = 0.78) and TAPSE (r 2 = 0.56) (each p < 0.001). Twenty-seven patients (63%) had severe RV dysfunction (RVEF < 35%). TAPSE < 15 mm predicted severe RV dysfunction with 92% sensitivity, 81% specificity, 89% positive predictive value and 86% negative predictive value. RVFAC < 30% predicted severe RV dysfunction with 96% sensitivity, 94% specificity, 96% positive predictive value and 94% negative predictive value.

The RVFAC is strongly related to RVEF, and RVFAC < 30% is a simple, sensitive and specific marker of severe RV dysfunction in PH patients. The prognostic implications of RVFAC still need confirmation.

A S229 C D E F G H Methods and Materials: All patients had DCE-US at baseline, D7, D14, D30, D60 and every 2 months. Each examination included a bolus injection of sonovue (Bracco ®) and registration of 3 minutes of raw linear data with an Aplio (Toshiba). Raw data were analysed with a mathematical model (patent PCT/IB2006/003742) to evaluate 7 parameters characterising the tumour perfusion curve. Response to treatment was evaluated every 2 months with RECIST criteria. In order to have sufficient follow-up data, the statistical analysis has to be performed more than 6 months after the inclusion of the last analysed patient. Inclusions were closed in March 2010. Results: A total of 539 patients have been included (mainly RCC (157) and HCC (107)); more than 2000 DCE-US examinations were performed. Preliminary results concerning 400 patients followed more than 6 months show that 5 parameters have a significant difference (P < 0.003) according to the response at 6 months. A cut-off between responders and non-responders was determined (decrease of 80% and 40% for AUC, respectively). Conclusion: Preliminary results (400 patients) confirm the usefulness of this tool to monitor anti-angiogenic treatments. The final result will allow us to present a complete analysis of 539 patients followed of 12 months. Purpose: Chemotherapy commonly causes liver injury through sinusoidal obstructive syndrome and steatosis. Chemotherapy-induced liver injury may make it more difficult to detect metastases secondary to reduced contrast between the injured liver and metastases. The aim of this study was to determine the accuracy of SPIO-contrast enhanced imaging in patients who have undergone chemotherapy prior to liver surgery. Methods and Materials: Local ethics committee approval was obtained. Thirty-one patients with hepatic metastases completing pre-operative chemotherapy were recruited. Images were reviewed independently by two blinded observers who identified and localised lesions with a four-point confidence scale. The alternative free-response receiver operator characteristic method was used to analyse the results, which were correlated with findings from surgery, intra-operative ultrasonography and histopathology. Results: The sensitivity in detecting colorectal metastases following chemotherapy was 78% and 76% for observers 1 and 2 (95% CI, 71%-85% and 68%-82%, respectively). The area under the alternative free-response receiver operator curves were 0.73 and 0.80 for observers 1 and 2, respectively. The inter-observer Kappa value = 0.41. Conclusion: Compared to previous published work, on chemotherapy naive patients, it is clear that the accuracy of SPIO-enhanced MR imaging in detecting colorectal metastases following chemotherapy is substantially reduced. It is therefore critical that all imaging, both pre-and during chemotherapy, is reviewed when reporting liver MRI prior to surgery. It is important that authors clarify whether patients are chemotherapy naive or not in further studies.

Functional MRI evaluation of liver tumour response after radiofrequency: short-and mid-term evolution of diffusion parameters T.-L.C. Lu, F. Becce, P. Frascarolo, A. Denys, P. Bize, S. Schmidt; Lausanne/CH Purpose: To evaluate the short-and mid-term evolutions of the apparent diffusion coefficient of lesions treated with RF, in order to determine if the ADC can be used as a marker of tumour response. Methods and Materials: Twenty patients were treated for a liver malignancy with RF and were examined on a 1.5 T/3.0 T machine with T2, gadolinium-enhanced T1 and diffusion sequences: before treatment (< 1 month), just after treatment (< 1 month) and midterm (3-6 months). The ADC was measured in the whole lesion and in the area with the most restricted diffusion (MRDA). The ROI size was also measured on the diffusion map. The Pearson/ANOVA tests were used. Results: All patients were successfully treated with complete disappearance of CE. The lesional size on T2 showed a negative evolution in time (p < 0.002). The ADC in the whole lesion showed a bell-shaped evolution (increasing just after RF, then decreasing, p = 0.02). The ROI size on the diffusion map followed a similar course (p = 0.01). For the MRDA, such evolutions were also found, but they were not significant. There was a negative correlation between CE and the ADC (p < 0.02) and between the lesional size on T2 and ADC (p = 0.03) in the whole lesion. There Conclusion: Morphological response assessment using different CT-measurements has no prognostic value concerning overall survival in patients with synchronous colorectal liver metastases. Larger studies need to investigate possible differences in patients with k-ras mutation.

Detectability of liver metastatic breast lesions with both the C-arm CT and MRI: assessment of lesion vascularisation at the first and last TACE E.C. Mbalisike, J. Gurung, A. Azizi, P. Farshid, B. Al hazaimeh, V. Jacobi, T.J. Vogl; Frankfurt a. Main/DE (embalisike@yahoo.com) Purpose: To assess the detectability of liver metastatic breast lesions with both C-arm CT and MRI, and determining TACE effects on lesion vascularisation after treatment. Methods and Materials: Both C-arm CT and CE-MRI were done on 10 female patients. Detectability was assessed using the contrast-to-noise ratio (CNR) of the C-arm CT at the first and fourth sessions. Lesion vascularisation was assessed quantitatively using a 4-tier grading system. Signal-to-noise ratio (SNR) of CE-MRI was assessed and tested with the C-arm CT (CNR). Calculations of the C-arm CT (CNR) and CE-MRI (SNR) were done as CNR= (ROI Lesion-ROI Liver)/SDnoise, and SNR= S/N (Noise (n) = SD liver/0.66), respectively. Statistical tests used here were the Wilcoxon´s matched pair test and the Spearman-Rang correlation test. Results: Using the CE-MRI, 29 lesions were assessed. 28 of these lesions were found by the C-arm CT. One hypovascular lesion missed by the C-arm CT showed similar signal with its surrounding normal liver tissues. Detectability of 26 lesions reduced significantly by their fourth C-arm CT session, though only 12 lesions lost their vascularization. C-arm CT contrast-to-noise ratio showed a P value of 0.005, while CE-MRI signal-to-noise ratio and contrast-to-noise ratio of C-arm CT showed P value of 0.036, and a Spearman-Rang coefficient of 0.391. Conclusion: 96.6% of metastatic liver lesions were detected by the C-arm CT using the MRI as the gold standard. This lesion detectability was significantly reduced by the fourth/last TACE session, signifying loss of lesion vascularisation. Purpose: To determine the influence of lesion vascularisation assessed by both intra-arterial C-arm CT (IACCT) and contrast-enhanced MRI (CE-MRI) on lesion response to TACE for liver metastatic colorectal carcinoma (CRC). Methods and Materials: 37 patients (mean age 58.1, SD 11.7) with 120 focal metastasis were assessed. 151 TACE sessions (mean session of 4.1/patient) were done. Assessment on vascularisation was done using four grades from hypovascular, heterogeneously hypervascular, homogeneously hypervascular and mixed vascular. Responses to therapy were assessed using MRI at 3 months and 6 months. Statistical tests used in this study were Spearman-Rang's test, and Wilcoxon´s-matched-pairs-test. Results: Using both CE-MRI and intra-arterial C-arm CT, 64.7% (n = 24) were grouped as hypovascular, 27% (n = 10) heterogeneously hypervascular, 5.4% (n = 2) mixed vascular and 2.7% (n = 1) homogeneously hypervascular. C-arm CT confirmed CE-MRI findings on lesion vascularisation in 33/37 patients. Spearman-Rang´s correlation coefficient used to compare pre-treatment volumes of intraarterial C-arm CT and MRI, obtained statistical correlation of 0.736 (P = 0.0001). Lesion response was assessed using Wilcoxon-matched-pairs-tests. P value of 0.003 was obtained at 3 months, and P value of 0.002 at 6 months. Conclusion: C-arm CT was able to confirm CE-MRI findings on vascularisation of lesion. Significant responses were seen in patients with CRC after therapy with TACE. 

A D E F G H Methods and Materials: DSL-6 A/C1 pancreatic tumour fragments were orthotopically implanted in 46 male Lewis rats. Three weeks after implantation, angiogenesis inhibition was initiated by intraperitoneal administration of bevacizumab (n = 13) or suramin (n = 19). The control group (n = 13) remained untreated. In 23 rats, dynamic albumin-(Gd-DTPA)-enhanced MRI at 1.5 T was performed 24 hours after the first treatment, the remaining animals were imaged after 4 weeks of therapy. Fractional plasma volumes (fPV) and vascular permeabilities (K PS ) were calculated based on the dynamic MRI data using a pharmacokinetic model. Results: fPV at 24 hours was similar among the bevacizumab, suramin and control group (7.3 ± 5.9%, 6.1 ± 3.3% and 6.1 ± 4.3%; p = 0.86). After 4 weeks, PV in suramin-treated tumours (2.7 ± 1.0%) declined significantly compared to the 24 hours measurements (p < 0.05) and to the control tumours (6.3 ± 4.0%; < 0.05), whereas changes in bevacizumab-treated tumours were not significant (4.0 ± 1.2%, p = 0.22). K PS at 24 hours was zero in bevacizumab-treated tumours and as such differed significantly from the suramin and control group (K PS = 0 ± 0, 0.062 ± 0.066 and 0.082 ± 0.039 mL/min/100 cc; p < 0.05). After 4 weeks no significant differences in K PS were detected among the 3 groups (0.035 ± 0.059, 0.049 ± 0.026 and 0.086 ± 0.088 mL/min/100 cc; p = 0.28) nor between the two time points. Purpose: To assess the performance of a new combination for neutral oral contrast for CT abdomen and CT Enterography. We wanted to compare the performance of a combination of methyl cellulose and sorbitol in water to that of commercially available 0.1% barium suspension, VoLumen. Methods and Materials: A total of 120 patients were included in the study, randomly divided into two groups. Group 1 received 1000-1500 ml of VoLumen and group 2 was given equal amounts of combination of 0.4% methyl cellulose plus 1.5 % sorbitol solution. Contrast-enhanced CT scan was performed on 64 detector row CT. Random, blinded evaluation of the studies was done by two of the four authors. Data for luminal distension in duodenum, jejunum and ileum were recorded along with assessment of mural details. Patient acceptance was also recorded along with cost effectiveness of each type of contrast. Student t test and Chi square test were applied to compare luminal distension between groups and within each group. Results: Our new combination contrast gave better luminal distension than VoLumen (p less than 0.05) but no significant difference for mural details. Moreover, it was much better tolerated and had a very significantly low cost.

The new combination of methyl cellulose and sorbitol as neutral contrast agent provides better luminal bowel distension than VoLumen. The significantly low cost and better patient acceptance makes this our recommendation, for use as a routine in all abdominal CT scans. were also positive correlations between the ROI size and ADC (p = 0.0008) and between CE and the size on T2 (p = 0.0002). The ADC in MRDA showed some non-significant correlations with other variables.

The lesions successfully treated with RF have a clear and predictable evolution in terms of T2 size, CE and ADC. ,600] and ADC mapping) before and after change in therapy. Exams were first analysed blindly and then PET/CT images were coregistered to T1 Gd MR images for lesion detection. SUV max and ADC were measured for the six largest lesions on MRI. The relationship between SUV max and ADC was analysed using Spearman's correlation. Results: Altogether, 24 lesions (15 hepatic and 9 non-hepatic) were analysed on both modalities. Three PET/CT lesions (12.5%) were initially not considered on ADC and 4 lesions on the second PET/CT were excluded because of hepatic vascular activity spillover. SUV max decreased from 7.2±7.7 g/mL to 5.9±5.9 g/mL (P = 0.53) and ADC increased from 1.2x10 -3 mm 2 /s ± 0.4 to 1.4x10 -3 mm 2 /s ± 0.4 (P = 0.07).

There was a significant association between SUV max decrease and ADC increase (rho= -0.64, P = 0.004). Conclusion: Changes in ADC from diffusion-weighted MRI reflect response of 18 F-FDG-avid GIST to therapy. The exact diagnostic value of DWI needs to be investigated further, as well as the effect of lesion size and time under therapy before imaging. Furthermore, the proven association between SUV max and ADC may be useful for the assessment of treatment response in 18 F-FDG non-avid GIST. Patients with luminal CD undergoing MRI and ileocolonoscopy were recruited. Patients drank 1600 ml of mannitol 2.5%. Coronal DCE-MRI was performed at 3 T during intravenous contrast injection of Gadolinium (TE/TR= 1.8/2.9 ms, FA 6°, 14 slices, dynamic scan duration 6.11 min, 450 dynamics, temporal resolution 0.81 sec). ME images were calculated. A region of interest (ROI) was drawn around the most abnormal region assessed on MRI and, if not present, a normal bowel loop and ME were averaged over these ROI. C-reactive protein and Crohn's disease activity index (CDAI) were determined. Also Crohn's disease endoscopic index of severity (CDEIS) was determined which evaluates bowel involvement. CDEIS evaluation included also a partial CDEIS per segment to assess different bowel segments. Correlations were measured with Spearman's test. Results: Sixteen patients were included (mean age 39 years range 24-72). Mean time between MRI and ileocolonoscopy was 15 days (range 2-32). Median CDAI was 329 (range 112-735). Median CDEIS was 5.5 (range 0.4 -39.0). Median Creactive protein was 22 (range < 1.1-95). ME had a significant positive correlation with CDEIS per-segment (r = 0.601, p = 0.014). All other per-patient indices (CDAI, CRP and CDEIS) had no significant correlation with ME, although for CDEIS there was a trend towards a positive correlation (r = 0.487, p = 0.056). Conclusion: ME measured with DCE-MRI is an objective non-invasive parameter to assess segmental disease activity in CD patients. Transmural histopathological scoring of acute inflammation (AIS) was performed at all locations (score 0-13). MRI parameters best predicting AIS were derived using multivariate analysis. The MRI activity index was prospectively applied to 26 Crohn's patients (mean age 32, range 13-69 years, 15 males) and correlated to terminal ileal biopsy scores of acute inflammation ("eAIS" score 1-6). Receiver operator characteristic curves were calculated. 

Using CT enterography to monitor Crohn's disease activity: a follow-up study Y. Wu, N. Hao, Y. Tang, F. Miao; Shanghai/CN (wuyw0103@hotmail.com) Purpose: To evaluate CTE values in CD patient follow-up through quantitative measures of CD-associated parameters. Methods and Materials: We prospectively included 40 known CD patients from 2008-2010.Endoscopy, CTE and laboratory tests (Hs-CRP, ESR) were performed in two weeks and CDAI index, Rutgeers score was calculated by experienced gastroenterologists. When CDAI > 150, active disease was considered. Endoscopic active disease was identified when Rutgeerts Score ≥ 1. CTE parameters, including bowel wall attenuation, bowel wall thickness, comb sign, perienteric fat density and luminal stenosis were evaluated by quantitative or semiquantitative methods. CTE.Hs-CRP,ESR ,Endoscopy and CDAI index were repeated 12 months later.

Twenty-one asymptomatic volunteers (10 males, 11 females, mean age 32 ± 8.4 years) were asked to eat a light dinner on the previous evening and to fast on the day of examinations. They underwent abdominal 1.5 T MRI using axial and coronal single-shot fast-spin-echo T2-weighted sequences before and after oral administration of 10 g of lactulose diluted in 125 mL of water. MRI acquisition was repeated every 15 min up to 180 min. Before each MR acquisition, breath gas-cromatography was performed (Microlizer DP Plus, Quintron, Milwaukee, IL). Correlation between presence of fluid bolus in the caecum at MRI and the detection of a hydrogen concentration peak was estimated. Results: Median oro-cecal transit time estimated by MRI was 135 min (range 113-136 min) while it was 128 min (120-146 min) by HBT. Excellent correlation between the two techniques was found (r = 0.910). Conclusion: MRI can be effectively used in the estimation of oro-caecal transit time. If confirmed on patients affected with gut dismotility and similar conditions, MRI could play a role in the diagnosis of a number of gastrointestinal diseases.

Test-retest reliability of dynamic physiological small bowel MR assessment of peristalsis A.S. Farghal, A.P. Toms, B. Kasmai, R. Greenwood, P. Malcolm; Norwich/UK (afarghal@gmail.com)

Purpose: A novel technique for quantifying small bowel peristalsis using measurements derived from dynamic small bowel MR has been developed in our institution in which peristalsis is defined using descriptive statistics summarising the profile of mean signal amplitude per voxel. The aim of this study was to assess the testretest reliability of this technique. Methods and Materials: Dynamic coronal MR sequences covering the whole abdomen (breath hold True FISP 21 frames) were acquired from 5 volunteers who fasted for six hours and drunk 1 L of water. MR data were analysed with an in-house software package, which calculated the mean absolute change in signal per voxel for each dynamic sequence. The total voxel activity (TVA) per slice was then calculated as a sum of all mean absolute changes in voxel signal. All studies were repeated with the same parameters and analysis with at least a two-week gap between studies. Descriptive statistics were performed on each study. Reliability was measured using a Pearson Product-Moment Correlation Coefficient comparing each slice in the two studies as well as total abdominal voxel activity.

The correlation coefficient demonstrated "almost perfect agreement" between first and second studies. The slice by slice correlation was r = 0.849 (95% CI 0.765 to 0.904), the total abdominal correlation was r = 0.9898 (95% CI 0.8483 to 0.9994).

The test-retest reliability of the technique described for quantifying small bowel peristalsis with dynamic MR is extremely good suggesting that this may be a robust technique for serial studies.

Global assessment of small bowel activity using dynamic MR with automated computerised method: protocol optimisation A.S. Farghal, A.P. Toms, B. Kasmai, A. Kerner-Bignell, P. Malcolm; Norwich/UK (afarghal@gmail.com) Purpose: The aim of this study was to optimise the MR protocol for evaluating global small bowel peristalsis in vivo using a simple automated method from dynamic MR. The minimal acquisition time, number of slices and study repetitions required to produce reproducible data were assessed. Methods and Materials: Twenty-five dynamic MR studies (True FISP, a single breath hold, ECG triggering) were obtained from three fasting healthy volunteers. The abdomen was covered by 10 dynamic coronal acquisitions. Parametric maps were automatically generated using in-house built software representing the mean absolute signal change in each pixel. Thresholds selected from histograms were used to decrease background noise. An index of peristalsis was generated by dividing the sum of the mean pixel values of the resulting parametric images divided by the number of pixels. ANOVA tests were performed to determine the shortest acquisition time, minimum number of slices and minimum number of study repetition required for a reproducible index of peristalsis. Results: There was no statistical significance between the 30, 25, 20 and 15 frames acquisition protocols. For acquisitions of 15 frames or less, there was significant variability in the data. Using half or less of the dataset by selecting alternate coronal slices results in significant differences in variance. There was no significant difference between a single acquisition and the combined data from up to five acquisitions. Conclusion: For optimal physiological small bowel dynamic MR studies, a single acquisition, with dynamic MR consisting of a minimum of 15 frames, covering the whole abdomen is satisfactory.

B C A D E F G H

Contrast-enhanced ultrasound in intestinal GvHD: microbubble penetration as a new diagnostic feature A.G. Schreyer, C. Friedrich, M. Jung; Regensburg/ DE (andreas.schreyer@klinik.uni-r.de) Purpose: Graft-versus-host-disease (GVHD) is an important complication in patients after allogeneic stem cell transplantation. Early detection is of paramount interest, because early therapy can improve the outcome. We evaluated contrast-enhanced ultrasound (CEUS) in patients with GVHD to assess typical imaging features. Methods and Materials: CEUS was performed in 9 patients with histologically proven GVHD. As a control 4 healthy volunteers and 6 patients with Crohn's disease (CD) were examined. We employed a high-resolution multi-frequency transducer (6-9 MHz) with contrast harmonic imaging and low mechanical index technique to preserve microbubble integrity. After the injection of 2.4 ml SonoVue® intravenously data were acquired and stored digitally. ROIs (region of interest) were manually placed over the surrounding mesenteric fat, bowel wall and bowel lumen. Movement compensation of the ROIs during the scan was performed. Time intensity curves were calculated based on the ultrasound signal and the maximum signal increase of each compartment was calculated.

Results: Patients with CD as well as those suffering from GVHD showed a significant contrast uptake of the bowel wall. Only patients with GVHD showed a transmural penetration of microbubbles into the bowel lumen (average maximum signal increase -GVHD: 7.88 dB; CD: 1.13 dB, healthy volunteer: 0.28 dB).

We assume that the damaged gut mucosal barrier in GVHD enables the microbubbles to penetrate through the bowel wall into the bowel lumen. We consider the penetration of microbubbles into the bowel lumen as a novel diagnostic feature for GVHD. Purpose: The vascular measurements for abdominal aortic stent graph endoprosthesis are often replicated by different health professionals groups. This study aimed to calculate the interobserver variability between three professional groups -4 radiographers (RD); 4 radiologists (RG); and 4 vascular surgeons (VS) -using angio-CT technique, according the pre-EVAR (endocvascular aneurism repair) measurement protocol. Methods and Materials: 480 observations from patients with abdominal aortic aneurism were analysed by 12 observers from 5 abdominal CT studies. Eight specific measures were applied (6 diameters and 2 distances) following the same pre-EVAR protocol. For statistical analysis Bland-Altman and analysis of variance test were used by Excel 2007 (Microsoft ® ) and analyze-it ® applications to evaluate the mean difference between observers. Results: The mean difference between observers was 1.55 ± 3.39 mm and 3.38 ± 5.39 mm for vessel diameters and distances, respectively. Pairs of observations differences showed that 75.1% were ≤2 mm and 4.2% ≥5 mm. Among professional groups was found a variation between RG and VS with -1.15 mm (95% CI: -1.597 mm, -0.702 mm) and between RD and CV with -2.7 mm (95% CI: -3.848 mm, -1.557 mm). There were no significant differences (p > 0.05) in results obtained between professional groups. The mean difference for the aneurysm diameter location was a 1.7 mm, less than the gold-standard considered (2.04 mm).

There was agreement between the three different professional groups.

The difference values we found could be considered as acceptable. The pre-EVAR measurements may be performed by radiographers and the replication is not justified unless when applied the multi-observational method for reliability purpose.

Results: Rutgeerts score and CDAI were correlated with CTE findings at baseline (P < 0.05). Follow-up results showed 18 (45%) patients get progression with Rutgeerts score, whereas 6 (15%) patients were found regression and 16 (40%) patients were stable. Similarly, using CDAI index progression was found in 14 (35%) patients whereas regression in 8 (20%) patients and 18 (45%) patients were found to be stable. Variations between two times Rutgeerts score was shown positively correlated with the variations of bowel wall attenuation, bowel wall thickness, comb sign, perienteric fat density and luminal stenosis.(Spearmen index 0.362-0.614, P < 0.05).CDAI variations were detected correlated with variations of bowel wall attenuation, bowel wall thickness, comb sign and perienteric fat density. (Spearmen Index 0.334-0.624, P < 0.05).Comb sign highest correlated with either CDAI (Spearmen Index = 0.624, P < 0.01) or Rutgeers score (Spearmen Index = 0.614,P = 1.7×10 -5 ). Conclusion: CTE is a valuable tool in Crohn's disease patients follow-up.Comb sign did best in monitoring disease activity. 

The results showed statistically significant improvements in the area under the curve values (AUC) of the juniour doctors when working with the RTs for both sets of images (wrist and CT) treated as random readers and cases (p < 0.008 and p < 0.0026, respectively). While the RTs' results saw no significant changes, their mean Az values did show an increasing trend when working in collaboration. Conclusion: Improvement in performance of juniour doctors following collaboration strongly suggests changes in the potential to improve accuracy of patient diagnosis and therefore patient care. Further training for juniour doctors in the interpretation of diagnostic images should also be considered. Decision making of juniour doctors was positively impacted on after introducing the opinion of an RT. Collaboration exceeds the sum of the parts; the two professions are better together. Purpose: To compare workflow efficiency between a conventional computed radiography (CR) system and a novel, portable, cassette-sized, and wireless flat-panel digital radiography (DR) system. Methods and Materials: Observational time-motion analyses were performed at one site at which CR, and the new portable DR system are used concurrently. The workflow steps of both systems were identified and categorised to facilitate comparison. The time required for examination preparation, patient positioning, exposure, post-acquisition processing, and for the examination as a whole, was recorded by a neutral observer. Timing differences between the CR and portable DR systems were compared, and all data were analysed using commercially available statistical software. Nine general radiographic exam types were selected with approximately 50 patients per exam type. Results: A total of 941 examinations (CR, n = 474; portable DR, n = 467) were timed in this study. Total examination time differences between CR and portable DR system (mean 26.44s; median 26.99s) were found to be statistically significant (P <.001), with DR proving faster than CR. The single largest contributor to the time difference between CR and portable DR was post-acquisition processing (mean 26.58s, median 25.91s), a composite of multiple individual steps, including cassette transport (CR only, mean 13.22s, median 12.74s), cassette readout (mean 10.15s, median 10.4s) and post-processing (mean 3.21s, median 3.11s). Conclusion: Overall radiographer time was significantly shorter when performing exam-related tasks with the novel, portable DR system than when performing comparable tasks with CR, a difference that appears to result largely from technology configuration.

Interobserver agreement among PGMI radiographer in quality classification of mammograms: a pilot study R. Gullien, J. Andersen, A. Haakull, E. Rostad, R. Hammond; Oslo/NO (raul@uus.no) Purpose: To evaluate the inter-observer agreement between internal and external radiographers with and without experience in PGMI classification of mammograms using the classifications recommended in the quality assurance manual for the Norwegian Breast Cancer Screening Program (NBCSP). Methods and Materials: Three PGMI radiographers, A = internal experienced radiographer, B = external experienced radiographer, C = internal inexperienced radiographer, reviewed and classified 480 screening mammograms from 120 women. Using the PGMI classifications were P = perfect, G = good, M = moderate and I = inadequate mammograms, each radiographer classified the mammograms independently, by scoring them as: P = 1, G = 2, M = 3, and I = 4. Inter-observer agreement was assessed (Kappa analysis). Results: The agreement between the PGMI radiographers was: κ = 0.40, 0.40, and 0.44, respectively, for A + B, B + C, A + C. This level of agreement is classified as fair to moderate according to Kappa classification. The exact agreement between A + B, B + C, A + C was, respectively, 57%, 56% and 61%. The best agreement appeared between the experienced and inexperienced internal radiographer (A + C), while there is less agreement between the two experienced (A + B). The lowest agreement was between the external experienced and the internal inexperienced radiographer (B + C). This indicate that there are a different competency levels between the three radiographers. Conclusion: Interobserver agreement with the PGMI quality classification in our pilot study was fair to moderate. This indicates that PGMI classification is a challenge in mammography screening, since there is room for individual assessments -despite guidelines.

Implementing evidence-based protocols for i.v. contrast media; the process, outcome and follow-up B. Vee, A. Taule, N. Kleven-Madsen; Bergen/NO (bente.vee@helse-bergen.no) Purpose: The department of radiology at HUH aims to provide patients with services based on best available evidence in all aspects of imaging. Based on risk assessment and ESUR guidelines for administering i.v. contrast media (IVCM), a need was recognised for systematic review of current practice. The objective was to 1) survey and 2) improve all processes involving the use of IVCM. The overall aim is to utilise and increase the competence of all professions involved in the various steps involved to provide the best achievable quality in the use of IVCM; evidence-based practice. Methods and Materials: All diagnostic use of IVCM; referral, radiology report, follow-up of sideeffects was reviewed. Main focuses were identifying precise, manageable measurement of kidney function (eGFR), patients at risk for CIN, NSF, thyroid patients, use of, i.e. metformin, classification and handling of sideeffects, responsibility (clinician, radiologist and radiographer) and patient information. Results: By reviewing routines and implementing new procedures, the process has created awareness on the use of IVCM so that patient care and safety are met. It has been challenging to clarify the responsibility for each professional participant. Continued staff information and education throughout the process was important. Conclusion: Evidence-based routines and procedures have been established, and are clearly communicated to all involved participants to achieve and maintain good practice. A successful implementation requires continuous focus (dynamic process) from dedicated staff to follow-up; implement new knowledge and adjust the daily routines. Quality indicators for evaluating the effect of implementation must be developed.

Effect of informed consent for administration of intravenous contrast material in the anxiety level of patients submitted to computed tomography exams K.B. Azevedo, A.F.C.L. Abrantes, L.P.V. Ribeiro, R.P.P. Almeida, C.A. Silva; Faro/PT (kbazevedo@ualg.pt) Purpose: Evaluation and relation of the state and trait anxiety level in patients before the delivery of informed consent with patients after that, and the evaluation of eventual relations between psychometric tests and sociodemographic characteristics.

The present study is quasi-experimental. The total sample of 161 patients was divided in a control group and an experimental group, and then state-trait anxiety inventory was applied to both, after the sociodemographic interview and always after the consent of participation of the patient. Results: At a 95 % confidence interval, the intergroup difference between means of state anxiety ranged from -0.929 and -0.810. The value of t student test was -28.701 (df = 159 and 2-tail sig = 0.000). On the other hand, the intergroup difference between means of trait anxiety ranged from -0.141 and -0.022. The value of t student test was -2.722 (df = 159 e 2-tail sig = 0.007). Application of Spearman's Ro revealed the correlation between values of state and trait anxiety based in a non-normal approximation: for control group 0.409 (sig = 0.000) and for experimental group 0.321 (sig = 0.004), whereas it did not reveal any statistically significant correlation between sociodemographic characteristics and psychometric tests. Conclusion: There exists a significant intergroup difference between the anxiety state and the anxiety trait. A significant correlation is verified between state and trait anxiety within each group, what is not verified between psychometric tests and sociodemographic characteristics. Purpose: In emergency departments (ED) junior doctors regularly make diagnostic decisions based on radiographic images [1]. This study investigates whether collaboration between juniour doctors and RTs impacts on diagnostic accuracy. Methods and Materials: Research was carried out in the ED of a university teaching hospital and included 10 pairs of participants. RTs and juniour doctors were shown 42 wrist radiographs and 40 CT brains and were asked for their level of confidence of the presence or absence of distal radius fractures or fresh intracranial bleeds, respectively, using ViewDEX software, first working alone and then in pairs. Receiver operating characteristic was used to analyse performance. Results were compared using one-way analysis of variance. Purpose: Follow-up imaging after stent-assisted coiling (SAC) of intracranial aneurysms is limited by signal loss in the stented vessel segment in MRI and by streak artefacts caused by aneurysm coils in MDCT. Searching for a non-invasive surveillance in this condition we propose a technique to minimise streak artefacts in MDCT by gated data acquisition and shifting the reconstruction window (gCTA).

B C A D E F G H

The effect of gCTA on artefact reduction was evaluated in a preliminary phantom study and compared with non-gated-CTA, TOF-MRA and DSA. To show translational feasibility gCTA and TOF-MRA were also performed on five patients treated with SAC as part of a follow-up protocol. The length of impaired vessel segments (LIVS) were compared and correlated with the stent length, the number of coils and the packing density. The assessment of treatment outcome (ATO) was compared with DSA as the standard of reference.

The phantom study revealed a distinct reduction of streak artefacts using gCTA. In MRI all stented vessel segments showed signal loss in both, in the phantom and in patients. The LIVS was 78% shorter in gCTA (4.86 ± 6.93 mm) as compared to TOF-MRA (21.82 ± 7.47 mm, p <.01). LIVS correlated in TOF-MRA the with the stent length, in gCTA with the number of coils. With regard to ATO gCTA and TOF-MRA correlated with DSA in 3/5 and 0/5 patients, respectively. Conclusion: Gated CTA is a promising technique to reduce the amount of artefacts induced by stent-assisted intracranial coils. Image quality and assessment of treatment outcome in patients with stent-assisted coiling is superior compared to TOF-MRA.

Radiographer led real time review of CT colonography: initial experiences R.J. Moore, D.P.G. Mullan, P. Healey, M. Hughes; Liverpool/UK (rachelmoore01@btinternet.com)

Purpose: To evaluate and report our experiences of role extension for radiographic advanced practitioners (GI) in CT colonography (CTC) techniques. Methods and Materials: Barium oenemas have been superceded by CTC in our institution. This reduced fluroscopic workload could lead to 'de-skilling' of GI radiographic advanced practitioners who are treained to masters level, with several years of specific experience. Dedicated CTC lists are performed with a seperate CT and GI radiographer in attendance. The GI radiographer assesses the adequacy of bowel prep, performs PR examinations, and continually analyses the quality of pneumocolon. There is focused real-time review of the first aquisition of images whilst the patient is prepared for the subsequent aquisition. If a tumour is discovered on real-time review, the GI radiographer will justify and perform a concurrent CT thorax for completion imaging.

Results: 401 studies were performed with real-time review between 1/6/10 and 1/9/10. This lead to same day CT thorax studies in 23 cases of previously undiagnosed colorectal tumours. Diagnostic quality pneumocolon was achieved in 97% of cases. Conclusion: Real-time review of CTC by a GI advanced practitioner leads to concurrent performing of CT thorax in appropriate cases. This leads to timely full body staging and reduces delays in appropriate cases. Understanding of colon anatomy and pathology allows the GI radiographer to safely achieve diagnostic quality pneumocolon in the majority of cases.

The effect of utilising SMS technology on patient care in radiology N.M. Mishah; Jeddah/SA (nabeel747@hotmail.com)

Purpose: To present the effect of utilising SMS (short message service) on patient care in radiology department since SMS technology was recently utilised to remind and/or confirm patient's appointment and to send patient preparations. The efficiency and effectiveness of this tool will be compared to other conventional telecommunication methods. Methods and Materials: Statistical data for the first half of year 2010 (6 months period) were retrieved from the system then, the electronic data were analysed in order to measure the percentage of SMS message delivery rate and its relation to better patient care and the list of no show patients (or) badly prepared patients for rebooking. Results: A tremendous improvement in patient care, attandence and communication if compared to the situation before implementing SMS. All the patients received the SMS (100% delivery rate) and no show list was decreased by 70% and bad patient preparation was decreased by 35%. Conclusion: SMS technology is available, easy to use, cost-effective, and efficient way to communicate with patients for better patient care and better resource utilisation.

Radiologists' responses to inadequate referrals K. Results: Ninety-five percent (344/361) reported daily or weekly actions related to inadequate referrals. Actions differed among subspecialties. The most frequent were contacting the referrer to clarify the clinical problem and checking test results/ information in the medical records. Both actions were more frequent among registrars than specialists and among hospital radiologists than institute radiologists. Institute radiologists were more likely to ask the patient for additional information and to examine the patient clinically. Factors rated as contributing most to prevent doubtful examinations were high risk of serious complications/side effects, high radiation dose and low patient age. Factors facilitating doubtful examinations included respect for the referrer's judgment, patient/next-of-kin wants the examination, patient has arrived, unreachable referrer, and time pressure. Conclusion: Radiologists frequently take action in response to inadequate referrals, mainly by searching for complementary information. Their actions seem to be mainly motivated by patient safety considerations. Vetting of referrals at arrival, easier access to referring clinicians, and sufficient time for radiologists to handle inadequate referrals, may contribute to improved use of imaging.

A S235 C D E F G H DS-CTA was analysed on a per-patient and a per-aneurysm basis with 3D-DSA as reference standard. Results: Of 506 patients, 107 patients (21.1%) had no aneurysms, while 399 patients (78.9%) had 452 aneurysms on 3D-DSA.DS-CTA correctly detected 450 aneurysms in 402 patients. Using 3D-DSA as the standard of reference, sensitivity and specificity of DS-CTA to detect intracranial aneurysms were 97.7% and 88.8% on a per-patient basis, 96.5% and 87.2% on a per-aneurysm basis. DS-CTA had sensitivities of 90.9%, 93.8%, 98.3%, 100% for the detection of aneurysms with the diameter of less than 3 mm, larger than 3 mm but less than 5 mm, larger than 5 mm but less than 10 mm, and larger than 10 mm. Excellent correlation was found between 3D-DSA and DS-CTA in the diagnosis of intracranial aneurysms (Kappa = 0.830, P < 0.001). Conclusion: On the basis of this large cohort, DS-CTA has a high sensitivity and specificity for the detection of intracranial aneurysms compared with 3D-DSA, even for the aneurysms with the diameter less than 3 mm.

Intravenous flat detector CT to visualise patency of intracranial stents: first results with an optimised protocol T. Struffert 1 , S. Kloska 1 , T. Engelhorn 1 , Y. Deuerling-Zheng 2 , S. Ott 1 , M. Doelken 1 , M. Saake 1 , A. Doerfler 1 ; 1 Erlangen/DE, 2 Forchheim/DE (tobias.struffert@uk-erlangen.de) Purpose: As stents for treating intracranial atherosclerotic stenosis may develop in-stent re-stenosis (ISR) in up to 30%, follow-up imaging is mandatory. Residual stenosis (RS) is not rare. We evaluated an minimal invasive optimised flat detector CT protocol with intravenous contrast application (i.v. FD-CTA) to conventional angiography to evaluate the stent lumen. Methods and Materials: In 12 patients treated with a self expandable stent follow-up imaging was performed using i.v. FD-CTA and intra-arterial angiography (DSA). Location of intracranial stenosis was M1 in 6, V4/BA in 5 and C5 segment in 1 patient. MPR, subtracted MIP and VRT reconstructions were used to correlate to DSA. Two neuroradiologists evaluated the images in anonymous consensus reading and calculated ISR or RS, respectively. Correlation coefficients and a Wilcoxon test were used for statistical analysis. Results: In 4 patients, no stenosis was detected. In 6 patients RS and in two cases ISR perfectly visible on MPR reconstructions of i.v. FD-CTA were detected. Wilcoxon's test showed no significant differences between the methods (p > 0.05). Purpose: Uncertainties about the frequency and the associated bleeding risk of recent silent ischaemia (RSI), incidentally found on pretreatment MRI, in candidates for thrombolysis require clarification because exclusion from therapy is a serious consequence for patients with such MRI findings.

We retrospectively analysed the fluid-attenuated inversion recovery (FLAIR)/diffusion-weighted imaging (DWI) before intravenous (IV) thrombolysis in 115 patients to search for MRI-defined RSI; these corresponded to well-developed FLAIR/DWI brain hyperintensities (RSI+), as distinct from the acute index ischaemia, which typically lacked FLAIR changes. Patients without such findings were assigned to the RSI-group. Groups were compared for baseline characteristics and for rates of symptomatic and asymptomatic haemorrhagic transformation (HT) using odds ratio (OR) and their 95% confidence interval (CI). Results: We observed RSI in 21 patients (18.3%). The mean (standard deviation) volume of RSI was 6.5 (12)mL (interquartile range 0.6-9). None of the baseline parameters differed between groups. There was no significant difference in rates of any type of HT between groups. Parenchymal haemorrhage type 1 (PH-1) or type 2 (PH-2) according to ECASS criteria occurred in 10 (11%) RSI-patients and in two (10%) RSI+ patients (OR=0.88; 95%CI=0.18-4.37). Symptomatic HT, defined according to NINDS criteria, occurred in 10 (11%) RSI-patients and in one (5%) RSI+ patient (OR=0.42; 95%CI=0.05-3.47). Conclusion: We found that 18.3% of acute stroke patients treated by IV thrombolysis in a stroke unit had RSI on pretreatment MRI. However, the presence of RSI was not associated with an increased risk of asymptomatic or symptomatic HT.

Circle of Willis single-source dual energy CTA: "metal artefact reduction software" in evaluation of vasospasm and residual aneurysm after endovascular approach for intracranial aneurysms S. Comelli, F. Fusaro, D. Savio, G. Vaudano; Turin/IT (fedefusaro@libero.it) Purpose: Single-source dual energy CT (Discovery 750HD, GE Med. Systems, WI) with a particular post-processing software allows to significantly decrease, even though not completely avoid, metal artefacts. We experienced it in patients previously treated for intracranial aneurysm with endovascular approach, either after coiling or flow-diverting device. Aim of the study is to evaluate vasospasm during ten days following treatment, and post-operative recurrent or residual aneurysms at mid-term (3 to 6 months) follow-up controls. Methods and Materials: From January to July 2010, 25 post-operative aneurysms (6 up to 28 mm in maximum diameter, mean 9.2 mm) in 24 patients (age 42-79, mean 58.3) have been studied with dual energy CT angiography (DECTA). 23 aneurysms were treated with detachable coils, 2 with pipeline flow-diverting stent (ev3, Irvine, CA). DECTA raw-data were post-processed with metal artefacts reduction software (MARS), using levels 64 KeV up to 72 KeV. Results: DECTA with MARS has proven to be diagnostic in vasospasm grading evaluation for 25/25 patients (100%). Compared with digital subtraction angiography (DSA, GE Innova 4100) performed in all patients after CT, MARS demonstrates 100% sensitivity and specificity in vasospasm evaluation. On mid-term follow-up controls, MARS significantly reduced metal artefacts, but does not seem to be enough accurate to detect 2 mm or smaller aneurysm residual (sensitivity 54%, specificity 68%). Purpose: To compare the diagnostic accuracy and reliability of TOF-MRA and CE-MRA at 3 T for detection and characterisation of untreated intracranial aneurysms. Methods and Materials: 3D-TOF-MRA and 3D-CE-MRA at 3 T were performed in 52 patients with suspected or known aneurysms sequently, prior to rotational subtraction angiography (RDSA). Accuracy, sensitivities, specificities, positive predictive value (PPV) and negative predictive value (NPV) on aneurysm-, vessel-, and size-based evaluation were calculated to detect or rule out intracranial aneurysms. The diagnostic relativity of VR 3D-TOF-MRA at 3 T and VR-DSA, 3D-CE-MRA at 3 T and VR-DSA was conducted, respectively, to compare the relativity of both two modalities to assess the intracranial aneurysms. Results: A total of 58 intracranial aneurysms were visualised on VR 3D-TOF-MRA. For aneurysm < 3 mm, one aneurysm was missed on VR 3D-TOF-MRA, while on VR 3D-CE-MRA, there were eight false-negative cases. Accuracy, sensitivity, specificity, PPV and NPV of VR 3D-TOF-MRA at 3 T for the detection of intracranial aneurysm < 3 mm were 97.7%, 94.7%, 100%, 100%, 95.7%, respectively, while on VR 3D-CE-MRA, 80.5%, 57.9%, 100%, 100%, 73.3%, respectively. For the detection of intracranial aneurysm ≥3 mm, there was no statistical difference between two modalities. Conclusion: 3D-TOF-MRA at 3 T is revealed to have overwhelming advantages over 3D-CE-MRA for the detection and characterisation of untreated intracranial aneurysms. 3D-CE-MRA is only superior to 3D-TOF-MRA for depicting the morphology of giant aneurysms with partial thrombosis or turbulent flow and in uncooperative patients with SAH. Thus, 3D-TOF-MRA at 3 T is recommended for detection and follow-up of untreated intracranial aneurysms. siphon was not achieved. In one case (3.03%), the patient did not cooperate nor was sedated enough for the procedure. There were two severe procedure-related complications (6.06%). In one case (3.03%), MCA rupture occurred after the ballooning of the stent. Immediate coiling was done, but the patient became comatose. In one case, ICH developed 2 weeks after the procedure. Mild restenoses occured in 2 patients without significant symptom changes during the follow-up period. Purpose: Acute stroke due to internal carotid artery occlusion has a poor prognosis. Clinical outcome is related to efficacy and timing of vessel recanalisation. We report our single center experience focusing on the adequate pre-procedural and post-procedural management.

We retrospectively reviewed 15 consecutive endovascular cases of acute internal carotid artery occlusions treated at our center between July 2009 and November 2009. Preprocedural symptoms were evaluated through NIHSS. Mechanical thrombolysis was perfomed through Penumbra system. Adjunctive intra-arterial pharmacological therapy (urokinase) was performed only if time to stroke was inferior to 6 hours. If thrombosis was associated with a significant vessel stenosis, a stent was deployed. All patients underwent post-procedural neurologic examinations and angio-MR performed at 3 and 6 months after the procedure. Results: Penumbra mechanical thrombolysis was performed in all cases. 86.6% of the treated vessels were revascularised to TIMI 2 of 3. Intra-arterial thrombolysis was associated with mechanical approach in 13 patients (86.6%). One patient died after the procedure (7%). In 5 patients (33.3%) a carotid stenting procedure was necessary. In 4 patients (26.6%) stenting of the middle cerebral artery was perfomed.

In the other patients we observed a significant improvement of symptoms (mean baseline NIHSS 19.5 ± 0.3; mean postprocedural NIHSS 3.5 ± 0.2, p < 0.0001). Conclusion: Mechanical thrombolysis associated with intra-arterial pharmacological therapy and stenting is an efficient treatment for acute stroke. 

The penumbra system -a mechanical device for the treatment of acute ischaemic stroke: a single center experience T. Kim, S. Kim, E. Kim, J. Lee, Y. Lim, M. Han; Suwon/KR (piglett82@hanmail.net)

Purpose: To analyse the feasibility, the safety and performance of the Penumbra System. Methods and Materials: A prospective, single arm, independently monitored and core laboratory adjudicated trial enrolled subjects with acute stroke, within 8 hrs of symptom onset and an angiographically verified occlusion (TIMI grade 0 or 1). The primary end point was revascularisation of the target vessel to TIMI grade 2 or 3. Secondary end points were achievement of mRs core ≤2, or ≥4-point NIHSS improvement score at 30-day follow-up, as well as mortality rate. Results: 20 target arteries of 18 patients were treated by the penumbra system. 14 cases of anterior circulation and 4 basilar artery occlusions. At baseline, mean mRS score was 4.5, and mean NIHSS score was 20. After procedure, 17 arteries were successfully revascularised (81%) to TIMI 2 or 3. At 30 day follow-up, 9 subjects (45%) had mRs core ≤2 or ≥4-point NIHSS. None of the patients developed sICH. The all-cause mortality rate was 22% (4 of 18), which is lower than expected in this severe stroke cohort, where 65% of the subjects at baseline had either an NIHSS score of more than 20 or a basilar occlusion. Conclusion: Penumbra system showed a high potential for recanalisation of acute thromboembolic occlusions of the large cerebral arteries. Clinical improvement can be observed in patients with ischaemic stroke using low-dose thrombolytic agents with adjunctive mechanical disruption of clot. Moreover, this strategy may reduce the risk of intracerebral haemorrhage observed with thrombolytics. . Six aneurysms were additional to ruptured aneurysms and four aneurysms were incidentally detected. Patients underwent endovascular treatment with coils (n = 5), balloon-assisted coiling (n = 10), stent-assisted coiling (n = 2) and parent artery occlusion (n = 1). Clinical and angiographic follow-up was done at regular intervals. Results: All but one aneurysm were technically successfully occluded. Two aneurysms had neck remnants that disappeared upon follow-up. Five procedural complications happened (28%). In two patients with SAH coil perforation occurred; one patient recovered and the other died. One patient died of SAH from another untreated aneurysm. Two patients with unruptured aneurysms had local thrombus formation that resolved without sequel. Upon follow-up 17 of 18 aneurysms (94%) remained occluded. One aneurysm with gradual re-opening was surgically clipped. Mean follow-up 11 months (range 0-24 months). Two patients (11%) had permanent neurological deficit. Two patients died of SAH complications. Conclusion: Endovascular occlusion of very small intracranial aneurysms poses a considerable treatment challenge with a significant morbidity/mortality rate, but can be performed in selected cases. 

The stent implantation was successful in 26 lesions (78.8%). We had 7 cases of failure (n = 7, 21.2%). We had two cases (6.06%) of microselection failure due to the tortuosity of the ICA. There was one case (3.03%) of selection failure due to tortuousity of aortic arch. In one case (3.03%), advance at the cavnernous

The effect of cervical spondylosis on vertebral arterial flow and its association with hearing impairment O. Kiritsi 1 , G. MIkroulis 2 , G. Noussios 3 , K. Tsitas 4 ; 1 Kozani/GR, 2 Thessaloniki/GR, 3 Serres/GR, 4 Katerini/ GR (kirolga@hotmail.com) Purpose: The aim of the present study was to evaluate the effect of cervical spondylosis on vertebral arterial flow and its association with hearing impairment. Methods and Materials: From April 2004 till February 2006 a prospective horizontal trial was performed. 56 individuals [mean age 53 ± 14 years] with a clinical diagnosis of cervical spondylosis were enrolled in the trial. A clinical investigation, including history, physical examination, x-rays and colour Doppler ultrasound, was carried out initially. 56 healthy individuals [mean age 29.34 ± 6.98 years] were used as control. Results: Sixteen controls and 3 group B participants dropped out before the radiological investigation was completed. Of the 53 participants 28 [52%] had hearing impairment and consisted group A. The rest 25/53 [48%] subjects scored 0 to the aforementioned test and formed group B. Group A demonstrated a total degeneration score of 227, while group B and controls showed a total score of 128 and 3, respectively. In group A, mean peak systolic velocity, mean end diastolic velocity and mean resistance index on the right was significantly lower compared to mean values in subjects without hearing disorders and healthy individuals. Pearson product-moment correlation demonstrated strong correlation between cervical spine spondylosis and vertebral artery flow parameters. Both the Mann-Whitney U test and the Pearson product-moment correlation coefficient depicted strong correlation between total degeneration score and hearing impairment in group A. Conclusion: In conclusion, a strong relationship was depicted between cervival spine spondylosis and vertebral artery flow parameters or hearing impairment.

Evaluation and comparison of vascular 18 F-sodium fluoride uptake and arterial wall calcifications using PET/CT P. Bannas, T. Derlin, S. Klutmann, J. Mester, G. Adam, M. Clausen, C. Weber;

Purpose: Calcification is a prominent feature of atherosclerotic plaques. Currently, F-18-sodium fluoride is extensively used for imaging of bone remodelling. There are no data concerning its potential for characterisation of atherosclerotic plaques. In this study the prevalence and distribution of vascular F-18-sodium fluoride uptake and arterial calcification in major arteries were examined with combined PET/CT. In addition, the topographic relationship between increased radiotracer accumulation and calcification was correlated. Methods and Materials: Image data of 75 patients (48 women, 27 men, age 65.2 ± 12.4 y) undergoing whole-body F-18-sodium fluoride PET/CT for assessment of bone metastases were evaluated retrospectively. F-18-sodium fluoride uptake and arterial wall calcifications in major arteries (left/right common carotid arteries, thoracic aorta, abdominal aorta, left/right common iliac arteries, and left/right femoral arteries) were analysed qualitatively and semiquantitatively, and compared. Results: F-18-Sodium fluoride uptake was observed at 254 sites in 57 (76%) of the 75 study patients, calcification at 1930 sites in 63 (84%) patients throughout all major arteries. The localisation was consistent with established atherosclerotic topography, e.g. increased uptake predominantly in the thoracic and abdominal aorta, and in the femoral arteries. The mean maximum standardised uptake value (SUVmax) of lesions showing regional uptake was 1.9 (range 0.8-3.5). Colocalisation of radiotracer accumulation and calcification could be observed in 223 (88%) uptake areas. However, only these 223 (12%) arterial calcification sites showed increased F-18-sodium fluoride uptake. Conclusion: Our data indicate for the first time the feasibility of F-18-sodium fluoride for imaging of calcified atherosclerotic lesions in humans.

Relationship between progression of arterial stiffness and atherosclerotic burden over time S. Oberoi, J.W. Nance, M. Meyer, J.M. Barraza, U.J. Schoepf;

Purpose: Both arterial stiffness and atherosclerosis are independent predictors of adverse cardiovascular events. They have also been shown to be independently associated with each other in cross-sectional studies, but the nature of the relationship is vague. The purpose of this study was to examine changes in arterial stiffness and atherosclerosis over time to determine if there is parallel progression of disease. Methods and Materials: 164 patients with suspected coronary artery disease underwent 2 coronary computed tomography (cCTA) scans an average of 12 ± 10 months apart. Arterial stiffness was represented by the aortic distensibility index Conclusion: Coronary DSCTA detected a high plaque burden in patients with type 2 DM. Plaques were mainly detected in the proximal segments of LAD. More calcified and less non-calcified plaques were observed. More attention should be paid to the significant prevalence of obstructive stenosis. Purpose: A reversal of endothelial dysfunction in T1D was previously described after pancreas transplant. CACs are vascular-committed bone-marrow-derived-cells expressing endothelial-and stem-cell markers, dysfunctional in T1D. We studied if restoration of endogenous-beta-cell-function with transplantation is associated with better CACs-function.

We studied with US the right brachial arteries of 14 insulin-independent-ITA, 18 T1D untransplanted patients and 14 controls. Endothelial-dependent-dilation (EED) was ultrasonically assessed by antero-posterior diameter increase of the right brachial artery during a post ischaemic hyperaemia, while endothelial-independent-dilation was studied by nitrate-induced vasodilation (NDD) as control. An occluding forearm cuff placed 5 cm below the antecubital fossa was inflated to 50 mmHg above systolic pressure for 5 minutes, then released to induce reactive local hyperaemia. Recordings were made 60 seconds after onset of hyperaemia and 3 minutes after sublingual administration of nitro-glycerine spray. The parameters of endothelial-function were correlated with Hba1c. Results: T1D showed a lower EED compared with healthy controls (T1D = 4.1±1.4% vs. C = 8.6±0.6%, p = 0.0001). Interestingly, ITA showed a normal EED being statistically higher compared with T1D group (ITA = 16.4±2.3%, p = 0.0007 vs. T1D). EED negatively correlated with HbA1c (R = -0.40, p = 0.01). No differences in NDD were found in the three groups studied (T1D = 17.3±3.0%, C = 23.6±4.5% and ITA = 19.2±3.5%; ns). Conclusion: Diabetic patients showed a lower percentage of EDD compared with healty controls, while after islet-transplantation a fully recovery of endothelial function is evident. Instead, no differences in NDD were found in three groups studied, confirming the hypotesis that the defect of CACs may reflect into a macroscopic alteration of endothelial function in patients.

Comparison of symptomatic and asymptomatic atherosclerotic carotid plaques using parallel imaging and 3 T high-resolution in vivo MR imaging J.M. Grimm 1 , A. Schindler 1 , C. Schmidt 1 , M. Dichgans 1 , C. Yuan 2 , C.C. Cyran 1 , M.F. Reiser 1 , K. Nikolaou 1 , T. Saam 1 ; 1 Munich/DE, 2 Seattle, WA/US (jochen.grimm@med.uni-muenchen.de) Purpose: To determine if high-resolution 3 T magnetic resonance imaging (hr-MRI) can depict differences between symptomatic and asymptomatic atherosclerotic carotid plaques. Methods and Materials: In this prospective monocentric observational study, 31 patients (7 females; 73 ± 8.3 years) with symptomatic carotid disease defined as ischaemic brain lesions in one carotid artery territory on diffusion weighted images underwent a carotid black-blood hr-MRI at 3.0-Tesla with fat-saturated pre-and post-contrast T1w-, PDw-, T2w-and TOF images using surface coils and parallel Imaging techniques (PAT factor = 2) within 7 days after symptom onset. All patients underwent extensive clinical workup (lab, brain MRI, duplex sonography, 24-hour ECG, transoesophageal echocardiography) to exclude other causes of ischaemic stroke. Prevalence of American Heart Association lesion type VI (AHA-LT6) and thrombus, status of the fibrous cap, presence and type of haemorrhage were determined in both carotid arteries based on previously published MRI criteria by one reviewer who was blinded to the clinical information. The Mc Nemar Test was used for statistical comparison. A p-value < 0.05 was considered statistically significant. Results: Symptomatic plaques had, compared to asymptomatic plaques, a higher prevalence of AHA-LT6 (65.5% vs. 10.3%; P < 0.001), ruptured/ulcerated fibrous caps (34.5% vs. 3.5%; P = 0.012), plaque haemorrhage (58.6% vs. 10.3%; P = 0.001), early subacute haemorrhage (27.6% vs. 0%; P = 0.008) and juxtaluminal thrombus (10.3% vs. 0%; P = n.s). Conclusion: 3 T hr-MRI is able to differentiate between symptomatic and asymptomatic carotid plaques, demonstrating the high potential of hr-MRI to differentiate between stable and vulnerable lesions and ultimately identify patients with low versus high risk for cardiovascular complications.

Purpose: So far the detection of abnormalities in ventricular size and shape are the most important sonographic clues in the early detection of developmental brain pathology. This study aims to analyse early foetal MR imaging signs in cases of lissencephaly. Methods and Materials: 26 foetal MRI examinations (1.5 Tesla) of 19 cases (mean age at initial MRI: 25.5 ± 4.3 gestational weeks -GW) with clinical, neuroimaging or neuropathological evidence for lissencephaly were retrospectively analysed. In 7 cases foetal MR and in 3 cases postnatal MR follow-up examinations including T1, multiplanar T2-weighted and diffusion weighted sequences were available. Overall image quality was rated excellent (54%;14/26), moderate (38%;10/26) and poor (8%;2/26). Results: An average mismatch between the morphologically estimated and actual gestational age of 5.1 ± 4.0 GW could be detected. 9/10 cases imaged before 24 GW showed abnormal opercularisation. Overall the characteristic pattern of foetal brain asymmetry was present in one third of cases. Signs of abnormal gyration were detected in 7/19 cases with 2/19 cases of Type 2 Lissencephaly showing an irregular cortical surface. The characteristic laminar organisation could not be discriminated in 60% of examinations on T2-w and in 40% of examinations on diffusion weighted sequences. Abnormally enlarged ganglionic eminence (6/19) and hippocampal malrotation (16/19) were commonly encountered. Infratentorial abnormalities included a wide posterior fossa (9/19), kinking of the brainstem (3/19) and an abnormally flat pons (9/19). Conclusion: Foetal MR detection of a variety of structural brain abnormalities can strengthen the confidence in the early in utero MR diagnosis of lissencephaly.

(ADI) and atherosclerosis by segment involvement score (SIS, defined as number of coronary artery segments per patient displaying atherosclerotic plaque). Changes in ADI and SIS between the scans were compared using linear and logistic regression accounting for differences in clinical and demographic baseline characteristics. Results: Age (p = 0.004), time between scans (p = 0.02), and an increase in SIS (p < 0.001) were associated with a decrease in ADI on univariate analysis. Increases in SIS remained a significant independent predictor of decreased ADI on both multivariate logistic regression (with change in ADI represented categorically) and multivariate linear regression (both p < 0.001). Conclusion: Progression of aortic stiffness and progression of atherosclerosis have a temporal relationship. While outcome studies are needed, a comprehensive assessment of vascular health with cCTA, including both indices of arterial stiffness and atherosclerotic burden, may improve risk stratification and therapy monitoring.

Use of 3D rather than 2D ultrasound results in more accurate determination of both endothelial and vascular function L. Kavanagh, C. Brown, P. Fitzgerald, M. Lee, A. Stanton; Dublin/IE (liamkavanagh@rcsi.ie) Purpose: Brachial artery flow-mediated dilatation (FMD) and nitrate-induced dilatation (NID), using 2D ultrasound, are non-invasive methods for assessing endothelial and vascular function. However, operator-dependence and poor reproducibility have limited the use of these assessments to research rather than clinical practice. We tested whether the use of 3D ultrasound could diminish operator dependence, and improve validity and reproducibility. Methods and Materials: 10 fasting males (18-80 years) attended the clinic twice. On each occasion the brachial artery (5 cm above the elbow) was imaged pre-and post-ischaemia (FMD), and pre-and post-sublingual glyceryl-tri-nitrate (NID), using a Phillips iU22 ultrasound system with L17-5 (2D) and VL13-5 (3D) probes. 2D cineloops and 3D volume sets were ECG-gated. Diastolic diameters (median of 3 R-wave images) were measured offline using AMS semi-automated edge-detection software. FMD and NID were calculated as the % changes in diastolic diameter with ischaemia and nitrates, respectively. Results: The use of 3D ultrasonic imaging resulted in improved reproducibility by comparison with 2D imaging -mean difference ± SD of differences between repeated measures for each of the assessments were as follows; 3D-FMD 0% ± 2.3% versus 2D-FMD 1.5% ± 5.9%; 3D-NID 0.7% ± 6.9% versus 2D-NID 2.7% ± 10.6%. Conclusion: Most likely due to avoidance of probe mal-alignment errors, use of 3D ultrasonic imaging rather than 2D results in more accurate determination of both endothelial and vascular function. This improved accuracy and operator independence is likely to facilitate more widespread clinical usage in patients at risk of cardiovascular events.

Study of total aorta and coronary arteries with prospective ECG-triggered 320-MDCT angiography Y. Li, Z. Fan, Z. Zhang; Beijing/ CN (athen06@hotmail.com) Purpose: To investigate the feasibility of prospective ECG-gating wide-volume protocol in CT angiography of whole aorta and coronary arteries Methods and Materials: Consecutively 55 patients suspected with acute aortic syndrome (AAS) underwent total aorta angiography with prospective ECG-gating wide-volume without heart rate control. The exposure window was set at 40-50% R-R interval (heart ≥ 70 bpm) or 70-80% R-R interval (heart < 70 bpm) within one forced heart beat. The parameters for the acquisition were 0.35 s/r, 120 kV, 450 mAs. The imaging quality of ascending aorta, aortic valve, and coronary artery was evaluated for motion artifact. The mean attenuation was measured in the ascending aorta, middle descending aorta and distal descending aorta. The radiation dose and contrast medium were recorded. Results: All of the examinations were performed successfully. The stair artefact could be found at thoracic part in patients with arrhythmia (16.4%). The imaging qualities were acceptable for evaluating the ascending aorta (100%), aortic valve (94.5%) and coronary arteries segments (97%). The mean attenuations were higher than 315 Hu at every segment of the aorta. The mean radiation dose was 20 ±4.7 mSv. The average agent was 88.2 ± 5 ml. Of the 55 patients (48 patients with aortic diseases, 3 paitents with aortic valve calcification), 21.8% patients had coronary artery diseases. In 2 patients with type A aortic dissection, right coronary artery came from false lumen. Conclusion: For patients suspected with AAS, prospective ECG-gating widevolume CTA of whole aorta has the potential to provide additional information about the coronary artery and aortic valve with low radiation exposure.

A S239 C D E F G H lobe was measured. The differences in PS in each area were assessed on each neonate. Ratios of the PS in the deep grey matter to PS in the cortical veins were compared with the corrected age, gestational age and age after birth. Results: There was a significant difference in PS among the regions (p < 0.01); pairwise comparison revealed that PS at the deep grey matter was significantly larger than that in the cortical veins (p < 0.01); PS in the rolandic area was significantly larger than that in the frontal or parietal area (p < 0.01). There was a negative correlation between the phase ratio and the corrected age (p < 0.01), but not correlated with gestational age and age after birth. Conclusion: A characteristic distribution of PS was identified in the neonatal venous system. The reduction of difference in PS in the vein at deep grey matter and cortex may reflect brain development. Purpose: MRI is the gold standard in the diagnosis and follow-up of pilocytic astrocytoma (PA). PA may appear as a cystic lesion with mural nodule, solid with intrinsic cysts or purely solid lesion; the solid part may enhance in a homogeneous or a heterogeneous fashion. Assessment of contrast enhancement (CE) pattern is essential for the diagnosis and follow-up. Sometime PA may mimic high grade tumour, and biopsy may not be always possible. Our purpose was to evaluate CE changes of PA with serial MR, and to assess these CE "fluctuations" as a possible new MRI sign for PA. Methods and Materials: Retrospective evaluation of 912 MR exams of 140 pts (with histologically proven PA) followed for an average of 4.5 ys with serial MRI. We selected 39 pts: off therapy, without NF1, without dimensional changes of tumour/ residual tumour. In pts with CE changes, tumour size and areas of CE were calculated with a cross-product. Descriptive statistics were calculated for continuous variables, effects of possible factors influencing changes of CE areas were tested. Results: Of the 39 pts without dimensional changes in tumour size, 12/39 showed CE changes in terms of increase/appear and/or decrease/disappear of the CE. 3/12 PA were infratentorial, 8/12 supratentorial and 1/12 endoventricular. There were no significant correlation between age, gender, tumour localisation, tumour size and CE areas. Conclusion: In our experience PA may show a variable CE during the years, in the absence of tumour/residual tumour dimensions change. CE fluctuation can be considered a MRI sign of PA, but not considered as an indicator of tumour progression or tumour response. 30-direction diffusion tensor imaging (DTI) at 3 Tesla. 3. Language characterized by paediatric neurologist. DTI was used for deterministic tract reconstruction. AF was categorized as left-only, right-only or bilateral. Language was categorized as intact (age-appropriate), mild-to-moderate impairment (some development, but delayed compared to piers) or profound impairment (non-verbal). Proportions were compared using Fisher exact test. Results: 30 patients met the criteria: 10 with polymicrogyria, 14 with focal cortical dysplasia (pathologically-proven), 1 lissencephaly, 4 schizencephaly, and 1 grey-matter heterotopia. 15 patients had intact language, 11 mild-to-moderate impairment and 4 profound impairment. 100% (10/10) of patients without a left AF had some degree of language impairment (6 mild-to-moderate, 4 profound); 25% (5/20) of patients with a left AF were impaired (5 mild-to-moderate). This difference was significant (p < 0.0001). 100% (15/15) of patients with intact language had an identifiable left AF. The frequency of absence of the left AF in patients with intact language was significantly less than that in patients with some degree of impairment (p < 0.0001). Absence of the left AF had 100% specificity [95%CI (75%, 100%)], 67% sensitivity and 100% positive predictive value [95%CI (68%, 100%)] for some degree of language impairment. Conclusion: These preliminary results suggest that absence of the left AF may be a specific biomarker of language impairment in patients with MCDs. NO (charlotte.de.lange@rr-research.no) Purpose: Changes in cerebral glucose metabolism may be an early prognostic indicator of hypoxic ischaemic injury. The aim of this study was to use dynamic FDG-PET for evaluating changes in cerebral glucose metabolism immediately after global hypoxia and the influence of the oxygen content in the resuscitation air. Methods and Materials: 16 anaesthesized piglets were subjected to global hypoxia using FiO 2 0.08 for 40 min followed by resuscitation with FiO 2 0.21 or 1.0 for 30 minutes and reoxygenation. Dynamic PET was performed by injection of 18 F Fluoro-2deoxy-glucose before and after hypoxia+resuscitation. T2-weighted MR images were acquired for anatomic imaging and coregistered with the FDG-PET images. ROIs were traced around the basal ganglia, cerebrum, cerebellum, cortex and white matter. Global and regional cerebral metabolic rates of glucose (CMRgl) before and after hypoxia+resuscitation was assessed by Patlak analysis for the two resuscitation groups. Statistical analysis was performed with paired t test with a significance level of p =.05. Results: CMRgl was significantly reduced from a mean baseline level (±1 SD) of 21.3±1.9 to 12.3±0.9 μmol/min/100 g after hypoxia+resuscitation, for all piglets (p < 0.01). There was no significant difference between the different regions or between the 21 and 100% group, but a trend of less decrease in CMRgl in white matter was observed. Conclusion: Global hypoxia causes immediate decrease of cerebral glucose metabolism in newborn pigs independent of resuscitation with 21 or 100% of O 2. Early regional and global cell death or a downregulation of the glucose metabolism are likely mechanisms. Purpose: Brainstem dysgenesis is a generic, umbrella, term that designates all patients with congenital dysfunction of cranial nerves and decreased muscle tone due to prenatal lesions or malformations of the brainstem. Cause is atributted to a vascular event occurring during early foetal life involving the developing brainstem tegmentum and provoking different clinical syndromes. A retrospective analysis of radiological-clinical findings in 24 children with brainstem dysgenesis is presented. Methods and Materials: A total of 24 patients with clinical brainstem dysgenesis were retrospectively reviewed. All patients were imaged at birth or shortly after and some of them were followed by MR imaging at older ages, including at that time DTI and tractography. Studies were performed using either a 1.5 or a 3.0-Tesla MRI scanner (Magnetom Avanto and Trio; Siemens, Erlangen, Germany). Images were reviewed independently by two paediatric neuroradiologists and findings were agreed on by consensus. Results: Brainstem tegmental hypoplasia was found in 11 of our 24 cases, being classified as symmetrical (n = 10), and asymmetrical (n = 1). One patient showed a tiny calcification within the pons on cranial CT. Supratentorial anomalies were found in 4 patients, mainly polymicrogyria (3) and ventriculomegaly (1). Miscellaneous imaging findings were agenesis/hypoplasia of cranial nerves, better depicted on CISS images. DTI-FT demonstrated additional findings beyond those seen with conventional MR imaging. Conclusion: Traditional MRI may be insufficient to evaluate brainstem hypoplasia or cranial nerves absence-hypoplasia, particularly during neonatal period. Modern sequences such as CISS or DTI-FT may depict greater detail providing important information in children with braistem dysgenesis.

Spatial dependency of phase shift in the cerebral venous system of neonates using susceptibility-weighted magnetic resonance imaging T. Niwa 1 , N. Aida 1 , H. Kawaguchi 2 , T. Obata 2 , Y. Tachibana 1 , R. Watanabe 1 , J. Shibasaki 1 ; 1 Yokohama/JP, 2 Chiba/JP (tniwa@kcmc.jp) Purpose: Phase analysis on susceptibility-weighted imaging (SWI) has been introduced to be related to blood oxygenation. The purpose of this study was to assess the spatial distribution of phase shift (PS) in the neonatal cerebral venous system using phase image on SWI. Then, 53/80 (66%) patients had surgery and underwent intraoperative electrophysiological studies. Analysis of US findings included exiting nerve root injuries with or without pseudomeningoceles and posttraumatic neuromas. Results: In patients who underwent surgery, the sensitivity and specificity of the US findings of posttraumatic neuromas in the interscalenic area were 89% and 95% respectively, using surgical findings and somatosensory-evoked potentials as the standard of reference. Dividing the brachial plexus in an upper and lower part, US was able to establish the level of injury in 70% of cases. Paraforaminal nerve root injuries and pseudomeningoceles were identified with US in only 12% and 8% of surgically-proven root avulsions. In the group of infants who were managed conservatively, US was 92% sensitive and 100% specific to detect neuromas using MR imaging as the standard of reference. In this group, posttraumatic neuromas were smaller in size (p <.0001) or absent (9 cases). Conclusion: US is promising for imaging brachial plexus birth injuries. The information provided has a prognostic value and is useful for the surgeon in a preoperative setting.

Why Purpose: To evaluate repeatability and reproducibility of 4 different ADC sampling methods in the healthy liver parenchyma. Methods and Materials: DWI breath-old echo-planar acquisitions (b=0-300-1000 s/mm²) were executed on a 1.5 T MR scanner. Two observers evaluated 30 healthy subjects (age range 18-> 65 years) DWI images, with 4 different sampling methods: one region of interest (ROI) per section (OR), one ROI per segment (OS), partial liver volume (PV), total volume (TV), considering, respectively 4%, 4%, 30% and 70% of the entire liver volume parenchyma. Data were reported per group (mean±SD). Statistical evaluation was done by ANOVA test. Results: ADC (10 -6 mm 2 /s) mean increases, enlarging ROI size: 1343±201 (OR), 1373±221 (OS), 1463±253 (PV), 1560±211 (TV). Left liver lobe was difficult to be evaluated, given the presence of artefacts. Repeatability was greater than reproducibility (ICC 0.80 vs < 0.45); both increase averaging 3 measurements of the lower-right liver lobe (ICC ≥0.87 and 0.82, respectively). Conclusion: Inter-observer variability was greater than intra-observer one. Reproducibility can be increased averaging measurements of the lower-right liver lobe, using a large-size ROI. ADC increment, observed using a large ROI, can be explained with the more pronounced contribution of perfusion and vessels.

Variation with age of the apparent diffusion coefficient (ADC) of the liver parenchyma: an in vivo versus in vitro study L.N. Mazzoni, F. Pasquinelli, C. Nardi, G. Belli, F. Regini, S. Colagrande;

Purpose: To evaluate mean and dispersion of liver ADC measurements varying the size of the region of interest (ROI) and the age of the subjects at two different b values, comparing results with phantom measurements B A S243 C D E F G H pathologic findings and imaging follow-up. Mann-Whitney test and a ROC curve were used for statistical analysis. Results: The mean ADC values (x10exp -3 mm 2 /sec) ± standard deviation were 1.77 ± 0.57 for haemangiomas, 1.25 ± 0.28 for adenomas; 1,257 ± 21 for CCC, 0.98 ± 0.21 for HCCs, 1,276 ± 0.299 for shunts, 1.6 ± 0.278 for regenerative nodules, and 2,676 ± 0.37 for cysts, 1,066 ± 0.26 for metastases. A good positive correlation was found between ADC value of liver lesions and reference standard (Spearman's rho =.67) with a significantly lower ADC value of malignant lesions compared to that of benign (p <.0001). 89% of lesions were correctly classified as malignant and benign at a threshold ADC value of 1.18.It is not possible to establish a cut-off point due to the wide variability into the benign lesion group. Conclusion: DWI using the respiratory-triggered SSEPI sequence could be useful to differentiate benign from malignant focal liver lesions. ADC value of malignant lesions is significantly lower than that of benign; however, it is not possible to establish a cut-off point.

Diffusion-weighted MRI for assessment of colorectal metastases: could it substitute Gd-EOB-DTPA-enhanced MRI? R. Cianci, A. Filippone, F. Sabatino, E. Pace, V. Bianco, A. Tartaro, A.R. Cotroneo; Chieti/ IT (r.cianci@rad.unich.it) Purpose: To evaluate the role of diffusion-weighted imaging (DWI) as a substitute for Gd-EOB-DTPA-enhanced MRI in the assessment of colorectal hepatic metastases.

We included 54 patients with ≤5 surgically proven colorectal hepatic metastases, who underwent Gd-EOB-DTPA-enhanced MR examination including DWI (b = 0-500 s/mm 2 ). All patients underwent liver resection after intraoperative ultrasound (IOUS) within two weeks from MRI. Two independent readers, with different experience in abdominal imaging, reviewed three image sets (DWI, Gd-EOB-DTPA, DWI and Gd-EOB-DTPA) to detect hepatic metastases using a four-grade scale. ROC analysis was performed and sensitivity was calculated for each reader and each reading session, using IOUS and histologic findings as reference standard. Results: There were a total of 115 liver metastases (mean diameter 2.3 cm, range 0.4-6.3 cm). The accuracy in detecting hepatic metastases was lowest for DWI alone (reader 1, 0.823; reader 2, 0.872), preceded by Gd-EOB-DTPA set (reader 1, 0.923; reader 2, 0.895) and combined set (reader 1, 0.985; reader 2, 0.994). The differences resulted statistically significant only between combined and DWI set (p = 0.034). For the less experienced reader, combined set showed a significantly higher sensitivity than DWI and Gd-EOB-DTPA set for lesions smaller than 1 cm, whereas no significant differences were observed for lesions bigger than 1 cm and for the experienced reader. Conclusion: DWI can not substitute Gd-EOB-DTPA-enhanced MRI in the assessment of hepatic metastases. Nevertheless, it seems to improve overall lesion detection and diagnostic confidence when combined with Gd-EOB-DTPA-enhanced MRI, especially for the less experienced radiologist.

Diffusion-weighted MR imaging for differentiation of benign from malignant lesions in the periampullary region S. Kim, H. Kwon, N. Lee, G. Kim, D. Kim, H. Seo; Pusan/KR (kh8310@naver.com) Purpose: The purpose of our study was investigating the value of DWI for distinguishing benign from malignant lesions of the periampullary region. Methods and Materials: Eighty-three patients underwent respiratory-triggered breath-hold DWI at 1.5 or 3.0 T MR. DWI was performed with three b factors (0, 500, and 800 s/mm 2 ), and apparent diffusion coefficients (ADCs) were also calculated in the periampullary lesions. On DWI, signal intensity (SI) of the periampullary lesions was compared with SI of the back muscle and pancreas. In addition, we compared ADCs of malignant from benign lesions at MR by the Mann-Whitney U test. Results: Seventy-eight patients were evaluated excluding 5 patients due to previous interventions or technical errors. Nine patients with malignant lesions and 23 patients with benign lesions were evaluated at 1.5 T. Thirteen patients with malignant lesions and 33 patients with benign lesions were evaluated at 3.0 T. On DWI with a b factor 800sec/mm 2 , 20 of 22 malignant lesions demonstrated slightly high SI (1/22, 4.5%) and significantly high SI (19/22, 86.4%). Fourteen of 56 benign lesions demonstrated slightly high SI (10/56, 17.9%) and significantly high SI (4/56, 7.1%). The mean ADC of malignant and benign periampullary lesions were 1.25±0.26×10 −3 mm 2 /s and 1.89±0.33×10 −3 mm 2 /s, respectively at 1.5 T MR (p =.00). The mean ADC of malignant and benign periampullary lesions were 1.39±0.35×10 −3 mm 2 /s and 1.96±0.35×10 −3 mm 2 /s, respectively at 3.0 T MR (p =.00). Conclusion: DWI can be reliable in the differentiation of benign from malignant lesions in the periampullary region.

Diffusion-weighted MR imaging after administration of different gadolinium contrast agents: is it still reliable? S. Kinner, L. Umutlu, G. Antoch, T.C. Lauenstein; Essen/DE (Sonja. Kinner@uni-due.de) Purpose: Diffusion weighted imaging (DWI) is an emerging technique for abdominal MRI. DWI is usually performed before intravenous contrast injection. However, it may be necessary or time-saving to acquire DWI data after gadolinium administration. We aimed to compare DWI before and after the administration of different gadolinium compounds. Methods and Materials: 15 patients with known HCC underwent liver MRI at 1.5 T (Magnetom Avanto, Siemens) including DWI on day 1 before and after administration of gadobutrol (Gadovist®) and on day 2 after administration of EOB-gadolinium-DTPA (Primovist®). Signal-to-noise ratios (SNR) and contrast-to-noise ratios (CNR) of HCC lesions and liver parenchyma were determined for all acquired DWI data sets. Furthermore, ADC values were calculated and compared using a Wilcoxon test. A p-value < 0.05 indicated statistical significance. Results: There were no statistically significant differences regarding SNR precontrast (mean: 43.3) after gadobutrol (43.2) or after EOB-gadolinium-DTPA (46.5). Similarly, no significant differences were found for CNR (average values: 34.7 vs. 32.1 vs. 35.7) nor for ADC-values (1.5 vs. 1.4 vs. 1.5 x 10 -3 ). Conclusion: There is no significant difference regarding DWI of HCC before and after intravenous injection of different gadolinium compounds. Hence, it is feasible to run DWI after gadolinium either as an alternative for unsuccessful pre-contrast DWI or as a gap filler to spare time in EOB-gadolinium-DTPA imaging. Purpose: Gd-EOB-DTPA-enhanced liver MRI is widely used for detection and differentiation of focal liver lesions. Diffusion weighted imaging (DWI) including apparent diffusion coefficient (ADC) measurements is increasingly utilised as a fast and, with limitations, quantitative method for liver lesion characterisation. DWI can be acquired before and after contrast agent injection. However, knowledge on Gd-EOB-DTPA-related influences on DWI of the liver is limited. Consequently, this single center prospective intraindividual comparison study was performed in order to systematically investigate this matter. Methods and Materials: Consecutive patients referred to standardised liver MRI (1.5 T, Gd-EOB-DTPA, 0.025 mmol/kg) for screening of metastases over 13 months were investigated. All underwent a breathold DWI sequence before and after contrast agent application (EPI-DWI, TR/TE (eff): 2100/62 ms, b-values: 0/800 s/ mm 2 ). Patients with treated liver lesions were excluded. Signal intensity of lesion, parenchyma and noise on DWI images as well as the ADC value was measured after identification by two observers in consensus using manually placed regions of interest. Reference standard was either histopathology or imaging follow-up determined separately by two radiologists. Data analysis included contrast-to-noise ratio (CNR) calculation and Wilcoxon signed-rank test. Results: 50 malignant and 39 benign lesions were identified. Neither CNR nor ADC values showed significant differences between pre-and postcontrast DWI. Both pre-and postcontrast ADC values differed highly significant between benign and malignant lesions (P < 0.001). Conclusion: The present study did not identify a significant influence of Gd-EOB-DTPA on DWI of liver lesions. This allows for individual tailoring of imaging protocols to clinical needs. Purpose: To evaluate the prognostic value of prospective ECG-gated dual source coronary CT angiography (DSCT-CA) for major adverse cardiac events (MACE) in patients with suspected coronary artery disease (CAD).

A total of 102 patients (58 males and 44 females, age 62 ± 11.7 years) with suspected coronary disease underwent prospective ECGgated DSCT-CA to assess the presence, severity and positive remodelling of coronary atherosclerotic plaques. The occurrence of MACE (cardiac death, myocardial infarction and unstable angina with hospitalisation) was evaluated in each patient during an average 28-month follow-up (24-32 months). CAD was defined significant when stenosis was >= 50% in any coronary artery. Purpose: Recent studies have established the prognostic value of cCTA for allcause mortality. However, the predictive value of cCTA specifically for cardiac events is less well documented. We hypothesize that extent and severity of coronary artery disease (CAD) as defined at cCTA are predictive of major adverse cardiac events (MACE).

We examined a consecutive cohort of 454 patients (233 women, mean age 57±years) who had undergone cCTA using either 64-slice or dual-source CT for atypical chest pain. Stenosis at cCTA was scored as minimal (< 30%), mild (30% -49%), moderate (50% -69%), or severe (≥70%) for each coronary artery. A > 12-month follow-up of MACE was assessed using Cox proportional hazards models adjusted for pretest CAD likelihood and risk factors. MACE were determined using the Social Security Death Index, medical record review and phone calls to patients and referring physicians. 

Ex-vivo classification of non-calcified coronary atherosclerotic plaque components on CT coronary angiography: impact of neighbouring attenuation and density thresholds E. Maffei, C. Martini, T. Arcadi, S. Seitun, E. Capuano, I. Romano, I. Romano, C. Lario, F. Cademartiri; Parma/IT (filippocademartiri@hotmail.com) Purpose: To assess the effect of neighbouring attenuation and density thresholds on attenuation values of non-calcified plaque by computed tomography coronary angiography (CTCA) in an ex-vivo model. Methods and Materials: We studied 7 ex-vivo left coronary artery specimens surrounded by olive-oil and filled with saline and 4 solutions with decreasing dilutions of contrast material: control (saline), 1/200, 1/80, 1/50 and 1/20. The CTCA protocol was: slice/collimation 32 x 2 x 0.6 mm, and rotation time 330 ms. The attenuation (HU) value of atherosclerotic plaques was measured for each dilution in: lumen, plaque (non-calcified coronary wall thickening), calcium and surrounding oil. Based on the measured CT attenuation values, plaques were classified as lipid-rich or fibrous using 30 HU, 50 HU and 70 HU density thresholds. The results were compared with one-way ANOVA-test and correlated with Pearson's test. Results: 500 cross sections with plaques were detected and assessed. Mean lumen (45 ± 38 HU to 669 ± 151 HU) and plaque (11 ± 35 HU to 101 ± 72 HU) attenuation differed significantly (p < 0.001) among the different dilutions. The attenuation of lumen and plaque of coronary plaques showed moderate correlation (r = 0.54; p < 0.001) while the correlation between plaque attenuations in neighbouring dilutions was good (r > 0.70). Using a 50 HU threshold, the prevalence of lipid-rich plaques would progressively shift from 441 (88%) to 136 (27%) with increasing lumen attenuation (p < 0.05). The same was observed using a threshold of 30 HU (from 75% to 19%; p < 0.05) and 70 HU (from 88% to 27%; p < 0.05). Conclusion: Classification of coronary plaques as lipid-rich or fibrous based on absolute CT attenuation values is significantly affected by the neighbouring attenuation and density thresholds.

Impact of metabolic syndrome on coronary artery stenosis and plaque characteristics assessed by 64-slice multidetector-row cardiac computed tomography E. Chun 1 , S. Lim 1 , S. Choi 1 , K. Jin 2 , Y. Lee 1 , C. Jeon 1 , Y. Bae 1 , J. Yoo 1 , H. Jang 1 ; 1 Seongnam-si/KR, 2 Seoul/KR (drejchun@hanmail.net) Purpose: We investigated plaque characteristics as well as coronary artery stenosis and plaque burden according to the presence of metabolic syndrome (MS) by 64-slice multidetector-row cardiac computed tomography (MDCT). Methods and Materials: Coronary artery stenosis (> 50% of diameter), plaque burden and plaque characteristics as well as coronary artery calcium score (CACS) were assessed by MDCT in 3,000 age and sex-matched asymptomatic individuals. Anthropometric parameters and metabolic profiles were also measured. Results: Subjects with MS had more coronary stenosis, multivessel involvement, greater plaque burden and higher CACS than subjects without MS (all, P < 0.01). Mixed or non-calcified plaques were also more prominent in subjects with MS than those without MS. After adjustment for confounding factors, MS was strongly associated with significant coronary stenosis, greater plaque burden and multivessel involvement. In plaque types, mixed plaque was associated with MS and correlated significantly with the Framingham risk scores. Conclusion: In asymptomatic subjects with MS, presence of mixed plaque, plaque burden and coronary stenosis were significantly greater than subjects without MS as assessed by MDCT. MDCT is an effective imaging modality for detecting coronary atherosclerosis and plaque characteristic in asymptomatic patients but future study on its clinical effect is warranted to further confirm its role.

Reproducibility of assessment of non-calcified coronary artery plaque burden across different image analysis platforms S. Oberoi, J.M. Barraza, C. Thilo, U.J. Schoepf; Charleston, SC/US (barraza@musc.edu) Purpose: Coronary CT angiography (cCTA) assessment of non-calcified coronary artery plaque (NCP) burden is gaining use for risk stratification and monitoring therapy. Sufficient reproducibility is essential for serial assessment. We investigated reproducibility of non-calcified plaque burden quantification across multiple commercial image-analysis platforms. Methods and Materials: For 47 patients with documented NCP on cCTA, automated quantification of NCP burden was performed on identical data sets using 3 different image analysis platforms (Aquarius, TeraRecon; Circulation, Siemens; Vitrea, Vital).

Results: 3 patients were excluded from the analysis. In 12 patients, coronary plaques were found in 46 slices on both CTA and MRI. These plaques were classified to 3 groups based on CTA: calcified plaques (n = 11), soft plaques (n = 23), and mixed plaques (n = 12). In MRI, the plaque burden, maximal wall thickness, SNR, CNR in the coronary walls containing plaques were greater than those of the normal coronary walls (0.83 ± 0.08 vs 0.73 ± 0.08, 1.88 ± 0.51 vs 1.51 ± 0.26 mm, 12.95 ± 2.78 vs 9.93 ± 2.31, 6.76 ± 2.52 vs 3.89 ± 1.54, respectively, p < 0.05). The luminal CSA at the plaque was smaller than that at normal coronary walls (2.50 ± 1.50 vs 4.72 ± 2.28 mm 2 , p < 0.05). The SNR in the soft plaque was significantly greater than those in calcified and mixed plaques (p < 0.05). Conclusion: Coronary wall MRI can identify coronary plaques in the proximal and middle segments, and has the potential to differentiate plaque types based on signal intensity.

Prognostic value of subclinic artery disease: correlation between MDCT-CA and carotid artery MRI R. Malago, C. Barbiani, S. Rainer, G. Caliari, A. Pezzato, P. Vallerio, A. Fratta Pasini, R. Pozzi Mucelli; Verona/IT (robertomalag@yahoo.it) Purpose: Acute cardiovascular events are related to type of coronary plaque rather than coronary artery stenosis degree. The aim of our study is to evaluate the prevalence of sub-clinic coronary artery disease (CAD) and relate plaque composition to carotid US and MRI. Methods and Materials: 30 consecutive patients (22 M, 8 F) mean age 67 ± 8 YO between 03 and 06/2010 underwent MDCT-CA and carotid artery US and MRI. All patients were on high level pre-test probability of CAD according to AHA risk score. For carotid arteries and coronaries type of plaque (calcific, mixed and fibrolipidic) and degree of stenosis was distinguished into significant (> 50% lumen narrowing) or not significant (< 50% lumen narrowing). Correlation of plaque composition and significance of stenosis was obtained by means of Spearman correlation. Results: 38 plaques (1.26/patient) were detected with US and MRI (27/38 fibrolipidic/mixed, 11/38 calcific). Mean Ca score obtained was 2205.5 (0-2893.8). 105 plaques were detected in MDCT-CA fibrolipidic (20/105) 19% mixed 29/105 (28%) and calcific 56/105 (53%). Correlation between US and MRI in plaque morphology was high (r = 0.95 for mixed, r = 1 for calcified and r = 0.8 for soft plaques). Correlation between the two MRI and MDCT-CA was r = 0.311, P = 0.101 for fibrolipidic plaque, r = 0.579, P = 0.001 for mixed plaque, r = 0.471, P = 0.010 for calcific plaque. Correlation for stenosis degree and between MRI and US was r = 0.8 and between the MDCT. CA and MRI technique was r = 0.518 P = 0.004. Conclusion: Plaque morphology and pattern of stenosis on carotid arteries can help to predict the presence of significative CAD.

Coronary atherosclerosis and coronary artery disease in patients with obstructive sleep apnoea: assessment with coronary CT angiography G.W. Rowe, R.B. Wilson, A.M. Armstrong, J.A. Abro, A.T. Parker, U. Schoepf, S. Sharma; Charleston, SC/US (wilsonro@musc.edu) Purpose: There is increasing evidence linking OSA with coronary artery disease (CAD). We aimed at using coronary CT angiography for studying the association between OSA and coronary atherosclerosis and CAD. Methods and Materials: 49 patients having documented OSA by polysomnogram and Berlin questionnaire underwent dual-source cCTA. They were matched with 46 patients with similar cardiovascular risk, but free of OSA. All patients underwent cCTA for atypical chest pain and/or prior equivocal physiological testing. In all 95 patients, two experienced observers in consensus analysed the cCTA data for the presence and degree of coronary artery stenosis. The association between OSA status, stenosis, and plaque type was tested using Χ 2 and Fisher's exact statistics. Results: There was no significant (p > 0.05) difference in demographics between patients with and without OSA. Global coronary artery calcium scores were not significantly different between the two groups (mean Agatston score 272 ± 422 with OSA versus 241 ± 415 without OSA, p = 0.5). However, in patients with OSA there was a significantly (p = 0.0013) higher prevalence of stenotic CAD and more extensive vessel involvement (6% one-vessel, 27% two-vessel, 22% three-vessel, 33% four-vessel CAD in OSA versus 15%/7%/13%/24% without OSA, p = 0.0017). Composition of stenotic lesions was notably different between the two groups, with a significantly (p = 0.0043) higher prevalence of non-calcified and mixed plaque in the OSA group. Conclusion: Our initial results suggest that patients with OSA have more active coronary atherosclerotic disease, with a greater degree of vessel involvement with stenosis and more stenotic non-calcified and mixed plaque than patients without OSA.

A D E F G H Methods and Materials: Seventy-seven patients ≥ 50 years (1-3 osteoporotic vertebral compression-fractures between T4 and L5) were enrolled in an open-label, prospective, randomised study (randomization 2:1 for CDKS (49patients/65 levels) or vertebroplasty (28patients/39 levels) treatment. Polymethylmethacrylate (PMMA)cement was used in both treatments. Cement-leakage and leak-locations were evaluated by CT and post-operative radiographs. Pain-levels were assessed pre-/ post-operatively, at 3 and 12 months using ten-point visual-analogue-scale (VAS). Disability was assessed pre-operatively, at 3 and 12 months using the Oswestry-Disability-Index (ODI). Radiographs and CTs were obtained at 3 months and final radiographs at 12 months. Anterior vertebral body height was determined from CT and heights were compared between post-op and 3 months. Radiographs and CTs were evaluated for refractures or new adjacent level-fractures at each follow-up. Results: CT provided increased resolution and additional axial-view for detection of more cement-leaks versus radiography. Treatment with CDKS significantly reduced number of levels with leaks compared with vertebroplasty (47.7%vs.71.8%,p = 0.0132). When leaks were classified by location, significantly fewer cortical and basivertebral vein-leaks occurred in the CDKS-group (p = 0.0158/p = 0.0217, respectively). Immediate pain-relief sustained throughout the study for both treatmentgroups. ODI-scores were similarly improved for CDKS and vertebroplasty at 3 and 12 months, compared with pre-operative levels. Examining height maintenance over time revealed that mean change in anterior vertebral body height was 5.90% (±11.36) for CDKS and 9.49% (±8.28) for vertebroplasty: CDKS may provide improved long-term height-maintenance (p = 0.2148). 

Last year, we prospectively compared 2 groups. Group A (36 patients, 75 ± 15 years) with vertebral fractures were evaluated on electronic baropodometer. Load distribution between right and left foot during standing and walking was recorded prior (group A1) and the day after (group A2) PV. Group B (30 patients, 42 ± 13 years) with no back pain or surgery record (normal population) was evaluated on the same electronic baropodometer. The two groups were compared by means of related samples Wilkoxon Signed Rank test. Results: Mean value of load distribution variation between rear and front of the same foot for the normal population was 9.45 ± 6.79% (standing) and 14.76 ± 7.09% (walking). Mean value of load distribution variation between rear and front of the same foot prior to PV was 16.52 ± 11.23% (standing) and 30.91 ± 19.26% (walking). Load distribution variation post-PV was 10.08 ± 6.26% (standing) and 14.25 ± 7.68% (walking). Comparison of load distribution variation among groups A1 and A2 or groups A1 and B is statistically significant (p = 0.001 and p = 0.011, respectively). Comparison of load distribution variation among groups A2 and group B is not statistically significant (p = 0.486). Conclusion: Comparative groups enable us to consider that PV apart from the already known pain reduction effect is efficient on equilibrium and load distribution improvement as well. Purpose: We determined whether the temperature of saline influences procedure performance and patient short-term outcome of an ultrasound-guided lavage technique for treating CTRC. Methods and Materials: 88 patients (49 females, mean age 47 ± 11.1 years) with painful shoulder and ultrasound diagnosis of CTRC were prospectically randomised. Group A (n = 44) underwent ultrasound-guided percutaneous treatment of CTRC (local anaesthesia, double-needle saline lavage and calcium aspiration, intrabursal steroid injection) using saline at room temperature, group B (n = 44) underwent the same treatment with saline at about 42 °C. Duration of procedure was recorded and ease of calcium dissolution was subjectively scored 1 (easy), 2 (intermediate), or 3 (difficult). Ultrasound appearance of the calcium deposit was judged as hard, soft, fluid. Visual analogue scale (VAS) was used for patients' discomfort. Results: Procedure duration was significantly lower (P =.030) in group B (544 ± 241 s) compared to group A (791 ± 311 s). Calcium dissolution was significantly improved (P =.034) in group B (median score 1) compared to group A (median score 2); in the subgroup with hard deposits (group A, n = 14; group B = 16), ease of calcium dissolution was larger and more significant (P <.001). VAS was significantly lower after treatments in both groups compared to baseline (group A, VAS before = 9.0 ± 0.6; VAS 1 month = 4.7 ± 0.5; VAS 3 months = 3.4 ± 0.4; group B, 9.1 ± 0.4, 4.5 ± 0.4, 3.3 ± 0.5, respectively, P <.001 for all). While four post-procedural bursitis were observed in group A, none of them was observed in group B. Conclusion: When treating CTRC, using warm saline solution could reduce procedure duration, improve calcium deposit dissolution, and reduce the occurrence of post-procedural bursitis. Purpose: To prospectively compare cement leakage-rates, leak-locations, painrelief and patient-disability for cement-directing kyphoplasty system (CDKS) and vertebroplasty in painful vertebral compression-fractures.

B A S247 C D E F G H

Pain palliation of bone metastasis: initial clinical experience using high intensity focused ultrasound therapy with magnetic resonance guidance A. Napoli, M. Anzidei, C. Catalano, M. Mastantuono, F. Zaccagna, L. Molisso, G. Cartocci, L. di Mare, R. Passariello; Rome/IT (alessandro.napoli@uniroma1.it) Purpose: To determine the efficacy of non-invasive high intensity MR-guided focused ultrasound (MRgFUS) treatment for palliation of bone metastasis pain in patients not candidated for External Beam Radiotherapy (EBRT). Methods and Materials: Under the IRB approval, 18 patients with 21 lesions underwent MRgFUS treatment using the ExAblate 2000 system (InSightec). Treatments were done in a single session, in an ambulatory setting. 12 patients underwent prior EBRT with a mean 6 months recurrent pain. In 6 patients, MRgFUS treatment was performed as first treatment modality. Effectiveness of pain palliation was evaluated at follow-up using the visual analog pain score (VAS) and measurable changes in analgesics intake. For tumour control perfusion, T1w images were obtained preand post-treatment in order to determine the non-perfused sonication-related area. Results: All patients and all lesions were treated. Mean follow-up time was 4 months. At baseline median, VAS was 7; it was 5 at 3 days, 3 at two weeks and 2 at one and four months. Wilcoxon's rank test showed a statistically significant difference between baseline and follow-up VAS values (0.01 < p < 0.038). No heating-related adverse event was recorded during this clinical application; patient medication intake was considerably reduced. Variable degree on non-perfused volume was observed after treatment, mainly within the pericortical region. Deeper penetration of the acoustic energy is at present desirable even if technically difficult to achieve with the current system. Conclusion: MRgFUS is a promising non-invasive treatment modality for successful palliation of bone metastasis pain in patients who are not candidate for EBRT.

Clinical application of CT-guided radiofrequency ablation for the treatment of metastatic bone neoplasms Z. Wang, L. Zhang, Y. Zheng, J. Gong, Z. Chen; Shanghai/CN (wzm0722@hotmail.com) Purpose: To investigate the clinical efficacy of CT-guided radiofrequency ablation (RFA) for the treatment of metastatic bone neoplasms. Methods and Materials: Under intravenous aneaesthesia. CT-guided RFA was performed in 20 patients with metastatic bone tumour. The degree of pain was evaluated at 24 hours, 3 and 6 months after the operation by brief pain inventory (BPI). Results: All patients were followed up for 6 months and survived so far. The average peak pain score before the operation was 8.1 (6-10), which decreased significantly to 6.1, 4.6, 3.3 and 3.0 at 24 hours, 1, 3 and 6 months after the operation, respectively (P < 0.001). The mean pain score before the operation was 63, which decreased significantly to 4.0, 2.3, 2.1 and 1.9 at 24 hours, 1, 3 and 6 months after operation, respectively (P < 0.001). After RFA treatment. The KPS scores of all patients increased while the CT values of the bones lesions decreased. No major complications occurred both during and after the operation. One patient with vertebral lamina destruction suffered from lower limb hypoesthesia after RFA procedure, and the lower limb sensation was restored within 48 hours after the injection of prednisone was employed. Conclusion: CT-guided radiofrequency ablation is a safe, effective, minimallyinvasive and up-to-date technique for the treatment of metastatic bone neoplasms with excellent anti-pain effect; its short-term response is sure and reliable. Methods and Materials: 48 disc segments of 41 patients scheduled for intradiscal thermal treatment, total disc replacement or spondylodesis were examined. An 1.0 T open MRI was used for instrument guidance and imaging. After primary disc puncture under guidance of fast interactive PDw TSE imaging, 1-2 ml of gadolinium contrast saline mixture was injected into the disc. The occurrence of memory pain during injection was recorded. Axial and sagittal T1w TSE images with and without fat-saturation were obtained. All MRI discograms were analysed by two readers, who were blinded to the clinical findings. Results: Overall, the placement of the puncture needle in the targeted disc was accurate under real-time MR guidance. Injections were technically successful in all cases. No major complications occurred. The mean procedure time was 17 min disability scale (ODI 0-50) was performed baseline, at one and six months after the procedure. We also carried out x-ray, CTms and MRI checks at one and six months. The use of analgesic drugs before and after the procedure was also evaluated. Results: we have obtained a progressive reduction of the pain in all the patients (VAS baseline: 8.36, at 1M 2.18, at 6M 2.17) and a good improvment of the quality of life (ODI baseline 24, at 1M: 11.7, at 6M: 9.3), a progressive reduction of bone oedema and a progressive inclusion of the bone cement. 100% of patients were used to take analgesic drugs before the implantation, while only 18% continued to do. No complications arose. Conclusion: Our preliminary studies prove that calcium sulphate could result in being an actual substitute of PMMA in the treatment of osteoporotic and traumatic vertebral fractures, especially in young patients. 

We performed PV procedures in 32 patients with metastatic disease, who presented at least one vertebral lesion with destruction of the posterior vertebral body wall or epidural extension of tumour. Amongst this patients, nine had signs of spinal cord or cauda equina compression. Procedural safety was evaluated by any post-procedure adverse events. The PV analgesic efficacy was evaluated by means of follow-up clinic reexamination or phone calls to assess change in patients' pain level prior and post procedure. The effectiveness was defined as at least a 50% improvement in pain as compared with baseline of prior procedure. Follow-up for assessments were performed after the procedure at 1 day; 1 week; 1, 3, and 6 months. Results: Fifty-three levels in 32 patients were treated in this group. Analgesic efficacy of PV was obtained in 97% patients (31/32) at 1 day and 1 week, 86% at 1 month, 89% at 3 months and 87% at 6 months after procedure. Leakage of PMMA was detected in 64% (33/53) treated vertebrae. No complications with clinical sequelae were encountered. Conclusion: Epidural involvement with or without neurological signs of spinal cord or cauda equina should not be a contraindication to performance of PV in patients with few other treatment options.

A prospective study comparing percutaneous decompression, combination of minimally invasive techniques and conservative therapy for the treatment of intervertebral disc herniation D.K. Filippiadis, G. Velonakis, A. Mazioti, A. Malagari, N.L. Kelekis, A. Kelekis; Athens/GR (dfilippiadis@yahoo.gr) Purpose: To compare efficacy (pain reduction and mobility improvement) among percutaneous disc decompression (PDD), combination of minimally invasive techniques and conservative therapy for intervertebral disc herniation treatment. Methods and Materials: During the last 4 years, we prospectively compared 3 groups of 15 patients, each suffering from sciatica due to intervertebral disc herniation. Group A underwent fluoroscopically guided percutaneous intervertebral disc decompression (PDD). Group B underwent a session where PDD was combined either to intradiscal discogel infusion, or to percutaneous nucleoplasty (plasma energy coblation). Group C underwent conservative therapy (analgesics, anti-inflammatory drugs, muscle relaxants, physiotherapy) for 6 weeks. Pain reduction and mobility improvement were recorded [clinical evaluation, 0-10 units Numeric Visual Scale (NVS)]. Results: Within a 24 months follow-up period, Group A patients presented a mean pain reduction of 6.60±3.31 NVS units (80.38% success rate), Group B patients a mean pain reduction of 5.73±2.60 NVS units (76.82% success rate) and Group C patients presented a mean pain reduction of 2.33±2.94 NVS units (34.30% success rate). No complications were noted. Comparison of pain reduction among Group A and Group C or among Group B and Group C is statistically significant (p=0.005 and p=0.004, respectively). Comparison of pain reduction among Group A and Group B is not statistically significant (p=0.43). Conclusion: Comparative groups enable us to consider minimally invasive techniques (solely performed) rather than conservative treatment as the treatment of choice for symptomatic small to medium-sized intervertebral disc hernias. There is no proof that combining minimally invasive techniques will improve the success rate. Purpose: To evaluate the cerebral vasoreactivity (CVR) using BOLD fMRI during carbogen inhalation with 7% CO 2 in Alzheimer disease (AD) and mild cognitive impairment (MCI). Methods and Materials: Thirty nine subjects were included to be investigated using BOLD fMRI (1.5 T) during a block-design carbogen inhalation paradigm, with a high concentration face-mask under physiological monitoring. Basal cerebral perfusion was measured using pulsed arterial spin labeling (PASL) with a Q2TIPS sequence. Data were analyzed using Matlab® and SPM5 with the group-average capnia responses as regressors and corrected for gray-matter partial volume. Results: Among participants, 12 subjects were excluded because of incomplete protocol, leaving 27 subjects without significant microangiopathy diagnosed for AD (n=9), MCI (n=7), and matched controls (n=11) to be analyzed. No adverse reaction related to the CO 2 challenge was reported. Carbogen inhalation induced a whole-brain BOLD signal increase, predominant in the gray matter. In patients, BOLD responses were decreased (0.36±0.13 %BOLD/mmHg in AD, 0.36±0.12 in MCI, 0.62±0.20 in controls). CVR impairments were diffuse but seemed predominant in posterior areas. Purpose: To determine the relationships between the precuneus atrophy (PCA) and common imaging parameters (hippocampal atrophy and microangiopathy) in patients with Alzheimer's disease (AD) and prodromal AD. Methods and Materials: 105 patients suspected of AD and prodromal AD according to the NINCDS-ADRDA and Petersen's criteria had a brain MRI, including T2-WI, T2*-WI, FLAIR, DWI, and 3D-GRE T1. Hippocampal atrophy was rated by two radiologists using the Scheltens's score (0-4), the microangiopahty with the Fazekas's score (0-3). The precuneus atrophy (PCA) was rated using an original visual score based on the antero-posterior enlargement of the cingular sulcus, ranging from -2 (anterior > posterior enlargement) to +3 (anterior < posterior enlargement).

The precuneus atrophy was defined by a PCA score ≥ 2. Thus, 2 groups were identified: PCA+ for patients with atrophy (n = 53), and PCA-for patients without (n = 52). No difference for gender and age was detected with: male/female (PCA+) = 24/29; M/F (PCA-) 23/29; age (PCA+) = 74.2 ± 8.0, age (PCA-) = 75.3 ± 10.3. Significant differences were detected for hippocampal atrophy with Scheltens's (PCA+) = 2.5 ± 0.9, Scheltens's (PCA-) = 2.9 ± 0.8 (p = 0.04), and for microangiopathy with Fazekas's (PCA+) = 0.6 ± 0.8, Fazekas's (PCA-) = 1.5 ± 0.9 (p < 0.001). The PCA score was negatively correlated with the severity of both the hippocampal atrophy (R Pearson = -0.26; p = 0.01), and the microangiopathy (R Pearson = -0.44; p < 0.001).

In patients with AD and prodromal AD, at least two distinct patterns can be identified based on the occurrence of the precuneus atrophy. When present, the microangiopathy is less severe, suggesting different pathophysiological mechanisms. This result may help to better distinguish mixed dementia from AD, and to constitute homogeneous populations in future therapeutics trials.

A S249 C D E F G H Results: Stuttering subjects showed increased ALFF in the left inferior frontal gyrus, superior, middle and inferior temporal gyri, ventral sensorimotor and Rolandic operculum cortices, and right superior and middle frontal gyri, and decreased ALFF in the bilateral paracentral lobules and supplementary motor areas, and the right occipital lobe. Stuttering subjects with and without family histories had differential distributions and patterns of abnormal ALFF relative to control subjects. Moreover, when directly compared with those without family histories, stuttering subjects with family histories showed significantly increased ALFF in the right auditory-and speech-related areas and decreased ALFF in the semantic-related brain areas. Conclusion: We suggest stuttering subjects with and without genetic propensity have differential alterations in brain spontaneous neuronal activity.

Functional MRI of regional brain responses to pleasant and unpleasant IAPS pictures: different networks F. Aldhafeeri, J. Alghamdi, V. Sluming; Liverpool/UK (F.M.Aldhafeeri@liverpool.ac.uk) Purpose: The main purpose of this study was to evaluate the efficiency of IAPS pictures and to investigate the neural substrates of positive and negative emotions Methods and Materials: We have scanned 14 healthy volunteers (8 males and 6 females) using Trio MRI scanner (Siemens, Germany). High-arousing pleasant pictures (N = 100) and high-arousing unpleasant pictures (N = 100) have been selected from IAPS. The experimental paradigm consisted of 10 blocks of emotionally-evocative visual stimuli in addition to neutral conditions in which blank screen with cross hair is presented. The stimulus duration was 21s, during which seven pleasant/unpleasant visual stimuli presented for 3s. Images were analysed using Brain Voyager QX software package (Brain Innovation, Maatricht, the Netherlands). We compared BOLD activation patterns to complex emotional pictures to determine if these stimuli would activate similar or distinct brain regions Results: Pleasant visual stimuli showed significant BOLD signal reduction in the visual cortex P ( Bonferroni corrected) < 0.05 with increased BOLD signal in the: PFC and posterior cingulated cortex. The unpleasnt visual stimuli demonstrated statistically significant, P ( Bonferroni corrected) < 0.05, BOLD signal reduction in the PFC, cingulate gyrus and cuneus. However, extensive increase in BOLD signal, P ( Bonferroni corrected) < 0.05, was observed in: visual cortex, parietal lobe, hippocampus, amygdale, temporal lobe and frontal lobe. Conclusion: Findings from our study prove the efficiency of IAPS pictures in promoting emotional centres in the brain. Although BOLD signal reduction is poorly understood, our findings demonstrate that positive and negative emotions can be distinguished using fMRI.

Functional connectivity MRI (FC-MRI) predicts linguistic performance after surgical treatment for drug-resistant epilepsy E. Pravata' 1 , M. Caulo 2 , C. Sestieri 2 , D. Mantini 2 , C. Briganti 2 , G. Colicchio 1 , A. Tartaro 2 , G. Romani 2 , C. Colosimo 1 ; 1 Rome/IT, 2 Chieti/IT (emanuele.pravata@gmail.com) Purpose: Functional connectivity (FC) of the language network (LN) is known to be reduced in the brain of epileptic patients and to positively correlate with linguistic performance. We investigated, in a group of patients with drug-resistant epilepsy, the existence of correlations between the pre-surgical LN FC and the post-operative linguistic outcome.

We retrospectively evaluated the FC of the LN of twelve right-handed patients with non-tumoral intractable epilepsy. FC was estimated between six common cortical regions for language in the frontal and temporal lobes of both hemispheres. Correlation between FC z-scores and post-operative (6+ months) neuropsychological follow-up vIQ scores (WAIS-R) was estimated using Pearson's test. Results: All patients experienced no seizures during the follow-up period. A negative correlation was found between the post-operative verbal-IQ scores and the preoperative mean inter-hemispheric FC (r = -0.56; p = 0.04) whereas no significant correlation was found with the mean intra-hemispheric FC (p > 0.05). Conclusion: A reduced FC of the LN in the dominant hemisphere of epileptic patients is correlated with a low verbal scores, as opposite a pre-operative high inter-hemispheric FC seems to predict a poorer linguistic outcome after surgery. Including an MRI assessment of the FC of the LN may help to better stratify the post-operative risk of language deficit.

were measured on coronal images of four continuous slices in which the trigeminal nerves from the root exit point to the distal cisternal segment were clearly shown. Results: A significant reduction in FA (t = -0.987, p < 0.001), as well as a significant increase in MD (t = 3.266, p = 0.002) were observed in the affected nerve of TN, while either health volunteers with NVC or normal subjects without NVC did not show any difference of FA and MD in both sides. Conclusion: The preliminary study of DTI demonstrates that trigeminal NVC may occur in asymptomatic patients, but functional abnormality is only detected in affected nerve with both TN and NVC. The application of DTI could be useful for diagnostic and therapeutic decisions in TN, and it could also help to understand its pathogenesis. Purpose: The aim of this study is to evaluate the optic nerve in multiple sclerosis patients with diffusion tensor imaging and to correlate fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values with visual acuity. Methods and Materials: Forty MS patients; 20 that had optic neuritis attack previously and 20 that had no optic neuritis attack before are included. As a control group 20 healthy volunteered patients without eye pathology were integrated. All patients were examined with 3 T MR equipment with 8-channel head coil with conventional sequences and DTI sequence FA and ADC values of both optic nerves were calculated in all patients. Findings were correlated with visual acuity. Results: In MS patients with optic neuritis attack previously, FA values of the affected optic nerves were significantly decreased compared with that of the control group (p < 0.001) and with clinically unaffected contralateral nerves (p < 0.01). Also significant correlation was found among FA reduction and visual acuity (p < 0.01). Conclusion: Axonal disruption of the affected nerve in MS patients can be demonstrated via DTI FA measurements.

Diffusion tensor imaging differences between prosocial and individualistic-oriented healthy subjects W. Van Hecke, G. Emonds, C.H. Declerck, C. Boone, P.M. Parizel; Antwerp/BE (wim.vanhecke@ua.ac.be) Purpose: To investigate white matter differences in subjects with a "prosocial" and an "individualistic" character orientation. These differences were examined using a voxel-based diffusion tensor imaging (DTI) approach. Methods and Materials: 28 healthy subjects, aged between 19 and 33 years, were recruited. Social value orientation was assessed by means of the "decomposed games measure". As a result, two groups were determined: 14 subjects with strong prosocial orientation, and 14 subjects with strong individualistic orientation. DTI data sets were acquired on a 3 T scanner with 64 diffusion directions, resolution of 2 x 2 x 2 mm 3 and a b-value of 900 s/mm 2 . All data sets were aligned to a DTI atlas with a non-rigid coregistration algorithm. After anisotropic smoothing with a kernel of 3 mm, FA values were compared between both subject groups in every voxel. Clusters of at least 10 voxels with a p < 0.005 were considered as significant. Results: FA differences between both groups were found in the inferior longitudinal fasciculus, thalamus, forceps minor, forceps major, fornix, superior longitudinal fasciculus, and corona radiata. Conclusion: Based on diffusion tensor measurements, there are differences in structural organisation of white matter fiber tracts between healthy subjects with a prosocial and individualistic character orientation.

Altered amplitude of low-frequency fluctuation in adult developmental stuttering C. Yu; Tianjin/CN (chunshuiyu@yahoo.cn) Purpose: Although developmental stuttering (DS) has been extensively studied using structural and task-based functional magnetic resonance imaging (fMRI), there is no study focused on the resting-state brain activity in this disorder or the effect of family history of stuttering on brain imaging alterations. Using the amplitude of low-frequency fluctuation (ALFF) as a measure of resting-state brain activity, we aim to investigate changes of the ALFF in adults with DS and the effect of family history of stuttering on this measure. Methods and Materials: Forty-nine adults with DS and 51 age-and gendermatched fluent controls were scanned using fMRI. In a voxel-wise manner, the ALFF was compared among stuttering subjects with and without family histories, and fluent controls. Purpose: In CT colonography chronic diverticulitis (CD) can mimic colorectal cancer (CRC) masses. In order to distinguish these two entities several findings were analysed. Methods and Materials: 318 symptomatic patients consecutively underwent CT colonography between June 2008 and August 2010. We included all patients with pathologic confirmation of CD (9) and CRC (26) on surgical specimens. The images were analysed by a consensus of two reviewers. The following findings were evaluated: length of the mass, diverticula included in the mass, growth pattern, luminal narrowing, presence of perilesional lymph nodes and mesenteric fat stranding. Results: There was a significant difference in median length of the mass between CD and CRC, 95 and 49 mm, respectively (p = 0.001). Diverticula included in the mass were visible in all patients with CD, but in none of the patients with CRC (p < 0.001). Furthermore, a curved growth pattern of the mass was noted in all but one patient with CD versus two out of 26 patients with CRC (p < 0.001). There were no significant differences between the groups with respect to median luminal narrowing (85 vs 77%, p = 0.5), presence of perilesional lymph nodes (p = 0.6) and presence of perilesional mesenteric fat stranding (p = 0.4).

To differentiate CD and CRC masses the length of the mass, inclusion of diverticula and growth pattern are important discriminating factors. Purpose: To investigate the value of double contrast-enhanced ultrasonography (DCEUS) in evaluation of perfusion status of the colorectal tumours, in which a kind of luminal contrast agent was used combined with SonoVue, and to assess the correlation between time-intensity (TIC) parameters and tumour angiogenesis.

Thirty-nine patients with colorectal tumours were examined preoperatively using DCEUS. Of the 39 tumours examined, 32 were adenocarcinoma and 7 were adenoma. The arrival time (AT), time to peak (TTP), peak intensity (PI) and area under the curve (AUC) in colorectal tumours were measured. The post-surgical specimens were assessed for microvessel density (MVD) and vessel endothelial growth factor (VEGF). Correlation between TIC parameters and the expression of VEGF or MVD were studied.

The mean values of AT, TTP, PI and AUC of the colorectal tumours were 14.32±11.36s, 30.61±18. 65s, 20.38±17.45 dB and 221.10±156.09, respectively. Both AUC and MVD were significantly higher in colorectal adenocarcinomas than in adenomas (all p < 0.05). There was a positive linear correlation between AUC and MVD in colorectal tumours and adenocarcinomas (r=0.686,p=0.0019; and r=0.736, p=0.0235, respectively). No correlation was found between VEGF and any ultrasound perfusion parameter. Purpose: Abnormal contrast enhancement on MRI is advocated as a biomarker for inflammation in colitis, although perfusion kinetics of normal colon are poorly described. The purpose was to assess enhancement patterns in normal colon and test for inter segmental differences. Methods and Materials: 11 patients (mean age 61) without prior history of inflammatory bowel disease underwent standard MR colonography followed by normal same day colonoscopy. Acquired sequences included a volumetric interpolated breath hold examination (VIBE) to encompass the whole colonic volume, performed three times prior to and every 30 seconds following intravenous Dotarem for 210 seconds. A region of interest was placed in each of 6 colonic segments and perfusion curves (averaged across all patients) were generated by plotting signal intensity [SI] (normalised to the maximum value in each segment) against time. Curve morphology was analysed and SI compared across segments using Mann Whitney statistics. Results: There was a distinct difference between the right colon (caecum and ascending) which showed typical type 2 curve morphology (peaking at 150 seconds followed by a plateau phase) and the left colon (rectum, sigmoid, descending and transverse) which demonstrated type 1 morphology (peak enhancement at 210 seconds, no plateau). There was a significant difference in normalised SI at 180 and 210 seconds between the right colon (mean 0.81) and left colon (mean 0.92), p = 0.02. Conclusion: Normal colon shows intersegmental differences in perfusion kinetics which should be considered when used as a biomarker for colonic inflammation. Purpose: Up to 40% of patients with colorectal cancer (CRC) develop metastases/ recurrence. Early and accurate identification is important because survival is improved when metastases/recurrences are curatively resected. The aim of this study is to determine which is the most accurate whole-body imaging (WBI) modality for detection of recurrent/metastatic disease during follow-up of patients with CRC. Methods and Materials: Medline and Embase were searched. Studies evaluating the accuracy of WBI with CT, PET-CT, MRI and/or PET for metastasis/recurrence detection during follow-up for CRC were included. Studies evaluating liver only or response to therapy were excluded. 2x2 contingency tables were extracted and summary receiver operator characteristics (sROC)-curves with area-under-thecurves (AUC) were constructed by a regression model. Subgroup analyses were performed to correct for heterogeneity. Results: 14 studies (published between 1997 and 2009) were included, some of which investigated multiple modalities. Grouping the articles according to modality, 12 investigated PET, 7 CT, 5 PET/CT and one MRI. Study populations comprised of patients with high suspicion for metastasis/recurrence. All studies used histopathology and clinical and radiological follow-up as reference method. PET-CT and PET both had an AUC of 0.94. CT had significantly lower performance: AUC 0.79, p < 0.05. After correction for sources of heterogeneity CT remained the modality with lowest diagnostic performance. Conclusion: PET-CT and PET have the highest diagnostic performance for the detection of metastasis/recurrence during follow-up of CRC patients with high suspicion for metastasis/recurrence. Particularly in these high-risk populations, CT has significantly lower diagnostic performance, partly due to its limitations in the detection of extrahepatic metastases. Purpose: Endometriosis infiltrating the bowel may simulate colorectal carcinoma in specific cases. The aim of this study was to evaluate MR DWI in differentiating endometriosis infiltrating the bowel from colorectal carcinoma. Methods and Materials: In 66 patients, MR DWI was added to the standard imaging protocol in patients visiting our outdoor MR clinic for the analysis of suspected or known deep infiltrating endometriosis (DIE). A total of 15 patients were diagnosed with DIE infiltrating the bowel. High b-value diffusion-weighted images were qualitatively assessed by two readers in consensus and compared with high b-value diffusion-weighted images in 15 patients evaluated for colorectal carcinoma. In addition, ADC values of lesions were calculated (b-values of 50, 400, 800 s/mm 2 ). Results: Endometriosis infiltrating the bowel showed low signal intensity on high b-value diffusion-weighted images, whereas colorectal carcinoma showed high signal intensity on high b-value diffusion-weighted images. Mean ADC value in endometriosis infiltrating the bowel (0.80 ± 0.06 x 10 -3 mm 2 /s) was significantly lower compared with mean ADC value in colorectal carcinoma (0.86 ± 0.06 x 10 -3 mm 2 /s), but with considerable overlap between ADC values. Conclusion: Only qualitative assessment of MR DWI may be valuable to differentiate between endometriosis infiltrating the bowel and colorectal carcinoma.

Rectal cancer: mucinous carcinoma on MR imaging indicates poor response to neoadjuvant chemoradiation K. Oberholzer, M. Menig, A. Hoetker, A. Kreft, T. Junginger, A. Heintz, H. Schmidberger; Mainz/DE (Andreas.Hoetker@uni-mainz.de) Purpose: To assess response of locally advanced rectal carcinoma to chemoradiation with regard to mucinous status and local tumour invasion found at pretherapeutic magnetic resonance imaging (MRI). Methods and Materials: Eighty-eight patients were included in this prospective study (advanced mrT3 and mrT4 carcinomas). Carcinomas were categorised by MRI as mucinous (mucin proportion > 50% within the tumour volume), and as nonmucinous. Patients received neoadjuvant chemoradiation consisting of 50.4 Gy (1.8 Gy/fraction) and 5-fluorouracil. Therapy response was assessed by comparing pretherapeutic MRI with histopathology of surgical specimens (minimum distance between outer tumour edge and circumferential resection margin = CRM, T and N category). value for T ≥ 3 of 17/17 (1.00, 95% CI, 0.81%-1.00), 6 of them (35%) being T4; no cases of IPF without ACD. Accuracy of ACD for T ≥ 3 was 40/44 (0.91, 95% CI 0.78-0.98), that of IPF 27/44 (0.61, 95% CI 0.46-0.76) (p < 0.001). Intra-and interobserver reproducibilities were excellent for ACD (k = 1; k = 0.93, respectively) and good for IPF (k = 0.69; k = 0.53). Conclusion: Both variables are highly reproducible. The presence of ACD outperformed that of IPF in predicting a T stage ≥3. The co-existence of ACD and IPF is highly suggestive of T ≥ 3, with about one-third of T4.

The impact of 18 F FDG-PET/CT in high and low risk primary rectal cancer C.J. Hunter 1 , A. Garant 2 , T. Vuong 3 , G. Artho 3 , R. Lisbona 3 , M. Abulafi 1 , G. Brown 4 ; 1 Croydon/UK, 2 Quebec, QC/CA, 3 Montreal, QC/CA, 4 Sutton/UK (chris_j_hunter@hotmail.com) Purpose: FDG-PET/CT is not currently recommended for routine staging of rectal cancer. We evaluated the diagnostic yield of FDG-PET/CT in MRI staged high and low risk rectal cancers. Methods and Materials: In a multicentre prospective observational trial, 236 patients with rectal cancer were recruited to a study evaluating FDG-PET/CT. All patients underwent MRI staging and were stratified into high and low risk (high risk: extramural venous invasion, extramural spread > 5 mm or T4, pelvic sidewall lymph nodes, N2 or involved circumferential resection margin). The primary aim was to evaluate detection rates of metastatic disease using FDG-PET/CT. The secondary aim was to determine the diagnostic yield of FDG-PET/CT according to MRI risk features. Results: Imaging data were available for 230/236 (97.5%) patients. Overall, FDG-PET/CT did not significantly increase detection of metastatic disease compared with computed tomography for the whole group 33/230 (14.4%, 95%CI 10.4-19.5%) versus 22/230 (9.6%, 95%CI 6.4-14.1%). FDG-PET/CT had a significantly higher diagnostic yield for distant metastases in MRI high risk rectal cancers, (28/136, 20.6%, 95%CI 14.7-28.2) versus (5/94, 5.3%, 95%CI 2.3-11.9, OR 6.4, p = 0.001) for low risk patients and also identified significantly more metastases not visible on conventional CT (OR 4.6, 95%CI 1.3-16.2, p = 0.001). In patients with high risk features on MRI the rate of distant metastases uniquely identified on FDG-PET/ CT was 13.2% (95%CI 8.5-20.0). Conclusion: It is possible to identify a high risk group using rectal MRI features with a high diagnostic yield on PET/CT. Patients in this group have significantly more metastases identified on PET/CT than CT alone.

Influence of FDG PET-CT on management decisions in patients with potentially operable metastatic colorectal cancer A.F. Scarsbrook, F.U. Chowdhury, R.H. Briggs; Leeds/UK (andrew.scarsbrook@leedsth.nhs.uk) Purpose: To assess the impact of FDG PET-CT in patients with potentially resectable metastatic colorectal carcinoma. Methods and Materials: All patients with potentially resectable metastatic colorectal undergoing FDG PET-CT between November 2007 and March 2009 were retrospectively analysed. Findings on PET-CT were compared to conventional imaging and impact on subsequent patient management was evaluated. Clinical impact was divided into: Major: Detection of occult disease or further characterisation of indeterminate lesion (s); Minor: Confirmation of suspected metastases; No impact. Outcome data was obtained from clinical and radiological databases. Results: 102 patients underwent PET-CT during the study period (mean age 67 years, range 27-85) including 94 patients with liver metastases, 5 patients with isolated lung metastases and 3 patients with limited peritoneal disease. In 31 patients (30%) PET-CT had a major impact on subsequent management. 16 patients had inoperable disease confirmed, previously unsuspected metastatic disease was detected in 9 patients, 3 patients had occult second primary tumours and 3 patients were down-staged. PET-CT had a minor impact in 12 patients (12 %) and no impact in 59 cases (58 %). The introduction of PET-CT into the pre-operative imaging algorithm had a statistically significant impact on patient management (p < 0.05, Chi squared test). At 18 month follow-up 17 patients remained radiologically disease free after liver (15/34) or lung (2/3) surgery. Conclusion: FDG PET-CT has a valuable role in selected patients with metastatic colorectal carcinoma by improving staging accuracy and characterising indeterminate lesions and can help triage patients to the most appropriate treatment.

A D E F G H (p > 0.05). At both field strengths, there was no correlation of the R2*-values of cortex and medulla with patients age (r = 0.02) or with the eGFR (r = 0.007). No differences in R2* were found between the different CKD-stages (medullary R2* range 27-33 s -1 at 3 T and 21-22 s -1 at 1.5 T, p > 0.05). Conclusion: Static BOLD-MRI measurements fail to discriminate between patients with various CKD-stages most likely due to superimposed external factors such as degree of hydration, medications and renal perfusion. For valuable clinical application of renal BOLD-MRI, dynamic pharmacodynamic studies with intraindividual assessment of signal variation seem mandatory

Purpose: The aim of this study was to assess the feasibility of dynamic contrastenhanced renal MRI at 7 Tesla with optimisation and implementation of a dedicated examination protocol. Methods and Materials: 12 healthy volunteers were examined on a 7 Tesla whole-body MR-system (Magnetom 7T, Siemens) utilising a custom-built 8-channel transmit/receive radiofrequency body coil suitable for RF-shimming. The following non-enhanced sequences were obtained: 1) T1w fs 2D FLASH, 2) T1w 2D in-andopposed phase, 3) True-FISP, 4) T2w TSE. For dynamic imaging, gadobutrol was injected intravenously and T1w 3D FLASH images were obtained pre-contrast and at 20, 70, and 120 sec delay. Qualitative image analysis was performed using a three-point scale (1 = poor, 2 = moderate, 3 = good quality). SNR and CNR of the renal cortex/medulla were measured for all sequences. Results: 2D FLASH imaging showed best overall image quality with an average score of 2.57 for overall image quality and being least prone to artefacts. Nonenhanced 2D and 3D T1w imaging revealed high signal intensity of renal vasculature (arteries: 2.7). Quantitative analysis showed highest CNR between cortex and medulla for arterial phase 3D FLASH imaging (CNR = 12.2), providing a statistically significant difference to all other sequences except for the in-and-opposed-phase and the fat-saturated 2D FLASH sequence. Conversely, equilibrium phase FLASH imaging yielded the weakest CNR score of 3.6. Conclusion: Our results demonstrate the feasibility and diagnostic potential of 7 Tesla kidney imaging, yielding an excellent delineation of anatomical details and non-enhanced vasculature with the potential of a non-enhanced angiographic application. Purpose: To assess the value of arterial spin labelling (ASL) perfusion magnetic resonance imaging (MRI) for characterisation of renal masses using histopathological findings as the reference standard. Methods and Materials: Forty-two patients (34 males, 8 females, mean age 60.1 ± 11.3 years) with suspected renal masses were prospectively included in this study. Pseudocontinous ASL images were acquired though the centre of the mass on a 1.5 T clinical MR scanner. Region-of-interest (ROI) analyses were performed to quantify the perfusion of the entire tumour (P whole ), the solid portion of the tumour (P solid ), as well as the maximum tumour perfusion (P max ). Results: Histopathologic analysis was available in thirty-four patients that were included in the final analysis. P whole and P solid of papillary RCC (32.5 ± 13.7 and 41.9 ± 18.4 ml/min/100 g, respectively) were significantly lower and P whole and P solid of oncocytomas (307.9 ± 74.3 ml/min/100 g) were significantly higher as compared with clear cell (167.5 ± 57.8 and 189.1 ± 47.7, respectively), chromophobe (161.5 ± 71.4) and unclassified (211.1 ± 36.4 ml/min/100 g) RCC (p < 0.01). P max of oncocytoma (488.8 ± 119.7 ml/min/100 g) was significantly higher than P max of papillary (111.3 ± 60.4 ml/min/100 g), unclassified (277.8 ± 78.3 ml/min/100 g) and chromophobe (280.3 ± 41.8 ml/min/100 g) RCC, but did not differ significantly from P max of clear cell RCC (332.2 ± 106.9 ml/min/100 g). Purpose: To assess whether static, single measurement renal blood oxygen leveldependent (BOLD)-MRI allows characterisation of the renal function.

In this IRB-approved study 246 consecutive patients (164 at 1.5 T, 82 at 3 T) underwent transversal BOLD-measurements with a GRE with 12 echo times (TR = 106 ms, TE = 5-33 ms, 1.4 x 1.4 x 4 mm³ spatial resolution). Using OsiriX 3.7 T2*-maps were generated in which ROIs were placed in the medulla and cortex of the kidneys. R2* was calculated as R2* = 1/T2*. Individual R2*-values were correlated to the patients' MDRD-eGFR, age and gender for 1.5 T and 3.0 T seperately. Patients were classified according to the National Kidney Foundation system of chronic kidney disease (CKD). Results: Measured R2* values at 3.0 T were on average 50% higher than at 1.5 T (p < 0.0001). At both field strengths no differences in R2* of cortex and medulla were found between male (R2* medulla 22.6s -1 at 1.5 T) and female (22.3 s -1 ) patients

Results: DW-MRI rated 23/52 kidneys as normal, 29/52 abnormal. Histology/ follow-up attributed pathologies in 41 kidneys. Among abnormal kidneys on DW-MRI (n = 29) only 5 showed also reduced/absent corticomedullary differentiation on morphology. Pathologies missed on morphological images were rejection (n = 5), vascular abnormalities (n = 5), reflux (n = 4), cyclosporine-A-toxicity (n = 3), chronic nephritis (n = 4), glomerulonephritis (n = 1), pyelonephritis (n = 1). In 12 kidneys DW-MRI/morphological imaging missed the pathology. This corresponds to a sensitivity of 72.5%, specificity of 100%, positive-predictive-value of 100%, negative-predictive-value of 47.8% and diagnostic-accuracy of 79% for DW-MRI. Pathologies missed on DW-MRI/morphology were cyclosporine-A-toxicity (n = 6), vascular abnormalities (n = 4), glomerulonephritis (n = 2) and rejection (n = 4). In 25 kidneys a total of 78 focal pathological lesions were only visible on DW-MRI with mean ADC-values beyond 1.88 x 10 -3 mm/s 2 . Renal parenchyma of abnormal kidneys without focal lesions on DW-MRI (n = 7) had mean ADC-values beyond a threshold-value of 2.00 x 10 -3 mm/s2 (mean: 1.83 +/ -0.13 x 10 -3 mm/s 2 ) indicating diffuse parenchymal injury whereas normal kidneys had a value of 2.03 +/ -0.13 x 10 -3 mm/s 2 . Conclusion: DW-MRI improves the diagnosis of focal and parenchymal pathologies in transplanted kidneys. Measurement of ADC-values may be a helpful diagnostic tool.

Comparison of three different in-phase and out-of-phase gradient-echo imaging techniques for the characterisation of adrenal lesions at 1.5 T: preliminary observations M. Ramalho 1 , R. de Campos 2 , V. Heredia 2 , R. Azevedo 2 , R. Semelka 2 ; 1 Almada/PT, 2 Chapel Hill, NC/US (jmpmramalho@gmail.com) Purpose: To compare the performance of three-dimensional (3D) gradient recalledecho (GRE) and magnetization-prepared (MP) GRE in-phase (IP) and out-of-phase (OP) imaging with standard two-dimensional (2D) GRE, for the characterisation of adrenal lesions, using previously reported quantitative evaluation methods. Methods and Materials: Forty-three consecutive subjects (21 males and 22 females; mean age, 59.5 ± 11.4 years) with 49 adrenal lesions constituted our study population. Adrenal signal intensity (SI) index, and adrenal-to-spleen, adrenal-toliver, and adrenal-to-muscle SI ratios were calculated for each lesion and compared between the three techniques by using independent samples student-t test. The area under the receiver operating characteristic (ROC) curve (AUC) for each evaluation method was determined and comparisons of independent ROC curves were performed between all sequences. Results: For all sequences, the mean adrenal SI index and SI ratios were significantly different between adenomas and nonadenomas (p < 0.002). For the 3D-GRE and MP-GRE techniques, the adrenal SI index and the modified adrenal-to-spleen ratio showed the higher AUCs, respectively, although not statistically significant. Different thresholds were recommended for discriminating adenomas from nonadenomas among the three techniques. Conclusion: 3D-GRE and MPGRE IP/OP sequences can be employed for adrenal lesion characterisation with comparable results with the reference standard of 2D-GRE. Different recommended thresholds should be selected accordingly to the used IP/OP technique and for the various evaluation methods.

Quantification of viscoelastic properties of the kidney using supersonic shear wave imaging in an experimental rat model of fibrosis induced by L-NAME M. Derieppe 1 , Y. Delmas 2 , C. Deminière 2 , S. Placier 1 , J.-L. Genisson 1 , M. Tanter 1 , C. Combe 2 , N. Grenier 2 ; 1 Paris/FR, 2 Bordeaux/FR (nicolas.grenier@chu-bordeaux.fr) Purpose: Chronic renal disease is characterised by progressive development of glomerulosclerosis and/or interstitial fibrosis. Renal fibrosis can be assessed by kidney biopsy only. The aim of our study was to validate the use of ultrasonic shearwave elastography in an experimental rat model of glomerulosclerosis induced by chronic administration of L-NAME (inhibitor of nitric oxyde synthesis). Methods and Materials: Four groups of rats were studied: a control group (n = 8), a group before and after 4 weeks of L-NAME administration (20 mg/kg/day) (H4, n = 8), a group before and after 6 weeks of L-NAME (H6, n = 15), and a group studied longitudinally (n = 9) before, after 4 weeks (L4) and after 7 weeks (L7) of administration. Elasticity of renal cortex was quantified through the shear modulus (μ) using a 8 MHz probe and ultrafast image sampling. After killing, urines were analysed for dosage of protein/creatinine ratio. Kidneys were removed for histological quantification of fibrosis.

Diffusion tensor imaging of the kidneys: influence of b-value and number of encoding directions N.C. Chuck 1 , O. Donati 1 , G. Steidle 2 , J. Hodler 1 , D. Nanz 1 , A. Boss 1 ; 1 Zurich/CH, 2 Tübingen/ DE (natalie.chuck@usz.ch) Purpose: The purpose of this study was to assess the dependence of image quality and diffusion tensor imaging (DTI) parameters (fractional anisotropy FA, mean diffusivity MD) on b-value selection and number of encoding directions in DTI of the kidney. Methods and Materials: Eight healthy volunteers consented to an examination on a 1.5 T whole-body MR scanner (Philips Achieva). After localisation, oblique coronal DTI data sets of the kidneys were acquired with systematic variation of b-value (300, 500 and 700 s/mm2) and number of diffusion-encoding directions (6, 15 and 32). A respiratory triggered (breathing pad) single-shot echo planar sequence (TR/ TE 267 ms/67 ms, parallel imaging SENSE factor 4, matrix size 124 x 128, scan time 2-6 minutes) was used for DTI data acquisition. Quality of FA and MD maps was assessed on a 4-point score by 2 readers. Results: In the visual read, the parametrical DTI maps obtained with all evaluated diffusion encoding schemes were attested good quality. Best scores were reached for encoding schemes with 32 encoding directions and b-values of either 500 or 700 s/mm2. FA values ranged between 0.42 and 0.46 (MD 1.5 -2.3 x10 (-3) mm (2)/s) for renal medulla and between 0.21 and 0.28 (MD 1.9 -2.9) for renal cortex. Conclusion: DTI of the kidneys with high b-values is feasible using parallel imaging techniques. Best image quality was achieved with 32 encoding directions, which, however, also required longest acquisition times.

Diffusion tensor imaging, tractography and arterial spin labelling in transplanted kidneys K. Hueper, M. Gutberlet, T. Rodt, W. Gwinner, F. Lehner, D. Hartung, M. Galanski; Hannover/DE (hueper.katja@mh-hannover.de) Purpose: Monitoring of kidney transplanted patients is important to early detect allograft damage and finally to ensure long-term allograft survival. In this study, we evaluated MR diffusion tensor imaging (DTI) and arterial spin labelling (ASL) for the detection of acute and chronic changes of allograft microstructure and perfusion. Methods and Materials: 10 kidney transplanted patients with differing degrees of allograft dysfunction and 8 healthy volunteers were examined using a fat-saturated echo-planar DTI-sequence (1.5 T MAGNETOM Avanto, 6 diffusion directions, b = 0.600 s/mm²) and an ASL-sequence (FAIR trueFISP, TI = 1,200 ms). Mean ADC and fractional anisotropy (FA) as well as specific perfusion rate were calculated and compared between healthy and transplanted kidneys. Tractography was performed for visualisation of diffusion properties. Results: The ADC in the cortex and in the medulla was lower in transplanted than in healthy kidneys (p < 0.01). FA in the cortex and in the medulla of healthy kidneys were 0.169 ± 0.009 and 0.417 ± 0.007. In allografts these values were significantly reduced to 0.122 ± 0.008 and 0.264 ± 0.014 (p < 0.001) and the perfusion was significantly decreased as well (p < 0.01). Tractography was able to visualise renal microstructure and functional units and thus allowed for demonstration of pathological changes in the allografts. Conclusion: The feasibility of DTI and ASL for assessment of renal transplants was demonstrated and differences in diffusion and perfusion properties were shown. Further investigations, e.g. comparison to kidney biopsy results and differentiation of subgroups are intended to evaluate DTI and ASL for routine clinical application. Examination of the high-temporal signal-intensity-time-curves allows the extraction of 4 different pharmacokinetic biomarkers (K trans ,k ep ,v e ,v p ), 5 different descriptive biomarkers (Wash-in, Wash-out, AUC, TTP, Peak enh ) and an absolute measurement of the dynamic change in the transverse relaxation rate (R 2 * ). In addition, the high-spatial resolution images yield an adequate assessment of tumour morphology. Conclusion: The novel split-dynamic sequence generates high-spatial resolution and high-temporal resolution images during a single CA-injection, providing both morphological features and contrast kinetics simultaneously.

Potential of diffusion MRI as a biomarker of low-risk DCIS M. Iima 1 , D. Le Bihan 1 , R. Okumura 2 , T. Okada 1 , K. Fujimoto 1 , S. Kanao 1 , S. Tanaka 1 , K. Togashi 1 ; 1 Kyoto/JP, 2 Osaka/JP (mamiiima1@gmail.com) Purpose: To evaluate the potential of breast diffusion MRI as a biomarker of lowgrade ductal carcinoma in situ (DCIS).

This study was a retrospective analysis of 19 patients (mean age 55 years) who underwent breast MRI examinations at 1.5 T, and informed consent was waived. MRI scans were performed with a four-channel breast array coil and included contrast-enhanced (Gadoteridol) T1-weighted and diffusion-weighted images (DWI) with b values of 0 and 1,000 s/mm 2 . All patients underwent surgical resection and were diagnosed with DCIS based on histopathology (6 low-grade; 7 intermediate-grade; 5 high-grade). Results: The mean apparent diffusion coefficient (ADC) values (and their 95% confidence intervals) were 1.43 (1.32-1.55)×10 -3 mm²/s, 1.28 (1.16-1.41)×10 -3 mm²/s, 1.20 (1.07-1.32)×10 -3 mm²/s and 2.10 (1.96-2.22)×10 -3 mm²/s, for low-, intermediate-, high-grade DCIS and normal breast tissues, respectively. The ADC value of high-grade DCIS was significantly lower than that of low-grade DCIS (p = 0.01), and there was a significant negative trend between ADC and tumour grade (p < 0.01). A receiver operating characteristic (ROC) analysis gave an ADC threshold of 1.3×10 -3 mm²/s for the diagnosis of high-grade DCIS (81% sensitivity, 62% specificity). Patients whose ADC values were above this threshold in all ROIs were identified as low grade with 100% specificity. Conclusion: These results suggest that quantitative DWI could be used to identify patients with low-grade, low-risk DCIS with very high specificity. Such patients could be spared from unnecessary invasive approaches, such as lumpectomy or mastectomy.

Breast MRI at 3 T: a pilot study to compare unenhanced MRI (DWI combined with a T2 IDEAL sequence) with contrast-enhanced MRI for the assessment of response to neo-adjuvant chemotherapy F. Pediconi, M. Telesca, V. Casali, F. Vasselli, E. Miglio, L. Ballesio, C. Catalano, R. Passariello; Rome/ IT (federica.pediconi@uniroma1.it) Purpose: To determine sensitivity and specificity of contrast agent free approach at 3 T using only diffusion weighted imaging (DWI) combined with T2w IDEAL images (ueMRI) and to compare findings with contrast-enhanced MRI (CE-MRI) for assessment of response to neoadjuvant chemotherapy. Methods and Materials: Twelve consecutive patients undergoing neoadjuvant chemotherapy for biopsy-proven breast cancer (> 2 cm lesions) underwent MR at 3 T prior to chemotherapy (every two cycles) and before surgery. MRI protocol comprised FSE-T2w-IDEAL and DWI sequences, and a Vibrant-3D-T1w sequence acquired before and after administration of gadobenate dimeglumine (MultiHance). Tumour response was classified using RECIST criteria based on tumour size at MR imaging. Tumor response at MR imaging after chemotherapy was compared with histological specimens obtained after surgery.Two blinded observers rated images from ueMRI first and then images from CE-MRI for the evaluation of tumour response. Lesion size, ADC values and FSE IDEAL descriptors were assessed. Results: Observers 1 and 2 determined sensitivities of 92% and 84%, respectively, for ueMRI and 100% (both observers) for CE-MRI. Similar specificity was noted by the two observers (82% for ueMRI; 86% for CE-MRI). Differences between methods and between observers were not statistically significant (P ≥ 0.09). Similarly, no significant differences in lesion size measurements were noted between sequences. Visibility of malignant lesions was considerably worse using ueMRI. The mean tumour ADC values for responders and non-responders were significantly different after the first II cycles of treatment. Conclusion: Combined DWI (for tumour density) and T2w IDEAL (for lesion morphology and size) sequences permit high sensitivity and specificity in evaluating tumour response to neoadjuvant chemotherapy although better performance is achieved with CE-MRI.ADC values may be useful for differentiating responders from nonresponders earlier during treatment.

Diseased rats showed an increased urinary protein/creatinine ratio (mean = 6.7 mg/mg) but moderate glomerulosclerosis at histological examination. Cortical elasticity was 4.55 ± 1.54 kPa in control kidneys. Compared to controls, it was increased in all pathological groups, except for H6: H4 = 6.87 ± 2.42 kPa (p = 0.046); H6 = 4.79 ± 1.08 kPa (p = 0.70); L4 = 7.18 ± 1.51 kPa (p = 0.012); L7 = 7.04 ± 1.55 kPa (p = 0.012). However, considering all animals, no correlation was observed between biological data, degree of glomerulosclerosis and level of cortical elasticity. Conclusion: Quantification of level of fibrosis by ultrasonic shearwave elastography is feasible. However, more experience is necessary to understand all the factors responsible for the increase of cortical elasticity. Using MIP reconstructions (FLASH sequences with 0.1 mmol/kg of gadobenate dimeglumine), two residents independently assigned a 5-point score at number, length, and diameter of vessels. In the presence of a mass-like lesion, -2 and -1 indicated marked and mild contralateral asymmetry; 0 symmetry; +1 and +2 mild and marked ipsilateral asymmetry. Asymmetry in negative examinations was assigned +1 or +2. Image analysis was repeated for axial, sagittal, and coronal MIP reconstructions. The score sum over number, length, and diameter of vessels and over the 3 planes was considered as VMAI. Histology/1-year follow-up served as reference standard. Spearman, ROC, Kruskal-Wallis, and reproducibility statistics. Results: 87 (50%) examinations showed malignant lesions (19 DCIS) and 25 (14%) benign lesions; the remaining 61 (36%) were negative. For the more experienced resident, ROC-AUC was 0.745: the VMAI threshold corresponding to a fixed specificity of 93% was 7.5 (sensitivity 38%; PPV 85%; NPV 60%). For invasive cancers, the median VMAI was 1 for G1 lesions, 5 for G2, and 12 for G3 (p < 0.001). The VMAI correlated with the diameter of malignant lesions (r = 0.507, p < 0.001). In only 6 (3%) cases the VMAI difference between the two residents was > 3. Conclusion: Our index is reproducible and correlates with both the diameter of malignant lesions and the histological grade of invasive cancers. Lesions with VMAI > 7.5 have a high probability (85%) of being malignant.

Split dynamic magnetic resonance imaging: a novel method for improved morphologic and functional assessment of breast tumours K.-I. Gjesdal 1 , E. Grøvik 2 ; 1 Ålesund/NO, 2 Oslo/ NO (k.i.gjesdal@medisin.uio.no) Purpose: One of the major issues concerning dynamic contrast-enhanced MRI of the breast is the application of either high-spatial resolution for the assessment of tumour morphology or high-temporal resolution for an adequate evaluation of contrast kinetics. These features can only be obtained separately using conventional MRI-techniques, with two injections of contrast agent (CA) causing higher cost for health care and inconvenience for the patients. There is a universal agreement that an optimal combination of both features may improve the diagnostic accuracy in the assessment of breast tumours. This work presents a novel method in which a high-spatial and a high-temporal resolution sequence are applied in an alternating fashion during a single CA injection, allowing a comprehensive assessment of tumour morphology and contrast kinetics, simultaneously.

The high-temporal resolution 3D T1 doubleecho EPI-sequence had the following key parameters: TR/TE1/TE2/FA = 42 ms/5.5 ms/23 ms/28°, temporal resolution = 2.8 s. The high-spatial resolution 3D T1-THRIVE sequence used the following parameters: TR/TE/FA = 5.4 ms/2.7 ms/10°, voxel size = 0.85*0.85*1.25 mm 3 , number of slices = 125, temporal resolution = 63s. Both sequences were performed with fat suppression.

A S255 C D E F G H response to NAC was assessed with change of tumour size on breast MRI using RECIST. The tumour response to NAC was defined as a > 50% decline of peak standard uptake value (SUV) on PET-CT based on EORTC. MRI and PET-CT findings were correlated with pathologic results, and their association with biologic markers (ER, PR, and Her2) was evaluated. Results: Complete, partial and no response was seen in 9 (45%), 8 (40%) and 3 (15%) patients on breast MRI, and 9 (45%), 10 (50%) and 1 (5%) patients on PET-CT scan, respectively. MRI was more accurate in predicting tumour response to neoadjuvant chemotherapy than PET-CT (55%). MRI was more accurate than PET-CT scan in cases of HER2-positive (100% vs 80%) and ER-positive (57% vs. 45%) breast cancer, but identical in cases of triple negative cases (60% predictive on both imaging modalities). However, these differences were not significant statistically (p > 0.05). Conclusion: Although there was no statistical significance, breast MRI for predicting tumour response to NAC could be more accurate than PET-CT scan, especially in cases of Her2-and ER-positive groups.

Diagnostic performance of breast MR imaging according to patients' characteristics and radiologists: before and after using computer-aided evaluation system for breast MR imaging B. Seo 1 , E. Pisano 2 , E. Cole 2 , D. Zeng 2 , K. Cho 3 , C. Kuzmiak 2 ; 1 Ansan/KR, 2 Chapel Hill, NC/US, 3 Seoul/KR (seoboky@korea.ac.kr) Purpose: To compare diagnostic performance of breast MR imaging before and after the introduction of a breast MRI computer-aided evaluation (CAD) system in an academic diagnostic breast imaging clinic. Methods and Materials: A total of 445 breast lesions (158 malignant, 287 benign) in 390 patients who underwent contrast-enhanced breast MR imaging at our institution between 2005 and 2008 were included. 258 of the 445 breast lesions (58%) were assessed using a breast MRI CAD system. One of four radiologists assessed morphologic findings and kinetics. In lesions that were evaluated with the CAD system, the lesion was considered positive for malignancy if a lesion had significant enhancement using the 100% threshold or washout enhancement pattern. We compared diagnostic performance of radiologists before and after the introduction of the CAD system according to patient characteristics and interpreting radiologist. Statistically, univariate and multivariate analyses were performed to compare diagnostic performance before and after introduction of CAD system. Results: The specificity increased after the introduction of the CAD system but not significantly (p = 0.798). There was no statistical difference in specificity before and after introduction of CAD controlling for each variable (p = 0.618). There was a decrease in sensitivity that approached significance after the introduction of the breast MRI CAD system but only for heterogeneously dense breasts (p = 0.0528). The difference in the sensitivity before and after introduction of the CAD system was not significant (p = 0.947). Conclusion: Based on our results, the use of a CAD system for breast MRI does not significantly improve diagnostic performance.

Potential of CAD to predict occurrence of distant metastasis in primary breast cancer M. Dietzel, P.A.T. Baltzer, R. Zoubi, C. Jerowski, M. Gajda, I.B. Runnebaum, W.A. Kaiser; Jena/DE Purpose: Breast-MRI is highly sensitive for the detection of breast cancer. Furthermore, it provides functional information on tissue vascularisation. Computerassisted diagnosis (CAD) is feasible to semi-automatically and to quantitatively assess imaging data. This breast-MRI investigation was performed to identify the potential of CAD to predict occurrence of distant metastasis in primary breast cancer. Methods and Materials: We enrolled 59 consecutive patients with primary breast cancer (data acquisition: 01.01.2004 to 31.12.2004 ). All received one staging breast-MRI examination at 1.5 Tesla according to international guidelines before therapy was initiated. Follow-up-interval and occurrence of distant metastasis was documented. For breast-MRI analysis dedicated commercial software was used (Brevis, cooperation of Siemens & Mevis, Erlangen/Bremen, Germany). It allows semiautomatic quantitative measurement of breast-MRI enhancement patterns (initial-enhancement, wash-out, peak-enhancement, time to peak enhancement [TTP] ). Results were evaluated to identify significant and independent parameters to predict occurrence of distant metastasis. Significant parameters were further analysed by Kaplan-Meier statistics. Results: Median follow-up time was 1560 days. 6 patients were diagnosed with metachronous distant metastasis during follow-up. Wash-out above 40% could be identified as significant and independent predictor for occurrence of distant me-

Value of MRI and diffusion-weighted imaging in evaluating early response of patients with locally advanced breast cancer undergoing recombinant human endostatin chemotherapy J.Q. Xu, Q.X. Jia, Y. Huan; Xi'an/ CN (junqingxu@126.com) Purpose: To evaluate the role and the performance of MR and diffusion weighted imaging (DWI) for assessing the response of patients to recombinant human endostatin (endostar, China) chemotherapy in local advanced breast cancer (LABC). Methods and Materials: 96 women with LABC (99 lesions) underwent routine MRI and DWI was carried out at 3.0 T before treatment and after three cycles of endostar chemotherapy. For each patient, analysis of the percentage change in ADC, volume and diameter after each cycle of chemotherapy were compared between the baseline and post-chemotherapy to predict the response to endostar. The residual tumour volumes were obtained using 3D maximum intensity projections (MIP) of dynamic enhanced scan and DWI map, respectively, and were compared with pathological findings to assess the accuracy of DWI in detecting and measuring residual tumour. All results were proved by histopathological findings. Results: There were 83 lesions responding to endostar, while 16 non-responders. After the third cycle, the sensitivity for differentiating responders from non-responders was 78% for volume and 62% for ADC. In the response group, there was a significant difference in ADC value between prior to and post-endostar (P < 0.005), while no significant differences were found in non-responders. A cut-off value for ADC, volume and diameter was obtained after the third cycles of endostar.

Conclusion: DWI appears to provide functional information regarding changes in ADC value of tumours due to endostar chemotherapy. DWI may be useful in monitoring the pathological response and evaluating the residual tumour after endostar chemotherapy. Purpose: To retrospectively evaluate changes in oedema in patients with advanced breast cancer undergoing neoadjuvant chemotherapy (NAC) and to determine whether reduction or absence of oedema can be used to predict final clinic-pathologic response. Methods and Materials: From January 2005 to 2010, 146 patients with advanced breast cancer associated with oedema were identified. Dynamic MR imaging (MRI) was performed in all patients before and during neoadjuvant chemotherapy. Results were compared with findings from pathologic assessment of surgical samples obtained after NAC. The breast MRI protocol comprised a precontrast T2w STIR sequence and a dynamic 3D T1w FLASH sequence acquired before and at 2-minute intervals after administration of gadobenate dimeglumine (MultiHance) at 0.1 mmol/ kg bodyweight. Patients were divided into 2 groups according to the oedema changes observed during NAC: patients with oedema that appeared reduced or absent and patients with no apparent variation of the oedema. Results: Of the 146 patients evaluated, 120 had invasive ductal cancer, 18 had lobular cancer, and 8 had mixed ductal/lobular cancer. 114/146 (78.1%) patients were deemed responders to NAC; in these patients, oedema was reduced or absent in 79/114 (69.3%) cases. The remaining 32 (21.9%) patients were considered nonresponders to NAC; in these patients, oedema was reduced or absent in 8/32 (25%) cases. There was a good correlation between the reduction of oedema on MR imaging during NAC and response to NAC (p < 0.003). How was correlation determined. Conclusion: Evaluation of oedema detected on MR imaging is a useful adjunct to findings from conventional T1w dynamic imaging to accurately assess the pathologic response of breast cancer after NAC. (clemens.cyran@med.uni-muenchen.de) Purpose: To investigate the effects of the multikinase inhibitor sorafenib on experimental prostate carcinomas in rats with dynamic contrast-enhanced (DCE-) MRI quantifications of tumour perfusion and endothelial permeability. Non-invasive MRI results were validated by immunohistochemistry.

Copenhagen rats (n = 20) implanted with subcutaneous prostate carcinoma allografts (MLLB-2) were imaged on days 0 and 7 using DCE-MRI at 3 T enhanced with gadobutrol. The treatment group (n = 10) received daily applications of 10 mg/kg/bodyweight Sorafenib (Nexavar ® , Bayer, Leverkusen, Germany) via gavage. Quantitative measurements of tumour perfusion (ml/100 ml/ min) and endothelial permeability (ml/100 ml/min) were calculated based on a two-compartment kinetic model. Tumours were excised on day 7 to undergo immunohistochemical staining for tumour vascularity (RECA-1), cell proliferation (TUNEL) and apoptosis (Ki-67). Results: Tumour perfusion in sorafenib-treated prostate carcinoma allografts declined significantly from days 0 to 7 (47.9 ± 36.9 vs. 24.4 ± 18.5 ml/100 ml/ min, p < 0.05). In the control group, tumour perfusion increased significantly from days 0 to 7 (37.6 ± 12.3 vs. 49.8 ± 15.0 ml/100 ml/min, p < 0.05). No significant effect on endothelial permeability was observed either in the therapy or in the control group (p > 0.05). Immunohistochemical measurements revealed significantly lower tumour vascularity in the therapy than in the control group (RECA-1 5.1 ± 1.9 vs. 23.1 ± 7.7, p < 0.05). In the sorafenib-treated tumours, significantly more apoptotic cells (TUNEL 427 ± 283 vs. 218 ± 312, p < 0.05) and significantly less proliferating cells (Ki-67 847 ± 307 vs. 1692 ± 469, p < 0.05) were observed than in the control group. Conclusion: Tumour perfusion quantified by gadobutrol-enhanced DCE-MRI can be applied as non-invasive surrogate parameter for monitoring the anti-angiogenic, anti-proliferative and pro-apoptotic effects of sorafenib on prostate carcinoma allografts as validated by immunohistochemistry.

tastasis (negative predictive value 100%, sensitivity 100%). Using this parameter Kaplan-Meier statistics were performed successfully (Logrank-test: P < 0.001). Conclusion: Semi-automatic analysis of breast MRI using CAD provides quantitative data, which can be used to predict occurrence of distant metastasis in primary breast cancer. Such data should be further investigated regarding risk stratification of patients and potential impact on the therapeutic approach.

Breast MRI for the prediction of lymph node metastases: systematic approach using individual descriptors and a dedicated decision tree To identify useful combinations of descriptors for the prediction of LNM a CHAID decision-tree was applied. Results: 7 of 17 descriptors were significantly associated with LNM. The most accurate were "skin thickening" (P < 0.001; DOR 5.9) and "internal enhancement" (P < 0.001; DOR ≤13.7). The CHAID decision-tree identified useful combinations of descriptors: skin thickening" plus "destruction of nipple line" raised the probability of N+ by 40% (P < 0.001). In case of absence of "skin thickening", "oedema" and "irregular margins" likelihood of N+ was 0% (P < 0.05). Conclusion: Our data demonstrate close association of breast MRI descriptors with nodal status. If present, such descriptors can be used -as standalone or in combination -to accurately predict lymph node metastasis and such to stratify prognosis of the patient. 

for the various parametric maps: Ktrans, slope, area under the curve (AUC) and maximum enhancement. The response categories were based on histopathological findings: responders (more than 50% necrosis) and non-responders (less than 50% necrosis). The differences in heterogeneity measures before and after treatment were compared for both response categories using a Wilcoxon signed rank test. Results: There is a clear correlation between pharmacokinetic and model free parametric maps. The heterogeneity measures computed from both type of maps were able to measure significant changes due to treatment in the response group (Coherence feature: for the Ktrans map p < 0.002; for the slope map p < 0.003; for the maximum enhancement map p < 0.004 for the AUC map p < 0.005), and concurrently measured no significant changes in the non-responder group. Conclusion: Heterogeneity features, computed from pharmacokinetic and heuristic parametric maps, show potential as a biomarker for monitoring tumour response in sarcomas.

Targeting human colon adenocarcioma with novel SPIO-loaded single unilamellar nanovesicles in xenograft mouse MR model 

In vitro: experimental designs of new forms of PEGylated SPIOSUV were performed with size control. LoVo-6-luc-1 human colon adenocarcinoma cells and L-O2 human hepatic cells were cultured in media containing different concentrations of SPIOSUV. MR imaging was performed and T2 relaxivity was assessed using T2-weighted mapping software package. In vivo: conventional SPIO, hydrophobic and hydrophilic forms of SPIOSUV were injected into BALB/C nude mice with LoVo xenograft (n = 3 x 6). Dynamic MRI was done and T2 relaxivity was measured in the tumour, liver and muscles. Results: In vitro: SPIOSUV was more superparamagnetic than conventional SPIO (p < 0.01). Higher T2 relaxivity of SPIOSUV with LoVo cells than L-O2 cells was shown (p < 0.01). Hydrophilic SPIOPN showed higher T2 relaxivity rate than hydrophobic form in vitro (p < 0.01). In vivo: T2-weighted MR signal of the human colon adenocarcinoma LoVo xenografts was significantly decreased with SPIOSUV than conventional SPIO in rodent model (p < 0.01). TEM confirmed SPIOSUV in tumour cell cytoplasm. Conclusion: We successfully manufactured two (hydrophobic and hydrophilic) forms of SPIOSUV. The feasibility of in vitro and in vivo MRI of the LoVo human colon adenocarcinoma had been shown. These new types of SPIOSUV may open the door to diagnosing early phenotype of human colon adenocarcinoma due to their unique structures, kinetics and biodistribution.

Imaging labelled tumour cell growth and lymphatic micrometastasis in mice xenograft model with MR Z.-H. Dong, T. Liu, F. Gao; Chengdu/CN Purpose: To investigate primary tumour formation and mechanism of tumour lymphatic micrometastasis in nude mouse using MRI with histology. Methods and Materials: Human colorectal cancer LOVO cells were labelled with USPIO. The labelling efficiency was evaluated by Prussian blue staining. Then the labelled and unlabelled LOVO cells (10 8 cells/0.05 ml) were inoculated in claw pad and groin area of nude mice. A 7T MR with T2, T2* of the tumours and surrounding lymph nodes was performed in the following 4 weeks. After imaging, tumour tissues and lymph nodes were collected and subjected to immunohistologic analysis, which include haematoxylin and eosin (H&E) staining, Prussian blue (PB) staining, CD68 staining, lymphatic vessel endothelial hyaluronan receptor (LYVE-1) staining and vascular endothelial growth factor C (VEGF-C) staining. Results: The labelling efficiency was 90%. MRI demonstrated hypointense regions where USPIO labelled tumour cells located; then the low signal increased gradually. The increase in signal of the periphery was more rapid than central parts of the tumour. MR T2* weighted image detected lymphatic micrometastasis at 8 days P.I., and lymphatic metastasis of tumour cells happened prior to exponential growth. Lymph node metastasis was verified by H&E staining and Prussian blue staining. Immunohistochemical analysis of primary tumour sections collected at 8, 28 days P.I. revealed existence of macrophages, lymphangiogenesis, Prussian blue positive tumour cells and overexpression of VEGF-C. Conclusion: The growth of tumour cells and lymphatic micrometastasis can be observed by MRI. There was a close relationship between tumour lymphatic metastasis and lymphangiogenesis.

A novel model demonstrating MRI molecular imaging in vivo E.A. Wallnöfer, G.C. Thurner, C. Kremser, H. Talasz, A. Abdelmoez, N. Klammsteiner, W. Jaschke, P. Debbage; Innsbruck/AT (Paul.Debbage@i-med.ac.at) Purpose: Tissue-specific MR imaging in living organisms requires accumulation of high gadolinium concentrations at specific target sites of the tissue. In this work nanoparticles, each bearing hundreds of gadolinium ions and targeted to oligosaccharides present in the glycocalyx at the endothelial luminal surface, are presented and vascular molecular imaging demonstrated in an animal model. Methods and Materials: Nanoparticles, bearing covalently attached gadolinium-DTPA chelates and covalently attached targeting groups, were prepared from human serum albumin. As targeting molecule the tomato lectin (LEA) was used. After intravenous injection of these nanoparticles into Sprague-Dawley rats MR images were acquired, from the first 30 seconds after injection onwards and then during a period up to 4 hours. Results: Nanoparticles with appropriate internal cross-linking produced strong MR signals from the vascular walls. The initial MR images after nanoparticle injection were blood-volume images. However, within 2-3 minutes the larger blood vessels appeared as ring like structures, indicating that the contrast agent was accumulating in the blood vessel walls at concentrations significantly higher than those in the blood. These results were confirmed by light microscopy showing accumulation of nanoparticles in the endothelium of microvessels, and in the endothelium, the media and the adventitia of arteries. Conclusion: All materials used to synthesise the presented nanoparticles are inexpensive and can be obtained or prepared in high purity. These inexpensive, high-quality nanoparticles, delivering strong MR signals from the vascular walls of living laboratory animals, offer the potential for improved vascular imaging, especially for detection of diseases affecting the endothelium. 

This study complied with national recommendations on animal experiment. Systemic injection of 1000 umol/kg of a macrophage-targetted contrast media (USPIO -P904, Guerbet, Aulnay, France) was performed in leptin deficient obese mice (n = 17), and their lean littermates (n = 17), at different ages (7, 10, 14 weeks). The effect of USPIO injection on adipose tissue signal variance observed at 4.7T MRI on GE T2 sequences was determined. MRI results were correlated to A TM quantification by F4/80 immunostaining, and pro-inflammatory markers quantification by RT-PCR (MCP1, TNF, F4/80). The amount of iron in adipose tissue was quantified by ferromagnetic resonance. Results: A significant increase in signal variance was seen in obese mice compared to lean mice (p < 0.0001) and non-injected obese mice (p = 0.0002). This was correlated to an increase of pro-inflammatory markers in AT (Pearson r = 0.5952), an increase of A TM quantification (Pearson r = 0.5749). Iron amount was increased in obese mice AT compared to lean animals (p = 0.0288). Progressive increase in signal variance was also demonstrated in aging animals, paralleling increased inflammation (p = 0.005). Conclusion: A TM inflammation detection and quantification is feasible on 4.7T MRI in an animal model of obesity.

Heterogeneity in DCE-MRI maps: a biomarker for tumour response? L. Alic, M. van Vliet, C.F. van Dijke, A. Eggermont, W. Niessen, J.F. Veenland; Rotterdam/NL Purpose: The heterogeneity in DCE-MRI maps provides essential information for the assessment of tumour response to treatment. We evaluated the ability of DCE-MRI heterogeneity to differentiate between responders and non-responders. Methods and Materials: 18 patients (males 10, females 8 median age: 56.5 y) with irresectable extremity soft tissue sarcoma were treated with isolated limb perfusion (ILP). DCE-MR images were acquired before and after ILP with a 1.5 T MRI scanner (Vision, Siemens, Erlangen, Germany) using a bolus injection of 0.1 mmol/kg gadopentetate dimeglumine. Perfusion parametric maps were calculated using a two-compartment pharmacokinetic model and a model free method. To validate these methods histology is the gold standard, exact co-localisation between histology and MR images is therefore a prerequisite. This co-localisation is complicated by deformation and shrinking of the tissue during histological processing. To meet this problem, 3D-histology stacks were created and registered to DCE-MRI images using non-rigid B-splines with a mutual information metric. Semi-quantitative and quantitative parameters (using a two-compartment model) were computed from the DCE data. For the evaluation, regions of interest (ROIs) consisting of vital and non-vital tumour tissue were drawn in histology and transformed to DCE-parametric maps. Results: In quantitative DCE-MRI parametric maps, the average values in vital and non-vital ROIs were significantly different for the standard Tofts Model. For the semi-quantitative parameters, a significant difference between vital and non-vital tissue was observed for the wash-in, wash-out, maximum enhancement and area under the curve (AUC). Whereas the Tofs Model and the wash-in parameters can discriminate between vital and non-vital tissue, the wash-out, maximum enhancement and the AUC showed large overlap between vital and non-vital tissue. Conclusion: Non-rigid registration is necessary for exact co-localisation of DCE-MRI data with histology. The pharmacokinetic DCE-parameters, and the heuristic parameter wash-in, discriminate between vital and non-vital tumour regions. In future, this method will be expanded to evaluate several quantitative DCE-MRI analysis methods.

A S259 C D E F G H (4 mmol/L) and iopamidol (300 mg iodine/mL) was injected. MR and CT images were analysed in a randomised fashion by two musculoskeletal radiologists. The arthroscopic findings represented the reference standard. Results: In detecting glenoid fractures, sensitivity and specificity were 100% and 96.4% for CTA, and 66.6% and 100% for MRA. In demonstrating ALPSA lesions, sensitivity and specificity were 92.9% and 100% for CTA, 100% and 91.7 % for MRA. For HAGL lesions diagnosis, sensitivity and specificity were 100% and 100% for CTA, 100% and 97.4% for MRA. In detecting middle and inferior glenohumeral ligamentous tears, sensitivity and specificity were 57.9% and 85.7% for CTA, 57.9% and 95.2% for MRA. In depicting glenoid cartilage lesions, sensitivity and specificity were 77.3% and 88.8% for CTA, 72.7% and 94.4% for MRA. No difference was observed for detection of rotator cuff tendons tears. Conclusion: CTA and MRA were equivalent in demonstrating labro-ligamentous and cartilaginous lesions associated with shoulder instability. CTA was superior in detecting glenoid rim osseous lesions that are known to be a decisional element in the surgical strategy. Hence, CTA may be considered a method of choice in the preoperative evaluation of shoulder anterior instability.

The clinical relevance of lateral acromial sclerosis, the 'sourcil' sign and other radiographic acromial features in rotator cuff disease T.D. Turmezei, W.A. Wallace, K.J. Fairbairn; Nottingham/UK (tom@diagnosticradiology.eu) Purpose: The purpose of this study was to investigate which radiographic acromial features could be used to reliably identify a full-thickness (FT) rotator cuff tear and what combination of these features might represent the 'sourcil' sign in a symptomatic population. Methods and Materials: A retrospective study was performed on 158 patients undergoing first-time surgery for rotator cuff disease taken from a single surgeon´s series over 3-and-a-half years. Dedicated antero-posterior gleno-humeral joint radiographs of 83 patients with impingement, 21 with a partial thickness (PT) tear and 54 with a FT tear were randomised and blindly assessed for seven individual acromial features, acromio-humeral distance (AHD) and acromial coverage index (ACI). PT tear and impingement subjects were combined (group 1) for comparison with FT tear subjects (group 2). Results: Lateral acromial inferior surface sclerosis was seen more commonly in group 2 (61.4%) than group 1 (51.4%) (odds ratio = 1.809) but not significantly (p = 0.162). Whole acromial inferior surface sclerosis and erosion was observed with a frequency of 0% in group 1 and 7% in group 2, giving a positive predictive value (PPV) of 100% for a FT tear. Age, AHD and ACI-adjusted odds ratios showed no change in significance of results. Conclusion: Lateral acromial inferior surface sclerosis is common in rotator cuff disease but is not a discriminator for a FT tear. For useful identification of a FT tear with plain radiography, the 'sourcil' sign could be described as a combination of whole acromial inferior surface sclerosis and erosion.

A. Horng 1 , R. Vieira 2 , M. Pietschmann 1 , M. Notohamiprodjo 1 , P. Müller 1 , M.F. Reiser 1 , C. Glaser 2 ; 1 Munich/DE, 2 New York, NY/US (annie.horng@med.uni-muenchen.de) Purpose: TSE sequences with Cartesian read-out of k-space are prone to artefacts caused by involuntary patient movements, which might result in non-diagnostic images. The recently developed T2-weighted sequence with rotating rectangular read-out of k-space (BLADE) is expected to reduce motion artefacts and improve image quality. This study compares BLADE to a fat-saturated TSE-sequence (cTSE) for image quality, anatomical detail depiction and confidence in detection of pathologies in shoulder-MRI. Methods and Materials: Moderately T2-w coronal and axial BLADE and cTSEsequences in identical positioning were acquired in 60 consecutive patients (TR 3000 ms/TE 45 ms/FOV 160/BW 260 Hz) at 3 T. Images were rated using a 5-point-Likert-Scale (1-bad to 5-good) for image quality, artefacts and clinically important anatomical structures (muscles/tendons of musculi supraspinatus/infraspinatus/ subscapularis, long biceps tendon, labrum, glenoid/humeral cartilage, inferior/ middle glenohumeral ligaments). Differences were calculated by a Wilcoxon test. Results: Average rating for image quality and artefacts was 4.6/4.8 for BLADE and 3.6/3.7 for cTSE. BLADE demonstrated significantly better image quality (85% better than cTSE, p < 0.01), less artefacts (78% less than cTSE, p < 0.01) and better depiction of anatomical structures (4.8 vs 4.7, p < 0.01) as compared to cTSE. Also, confidence of diagnosis for pathologies was higher for BLADE (4.8 vs 4.6) than for cTSE (p < 0.01).

Detection of hyaline cartilage lesions of the glenohumeral joint: a comparative study of multiple detector CT-arthrography and MRarthrography at 1.5 T, with arthroscopic correlation P. Omoumi, A. Rubini, B.C. Vande Berg, F.E. Lecouvet; Brussels/BE (patrick.omoumi@uclouvain.be) Purpose: To compare the diagnostic performance of MDCT arthrography (CTA) and MR arthrography (MRA) at 1.5 T in detecting hyaline cartilage lesions of the shoulder joint, with correlation to arthroscopy. Methods and Materials: CTA and MRA examinations prospectively obtained in 57 consecutive patients (mean age 47 ; 23 females, 34 males) during the same arthrographic procedure were evaluated for glenohumeral cartilage lesions. Two radiologists independently analysed the cartilage surfaces of the humeral head and the glenoid fossa in 9 anatomical surface areas. Arthroscopy, performed within one month of the imaging, was taken as the gold-standard. We determined the sensitivity and specificity of CTA and MRA for detecting cartilage lesions with substance loss, the Spearman correlation coefficient between CTA/MRA and arthroscopic gradings, as well as κ statistics for assessing interobserver reproducibility. Results: CTA showed significantly higher sensitivity than MRA for the detection of cartilage lesions with substance loss (81 vs. 66% for grade ≥ 2 lesions, respectively [p < 0.001]) (91 vs. 64% for grade 4 lesions, respectively [p < 0.001]. The specificity for the detection of grade ≥ 2 lesions was not statistically different (96% for both techniques). Spearman correlation coefficients between CTA/MRA and arthroscopic grading of cartilage lesions were 0.773/0.689, respectively. Interobserver agreement with CTA and MRA was moderate to substantial for grading all articular surfaces and for detecting lesions with substance loss (κ = 0.656 and 0.588, respectively). Conclusion: CTA and MRA show similar diagnostic performance for detecting cartilage lesions with substance loss in the entire glenohumeral joint at 1.5 T, although CTA has higher sensitivity than MRA.

Evaluation of rotator cuff lesions: a comparative study of multi-detector CT-arthrography and MR-arthrography at 1.5 T, with arthroscopic correlation P. Omoumi, B.C. Vande Berg, F.E. Lecouvet; Brussels/BE (patrick.omoumi@uclouvain.be) Purpose: To compare the diagnostic performance of MDCT arthrography (CTA) and MR arthrography (MRA) at 1.5 T in detecting rotator cuff tears, with correlation to arthroscopy. Methods and Materials: CTA and MRA examinations prospectively obtained in 58 consecutive patients (mean age, 45 [16-72] ; 26 females, 32 males) during the same arthrographic procedure were independently evaluated for rotator cuff tears by two radiologists. Arthroscopy, performed within one month of the imaging, was taken as the gold-standard. We determined the sensitivity and specificity of CTA and MRA for detecting rotator cuff tears (supraspinatus (SS), infraspinatus (IS), subscapularis (SC), long tendon of the biceps brachii (Bc)), the Spearman correlation coefficient between CTA/MRA and arthroscopic gradings, as well as the interobserver reproducibility using K statistics. Results: For partial and complete tears considered together, there was no statistical difference in sensitivity or specificity between CTA and MRA (p < 0.05) (sensitivity: SS: 89/93, IS: 89/100, SC: 100/100, Bc: 70/80%, respectively; specificity: SS: 97/87, IS: 76/80, SC: 90/90, Bc: 100/100, respectively). For complete tears of the SS tendon, both CTA and MRA showed a sensitivity and specificity of 100%. Spearman correlation coefficients between CTA/MRA and arthroscopic detection of SS tears were 0.951 and 0.928, respectively. Interobserver agreement for the detection of rotator cuff tears was almost excellent with CTA (κ, 0.862-1.000) and substantial to almost excellent with MRA (κ, 0.741-0.914). Conclusion: CTA and MRA have similar diagnostic performance for detecting rotator cuff tears at 1.5 T. CTA has slightly better reproducibility. 

Optimisation of b-value for diffusion tensor imaging of the median nerve at 3.0 T and of reconstruction parameter for fiber tractography R. Guggenberger 1 , D. Markovic 1 , P. Eppenberger 1 , D. Nanz 1 , M. Wyss 1 , A. Chhabra 2 , K. Pruessmann 1 , J. Hodler 1 , G. Andreisek 1 ; 1 Zurich/CH, 2 Baltimore, MD/US (roman.guggenberger@usz.ch)

Recently, several studies on diffusion tensor imaging (DTI) of the median nerve have been published. However, various imaging and reconstruction parameters were used. Thus, the purpose of this study was to systematically assess the optimal b-value for DTI at 3.0 T and reconstruction parameter for fiber tractography. Methods and Materials: Local ethical board approved study with written informed consent from 45 healthy volunteers (30 women, 15 men; mean age, 41 ± 3.4 ys) who underwent DTI of the right median nerve at 3.0 T (Achieva, Phillips, Best, the Netherlands) using a dedicated 8-channel wrist coil. A single-shot echo-planarimaging sequence (TR/TE 10123/40 ms) was acquired at four different b-values (800, 1000, 1200, and 1400 s/mm 2 ). Fiber tractography was performed by two independent readers using dedicated software (Phillips) and four different tracking algorithms which included different FA and reconstruction angle thresholds. Fiber tractography was then evaluated quantitatively and qualitatively. Results: DTI acquisitions at higher b-values (1200 and 1400 s/mm 2 ) resulted in significantly longer fibers at tractography when compared to tractography at b-values of 800 and 1000 s/mm 2 (p < 0.001). Overall fiber quality (fiber length, homogeneity, density and accordance with anatomy) was best at a b-value of 1200 s/mm 2 (p < 0.001). Tracking algorithms using a FA and angle threshold of 0.2 and 10°, respectively, were significantly better than other reconstruction algorithms. Conclusion: At 3.0 T, the optimal b-value for DTI is 1200 s/mm 2 and the optimal reconstruction parameter for fiber tractography includes a minimum FA and angle threshold of 0.2 and 10°, respectively.

The radiographer's role in optimising radiation exposure Purpose: The increasing frequency of CT leads to concerns about radiation risks. However, the consequences of not performing the CT examination thus missing injuries and disease have an impact upon patient management. For justified examinations, dose optimisation is needed. CTDI and DLP are used to indicate dose and quantify improvements achieved through optimisation. The aim of this study was to optimise mA; kV and pitch, while maintaining image quality and reducing dose. Methods and Materials: Current CT scan parameters for head, abdomen and chest examinations were collected from four CT suites. These protocols were used to image a Catphan®600 CT QA phantom whereby image quality was assessed. Each of mA; kV and pitch was then systematically reduced until the contrast resolution (CR), spatial resolution (SR) and noise were significantly lowered. The Catphan®600 images were evaluated by two expert observers. The protocol considered as the optimisation threshold was the one just above the setting that resulted in a significant reduction in CR and noise but not affecting SR at the 95% confidence interval. Results: Using optimised parameters, CTDI and DLP were significantly reduced (p < 0.001) by 0-49% and 0-51%, respectively. Below the optimisation threshold, CR and noise were significantly affected (p 0.083).

To affect an optimisation strategy we have established the limits at which image quality is detrimentally reduced by dose reduction. This work established the mA, kV and pitch changes to achieve the optimisation threshold, reducing dose to the limit at which image quality is significantly decreased.

Rotating read-out of k-space (BLADE) yields significant reduction of motion artefacts and better depiction of anatomical details in comparison to cTSE sequences used in musculoskeletal radiology. Thus, BLADE provides a promising alternative with possible benefits in image quality for the examination of children, critically ill or claustrophobic patients, who express a higher probability for motion artefacts. 

A S261 C D E F G H Conclusion: More explicit guidelines regarding radiographer training and bi-annual individual mammographer technique and image quality review must be enforced including greater training of radiographers performing mammography to standardise undertaking of mammographic projections with regard to achievable compression depth, application of compression force and MGDs delivered to the breasts of Irish women attending symptomatic breast services. MGD is proposed at the 95 th percentile for 55-65 mm breast compression for full field digital mammography units at 2.4 mGy and for film-screen units at the mean MGD in line with other European publications as 2.64 mGy.

Impact of an optimization training program for radiological institutes on the reduction of CT radiation dose S.T. Schindera, G. von Allmen, R. Treier, P.R. Trueb, C. Nauer, P. Vock, Z. Szucs-Farkas; Berne/CH (sschindera@aol.com) Purpose: To assess the efficacy of an optimisation training program for radiologists and technologists on the reduction of CT radiation dose. Methods and Materials: Ten radiological institutes (seven public and three private institutes) were counselled by a radiation consulting team, including a radiologist and technologist with extensive experience in optimisation of CT protocols. The optimisation training program contained a small group teaching on protocol optimisation and a lecture on radiation dose reduction strategies. The radiation dose was assessed before and after training by the dose-length product (DLP) of five adult CT protocols (sinus, brain, chest, pulmonary angiography and abdomen). Only the DLP values of average-sized patients were included in the study. The mean DLP values were compared with the dagnostic reference levels (DRLs) of Switzerland. Results: During the pre-and post-optimisation period the DLP values of 180 CT examinations each were collected. The average reduction of DLP in the post-optimisation period was 34% for sinus (170 vs. 112 mGy•cm, P < 0.001), 13% for brain (1005 vs. 874 mGy•cm, P < 0.05), 30% for chest (443 vs. 311 mGy•cm, P < 0.05), 40% for pulmonary arteries (364 vs. 217 mGy•cm, P < 0.001) and 4% for abdomen (575 vs. 555 mGy•cm, P=0.60). The post-optimisation DLP values of the sinus, brain, chest, pulmonary arteries and abdomen were 68%, 13%, 22%, 52% and 15%, respectively, below the DRL which is 350, 1000, 400, 450 and 650 mGy•cm, respectively. Conclusion: The optimization training program is a very effective method to achieve substantial radiation dose reduction in CT independent of the scanner type.

Reducing dose to the patient through PA techniques for the clavicle M.F. Mc Entee, C. Kinsella; Dublin/IE (mark.mcentee@ucd.ie) Purpose: This study compares dose and image quality during PA and AP radiography of the clavicle to determine the optimum technique and quantify and difference. Methods and Materials: The methodology involved a cadaver-based dose and image quality study recording dose to the thyroid, eye, breast, entrance and exit surfaces. An image quality VGA study analysing AP and PA projections was also carried out. Results: A statistically significant 56.1% (p < 0.001) and 56.5% (p < 0.001) reductions to breast tissue and 62.3% (p < 0.001) and 78% (p < 0.001) reductions in the thyroid dose with employment of PA and PA15 caudal projections. Reductions of 28.5% (p < 0.08) and 11.8% (p < 0.12) in the doses to the eye were demonstrated for the PA. Differences in entrance-surface and exit doses were deemed nonsignificant. A 5.9% (p < 0.39) reduction in overall image quality was seen with PA positioning. Reductions in image quality were evaluated to be nonsignificant at 95% (AP vs PA (p < 0.06), AP15 vs PA15 (p < 0.28).

The current study demonstrated consistent significant dose reductions to the thyroid and breast when PA projection is chosen over the AP projection. The authors recommend the implementation of PA positioning for clavicle radiography.

Patient's dose optimisation in projection lumbar spine examination with CR A. Henner 1 , A.-R. Pietilä 2 , H. Pöyskö 2 , J. Sergejeff 1 , H. Paso 1 ; 1 Oulu/FI, 2 Haukipudas/FI (anja.henner@oamk.fi)

Purpose: The purpose of this study was to find out methods to evaluate the adult patient's dose in projection lumbar spine examination with computed radiography (CR) in health center after the change of procedure. Methods and Materials: Data from lumbar spine ap on Bucky table were collected regularly as said in state 423/2000. In spring 2009 the procedure was changed due to the order of the central hospital. The radiographers find out increase in patient dose with AEC (automatic exposure control) when changing the procedure from Bucky table (patient lying on the table) to wall-stand (patient standing). The

Optimisation and dose reduction in CT of the skull M. Monteiro, A. Catarino, E. Fernandes; Coimbra/PT (mmonteiro@estescoimbra.pt) Purpose: This study aims to assess the CT scan parameters: noise, spatial resolution and low contrast resolution by reducing only the mA 10% to 40% in order to find a compromise between quality image and the dose administered. Methods and Materials: To collect data we used CT equipment Light Speed -GE 4-slices and a Light Speed VCT-GE 64-slices, and a LightSpeed™ 7.X phantom that simulates the density of the skull using the usual protocol for examinations of skull images obtained from both CT equipment in Sections 1 and 2 of the phantom. Subsequently reductions were made in the value of mA from 10% to 40% keeping constant the time of scan and kV. Results: Based on a QA phantom using a FOV similar to the skull, we conclude that for both CT equipment in the images of 2.5 mm for the study of the posterior fossa, we can reduce to 20% of the value of mA, for images related to the study of supra-tentorial region the mA can be reduced by 30% without loss of image quality. In both CT equipment the limiting factor reduction mA was the noise parameter. Conclusion: We find that for the equipment 4 slices CT we can reduce the value of DLP by 20%, in 64 slices CT equipment we can reduce the value of DLP in about 27%, compared with the dose protocol commonly used for examining the skull on both CT. Purpose: DRLs are dose levels that should not normally be exceeded, where good practice is applied. EU directive (97/43/EURATOM) makes the implementation of DRLs mandatory for member states. The current work establishes DRLs for three common CT examinations as the first step towards CT dose optimisation nationwide.

The study included 57% of CT units in Malta, including three CT manufacturers. CT head, chest and abdomen were included being the most frequent CT examinations identified through a survey based on NRPB W67 guidelines. CTDI and DLP were obtained from the units' display console for 276 CT examinations, while effective dose (E) was estimated using ImPACT CT dose calculator. Results: Maltese 3 rd quartile CTDIw values for abdomen: 11.7 mGy; chest: 12.5 mGy; head: 36.1 mGy and DLP values for head: 736 mGycm were well below EU recommendations. However, 3 rd quartile DLP values for abdomen: 539.4 mGycm and chest: 492 mGycm were fairly higher indicating the need for dose optimisation in these areas. Displayed CTDI and DLP correlate very well to calculated values using the ImPact method (r> 0.7), indicating the use of both as appropriate CT dose descriptors. Variations in choice of scan parameters, scanner type and design are main reasons for significant differences in mean CTDI, DLP and E between hospitals. Conclusion: 3 rd quartile levels have been established for three CT examinations in Malta. Further work to minimise dose without affecting image quality and bring all DRLs below EU levels is also needed. DRLs should also be established for other CT examinations.

Optimisation of compression, image quality and radiation dose in mammography in Ireland D. O'Leary, L. Rainford; Dublin/IE (desiree.oleary@ucd.ie) Purpose: The study objective was to find achievable, objective and standardised compression depths to assist radiographers to achieve the lowest possible radiation dose with highest diagnostic image quality in mammography within the symptomatic breast service in Ireland.

This large quantitative and qualitative study of symptomatic breast units within the Republic of Ireland, collected image quality, compression and radiation dose data (n = 4790 images). The data were analysed using mathematical modelling and SPSS. Results: Objective compression levels could not be proposed due to higher than expected inadequate image quality rates (6.2%) and inconsistent compression forces; other factors may also impact on the compression levels attained which cannot be teased out by mathematical modelling of the current data set. Compression forces are low affecting image quality; a significant finding is that greater compression force by 11-15 N is needed to achieve perfect images. Mean glandular doses (MGDs) received by perfect images are significantly lower than inadequate images.

Purpose: To quantify the inherent properties, the potential, and the limitations of x-ray phase contrast imaging (PCI). Methods and Materials: The refractive index of materials can be characterised by its real part (phase shift) and imaginary part (absorption). PCI visualises the real part of the refractive index by exploiting the differential phase shift. It is believed patient's dose was evaluated with ESD and after that the manual settings were taken in use. The next step was to adjust the sensitivity of the AEC by the vendor. Results: The doses increased significantly when changing to the stand wall. The radiographers decided to change to the manual settings and reached the earlier or lower dose level. Later the sensitivity of the AEC was increased and after that the use of AEC was possible without any dose increase. Conclusion: The radiographers must be aware of the dose levels. Doses must be followed up regularly, especially when changing the procedures. The decrease in dose can be made by easy and cheap ways. The key person in dose optimisation is the radiographer. Working in this way demonstrates very high commitment to the radiation safety culture.

Mobile chest radiography on special care baby units: can we improve our practice? K. Matthews, M. McEntee; Dublin/IE (kate.matthews@ucd.ie) Purpose: To analyse potential for improvement in mobile chest radiography on neonates and infants by investigating: (1) does referral information provide evidence for justification;

(2) what national diagnostic reference levels (DRLs) are appropriate;

(3) could practice be improved by modifications drawn from survey data analysis? Methods and Materials: Data were collected on 64 examinations in six hospitals: two tertiary referral, two neonatal and two general, representing 29% of hospitals nationwide providing mobile radiography to a Special Care Baby Unit. Radiation dose was recorded with calibrated DAP meters in centiGray*cm 2 ( cGy*cm 2 ). Radiographers recorded the research data. Results: Referral information was generally complete, although some requests had sparse or no clinical details. DAP dose correlated poorly with age, but reasonably with weight (r = 0.464). At specific weights, DAP dose varied within and between hospitals by factors up to 17 and 24, respectively. Discounting aberrant dose outliers, there was never greater than five times variation between hospitals. Small samples at particular weights impacted onto regression analysis of influences on DAP. However, trends indicate that technically, field size, kVp and mAs have the greatest influence. Conclusion: The majority of mobile chest examinations have reasonable clinical justification. Weight-specific, Irish DRLs in the range 0.53 to 1.35cGy*cm 2 are proposed for babies under one year and up to 6 kgs. Survey doses are comparable to published data. Unnecessary variations in DAP are apparent. Analysis of DAP variations suggests that more objective matching of baby weight to kVp and mAs could considerably reduce the DAP dose without detriment to image quality.

Radiographer's role in the optimisation of MR-guided HDR prostate brachytherapy C. Vandulek, J. Pall, F. Lakosi, G. Antal, A. Kovacs, J. Hadjiev, I. Repa; Kaposvár/HU (cvandulek@gmail.com) Purpose: MRI provides excellent visualisation of the prostate and surrounding tissues, making it the choice for guiding and monitoring brachytherapy. The purpose of this study was to optimise the initial protocol of transperineal MR-guided highdose-rate prostate brachytherapy. The radiographers played an important role in the fine-tuning of the methodology. Methods and Materials: The procedures were performed using an open, 0.35T MRI scanner (GE Healthcare). Study volunteers were scanned comparing signalto-noise ratio (SNR) of the 9-inch GP coil and abdomen coil. Volunteers were scanned feet first in lateral decubitus and supine position to study patient comfort and ergonomics. Standard and modified T1 and T2 MRI sequences were compared in respect of artefacts and SNR. Results: The tested and modified imaging protocol allows clear definition of the prostate, coaxial needles and catheters. The results of the study justified the transition in methodology from GP surface coil to abdomen coil, change in positioning from lateral decubitus to supine position. Patients confirmed improvements in comfort and short-term side-effects. Modified pulse sequences resulted in improved SNR and reduction of artefacts. Conclusion: The initial methodology implemented for MR-guided prostate brachytherapy has been further developed leading to improvements in patient comfort, ergonomics, patient access and SNR. The radiographers played an important role in the optimisation of the protocol contributing to the success of the study. Our results facilitate us to implement the optimised techniques in further MR-guided brachytherapy protocols other than the prostate. Published studies have shown that in patients with small cross-section, using lower kVs (80 or 100 kV) allows acquiring data with a similar image quality but lower dose. In our study we assessed the impact of 70 kV on image quality and dose. Methods and Materials: Measurements were performed on a prototype system (Definition AS+, Siemens), which was modified to allow scanning at 70 kV. To assess image quality, measurements with different sized phantoms simulating paediatric patients were performed. Noise was determined with ROI measurements. To assess contrast, probes with diluted iodinated contrast agent were placed in central and peripheral positions of the phantoms and the enhancement measured. Dose was determined by CTDIvol measurements on CTDI phantoms (diameter 8 and 16 cm). In addition, patient dose for a CT of the abdomen was estimated using a paediatric anthropomorphic phantom and TLD measurements. Results: In case of the 8 cm phantom, for the same CTDIvol, noise does not change significantly for lower voltages. However, iodine contrast increases from 427 HU at 120 kV to 998 HU at 80 kV and 1125 HU at 70 kV. Differences in effective dose between 70 and 80 kV for the abdominal scan were strongly correlated with respective the difference in CTDIvol (< 10%). Conclusion: For small patients a further reduction in tube voltage from 80 kV down to 70 kV allows either to acquire images with an increased iodine contract (+13%) or a dose reduction of 28% for the image quality in terms of iodine contrast to noise.

Dual-energy spectral CT (DESCT) allows selection of the optimal monochromatic energy for imaging the instrumented spine and improves diagnostic quality over traditional imaging J.M. Kessler, M. Ellestad, J. Rios, E.G. Stein, L. Tanenbaum; New York, NY/US (lawrence.tanenbaum@mountsinai.org) Purpose: Leveraging unique atomic densities of different elements, and the ability of modern CT systems to image at multiple energies, dual-energy-spectral CT (DESCT) provides novel imaging information for CT diagnosis of spine disease. Analysis of DESCT scan spectral Hounsfield signal behaviour in instrumented spines allows selection of appropriate energy levels that minimise artefact and optimise image quality. We sought to determine energy levels that maximise signal-to-noise ratios (SNRs) for instrumented spines and compared these optimal images with those from traditional polychromatic 140 kVp and routine monochromatic 70 keV images. Methods and Materials: Ten instrumented spine patients were imaged on a GE CT750HD using DESCT. Data were processed and spectral Hounsfield unit curves (SHUCs) were generated from region of interests placed in spinal canals adjacent to hardware. An optimal SNR keV level was identified and a set of multiplanar images were generated at that level. These images were compared with standard 70 keV monochromatic and 140 kVp polychromatic sets and evaluated for artifact obscuring the spinal canal, and overall diagnostic quality. Results: Evaluation of SHUCs demonstrates that optimal SNRs were obtained at energy levels 90 -95keV. Comparing monochromatic images produced at optimal energy levels to standard monochromatic 70 keV and 140 kVp polychromatic images significantly reduces spinal canal noise and artefact. Furthermore, these images remained adequate for assessing other findings (e.g. degenerative disease). Conclusion: Monochromatic images of instrumented spines obtained at 93 keV (range 90-95 keV) produced images with fewer artefacts and better hardware visualisation, with superior overall quality.

Performance of static MDCT beam-shaping filtration for elliptical subject geometries in dependence of projection angle L.-S. Veloza, H.-U. Kauczor, W. Stiller; Heidelberg/ DE (stella.veloza@med.uni-heidelberg.de) Purpose: MDCT beam-shaping filters modulate x-ray beam intensity across the fan, compensating for decreasing absorption due to shorter x-ray path lengths through patient body periphery compared to its center. To investigate projection-angle dependence of detector statistics for elliptical subject geometries, performance of experimentally measured filter geometry and theoretical filter geometry adapted to elliptical specimen has been compared. Methods and Materials: Using a Compton spectrometer, primary MDCT x-ray spectra (120 kV p ) have been measured at different angular distance (0°-21°) from the central ray of the x-ray fan. These have been used to calculate the geometrical shape of the aluminum bowtie filter. A theoretical filter model was determined for elliptical subject geometries. For both shape models, detector statistics after x-ray to provide additional information on tissue properties compared with conventional absorption-based imaging (AI) techniques. We have developed a mathematical formalism to assess the contrast-to-noise ratio (CNR) per dose analytically. This allows for a direct performance comparison of PCI versus AI. The mathematical model was validated by numerical simulations and compared to experimental data, particularly, with regard to CT. Results: 1. PCI shows a fundamentally different noise power spectrum (NPS). PCI images show strong statistical fluctuations on growing length scales. The predicted characteristics are confirmed by simulations as well as experimental data. Consequently, the performance of PCI deteriorates with decreasing spatial resolution. 2. Due to the measurement of periodic quantities, information propagation collapses at low dose. Simulations and experimental findings are in agreement with this result. This implies a minimally required radiation dose. In contrast, AI does not collapse at low dose. Conclusion: To achieve a benefit of PCI versus AI a minimum spatial resolution is necessary. This implies increased radiation dose if the resolution to break even is higher than that of a specific diagnostic application today. On the other hand, PCI cannot be performed at arbitrarily low dose as a matter of principle. 

We retrospectively analyzed 106 PET-CT studies of 63 consecutive patients diagnosed with head and neck cancer referred to our institution from 01/01/2006 to 07/31/2009. The protocol included PET-CT from the top of the head to the mid-thigh without intravenous contrast, followed by a head and neck PET-CT after intravenous contrast. The SUV values were calculated on the lesions (SUV LESION ) and on the great vessels (SUV VESSEL ) before and after contrast administration. Results: Patients' average age were 56.5 years and 53 were male. From the 106 studies analysed, 59 showed areas of abnormal FDG uptake on the head and neck, suggestive of neoplastic disease. Analysing the 59 positive studies, there was an absolute variation on the SUV LESION of 1.28 ± 1.47 after the contrast enhancement, with a percentage change in the SUV LESION of 16.5 ± 16.8%. The average value of the SUV LESION between the studies performed before and after the contrast raised from 6.64 to 7.65 (p = 0.191), not achieving statistical significance. Including all the studies performed, there was an absolute variation on the value of the SUV VESSEL of 0.05 ± 0.22 and a percentage change of 3.6 ± 12.0% after contrast enhancement. The average value of the SUV VESSEL ranged from 1.82 before the contrast to 1.86 after the contrast (p = 0.389), also without statistical significance. Conclusion: Intravenous contrast does not alter significantly the value of SUV, thus it can be used in the PET-CT studies for attenuation correction.

Dose reduction by on-line tube voltage modulation (TVM) in clinical conebeam CT M. Baer, M. Kachelrieß; Erlangen/ DE (matthias.baer@imp.uni-erlangen.de) Purpose: To evaluate the potential benefits of anatomy-dependent tube current and tube voltage modulation. Methods and Materials: All CT manufacturers recognised the dose reduction potential of patient-dependent tube current modulation (TCM) and provide manufacturer-specific implementations. There are two drawbacks: the dose is estimated to be proportional to the tube current and the tube voltage always remains at a constant level. We derived a new approach that both accounts for the actual patient dose and that modulates the tube current as a function of the patient anatomy. The potential of combining TCM with TVM was evaluated in a simulation study using clinical patient data. Patient dose was estimated for each projection by analysing the rawdata using a Monte Carlo-calibrated polychromatic deterministic model. Rawdata were water precorrected using the approach of reference . The dose required to obtain a specified image quality (spatial resolution and noise) was recorded to quantify the value of combined TCM and TVM. Results: With TVM noise variance, respectively, dose can be lowered by up to 18% for both the hip and the shoulder region compared to TCM at a constant tube voltage of 120 kV. CT image quality is not impaired by TVM. Conclusion: Combining TVM with TCM reduces dose by 10% to 20% compared to TCM alone. It can therefore be an important contribution to the world-wide efforts to reduce patient dose. Purpose: The MR-guided High Intensity Focused Ultrasounds (MRgFUS) is a rapidly developing methodology for lesion treatments. The purpose of this work is to describe the methodology and characterise the device from physical point of view. Methods and Materials: The ExAblalte 2000v.4.2 (Insightec (IL)) is integrated with MRI-GE-Signa-HDxt-Advantage-1.5 T. The system controls the position of the US beam and the real time temperature of the tissue using the MRI system. The thermometry is based on PRF shift. We studied geometrical characteristics of a single sonification beam (3-4 cm lengths, half centimeters diameter) using specific HiFU gels. We investigated the relationship between temperature and the phase shift with an independent heating source and high accuracy fibre optics temperature sensors by means the α value evaluation. Also, T1 method for temperature evaluation was analysed. Before every treatment, phantom quality control protocol (DQA) was applied: the constancy of size focus, energy, temperature was tested.

The results show good correlation between chemical shift and temperature. We estimate the α value as -0.0109±0.0002 ppm/°C. Our data are in agreements with literature. Using PRF method, our precision of the single pixel readout temperature is 4.2 °C; the accuracy is 0.26 °C (ΔT=10 °C) and 0.65 °C (ΔT=20 °C). Using T1 method, our precision of the single pixel readout temperature is 9.1 °C; the accuracy is 1.7 °C (ΔT=10 °C) and 7 °C (ΔT=20 °C). This lack of precision and accuracy of the T1 method should be explained with a choice of TR in the MRI protocol. DQA analysis: the stability of the physical parameters for the treatment is inside the experimental errors. Munich/DE (cwallner@med.uni-muenchen.de) Purpose: To evaluate the use of iterative reconstruction technique for dose reduction and image quality improvement in computer tomography (CT) of the paediatric chest. Methods and Materials: Low-dose paediatric chest CT scans using the adaptive statistical iterative reconstruction (ASIR) method and the standard paediatric LDCT protocol were performed on a GE Discovery CT 750HD. Natural sponges as lung equivalent saturated with iodinated contrast media inserted within a chicken specimen simulating a neonatal patient and within a turkey specimen as an infant phantom were used. For the neonatal studies tube current was 80 kVp, for the infant studies 100 kVp. The ASIR reconstruction level was 40% for both the neonatal and the infant examinations. The noise index was identical for all protocols with a value of 45.

The radiation dose delivered during the LDCT scan without IR was 1.64 mGy CTDI (25.25 mGy*cm DLP) for the neonatal phantom and 4.06 mGy CTDI (100.79 mGy*cm DLP) for the infant phantom. For the neonatal phantom the dose delivered using LDCT with IR was 0.93 mGy CTDI (14.30 mGy*cm DLP) and for the infant phantom 2.29 mGy CTDI (57.92 mGy*cm DLP). Dose reduction was 43.3% for the 80 kVp neonatal protocol and 43.6% for the 100 kVp infant protocol. Conclusion: With iterative reconstruction technique compared to a standard LDCT paediatric chest protocol we found the patient radiation dose being significantly reduced up to 44% when using the low-dose IR method. The resulting diagnostic image quality was similar for both methods.

passage through a homogeneous elliptical phantom (16.0 cm semimajor axis, 12.5 cm semiminor axis, 8.0 cm thickness) of polypropylene, PMMA and water was Monte-Carlo simulated with Geant4. The number of surviving quanta after xray passage through beam-shaping filter plus phantom was counted by a detector array positioned for projections a.p. to lateral in 15°-steps. Results: At the fan edge, measured filter shape is systematically too thin: for lateral projections maximum difference is 28% compared to the theoretical filter model for elliptical subject geometries. Performance of the latter varies with projection angle and thus requires adaption of geometry to view angle. Conclusion: Static beam-shaping filter geometry and its material composition should be reevaluated to reduce signal and thus patients' skin dose for lateral projections. Dynamic beam-shaping concepts could allow for radiation exposure reduction in MDCT.

Three Here, we compared two-dimensional (2D) and three-dimensional (3D) FLAIR sequences concerning their proneness to these artefacts, and their sensitivity and specificity for SAHs.

The following four FLAIR sequences were evaluated:

(1) 2D FLAIR at 1.5 T, (2) 2D FLAIR, (3) 2D PROPELLER (= periodically rotated overlapping parallel lines with enhanced reconstruction) FLAIR, and (4) 3D FLAIR (Cube-FLAIR) at 3 T. All sequences were applied in five healthy volunteers; sequences 2 and 4 were also applied under routine conditions in ten patients with focal epilepsy, and in five patients with SAH. Analysis was done independently by two neuroradiologists. For volunteers and epilepsy patients the presence of flow artefacts in ventricles and cisterns was evaluated and scored on a four-point scale.

Mean values were calculated and compared using paired t-tests. Sensitivity and specificity of sequences 2 and 4 for SAH detection were determined.

Results: Cube-FLAIR showed almost no CSF artefacts at all, either in the volunteers or in the epilepsy patients, and therefore was significantly superior to any other FLAIR (p < 0.001). Sensitivity and specificity of the 3 T 2D FLAIR for SAHs were 68.8% and 90.3%, respectively, while the Cube-FLAIR showed a sensitivity of 96.9% and a specificity of 100%.

The Cube-FLAIR allows FLAIR imaging with almost no CSF artefacts, and thus seems to be very helpful to detect SAHs.

To evaluate flow-sensitive 4D MRI for the analysis of haemodynamics in patients after repair of tetralogy of Fallot (TOF), particularly with regard to altered flow patterns in the pulmonary system and their correlation with pulmonary vascular geometry. Methods and Materials: ECG and respiration synchronised flow-sensitive 4D MRI with 3D velocity encoding (TE = 2.4 ms, TR = 4.8 ms, venc = 200 cm/s, spatial resolution~2.5 mm³, temporal resolution = 38.4 ms) was acquired in ten patients after repair of TOF (mean age 12 years) and in four healthy controls (mean age 26 years). Analysis included the assessment of vascular geometry (vessel diameters and branching angles) in MR angiography data, evaluation of haemodynamics by 3D blood flow visualisation (particle traces) and quantification in the aorta, in the pulmonary trunk (TP) and left (lPA) and right (rPA) pulmonary arteries. Results: Systolic peak velocity in the TP was significantly higher in patients (1.9m/ s±0.7m/s) than in controls (0.9m/s±0.1m/s). Regurgitation fraction in patients was 29%±11% and no regurgitation was measured in controls. A higher blood flow in the rPA of the TOF patients was measured: rPA/lPA flow ratio of 2.6±2.5 versus 1.1±0.1 in controls. Correlation analysis detected relationships between the severity of vortices in the pulmonary system and vessel size (r = 0.79) and between the regurgitation fraction in the rPA and the pulmonary arteries branching angles (r = 0.73). Conclusion: 4D MR flow analysis showed abnormal flow patterns in the entire group of TOF patients which correlated significantly with vascular geometry. 4D MRI has the potential to demonstrate the complex nature of pulmonary morphology and haemodynamics in corrected TOF patients. was performed in nine patients after repair of TGA and in four healthy controls (mean age 12.2 years vs. 26 years). Blood flow was visualised by time-resolved 3D particle traces emitted from planes in the ascending aorta and the pulmonary trunk (TP). Visual grading was performed using a 3 grade ranking scale (0-2) with regard to the extent of flow vortices, helices, retrograde flow and presence of high peak velocities in the aorta and pulmonary system (v > 1.5m/s). Results: Six patients had vortices in the TP, mean vortex severity was graded as 0.78. Vortex formation was less pronounced in the right (rPA) and left pulmonary artery (lPA) (0.33 and 0) and aorta (0.22). Helices were observed in the lPA and aorta with an average grading of 0.78. Retrograde flow was rare in patients: 0.33 in TP and 0.22 in lPA. No reflux was detected in the rPA and aorta. Peak velocities > 1.5 m/s were seen in TP, rPA and lPA in all nine patients, in the aorta in 7/9. No vortices, helices, reflux or accelerated flow were detected in volunteers.

Methods and Materials: 45 Patients (67% female, mean age = 53.7) with clinical suspicion of an arthritic joint disease were examined with an ICG-enhanced optical imaging system (Mivenion GmbH, Berlin, Germany). The degree of inflammation in the carpal, metacarpophalangeal, proximal and distal interphalangeal joints of both hands was graded objectively by four independent radiologists on a 4-pointordinate scale (0: no inflammation, 1: mild, 2: moderate, 3: severe). Results were correlated with MRI (3 T Verio, Siemens, Germany) as standard of reference. Results: Of the 1350 joints of the evaluated 45 patients, in MRI a total of 608 (45.0%) joints showed mild (37.8%), moderate (6.4%) or severe (0.8%) inflammation. In OI an average of 414 (30.7%) joints showed mild (61.2%), moderate (24.6%) or severe (14.3%) inflammation. Using MRI as standard of reference, OI had a sensitivity of 43.5% (35.3% -52.1%), a specificity of 80.3% (74.8% -84.8%), and accuracy of 65.7% for detection of active inflammation in patients with RA.

The evaluated ICG enhanced OI system showed substantial limitations in the detection of active inflammation in patients with arthritic finger joints in comparison to MRI as standard of reference. Despite its advantages of its minimal invasiveness, lower costs and easy operability the OI system appears to be of limited value for the clinical routine.

Do we really need the full dose of multihance for contrast-enhanced MR imaging for the assessment of synovial enhancement in patients with rheumatoid arthritis? C. Purpose: To investigate the diagnostic accuracy of a half-dose of multihance (Bracco, Italy) in the assessment of synovitis in hand and finger joints in RA.

Thirty patients with early rheumatoid arthritis underwent 3 T high-resolution MRI. Coronal T1-weighted TSE and STIR sequences and transversal T2-weighted SE sequences were acquired before contrast media. Coronal T1-weighted TSE sequences with fat saturation were obtained after 0.05 mM/kgKG multihance (half-dose = HD) with a delay of five minutes, and, five minutes later, additional coronal T1-weighted TSE sequences were obtained after a second application of 0.05 mM/kgKG multihance (full dose = FD) with a delay of five minutes. The contrast enhancement in the region of interest was measured in inflamed synovial tissue in exactly the same slice on half-and full-dose contrastenhanced sequences. Differences and the correlation of signal intensities on HD and FD contrast-enhanced sequences were compared using the paired t-test for statistical analysis as well as the Bland-Altman plot. Results: A statistically significant difference was found for half-and full-dose multihance in contrast-enhanced synovial tissue (mean: 1184.527 vs. 1363.137; p < 0.001). As the signal intensities showed a high correlation between HD and FD (r = 0.779), the formula FD = 375.858 + 0.833*HD can be used to convert HD to FD. Conclusion: In inflamed synovial tissue, the post-contrast signal intensities showed a high correlation between HD and FD. Therefore, HD multihance is sufficient for the assessment of synovial enhancement in patients with early rheumatoid arthritis with 3 T MRI. No additional relevant information for clinical purposes can be obtained by giving the FD multihance.

Enthesitis and related changes in the knees in seronegative spondyloarthropathies: MRI control study Y. Ragab, Y. Emad, H. Hamza; Cairo/EG (yragab61@hotmail.com)

Purpose: To evaluate the role of magnetic resonance imaging (MRI) in the detection of knee changes in patients with seronegative spondyloarthropathy (SpA). Methods and Materials: 56 patients including 30 with psoriatic arthritis, 5 with ankylosing spondylitis, 5 with reactive arthritis, 5 with ulcerative colitis (UC), 5 with Crohn's disease, and another 6 with skin psoriasis. Controls were 20 healthy subjects without knee complaints. MRI was performed in all participants including Gd-DTPA enhancement. Results: Both knees were studied in 45 (80.3%) patients and one knee in 11 (19.6%). MRI showed evidence of bone marrow oedema in 13 (23.2%) patients, cartilaginous erosions in 18 (32.1%), and bone erosions in 9 (16.1%). Enthesitis was found in medial collateral ligaments in 18 (32.1%), lateral collateral ligaments in 8 (14.3%), posterior cruciate ligaments in 3 (5.35%), patellar tendon in 18 (32.1%), biceps femoris insertion in (5.35%), medial patellofemoral ligaments (MPFL) in 5 (8.9%) and lateral patellofemoral ligament in 1 patient (1.8%). In the UC and Crohn's patients (n = 10), 2 had bone erosions and 5 had enthesitis. In the skin psoriasis group (n = 6), one had bone marrow oedema; enthesitis was detected in 5 at the patellar tendon insertion and in one in the MPFL. Purpose: To assess misregistration artefacts in dGEMRIC of finger joints and to evaluate the effectiveness of motion correction.

In 40 cases dGEMRIC of metacarpophalangeal joint II was performed. Imaging used a dual flip angle approach (TE 3.72 ms, TR 15 ms, flip angles 5° and 26°). Two sets of T1 maps for were calculated for dGEMRIC analysis from the imaging data for each patient: one with and one without motion correction. To compare image quality, visual grading analysis and precision of dGEMRIC measurement of both dGEMRIC maps for each case were evaluated. Results: Motion artefacts were present in 82% (33/40) of uncorrected dGEMRIC maps. Motion artefacts were graded as severe or as rendering evaluation impossible in 43% (17/40) of uncorrected dGEMRIC maps. Motion-corrected maps showed significantly less motion artefacts (P < 0.001) and were graded as evaluable in 97% (39/40) of cases. CV was significantly lower in motion-corrected images (0.176 ± 0.077), compared with uncorrected images (0.445 ± 0.347) (P < 0.001). Motion-corrected dGERMIC was different in volunteers and patients (P = 0.044), whereas uncorrected dGEMRIC was not (P = 0.234). Conclusion: Motion correction improves image quality, dGEMRIC measurement precision and diagnostic performance in dGEMRIC of finger joints. Methods and Materials: 29 patients with active RA, defined by a DAS28 higher than 3.2 and naïve for biologic therapy, were treated with etanercept added to previous DMARD therapy during a year. Clinical evaluation, laboratory tests, and MRI of the dominant hand were performed before and after one year of treatment. Paired sets of h-MRI were read blinded for time-sequence by 2 experienced readers following RAMRIS method, tenosynovitis was also scored. Regression or repair was studied in three different ways: 1. The observed yearly progression rate was lower than the estimated rate, 2. RAMRIS score for erosion after one year was lower than the baseline score, or 3. Change of RAMRIS score for erosion exceeding the smallest detectable difference (SDD) and minimal detectable change (MDC < 20%). Results: 29 patients (25 women) started therapy with etanercept combined with DMARDs. At baseline, the mean DAS28 was 4.7 ± 1.1 decreasing to 2.9 ± 1.4 (p < 0.001) after one year. All the other clinical parameters also showed significant decreases. Mean RAMRIS score for erosions did not change, but all the other MRI parameters showed a significant decrease. 

An indigenous MRI scoring system for differentiating advanced femoral osteonecrosis from tubercular arthritis: myth or mystique? V. Venugopal, I. Afshan, E. Ullah, I. Ahmad, S. Wahab; Aligarh/IN Purpose: A MRI-based scoring system for differentiating advanced osteonecrosis of femoral head from tubercular arthritis is introduced. Methods and Materials: MRI findings in 18 hips with tubercular arthritis and 36 hips with advanced osteonecrosis of femur were reviewed retrospectively. Confirmation of tuberculous arthritis was based on ELISA and/ or synovial biopsy. Osteonecrosis was confirmed by histopathology and only Arlet and Durroux types 3 and 4 were included in the study. The findings were analysed with emphasis on the changes in femoral head marrow, joint cavity, synovium, acetabulum and contrast-enhancement patterns.A score of two was assigned for presence of each of T2 hyperintensity of femoral head, synovial hypertrophy, articular cartilage erosion, unilateral involvement of femoral head, acetabular oedema/sclerosis and enhancement of involved head. A score of one was assigned for each of joint effusion, oedema of adjacent marrow and enhancement of adjacent soft tissue. A cut-off value of 10 out 15 points was considered to be positive for TB arthritis. Results: 16 out of 18 cases of TB arthritis were correctly identifiable on the basis of this scoring system. The two remaining cases had a score of 9. None of the AVN cases scored more than 10. A score of 10 for had a sensitivity 88.89 % and specificity 100% the positive diagnosis of tuberculous arthritis.Positive and negative predictive values were 1 and 0.94, respectively. Conclusion: This indigenous scoring system is very effective in solving the imaging dilemma pertinent to endemic regions of differentiating advanced femoral osteonecrosis from avascular necrosis.

Dual energy "virtual non-calcium" technique: detecting posttraumatic bone bruises of the knee with computed tomography G. Pache 1 , B. Krauss 2 , P. Blanke 1 , S. Bulla 1 , M. Langer 1 , T. Baumann 1 ; 1 Freiburg/DE, 2 Forchheim/ DE (gregor.pache@uniklinik-freiburg.de) Purpose: To evaluate bone marrow assessment with a dual energy CT "virtual non-calcium" technique utilising MRI proven bone bruises as a model for typical marrow pathology so far indiscernible via CT. Methods and Materials: In this prospective, institutional review board approved the study; 21 patients with an acute knee trauma underwent dual energy CT (Somatom Definiton, Siemens) and MRI scans. The application LiverVNC (Syngo Dual-Energy, Siemens) was used to virtually subtract calcium from the images. Presence of bone bruise was rated by 2 radiologists on a 4-point scale for 6 femoral and tibial regions each. HU measurements were performed in the same regions. Consensus reading of independently randomised MRI scans served as the standard of reference. Image ratings and HU values were subjected to receiver-operator-characteristic (ROC) analysis. Results: MRI revealed 59 bone bruises in a total of 236 regions (19 of 114 femoral, 40 of 122 tibial, 16 excluded due to artifacts). Visual rating revealed area-under-thecurve (AUC) values of 0.886 (0.897) in the femur and 0.974 (0.953) in the tibia for observer 1 and 2, respectively. For HU measurements the respective AUC values were 0.922 and 0.974. If scores of 1 and 2 (strong/mild bone bruise) were counted as positive, sensitivity and specificity were 86.4% (86.4%) and 94.4% (95.5%) for observers 1 and 2, respectively. Kappa analysis demonstrated good to excellent agreement (κ-value: 0.78/0.87 femoral/tibial). Conclusion: Dual energy CT "virtual non-calcium" technique can subtract calcium from cancellous bone, allowing bone marrow assessment and making posttraumatic bone bruises of the knee detectable with computed tomography.

The initial assessment of clinical value of gouty arthritis by DSCT dualenergy technique D. Han, L. Wu, B. He; Kunming/ CN (kmhandan@sina.com) Purpose: To explore the diagnostic value of gouty arthritis from crystal deposition of uric acid inside joints and related soft tissues by dual-source CT dual-energy mode (DECT) technique. Methods and Materials: Plain DECT scans under gout program had been done on 132 patients that include 50 cases secured diagnosed, 37 suspected and 45 as control, respectively. The output images were sent to DE gout software, which can display the crystal deposit of uric acid on target positions by colour marked MPR and VRT images. The diagnosis of gouty arthritis was made based on DECT findings and the clinical manifestation.In the control group, colour signs were believed to be false positive.

Entheseal-related changes can be found in patients with skin psoriasis or inflammatory bowel disease with no clinical arthritis. Enthesitis of the knee joint may be an early sign on MRI and a characteristic finding in the SpA group of diseases.

In vivo monitoring of septic arthritis using macrophage MR imaging G. Bierry, S. Lefevre, D. Ruimy, F. Jehl, A. Neuville, J.-L. Dietemann; Strasbourg/FR (guillaume.bierry@chru-strasbourg.fr) Purpose: To prospectively evaluate macrophage imaging using ultrasmall superparamagnetic iron oxide (USPIO)-enhanced MR imaging for the in vivo non-invasive monitoring of experimental infectious arthritis under antibiotic therapy. Methods and Materials: Unilateral knee infection was induced by intra-articular injection of Staphylococcus aureus in 12 rabbits randomly assigned in 2 groups. All animals were explored by MRI at the acute phase before and after USPIO administration. Animals of group 1 (n = 6) were then killed. Animals of group 2 (n = 6) received successful antibiotic treatment. Two MRI sessions were then performed, before and after UPSIO administration (n = 6) and animals were then killed. MRI protocol included unenhanced and USPIO-enhanced T2*GRE, unenhanced and gadolinium-enhanced FS T1 SE sequences. MR data were qualitatively and quantitatively analysed and compared with bacteriological and histopathologic findings. Results: At acute phase: an intense synovitis with marked signal increase of the synovium on Gd-enhanced FS T1 image was observed in all animals, associated with areas of signal loss within infected synovium on USPIO-enhanced T2*GRE images reflecting an intense infiltration of USPIO-loaded macrophages. At healing phase: significant reduction of areas of the synovium that presented with signal loss on T2*GRE images as compared to acute phase, due to a significant reduction of macrophage infiltration. At the opposite, signal increase on Gd-enhanced FS T1 image remained unchanged. Conclusion: In contrast to gadolinium-enhanced sequences, macrophage imaging using USPIO-enhanced MRI can accurately demonstrate resolution of experimental bacterial joint infection. Purpose: To evaluate WB-MRI using DWI as a new specific modality for the aetiologic workup of fever of unknown origin (FUO) without radiation exposure and less technical effort in comparison to scintigraphy/PET-CT. Methods and Materials: 20 patients with FUO were examined at 3/1.5 T (Siemens Magnetom Trio/Avanto) using a combined protocol with coronal WB diffusionweighted sequences (EPI-SE, 2.0 x 2.0 x 4.0 mm), axial WB TimCT HASTE STIR sequences (2.0 x 1.0 x 6.0 mm³) and ce T1-weighted GRE -sequences as reference. In 8 patients 67 Gallium-scintigraphy (Ecam, Siemens) was performed as standard of reference, in 5 patients 18 F-FDG-PET-CT (Biograph mCT, Siemens). In 3 adolescents ultrasound correlation was available. Scintigraphy/PET-CT and WB MRI were read by two blinded readers. Sensitivities/specificities were assessed. Results were compared, findings were clinically assured and modalities were reevaluated after diagnosis. Results: In 14 patients 17 fever foci/fever causing entities found were found by WB-MRI. In 6 patients no focus was found by WB-MRI and scintigraphy /PET-CT and clinical follow-up was negative. In 4 patients (20%) with positive MRI scintigraphy was negative, in these patients clinical assessment confirmed the MR diagnosis. Sensitivity of DWI was 94%, specificity 100%, PPV 100%, NPV 80%. In 3 patients negative DWI ruled out suspected infection from STIR and ce GRE images. Conclusion: WB-MRI excellently depicts fever foci and shows higher detection rates than the current standard of reference. DWI has a high sensitivity, specificity and PPV and offers additional functional information to standard morphological imaging.

A D E F G H There was no significant difference between the mean values of segment 5/6 and 7/8. There was no inter-and intra-observer variability (p=ns). Shear velocity weakly correlated between BMI (r=+0.05, p=0.04), and age (r=-0.09, p=0.002). A stronger correlation was observed with CB distance (r=-0.21, p=< 0.0001). Conclusion: With experienced radiologists, ARFI imaging is a robust and reproducible method for quantifying liver stiffness. It is reliable when taking shear velocity measurements from segments 5/6 or 7/8 but segment 3 should not be used. The influence of CB distance should be considered in future study design.

Haemodynamic changes occurring in liver cirrhosis: evaluation by a contrast-enhanced ultrasound dedicated software and potentially clinical applications T. Abbattista, F. Ridolfi, E. Brunelli, P. Busilacchi; Senigallia/IT (t.abbattista@tin.it)

Purpose: The early enhancement of hepatic veins (HVAT) distinguishes cirrhosis from chronic liver diseases. The aim of this study was to determine if a contrastenhanced ultrasound dedicated software (Quontrast, Bracco, Italy) could be used to give a map of haemodynamic changes of hepatic parenchyma and to differentiate liver cirrhosis from chronic hepatitis. Methods and Materials: CEUS (Sonovue, Bracco S.p.a., Milan, Italy) was performed on 17 patients with liver cirrhosis, 20 patients with chronic hepatitis B/C and 10 healthy controls. Qontrast was used to obtain maps of the enhancement of hepatic parenchyma, including a main hepatic vein. Post-processing enhancement intensity was analysed to determine TTP (time to reach peak of enhancement) and peak (enhancement peak quantity). Results: TTP was significantly reduced in patients affected by liver cirrhosis both in respect of patients with chronic hepatitis (69.2±12 vs. 81.3±15.3 seconds, p=0.01) and to control patients (86.3±20.3 seconds, p = 0.007). Peak was higher in patients with chronic hepatitis (25.5±8.7 with respect to 22.4±6.4 in patients with liver cirrhosis and 18.9±7.1 in control patients): this difference approched but did not reach the statistical significance (p = 0.055). Quontrast revealed a high reproducibility: inter and intraobserver kappa coefficients for TTPs were 0.77 and 0.85, respectively. Conclusion: Early TTPs seem to be predictive of liver cirrhosis and confirm the analysis of HVAT. This should be due to intra-hepatic shunts developed during chronic injury. The trend in reaching the highest peak of enhancement in chronic hepatitis should be due to wider vascular bed present in chronic liver diseases of viral origin.

Effects of contrast agent and outer volume saturation bands on water suppression and shimming of hepatic single-volume proton MR spectroscopy at 3.0 T L. Xu; Guangzhou/CN Purpose: To determine whether administration of gadolinium diethylenetriamine pentaacetic acid (Gd-DTPA) and placement of the outer volume saturation bands significantly affect shimming and water suppression on hepatic MRS prescan adjustments on a 3.0 T system.

Region of interest (ROI) of 2 cm×2 cm ×2 cm were carefully positioned in the region of the middle portion of the right hepatic lobe. 32 patients were examined before and after administration of Gd-DTPA with and without outer-volume saturation bands. Linewidths (FWHM) and water suppression were obtained. A paired t-test for comparison of means was used.

Results: (1) The group with the outer volume saturation bands has slightly better water suppression effect than the group without outer volume saturation bands before administration (94.0±2.4%, 93.2±2.8%, t=3.763, P=0.001).

(2) The group with the outer volume saturation bands has better water suppression effect than the group without outer volume saturation bands after administration (91.2±2.2%, 88.9±3.0%, t=10.811, P < 0.000).

(3) Both shimming (20.6±5.7 Hz, 19.5±5.8 Hz, t=-2.137, P=0.041) and water suppression effects (91.2±2.2%, 94.0±2.4%, t=8.649, P < 0.000) were decreased on enhanced MRS prescan adjustments. Conclusion: Placement of the outer volume saturation bands is helpful to improve water suppression both before and after contrast agent administration. Gd-DTPA exerts an adverse effect on water suppression and shimming of hepatic singlevolume proton MR spectroscopy at 3.0 T.

Uric acid crystal deposits were detected in 71 patients by DECT. Combing with the clinical manifestation, 50 cases were secured diagnosed and 21 suspected. The detection rate was 81.6%. 69 patients treated as gouty arthritis have showed positive response; the remaining two patients were confirmed by detecting uric acid in the joint effusion from surgery. The sensitivity was 100%. The false-positive rate was 6.7% in control group. The detection rate was higher in patients with increased blood uric acid than those with normal blood uric acid (P < 0.01). However, uric acid crystals could be detected in joints and related soft tissues with normal blood uric acid (23/35.66%). Conclusion: DECT technique has relatively higher sensitivity and specificity in detecting uric acid crystal in joints and related areas. It is quite useful in the diagnosis and differentiation of gouty arthritis.

1H MR spectroscopy; HR-MAS spectroscopy on biopsy specimen. Measures of association between LFC and histological steatosis percentage were calculated using univariate linear regressions and Pearson r coefficient. Respective ROC curves were used to compare specificity and sensibility of in-vivo MR methods in diagnosis of steatosis (cut-off value 5%) and in quantitative evaluation of steatosis severity (cut-off value 33%). Results: High association was found between LFC and histological steatosis percentage (fat-suppressed r = 0.80, p = 0.0005; dual phase r = 0.89, p < 0.0001; multiecho r = 0.94, p < 0.0001; spectroscopy r = 0.79, p = 0.002; HR-MAS r = 0.71, p = 0.006). According to the ROC curves, in-vivo MR techniques had high sensitivity and specificity in diagnosis and in quantitative evaluation of steatosis severity (area under the curves from 0.81 to 0.97), but no statistically significant difference was found between the four techniques (p = 0.39 and 0.62). Conclusion: MR imaging can be considered a valid non-invasive alternative to biopsy for the diagnosis of steatosis and quantification of liver fat content. In addition, the small sample size may be responsible for the lack of significant difference between MRI techniques, but Multiecho seems to be the most promising.

In vivo 1 H MR spectroscopy in the early diagnosis and staging of liver fibrosis induced by viral hepatitis B W. Li, N. Zhu, X. Zhao, K. Chen, W. Chai; Shanghai/ CN (lwxyjck@126.com) Purpose: To explore the value of 1 H MRS for the early diagnosis and staging of liver fibrosis induced by viral hepatitis B. Methods and Materials: 24 subjects underwent 1 H-MRS of liver. Among them, 5 subjects were volunteers without hepatic disease and the other 19 subjects were patients with liver fibrosis who were further divided into two groups according to the Ishak system and whether it should be treated: 12 cases were slight liver fibrosis (F0~2) and the other 7 cases were advanced liver fibrosis (F3~6), the area under the peak of choline (Cho), lipid (Lip) and the relative Cho-to-Lip ratio (Cho/Lip) were calculated by GE spectra analysis software.

The areas under the peaks of Cho, Lip and the ratio of Cho/Lip tended to increase gradually with the increasing degree of liver fibrosis (Cho: 4.99 × 10 -3 , 29.86 × 10 -3 , 233.52 × 10 -3 ; Lip: 45.05 × 10 -3 , 328.26 × 10 -3 , 406.62 × 10 -3 ; Cho/ Lip: 0.10, 0.23, 0.51). There were significant correlations between Cho, Lip, Cho/ Lip and the degree of liver fibrosis (P < 0.05). The AUC of Cho, Lip and Cho/Lip for diagnosis of slight liver fibrosis were 0.95, 0.83 and 0.75, their specificities were 80%, 100% and 80%, sensitivities were 100%, 75% and 83.3% respectively. The AUC of Cho, Lip and Cho/Lip for diagnosis of advanced liver fibrosis were 0.87, 0.61 and 0.80, their specificities were 100%, 33.3% and 66.7%, sensitivities were 71.4%, 100% and 85.7%, respectively. Conclusion: The metabolite (the area under the peak of Cho, Lip and the ratio of Cho/Lip) analysed with 1 H-MRS is feasible, non-invasive makers in the early diagnosis and staging of liver fibrosis.

In-phase and out-of-phase gradient-echo imaging in abdominal studies: intra-individual comparison of three different techniques M. Ramalho 1 , V. Heredia 1 , R.O.P. de Campos 1 , R.M. Azevedo 1 , B.M. Dale 2 , R.C. Semelka 1 ; 1 Chapel Hill, NC/US, 2 Morrisville, NC/US (miguel-ramalho@netcabo.pt) Purpose: To compare three T1-weighted in-phase and out-of-phase (IP/OP) gradient-echo imaging techniques in an intra-individual fashion, and to determine whether advantages exist for each of these sequences for various patient types. Methods and Materials: Institutional review board approval was obtained for this retrospective HIPAA-compliant study with waived informed consent. 118 consecutive subjects (74 males, 44 females; mean age 53.9 ± 13.8) who underwent a standard abdominal MRI protocol at 1.5 T, between January and February 2010, including all three different IP/OP sequences [two-dimensional spoiled gradient-echo (2D-SGE), three-dimensional gradient-echo (3D-GRE) and magnetisation-prepared gradient-recall echo (MP-GRE)] were included. Two different reviewers independently and blindly evaluated IP/OP sequences to determine image quality, extent of artifacts, lesion detectability and conspicuity, and subjective grading of liver steatosis for the various sequences. Qualitative and quantitative data were subjected to statistical analysis. Results: Respiratory ghosting, parallel imaging and truncation artefacts as well as shading and blurring were more pronounced with 3D-GRE IP/OP imaging. Overall image quality was higher with 2D-SGE (p < 0.05). Detectability of low-fluid content lesions was lower with IP/OP MP-GRE sequences. MP-GRE sequences had the lowest SNRs (p < 0.001). Liver-to-spleen and liver-to-lesion CNRs were significantly lower with 3D-GRE and MP-GR, respectively (p < 0.001). Fat liver indexes showed strongly positive correlation between all sequences.

Multiecho MR sequences and high-resolution magic angle spinning (HR-MAS) ex-vivo spectroscopy in the qualitative analysis and differentiation between steatohepatitis and steatosis G. Besutti, G. Ligabue, L. Nocetti, F. Fiocchi, C. Stentarelli, A. Mucci, P. Loria, G. Guaraldi, P. Torricelli; Modena/IT (gligabue@sirm.org) Purpose: To compare multiecho gradient-echo MR (magnetic resonance) sequences in the differentiation between steatohepatitis and steatosis and to describe HR-MAS spectra of liver biopsy showing steatohepatitis or steatosis. Methods and Materials: Fourteen patients with indication for biopsy assessment of steatosis underwent liver biopsy (reference standard) and MR imaging. Liver biopsy of both viral and metabolic steatosis were classified using NAFLD activity score (NAS) which depicts the degree of necro-inflammatrory activity allowing to differentiate between steatohepatitis and steatosis. Besides liver fat content (LFC), multiecho sequences were also used to calculate water and fat relaxation times (T2*), which are influenced by microenvironmental characteristics, so potentially associated with necro-inflammatory activity. Relation between each multiecho parameter (LFC/T2*water/T2*fat) and NAS was estimated using univariate linear regression and Pearson coefficient. A fragment of biopsy specimen was analysed through HR-MAS to obtain metabolic tissue characterisation. Results: Association was found between: NAS and LFCmulti (r = 0.7; p = 0.006), NAS and T2*fat (r = -0.73, p = 0.063, ns, T2*fat was available for 7 patients only). No correlation was found between NAS and T2*water. HR-MAS spectra showed tissue metabolic heterogeneity, with particular regard to the contents of free glucose, alanine, glutamine/glutamate and phospholipids. Conclusion: This pilot study describes multiecho parameters associated with histological necro-inflammatory activity, allowing to study the potential capability of MR to differentiate between steatohepatitis and steatosis. Description of HR-MAS spectral heterogeneity in NAFLD and NASH may allow to find biochemical indicators of steatosis progression to be used in differentiating between steatohepatitis and steatosis in spectra acquired with in vivo MR Spectroscopy too.

Accurate quantification of liver fat by H-MR spectroscopy and its correlation with risk of coronary heart disease and diabetes mellitus A.C. Dhapare, A. Kohli, D. Rajput, M. Singh; Mumbai/ IN (amoldhapare@yahoo.com) Purpose: To evaluate the effectiveness of MR spectroscopy as a non-invasive tool for accurate detection of fat in liver and also to evaluate the risk of coronary heart disease and diabetes mellitus in individuals with fatty liver. Methods and Materials: Spin echo sngle voxel proton MR spectroscopy with TR 2000 ms and TE 30 ms with a voxel size of 20 x 15 x 15 mm without breath holding is performed. Water peak is seen at 4.7 ppm and fat peaks between 0.9 to 2.1 ppm. Percentage of fat is calculated from the area under the peaks. Around 80 healthy subjects in age group 25-50 years were scanned on 3 T Verio Siemens machine. Subjects with liver fat more than 5% were advised complete lipid profile along with fasting sugar and fasting insulin blood levels to calculate insulin resistance. Presence of obesity was also evaluated by measuring the waist circumference and BMI. Results: MR spectroscopy was accurate in measuring liver fat content. The startling finding was high incidence of hyperlipidaemia and insulin resistance in non obese subjects with fatty liver as compared with obese individuals without fatty liver with high incidence even in young individuals. Conclusion: There is high incidence of fatty liver in general population with rising incidence in younger age group. It is emerging as a single most important risk factor for development of coronary heart disease and diabetes mellitus. Thus, MR spectroscopy proves to be an effective non invasive modality for early detection of fatty liver and preventing its complications. Purpose: To determine the radiologic evolution in patients with H1N1 infection and findings associated with admission on the intensive care unit (ICU) and development of acute respiratory distress syndrome (ARDS). Methods and Materials: 104 patients (53 males, 51 females; median age 40 years, range 15-96 years) with microbiologically confirmed H1N1 infection seen from July to December 2009 who underwent chest radiographs at the emergency department were studied. Radiographs were evaluated for the pattern (alveolar, interstitial, alveolointerstitial), distribution and extent (focal, multifocal, and diffuse), and findings of ARDS. Radiographs were reviewed by 2 observers who reached a consensus decision. Results: 42 patients had abnormal initial chest radiograph and at least one more radiograph was obtained. Abnormalities improved and resolved in most patients in the first 3 weeks. Progression of findings was observed within the first 48 hours in 15 cases (14.4%). 17 patients (16.3%) required admission in ICU and mechanical ventilation. All of them showed an abnormal initial radiograph (16 alveolar o alveolo-intersticial and 1 interstitial pattern). Evolution to ARDS was found in 8 cases (6.7%). All of them had on the initial radiograph alveolar multifocal involvement (p = 0.001) with bilateral lesions in 4 cases.

In the multivariant analysis, a higher number of lung zones involved, the alveolar pattern and quickly progression were identified in patients requiring ICU admission and development ARDS.

The risk is higher for those showing alveolar pattern, a higher number of lung zones involved on the initial radiograph and quickly initial progression. Purpose: To retrospectively describe initial thoracic CT and radiographic findings of H1N1 influenza virus in immunocompromised patients and to determine whether initial radiologic findings can help triage patients who are at risk for an adverse clinical outcome.

The study population consisted of 33 patients with confirmed H1N1 infection suffering from an underlying immunocompromising disease who underwent initial chest radiographs. Three radiologists evaluated all initial radiographs (n = 33) and CT scans (n = 14) for type and pattern of opacities, distribution and extent of the observed abnormalities. Adverse outcome measures were defined as the need for mechanical ventilation, extracorporeal membrane oxygenation or death. Results: The initial radiograph was abnormal in 19 of 33 and the corresponding CT in 13 of 14 of the patients. 10 patients showed an adverse outcome resulting in a death rate of 50% (n = 5). 14 patients revealed no abnormal radiographic findings. None of them had an adverse outcome. The most common radiographic findings included the following: consolidation (74%), ground-glass opacity (63%), reticular pattern (32%), tree-in-bud pattern (31%). The distribution was bilateral in 58% and affected predominantely the lower lung zones. A bilateral combination of consolidation and peribronchovascular ground-glass opacities affecting more than three lung zones was most predictive for a severe clinical course.

The majority of immunocompromised patients with pulmonary manifestation of H1N1 infection reveal abnormal radiologic findings. Early presumption of an adverse clinical outcome can be made by initial radiography and MDCT scans depicting multizonal bilateral consolidations and ground-glass opacities. Purpose: To investigate DWI as tool to detect and quantify inflammation in CF lung disease. Methods and Materials: 33 patients had a chest CT and MRI, including a morphological (T2w BLADE: TR/TE/alpha/TA: ∞/28 ms/180º/18 s; slice thickness: 4 mm) and DWI (TR/TE/FA: 5632, 12/83 ms/90º; slice thickness 5 mm) study on the same day. CT and morphological MRI scored using a validated CF-CT score and equiva-Conclusion: All sequences resulted in adequate IP and OP information and image quality. 2D-SGE remains the best approach for IP/OP imaging. The good image quality of MP-GRE sequences acquired in a free breathing manner should recommend its use in patients unable to suspend breathing.

Diagnostic accuracy of post-mortem whole-body magnetic resonance imaging in comparison with conventional autopsy F. Henes, K. Püschel, G. Adam, M. Lorenzen; Hamburg/ DE (f.henes@uke.uni-hamburg.de) Purpose: Evaluation of diagnostic performance of whole-body magnetic resonance imaging (MRI) for detection of pathologic findings and prediction of causes of death as an alternative to conventional autopsy (CA).

In a prospective cohort study, thirty consecutive deceased patients underwent whole-body MRI (Philips Achieva 1.5-T, Best, Netherlands) prior to CA. In each cadaver coronal images of head and neck, thorax, abdomen, pelvis and lower extremities were acquired using a 3D-GRE-and a T2-STIR-sequence (TE: 3.3 ms, TR: 93 ms, FOV 400 and TE: 70 ms, TR: 12.6 s, IR: 190 ms; FOV 400, respectively). Subsequently, conventional autopsy was performed within 24 h. MR images were evaluated by two radiologist in consensus. Pathologic findings were documented and causes of death were predicted for each patient. Finally, results were correlated with autopsy findings.

In 25 patients (83%) MRI and CA were in agreement on the cause of death. In particular, in the detection of cardiac causes of death, post-mortem MRI achieved a high diagnostic accuracy (sensitiviy 84%, specificity 100%). MRI failed to demonstrate causes of death in 5 patients (1 pneumonia, 1 gastric ulcera, 1 alcohol intoxication, 2 myocardial infarction). With regard to the secondary findings a high finding agreement, > 95% was observed for pleural and pericardial effusion, ascites and rip fractures. Finding agreement for pulmonary emphysema, pulmonary oedema and renal infarction emerged to be poor (23%). Conclusion: Post-mortem whole-body magnetic resonance imaging achieves a high percentage of major finding agreement in comparison to conventional autopsy. Purpose: To assess the initial radiographic findings in patients with H1N1 infection and compare them with those in patients with seasonal influenza. Methods and Materials: 95 patients (50 males, 45 females) with microbiologically confirmed H1N1 infection seen from July to November 2009 who underwent chest radiographs at the emergency department formed the study group. The control group was formed by 98 patients (65 males, 33 females) with suspected seasonal flu seen from January to February 2009 at the same department. The initial radiographs were evaluated for the pattern (alveolar, interstitial, and alveolointerstitial), distribution and extent (focal, multifocal, and diffuse as the volumen of one lung). Radiographs were reviewed independently by 2 observers who reached a consensus decision.

The study group showed a lower mean age (40.2 vs 50.9 years; p < 0.001) and more underlying comorbidities (48 vs 35 patients; p < 0.001). The initial radiograph was abnormal in 52 (54.7%) patients in the study group and 41 (41.8%) in the control group. The most common finding in the study group was multifocal patchy consolidation (41.2%, p < 0.001) whereas in the control group was lobar consolidation (43.9%). Fourteen patients had involvement of more than 2 lobes and 5 patients showed diffuse involvement in the study group and none of the patients in the control group had more than 2 lobes affected. Conclusion: Radiographs reveal abnormalities in more than half of the patients with H1N1 infection requiring admission. Multifocal patchy alveolar consolidation is the most frequent pattern in contrast with seasonal influenza.

A S273 C D E F G H Conclusion: Automatic airway detection and segmentation succeeded to measure bronchial wall thickening and wider bronchial diameters in CF as compared to NORMAL. Moreover, a lack of bronchial tapering could be detected for distal bronchi in CF. Automatic computational results might be used for quantitative monitoring of the disease instead of manual scoring.

Radiologic manifestations of tuberculosis in HIV-positive patients B. Saidi, M. Bakhshayesh-Karam, P. Tabarsi, S. Zahirifard, M. Ghofrani; Tehran/IR (bahare_saidi@yahoo.com) Purpose: The chest presentation of tuberculosis in HIV positive differs from that of immuno-competent patients.The association of CD4 count and radiologic pattern has been previously investigated. The aim of this study was to review the pulmonary manifestations of tuberculosis in HIV-positive patients at our centre. Methods and Materials: The radiographs of 52 male HIV-seropositive patients (mean age 39.9±9.5) with sputum or bronchoalveolar lavage, culture or PCR proven tuberculosis who were referred to our center from April 2005 to January 2010 were studied. The CD4 count and ADA levels were available for 43 patients. The radiographs were reviewed regarding pattern 1. postprimary 2. the patterns not characteristic of postprimary: pleural effusion, hilar or mediastinal lymphadenopathy, middle lobe or basal lower lobe involvement (atypical pattern) 3.miliary. The patterns were correlated with CD4 level. Results: Seven (13.5%) patients died. Thirty-six (70.6%) patients had CD4 count less than 200. 19 (36.5%) had post-primary pattern tuberculosis on x-ray, 23 (44.2%) had atypical and 10 (19.2%) had miliary pattern. 10 (19.2%) had cavity, 15 (28.8%) had bronchiectasis, 5 (9.6%) had atelectasis, one had pneumothorax and one hydropneumothorax.12 (23.1%) had pleural effusion, 14 (26.9%) had hilar or mediastinal adenopathy. The mean CD4 count in the post-primary pattern was 176.8±193; in the atypical pattern 137.05±133 and miliary pattern 46.56±55, the difference was statistically significant between the groups (p < 0.05). Conclusion: The atypical pattern and miliary tuberculosis are common observations in HIV-positive patients. The CD4 counts were generally low in our patients; however, the mean CD4 count was significantly lower in the miliary pattern. Patients with MAC infection were older. M. xenopi patients had a higher rate of any malignancy (58% versus 11%, p = 0.002) and pulmonary malignancy (25% versus 5%, p = 0.08). Patients with M. xenopi more commonly had cavities (46% versus 16%, p = 0.02), ≤5 mm nodules (88% versus 58%, p = 0.02), and > 5 mm nodules (67% versus 46%, p = 0.16). M. xenopi disease also more often demonstrated the so-called "classic fibrocavitary type" of radiological pattern (88% versus 21%, p = 0.0002). MAC was more often associated with bronchiectasis (60% MAC vs 33% M. xenopi, p = 0.057) and a nodular-bronchiectatic pattern (79% MAC vs 11% M. xenopi, p = 0.0002). Emphysema was more prevalent in patients with M. xenopi (50% versus 20%, p = 0.02). Conclusion: Compared with MAC, patients with M. xenopi tend to be younger and more often have malignancies and emphysema.m. xenopi is more often associated with cavities, a "fibro-cavitary" CT pattern and nodules. MAC is more often associated with bronchiectasis and a "nodular-bronchiectatic" CT pattern. lent CF-MRI scoring system; DWI scored using a newly developed score. Overlap between DWI and morphological MRI assessed with overlay techniques. DWI score was correlated to clinical and radiological parameters of disease severity: FEV1, body mass index (BMI), CT and MR scores and CT bronchiectasis score (CT-BE). Patients with DWI hot spots were compared to those without. Results expressed as mean (range). Scores expressed as % of the maximal score. Results: Mean age 24.6 years (6 to51); FEV1 78.64 (38 to 113); DWI score 9.89 (0 to 37); CT scores 22.88 (3 to 49); MRI scores 20.65 (7 to 40). Total DWI score correlated to CT score (r = 0.66, p = 0.0001), MRI score (r = 0.599, p = 0.001) and CT-BE score (r = 0.615, p = 0.001). FEV1 of patients without DWI hotspots was higher (p < 0.0001) compared to patients with hot spots and DWI score correlated negatively with FEV1 (r = -0.641; p < 0.0001). No correlation was found between BMI and DWI score. Diffusion pattern only in part overlapped structural abnormalities on morphological MRI or CT. Conclusion: DWI could be used to localise and quantify lung inflammation in CF. DWI is able to distinguish between hot and cold spots, that only partly overlap abnormalities on MRI and CT.

Comparison of chest-MRI to chest-CT to monitor cystic fibrosis (CF) lung disease P. Ciet 1 , G. Serra 2 , S. Purpose: The most important components of CF lung disease are bronchiectasis (BE) and trapped air (TA). CT is the current gold standard to diagnose and monitor BE and TA. MRI has been suggested as a radiation free alternative for CT, but its spatial resolution is inferior to CT. The purpose of our study was to compare the sensitivity of CT vs. MRI to monitor BE and TA in CF. Methods and Materials: 39 stable CF patients (20 female; mean age 16.7 years) had a chest-CT and -MRI performed on the same day. MRI (Siemens Avanto): BLADE proton density (PD) transversal (TR/TE/alpha ∞/28 ms/180º). CT (Siemens Somatom): low dose; volumetric; end inspiration and end expiration. CTs and MRIs were anonymised and scored in random order by 2 independent observers using the CF-CT score and an equivalent CF-MRI score. Scores are expressed as % of the maximal score. Statistics: Pearson; intra-class coefficient (ICC); Bland-Altman plots; results mean (range). Purpose: Major findings in multi-detector computed tomograms (MD-CT) of the chest of patients suffering from cystic fibrosis (CF) are bronchial wall thickening, mucous obstruction, and bronchiectasis. Aim of this study was to evaluate fully automatic airway detection and quantitative analysis of bronchial wall thickness on thin-section MD-CT in patients with CF compared to a control group without airway changes scanned for other reasons (NORMAL). Methods and Materials: MD-CT (4 x 1.25 mm, 70 mAs, 120 kV) acquired from 22 CF patients and 33 NORMAL were subjected to a self-written airway analysis tool (YACTA). Automatic segmentation of the whole bronchial tree was performed and total bronchial diameter (TD), wall thickness (WT), and percentage of wall thickness of TD (WP) were calculated for each bronchial generation separately. Results: In CF patients, a mean of 768 measurements of the segmented bronchial tree compared to 373 in NORMAL group could be performed (1 st -14 th generation, p < 0.0001). WT was significantly increased in CF from 2 nd -9 th generation bronchi (mean 2.1-1.1 mm in CF vs. 1.8-0.7 mm in NORMAL, p < 0.0001). TD was larger in 5 th -9 th generation bronchi in CF (mean 8.8-7.0 mm vs. 7.3-6.4 mm, p < 0.05-0.0001). Consecutively, WP was increased in CF from 2 nd -9 th generation (mean ranges 41-61% vs. 37-47%, p < 0.05-0.0001). Purpose: To evaluate technical feasibility and outcome of stent placement in sub-/ acute complicated Stanford type B dissection. Methods and Materials: 12 patients (1 female, range 44-71 years) with aortic dissection suffering from severe gastrointestinal malperfusion and claudication underwent uncovered stent placement (diameter 7-25 mm, length 40-100 mm) into the aorta to achieve sufficient perfusion. Additional stents were placed into the visceral arteries in 3 patients. Results: Thoracoabdominal stent placement yielded to considerable clinical improvement in 11 of 12 patients. Additional stents (up to four) were placed in seven patients (1 celiac, 1 mesenteric, 2 renal, 6 iliac). Follow-up CTA showed near complete collapse of four stents (diameter 9-25 mm, length 100 mm) after one week. Catheterisation and balloon dilatation of all four stents were possible. But all cases showed re-collapse on follow-up CTA. However, duplex-ultrasound and CTA showed perfusion of the superior mesenteric artery in all patients. Follow-up ranged from 3 months up to 3 years (average 1.7 years). During this time, one patient required an iliaco-mesenteric bypass for treatment of chronic mesenteric ischaemia, one month after stent placement. All other patients did well without further interventional or surgical therapy. Conclusion: Thoracoabdominal stent placement in sub-/acute complicated Stanford type B dissection is technically feasible and can be considered as a minimal-invasive therapy with similar clinical results compared to the placement of an endoprosthesis. Stent size should be adapted to the diameter of the true aortic lumen to avoid stent collapse, although even collapsed stents maintained sufficient perfusion.

Reliability and interobserver variability of preoperative measurements for thoracic endovascular aortic repair: semiautomatic centerline analysis versus manual measurement techniques Purpose: Reliable preoperative measurements are crucial for successful thoracic endovascular aortic repair (TEVAR). The purpose of this study was to test whether reliability and interobserver variability of preoperative measurements for TEVAR are improved by means of semiautomatic centerline analysis compared to manual assessment on axial slices and double oblique multiplanar reformations (MPR). Methods and Materials: Preoperative CT angiographies of 30 patients with thoracic aortic disease (mean age 66.8 ± 11.6 years, 23 men) were analysed by one blinded vascular expert (reference standard) and three blinded non-experts. Maximum aortic diameters were measured at four positions relevant to TEVAR using three measurement techniques (manual axial slices, manual MPR, semiautomatic centerline analysis). Reliability was calculated as the absolute measurement deviation (AMD) from the reference standard and interobserver variability as the coefficient of variance (CV). Results: Mean AMD was 7.3 ± 7.7% (axial), 6.7 ± 4.5% (MPR), and 4.7 ± 4.8% (centerline). Mean CV was 5.2 ± 4.2% (axial), 5.8 ± 4.8% (MPR), and 3.9 ± 5.4% (centerline). Both AMD and CV were significantly lower for centerline analysis compared to axial technique (p = 0.001/0.042) and MPR (p = 0.009/0.003). Deviations from reference standard were above 1 mm in 66% (axial), 64% (MPR), and 42% (centerline). A stent graft diameter with ≥4 mm difference from the reference standard was chosen in 20% based on axial technique or MPR, and only in 8% based on centerline analysis. Conclusion: Semiautomatic centerline analysis provides the most reliable and least variable diameter measurements in TEVAR patients. MPR should be performed by vascular experts due to significantly lower reliability and significantly higher interobserver variability among non-experts.

Predictive factors for endoleaks after thoracic aortic aneurysm endograft repair A. Di Massa, M. Colosimo, E. Cotta, M. Mangini, G. Piffaretti, G. Carrafiello, C. Fugazzola; Varese/IT Purpose: Our prospective investigation aimed to determine and analyse the incidence and the determinants of endoleaks after thoracic stent graft. Methods and Materials: Sixty-one patients affected by thoracic aortic aneurysms were treated. The study cohort contained 54 men, with a mean age of 63.6 ± 17.9

Pleuropulmonary paragonimiasis: why is this disease delayed or misdiagnosed on CT? S.K. Kim, G.Y. Jin; Jeonju-si/KP (inacien7777@paran.com)

Purpose: We investigated CT findings of pleuropulmonary paragonimiasis and evaluated the cause of delay or misdiagnosis of paragonimiasis.

This study was performed prospectively with informed questionnaire from January 2009 to June 2009 in four tertiary hospital. Electronical survey with several topics (personal information, method of diagnosis, kinds of CT machine, CT findings, extrapulmonary diagnosis, first impression) was done in each hospital for the patients who were diagnosed as pleuropulmonary paragonimiasis from 2003 to 2008. The data were collected and analysed at one hospital. Results: 103 patients (58 men and 45 women; age 46.1 ± 14.6) from four hospitals were enrolled. The main lesion was at upper lobe in 58%, lower lobe in 35% and right middle lobe in 7%, with peripheral distribution in 78%, central distribution in 14%, and both in 8%. CT findings were non-cavitary nodule in 32.9%, cavitary nodule in 23.4%, pneumonic consolidation in 26.5%, lung mass in 8.5%, linear density in 5.3% and others in 3.2%. Other superimposed findings were bronchiectasis in 53%, cyst in 29%, bronchial wall thickening in 15% and mucoid impaction in 3%. Worm tract sign, which is known as typical finding of pleuropulmonary paragonimiasis, was only present in 18.5%. First impression on CT was paragonimiasis in 46%, pneumonia in 18%, tuberculosis in 14%, lung cancer in 7%, and others in 15%. The diagnostic accuracy of CT was 43.7%. Conclusion: Diagnosis of pleuropulmonary paragonimiasis is often delayed or missed especially when the lesion is at upper lobe, which occasionally mimics tuberculosis or malignancy.

A S275 C D E F G H time for lysis was 41 hours. Median total dose of thrombolytic drug was 44 mg rtPA (26-79). In three patients an iliac vein was stented to remove a flow obstructing lesion and to avoid recurrent thrombosis. During 7 months of follow-up early rethrombosis occurred in 4 patients. One patient developed a haematoma in the calf which was treated conservatively. Conclusion: Pharmaco-mechanical lysis using the EKOS ® EndoWave™ system appears to be a safe and viable alternative option for treating caval and ileo-femoral DVT, warranting further studies.

Experience with superior vena cava filters F. Pollice 1 , P. Pollice 1 , R. Muller 2 ; 1 Andria/IT, 2 Paris/IT (francesco.pollice@tiscali.it) Purpose: Therapy to prevent pulmonary embolism (PE) resulting from upper extremity deep venous thrombosis (UEDVT) remains controversial despite an increasing incidence of DVT of upper extremity origin. The purpose of this study was to evaluate the results of 72 superior vena cava Greenfield filters (SVC-GFs) placed at risk for PE arising from UEDVT. Methods and Materials: During the past 78 months, we placed SVC-GFs in 72 patients with UEDVT in whom anticoagulation was either deemed contraindicated (n = 67)or proved ineffective in preventing recurrent PE (4=) or extension of the thrombus (n = 1).There were 25 male (35%) and 47 (65%) female patients whose ages ranged from 25 to 99 years. Follow-up ranged from 10 days to 78 months. Sequential chest radiographs revealed no filter migration or displacement in 26 patients. Results: Thirty-four patients died in the hospital of causes unrelated to the SVC filter or recurrent thromboembolism (mean time to death, 20 days). Follow-up of the surviving 38 patients ranged from 1 month to 78 months (mean 22 months); none of these patients were seen with any evidence of PE. One SVC-GF was incorrectly discharged into the innominate vein and left in place. This vein remains patent 2 months after insertion without evidence of filter migration. Conclusion: We think that insertion of SVC-GFs is a safe, efficacious, and feasible therapy and may prevent recurrent thromboembolism in patients with UEDVT who are resistant to anticoagulation or have contraindications to anticoagulation.

Concomitant use of endovenous laser and foamed sclerosant for treatment of lower limb varicosities M. Bhalla, N. Bhalla; Ahmedabad/ IN (namratamanav@hotmail.com) Purpose: To report mid-term results (upto 3 years) of concomitant use of two modes of treating varicose veins -endovenous laser treatment for Great Saphenous Vein (GSV) reflux and sclerotherapy for incompetent perforators and accessory venous channels. Methods and Materials: 253 lower limbs (76 Left,43 Right,134 billateral) of 186 patients (66 women,120 men;mean age 45.5years) were treated with concomitant use of endovenous laser and sclerotherapy. 5 patients (18 legs;7%) had ulcers. GSV was mapped using 10MHZ linear Ultrasound (US) probe. Following local anaesthesia, 0.035 inch J-Guidewire was introduced in GSV from lowest point of reflux. A 5 F sheath and 610μm laser fibre was used. Laser energy was delivered intraluminally at 14 W continous mode. Fibre was uniformly withdrawn. Sclerosant, mixed with air to produce foam, was introduced through needle in perforators and additional venous channels under US guidance. All patients were advised ibuprofen and class II stockings. Follow-up was done at 1 week, 3.6,12 ,24 and 36 months. Results: 98% laser-treated GSV remained ablated at 3-year follow-up. 7 % limbs in first year and additional 3% limbs in next 2 years (i.e. 10 % limbs in 3 years) required one more sclerotherapy session for new incompetent perforators. No complications were noted.90% patients complained of tightness along the course of treated GSV. Patients resumed routine activities immediately. Conclusion: Laser has already proven its use in treating GSV. Sclerotherapy is effective in dealing with perforators and provides rapid healing of chronic venous ulcers. Concomitant use of both modes limits the need of subsequent procedure.

A comparison of standard dual-tip and split lumen heamodialysis catheters H. Lee, S. Park, Y.-I. Jo, I. Chang, I. Yun; Seoul/ KR (20080136@kuh.ac.kr) Purpose: To compare a newer split lumen haemodialysis catheter with the standard dual-tip haemodialysis catheter. Methods and Materials: From August 2005 to December 2009, among patients who were referred for tunneled haemodialysis catheters, the patients who underwent standard dual-tip haemodialysis catheter (DTC) insertion or split lumen years.The follow-up imaging protocol included chest radiographs and triple-phase computed tomographic angiography performed at 1, 4 and 12 postoperative months and annually thereafter. Results: Median follow-up was 32.4 months (range: 1-96 months). Endoleaks were detected in 9 (14.7%) patients, of which 7 were type 1. Five endoleaks were detected at 30 postoperative days, and the other 4 developed with a mean delay of 12 months. Endovascular or hybrid interventions were used to treat the endoleaks. Secondary technical success rate was 100%. Multivariate analysis demonstrated that the diameter of the aneurysmal aorta (odds ratio 1.75.95%confidence interval 1.07-2.86) and the coverage of the left subclavian artery (odds ratio 12.05, 95% confidence interval 1.28-113.30) were independently associated with endoleak development. The percentages of patients in whom reinterventions were unnecessary were 94.6% ± 3.0%, 88.3% ± 4.5%, and 85.4% ± 5.2%, at 1, 2, and 5 years, respectively. The actuarial survival estimates at 1, 2, and 5 years were 85.2% ± 4.6%, 78.1% ± 5.4%, and 70.6% ± 6.4%, respectively. Conclusion: The diameter of the aneurysmal aorta and the position of the landing zone are independent predictors of endoleak occurrence after thoracic stent-graft procedures. A careful follow-up program should be considered in patients in whom these indices are unfavourable, because most of the endoleaks may be successfully and promptly treated by additional endovascular procedures.

Stenting as an effective endovascular treatment of superior vena cava syndrome: review of hundred and fifty-four cases. Single-centre experience S. Kudrnova, B. Nemes, E. Rimely, K. Hüttl; Budapest/HU (skudrnova@gmail.com) Purpose: VCSS constitutes a life-threatening clinical manifestation of central vein obstruction mainly due to external compression by malignancies. Prompt restoration of VCS patency is required to enable diagnostic and therapeutical procedures ameliorating and prolonging patients` lives. VCS stenting represents an effective alternative to insufficient or high-risk best medical treatment, irradiation or opensurgery. We evaluate its technical success, clinical effectiveness, patency rate and patients´ overall survival period in a large group of patients. Methods and Materials: 154 patients (age range 23-87 years) underwent endovascular treatment for SVCS in our centre between November 2002 and May 2009 (malignant etiology in 143 cases, benign in 11). SVC was primarily stented in 145, balloon dilated in 1 and interventional attempt was unsuccessful in 8 cases. Patients were invited back upon recurrence of symptoms. Results: All successful interventions led to immediate relief of symptoms. No periprocedural mortality occurred, but 1 SVC rupture, 1 stent fracture and 1 haemopericardium. Reocclusion occurred in 11 cases (including 5 and 3 times in a single patient), within a time range of 2 days-2.5 months (thrombolysis reintervention). Restenosis was reported in 7 cases (including 5 and 2 times in a single patient), within a time range of 2 weeks-9 months (stent-in-stent reintervention). Survival time follow-up could be obtained in 94 (64%) patients so far (1 week Survival Rate 93%, 1 month SR 83%, 6 months SR 45%, 1 year SR 10 %, 3 year SR 3%). Conclusion: Stenting has in our centre proved to be a highly technically successful, clinically effective, low-morbidity, low-mortality, low-restenosis treatment option for VCSS.

Catheter-directed thrombolysis in caval thrombosis with the EKOS catheter: preliminary results A.H. Mahnken 1 , J. Grommes 1 , C. Plumhans 1 , S. Langer 1 , C. Wittens 2 ; 1 Aachen/DE, 2 Maastricht/NL (mahnken@rad.rwth-aachen.de) Purpose: Currently, one in four patients with primary iliofemoral deep vein thrombosis (DVT) develops a post thrombotic syndrome within one year. Pharmacomechanical thrombolysis is a viable option in the treatment of DVT. The aim of this study was to access the feasibility of new pharmaco-mechanical thrombolysis system in DVT with involvement in inferior vena cava. Methods and Materials: 12 patients (7 males, mean age 44 [5-79] years) with severe ilio-femoral and caval thrombosis were treated using the EKOS ® EndoWave™ Peripheral Infusion System. The system combines a proprietary multiple-side-hole drug infusion catheter with a guidewire-exchangeable ultrasound core for simultaneously infusing therapeutic agents and delivering high-frequency, low-power ultrasound energy. Lysis was initiated with a bolus of 2.5 mg rtPA followed a continuous infusion of 1 mg trPA/h. All patients were treated with an additional continuous intravenous infusion of 1000 U Heparin/h. Results: Complete recanalisation was achieved in 10/13 patients. In one patient partial recanalisation and in two patients no recanalisation was achieved. Median Purpose: In breast cancer accurate and non-invasive assessment of prognosis might potentially impact both patient management and therapeutic approach. This observational-unicenter study was conducted to identify the potential of magnetic-resonance-mammography (MRM) to predict disease-related death in primary breast cancer. Methods and Materials: In this IRB-approved investigation, standardised protocols and study design were applied according to international guidelines (T1w-FLASH; 0.1 mmol/kg BW Gd-DTPA; T2w-TSE; 1.5-Tesla). All patients with invasive breast cancer and subsequent surgicopathological verification (no secondary/recurrent breast-cancer; study-duration: 24 consecutive months) were clinically followed up. Interval between initial staging-MRM and last follow-up, as well as occurrence of "disease-related death" were documented. For quantitative analysis of tumour-vasculature dedicated computer-assisted-diagnosis software was used. Pharmacokinetic parameters were automatically calculated for every tumour on a voxel by voxel basis and correlated with occurrence and time interval to "diseaserelated death" (Cox proportional-hazards-regression: CR). Results: 839 patients were enrolled for MRM imaging. 122 were eligible for CR analysis. Median follow-up time was 52 months. In 15/122 patients "disease-related death" occurred. Survival time analysis by CR was possible with 4 CAD-parameters on a high level of significance (P < 0.001), demonstrating appropriate overall model fit (2 Log-Likelihood: 138.2; Chi-square: 37.7). haemodialysis catheter (SLC) insertion were enrolled. During haemodialysis with those haemodialysis catheters, initial catheter dwell times and catheter-related complications were compared between the two groups. Results: A total of 164 patients were enrolled with 80 SLC insertions and 133 DTC insertions. Initial catheter dwell time was 71.94 days (range 4-277 days) in SLC and 68.55 days (range 1-292 days) in DTC (p = 0.76). In terms of catheterrelated complications, catheter migration was detected in 8 patients (10.5%) and 11 catheters (12.4%) in the SLC group and in 2 patients (1.7%) and 3 catheters (2.0%) in the DTC group (p = 0.0026). Conclusion: SLC did not extend the initial catheter dwell time compared to DTC. Furthermore, SLC was more prone to catheter-related complications, particularly catheter migration, than DTC.

Effectiveness of therapeutic lymphography on lymphatic leakage E. Alejandre Lafont, C. Krompiec, W.S. Rau, G. Krombach; Giessen/ DE (enrique.alejandre-lafont@radiol.med.uni-giessen.de) Purpose: To investigate if therapeutic lymphography is a reliable method to treat lymphatic leakage when conservative treatment fails and to investigate which parameters influence the success rate. Methods and Materials: Between 08/1995 and 1/2008, 50 patients with lymphatic leakage in the form of chylothorax, chylous ascites, lymphocele, and lymphatic fistula had undergone conventional therapeutic lymphography after conservative therapy had failed. Of these 50 lymphographies, 7 could not be statistically evaluated in our retrospective study: 1 patient had died of cancer a day after lymphography; 6 were excluded due to various technical problems. The remaining 43 patients were evaluated. Therapeutic success was evaluated and correlated to the volume of lymphatic leakage (more or less than 500 ml/day), as assessed by drainage. Results: In nearly 78.8% of patients, the location of the leak could be detected, and surgical intervention could be planned when therapeutic lymphography failed. The lymphatic leak could be completely occluded in 70% of patients when the lymphatic drainage volume was less than 500 ml/day. Even when lymphatic drainage was higher than 500 ml/d, therapeutic lymphography was still successful in 35% of patients. The overall success rate in patients with failed conservative treatment was 51.2%. Success did not depend on other factors such as age, sex of the patient, or time elapsed between lymphatic injury and intervention. Conclusion: Therapeutic lymphography is an effective method in the treatment of lymphatic leakage when conservative therapy fails. The volume of lymphatic drainage per day is a significant predictor of the therapeutical success rate. Purpose: Breast cancer is a major health concern for women. Its early diagnosis relies on the detection and characterization of microcalcifications and/or masses. Observer studies related to the detectability of lesions are an important topic of research. Therefore, the availability of a database representing different types of lesions while preserving the morphological shape is crucial. We report on creating a database of 3D models of microcalcification clusters to be further simulated into projection images of 2D, breast-tomosynthesis and breast-CT. Methods and Materials: Biopsy specimens containing microcalcification clusters were collected and imaged by a micro-CT scanner. A segmentation procedure was applied to every reconstructed image to build the 3D model. Several algorithms implemented by the computer-aided diagnosis (CADx) that preserve the morphology of microcalcifications have been investigated, but failed to segment perfectly due to the different anatomical background of the micro-CT images compared with mammograms. Therefore, an adapted segmentation procedure was developed resulting in a good morphological segmentation. However, it suffered from false positives that were enhanced by applying the top-hat routine.

Machine learning for the multivariate diagnosis of malignant tumours: development and testing of a dedicated software for the automatic selection and optimisation of classification algorithms A. Dietzel 1 , P.A. Baltzer 2 , M. Dietzel 2 , M. Bogdan 1 , W.A. Kaiser 3 ; 1 Tübingen/DE, 2 Jena/DE, 3 Boston, MA/US Purpose: Machine-Learning is a promising field to solve the challenging task of multivariate diagnosis of malignant tumors. However, there is a multitude of algorithms usually requiring adjustment of a significant range of parameters to the specific radiological setting. Usually, such adjustments are done empirically by computer-scientists. In order to decrease operator related-bias and to simplify clinical application we developed a software to automatically apply and adjust Machine-learning algorithms on quantitative imaging data. Finally, the software was tested on a large clinical dataset. Methods and Materials: A software tool was implemented to automatically test different Machine-learning algorithms and adjustments for the diagnosis of malignancy. Evaluation methods were implemented as default (e.g. multiple k-fold Cross-Validation). This software was tested on a clinical dataset. It included 329 benign and malignant breast lesions imaged by MRI (imaging-protocols according to EUSOBI-guidelines; standard-of-reference: histological verification). All lesions were characterized by a commercially CAD-software using selected standard enhancement-parameters. Based on this dataset (input variables: enhancement-parameters; classification variable: histology) different Machine-learning algorithms were tested and fine-tuned applying >> 100 network parameters using the newly developed software. Results: The software identified a medium sized network as most appropriate (one hidden-layer containing 5 neurons; learning-epochs: 50; learning-rate: 1.1). Test results demonstrated significant potential for the multivariate diagnosis of malignant tumors (p < 0.001) at moderated accuracy (median AUC[Area-under-the-Curve]: 0.7249; Confidence-Interval: 0.70122-0.73923). Conclusion: A computer tool for testing different Machine-learning settings was successfully implemented and tested on quantitative radiological data. Optimal parameter settings could be identified and appropriate performance was reached.

MSCT follow-up in patients with malignant lymphoma: improved therapy response classification using semi-automated volumetry compared to unidimensional measurements M. Puesken 1 , B. Buerke 1 , N. Kohlhase 1 , V. Dicken 2 , C. Schuelke 1 , W. Heindel 1 , J. Wessling 1 ; 1 Münster/DE, 2 Bremen/DE Purpose: Impact of semi-automated volumetry compared to unidimensional measurements on therapy response classification in CT follow-up of malignant lymphoma. Methods and Materials: MSCT scans of 63 patients with malignant lymphoma prior to therapy (baseline) and after 2 cycles of chemotherapy (follow-up) were included. A total of 313 target lymph nodes (56 cervical, 131 thoracic and 126 abdominal) were evaluated by two radiologists independently. Long axis diameter (LAD), short axis diameter (SAD) and volume were determined manually and using semi-automated segmentation software. Relative interobserver difference (RID) and time for manual and semi-automated segmentation were evaluated. Therapy response was calculated for each parameter based on "IWC" lymphoma-guidelines and "RECIST"-adapted guidelines. Mean of metric and volumetric measurements served as the reference standard. Statistical analysis encompassed intraclass correlation coefficients (ICC), t-and McNemar-test. Results:Over all regions mean lymph node size in baseline/follow-up was 23.8 ± 10.3 mm/17.0 ± 9.2 mm for LAD and 7.2 ± 13.5 ml/3.4 ± 9.9 ml for volume. RID was consistently low in baseline and follow-up with high ICC > 0.96 for semiautomated measurements. Mean evaluation time for semi-automated segmentation without need for correction was shorter (16.6 ± 11.7 sec) than for manual measurements (29.0 ± 14.5 sec). In 65% of all lymph nodes correction was necessary and evaluation time increased to 39.5 ± 25.9 sec. Regarding therapy response, semi-automated volumetry obtained significantly more accurate classifications than semi-automated and manual LAD and SAD (e.g. volume 87.8% vs. semi-automated LAD 83.8%, manual SAD 78.9%, all p < 0.05). Conclusion: Semi-automated lymph node volumetry is more accurate for therapy response classification in patients with malignant lymphoma as compared to established LAD.

Our data demonstrate the potential of MRM to significantly predict disease-related death in primary breast-cancer. As results were extracted from a routine staging examination, MRM non-invasively provides not only diagnostic information but also quantitative outcome data at one step. Future studies should address the impact of these findings on patient management and therapeutic approach.

Development and testing of an artificial neural network for the noninvasive differentiation of papillomas and fibroadenomas based on breast MRI characteristics P.A. Baltzer 1 , M. Dietzel 1 , A. Dietzel 2 , T. Gröschel 1 , R. Zoubi 1 , M. Bogdan 2 , W.A. Kaiser 3 ; 1 Jena/DE, 2 Tübingen/DE, 3 Boston, MA/US Purpose: Non-invasive differential diagnosis of papillomas and fibroadenomas is of potential clinical importance, as it might aid in the indication for surgical resection of benign breast masses. The aim of this study was to design an artificial neural network (ANN) to predict presence of papillomas based on magnetic-resonancemammography (MRM) findings. Methods and Materials: 103 fibroadenomas and 83 papillomas were included. All were imaged by MRM applying standardised protocols and study design (T1w-FLASH; 0.1 mmol/kg BW Gd-DTPA; T2w-TSE; histological verification after MRM; IRB-approval). Expierenced radiologists without knowledge of final histological diagnosis evaluated this dataset using predefined MRM-descriptors. Based on such an ANN was developed to differentiate papillomas and fibroadenomas (The-MathWorks/ Inc., Feed-Forward-Architecture/Resilient Back-propagation-Algorithm). Performance of this ANN was evaluated 19 times independently and quantified on an unknown testing sample applying five-fold cross-validation (AUC: the area under the curve). Results: The ANN was able to significantly differentiate fibroadenomas and papillomas (P < 0.001). One hidden layer including 10 neurons at a learning rate of 1, 2 and scaling-factor of 0.85 was identified as the optimal network-topology. This ANN reached a maximum AUC of 0.726 for the differentiation of papillomas and fibroadenomas (standard error: 0.0367; 95% confidence-interval 0.656 to 0.789). Conclusion: On unknown data, the optimised neural network showed significant performance and moderate diagnostic accuracy. Accordingly, multivariate differentiation of papillomas and fibroadenomas is possible based on MRM-criteria and might be used to non-invasively aid clinical decisions for surgical resection of benign breast masses.

Automated quantitative computer-assisted analysis of elastography to discriminate breast lesions A. Malich, A. Kott, R. Gorna; Nordhausen/ DE (ansgar.malich@shk-ndh.de) Purpose: This study aimed to verify the diagnostic value of strain ratio as a quantitative elastographic parameter in ultrasound of breast lesion and related influencing factors. Methods and Materials: 180 breast lesions were analysed sonographically (14 MHz probe) by two experiences radiologists in consensus including elastography analysis (Ueno-scale) and computer-based calculation of strain ratio. Values were related to size and pathological outcome. The following size-related groups were used (largest available diameter): S1: < 5 mm; S2< 10 mm; S3< 15 mm; S4< 20 mm; S5> 30 mm. Mean overall size was 10.4 mm. ROC-analysis for cut-off values was performed. Results: Strain ratio of malignant versus benign lesions was 6.36 versus 2.27. Mean strain ratios of benign lesions reflecting pathology were 0.98 (fibrolipoma); 1.05 (adenosis); 1.31 (fibrosis); 1.58 (cysts) 1.53 (intraglandular lymph nodes); 2.40 (fibroadenoma); 3.31 (calcifications-liponecrosis); 2.49 (papillomata); 4.44 (scars); and 2.06 (remaining benign entities). Size-related analysis of strain ratio was calculated for malignancies/fibroadenoma/cysts/other benign lesions as S1: 1.66/1.74/1.44/0.91; S2: 2.50/2.08/1.22/2.95; S3: 6.72/2.52/2.20/3.03; S4: 10.25/3.76/0.05/3.09; S5: 4.75/8.00/2.09/1.45. Best performing cut off values are (according to ROC-analysis) S2: 2.4; S3: 2.8; S4: 3.9. Conclusion: Strain ratio is influenced by size and histopathology. In the diagnostically most relevant group of 5-20 mm lesions, malignant lesions are characterised by a higher strain ratio versus all other entitites. Large malignant lesions are typically characterized by a lowered SR versus fibroadenomata (probably due to necrotic liquid components in cancer versus macrocalcifications in fibroadenomata). Typically fibroadenomas, scars and papillomas are characterised by increased SR-values as well. Elastography is of diagnostic use in the differential diagnosis of breast lesions.

A D E F G H 3 -Functool; GE Healthcare Technologies); blood volume (BV), blood flow (BF), mean transit time (MTT) and permeability surface (PS) and ADC values for selected region of interest were measured. Signed rank test was used. Results: Residual masses were observed in all patients (12/12) and in 48% of subjects at restaging time (10/21). There were significant differences in PS (p < 0.01) between vital tumour and tissue fibrosis. Six CT perfusion and 5 WB DWI studies resulted negative even before FDG-PET studies. Conclusion: CT perfusion and WB DWI offer functional tissue characterisation as well as FDG-PET; WB DWI can spares radiation to young subjects and CT perfusion can show better morphologic appearance. They can represent new alternative strategies in evaluating response to treatment in early assessment of HD. Purpose: Appendiceal changes in cystic fibrosis (CF) are well known but there is no previous study with systematic description on computed tomography (CT). Our objective was to describe the appearance of the appendix of adult CF and to see if it can be linked to clinical symptoms.

In a retrospective study of abdominal and pelvic CT scans of adult patients with CF: 55 CT scans from 55 patients (18 to 52 years old), cause of examination and diagnosis were noted. There was one case of colonic perforation with appendiceal micro-abscesses; for the other cases, there was no proven acute appendicitis. Findings studied: maximum appendiceal outer diameter and intraluminal fluid diameter, mucocele, appendicolith, peritoneal fat infiltration and effusion. Results: Appendiceal diameter was enlarged (10.9 ± 2.6 mm). There was a mucocele in 23/55 and appendicolith in 2 cases. Peritoneal fat infiltration was present in 9/55 and effusion in 11/55 cases. Abdominal pain was mostly due to distal intestinal obstruction syndrome (DIOS) (χ² = 17.6; p < 0.05). Appendiceal diameter did not correlated with age (R² < 0.001; p = 0.88) but did strongly with wall thickness (R² = 0.23; p < 0.05). No difference of diameter between the DIOS and the non-DIOS groups (p > 0.05) was observed. Conclusion: There is a high variability of the CT appearance of the appendix in adult CF patients. Appendiceal diameter is routinely enlarged but is age and symptomindependent. The CT diagnosis of acute appendicitis in this population is therefore not reliable. Clinical examination and comparison with previous CTs are essential. Purpose: As MRI is introduced for diagnosing appendicitis, we evaluated the effect of direct feedback on the sensitivity and specificity of MRI for appendicitis in inexperienced radiologists and residents. Methods and Materials: Twelve inexperienced readers (6 radiologists, 6 residents) evaluated 102 abdominal MRI's for appendicitis receiving direct feedback (final diagnosis) after each examination. The readers recorded the diagnosis and reading time per case. Prevalence of appendicitis was 44%; cases were presented in random order. Mean sensitivity, specificity and time were calculated for the first and the last set of 25 scans. Results: In radiologists, sensitivity for appendicitis improved from 73% in the first set of 25 cases to 88% (95% CI: 71% to 95%) in the last set of cases. In residents, the sensitivity increased from 73% to 87% (95% CI: 67% to 96%). The specificity in radiologists was 88% in the first and 91% (95% CI: 83 to 96%) in the last set, versus 87% and 93% (95%CI: 84% to 97%) in residents. None of these differences reached statistical significance. The mean reading time decreased significantly from 231 seconds per scan in the first set to 140 seconds in the last set for radiologists (p < 0.001) and from 480 seconds per scan to 339 seconds for residents (p < 0.001).

A new algorithm for automatic segmentation of the left ventricular cavity in cardiac MRI M. Mazonakis, K. Pagonidis, E. Grinias, J. Damilakis; Iraklion/GR (mazonak@med.uoc.gr) Purpose: To present an algorithm for automatic segmentation of the left ventricular (LV) cavity in MR images. Methods and Materials: Short-axis cine cardiac MR images were acquired from sixteen consecutive patients with known or suspected coronary artery disease. The automatic segmentation involved the classification of image pixels in three classes according to their grey value. The initial regions of high confidence were determined per class based on a histogram analysis and, then, the regions were grown using a flooding process. The algorithm could automatically detect and correct segmentation errors. The corrected segmentation map was presented to the user to select the LV endocardial region. The endocardial contours extracted from MR images were used to calculate the end-diastolic volume (EDV), end-systolic volume (ESV) and ejection fraction (EF). The automatically determined LV parameters were compared with the reference values obtained by manually tracing the endocardial borders. Results: Bland-Altman analysis showed a small underestimation of LV volumes (EDV: mean difference = 7.8±9.5 ml; ESV: mean difference = 2.3±4.6 ml) and a slight overestimation of EF (mean difference = 0.6±4.8 %) with the automatic segmentation algorithm compared to manual tracing. The differences between the two segmentation methods were not statistically significant (EDV: p = 0.06; ESV: p = 0.08; EF: p = 0.45). The parameters determined by automatic segmentation correlated well with those obtained by manual tracing (EDV: r = 0.86; ESV: r = 0.92; EF: r = 0.81). The automatic image analysis time was less than 1.5 min per patient.

The new algorithm enables the accurate assessment of LV volumes and EF from MR images with great saving in labour. Purpose: In oncological therapy monitoring, an estimation of the change in size of a tumour is an important criterion for the assessment of treatment success. This requires finding correspondences between lesions in the baseline and follow-up CT scans, which can be a tedious task. We present an automatic tracking algorithm that identifies the positions of the corresponding follow-up lesions for a given set of segmented baseline lesions.

The two CT scans are coarsely aligned by a rigid registration which defines a search area in the follow-up image. Here, lesion candidates are detected by grey value and shape features. Finally, the candidate with the highest cross-correlation to the baseline lesion is chosen. We evaluated our method by applying it to 711 lesions (220 lung nodules, 263 liver metastases, and 228 lymph nodes) in CT scans of 172 patients from four different sites. It was considered successful when the point was inside an expert segmentation that was available for the follow-up lesions.

The tracking was successful for 95% of the lung nodules, 88.6% of the liver metastases, and 83.8% of the lymph nodes. In cases of failure, the mean distance to the correct lesion was 4.7 mm. The average computation time was 3.5 s.

We developed a robust and fast method for automatic lesion tracking.

In combination with a segmentation algorithm, it can accelerate the radiological workflow by allowing fully automatic precomputations of follow-up examinations which only have to be confirmed or corrected by the radiologist.

New strategies for assessment Hodgkin lymphoma during and after therapy: preliminary results in comparison CT perfusion, whole body DWI and FDG-PET A. Fausto, M.A. Mazzei, F.G. Mazzei, A. Fabbri, F. Lauria, L. Volterrani; Siena/IT (afausto@sirm.org)

Purpose: To assess the advantage of new techniques as CT perfusion and WB DWI versus FDG-PET in evaluating treatment response for Hodgkin disease (HD) during and after therapy. Methods and Materials: 21 patients with HD underwent to CT perfusion (100 ml contrast media, 4 ml/s, 64-row VCT, GE Healthcare), WB DWI (b = 600, 1.5 T Signa HDx scanner, GE Healthcare), and FDG-PET at early assessment and restaging time during and after induction therapy (n = 9), only at restaging time after induction therapy (n = 9) and at staging, early assessment and restaging time prior, during and after induction therapy (n = 3), for a total of 36 studies. All CT perfusion and DWI studies were analysed using commercially available software (CT Perfusion B A S279 C D E F G H Methods and Materials: 13 Sprague Dawley rats were anesthetised and a loop (3-0 gut) was tied loosely around the SMA, without occluding the artery. 3 days later, each rat underwent micro-MR (Bruker Biospin) abdominal scans before squeezing the loop and after that at different timepoints (4 and 8 hrs). Then the entire bowel was removed and processed for histological analysis on HE-stained sections. Results: One rat was excluded from the analysis because it showed an incomplete occlusion in angio-MR sequences 8 hrs after SMA ligation. Basal micro-MR scans showed no gas in the abdomen and no mesentery irritation. Rare-T2 micro-MR sequences identified several signs of vascular occlusion such as loops dilation, bowel wall thinning and peritoneal fluid. These changes were confirmed by ex-vivo histological examination. Conclusion: This animal model could represent a reproducible tool to evaluate by imaging the evolution of ischaemic gut due to acute arterial occlusion and to test the effectiveness of new therapeutic modalities.

Conventional defecography versus MR defecography for diagnosing obstructive defecation syndrome: is patient exposure to radiation still justified? H. Neubauer, B. Hartung, F. Wendel, T. Pabst, W. Kenn, D. Hahn; Würzburg/DE (gwendolinhartung@yahoo.de) Purpose: High-resolution MR defecography (MRD) is established for imaging obstructive defecation syndrome. Referral to conventional defecography (CD) is still common, as radiographic images appear more easily interpretable to clinicians. We compared both modalities for feasibility and diagnostic value in clinical rountine. Methods and Materials: Seventy-two patients (68 females, mean age 53±13) referred to conventional defecography (oral, vesical, vaginal and high rectosigmoidal contrast, temporal resolution 0.5s) also underwent MR defecography (1.5 T, highres TRUFI, three sagittal 8 mm slices, rectal contrast 200 ml ultrasonography gel, temporal resolution 1.2s) with imaging at rest, retention, during evacuation and post-evacuation. All examinations were read by two radiologists in consensus. Results: Periprocedural problems include failure to evacuate (MRD n=11, CD n=2), spontaneous loss of rectal contrast (MRD n=7, CD n=14) and insufficient opacification of the rectosigmoid junction (CD n=22). Diagnostic accuracy for intussusception, enterocele, rectocele, cystocele and sigmoidocele was 85%, 97%, 100%, 85% and 98% (MRD) and 97%, 89%, 92%, 97% and 88% (CD). Kappa values ranged 0.42~0.65. Of 39 cases with intussusception, 9 were seen on CD and 2 on MRD only. Of 9 cases with sigmoidocele and 22 cases with enterocele, 56% and 22% were visible on MRD only. Mean dose area product for CD was 11.359 ± 8292 cGy/cm2. 25T field (G-SCAN, Esaote, Italy). The magnet table was provided with a tilting mechanism going from 0° to 90° with two steps and allowed the evaluation both in supine and orthostatic position. Images acquired on the sagittal plane were used for quantitative analysis and were evaluated separately by two radiologists to establish inter-observer concordance. Conventional Defecography (CD) was performed in all patients. Results: The comparison between CD and MR-D showed statistically significant differences in the evaluation of anterior rectocele during evacuation in both positions (inaccuracy of 0.43 cm for overestimation by MR-D in orthostatism and inaccuracy of 0.81 cm for underestimation by MR-D in the supine position) and of ano-rectal junction descent from pubo-coccygeal line (PCL) during evacuation, only in the supine position (with inaccuracy of 1.66 cm of underestimation by MR-D). The value of intra-observer intra-class correlation coefficient (ICC) ranged from good to excellent; the inter-observer ICC ranged from moderate to excellent. Conclusion: MR-D is more accurate in orthostatic than in supine position for the evaluation of pelvic floor disorders.

While reading time for inexperienced MRI readers decreased significantly with systematic feedback in a series of 102 suspected appendicitis cases, sensitivity and specificity did not improve significantly, and diagnostic accuracy did not reach the level of experts.

In vivo visualisation of SPIO-loaded surgical mesh implants by MRI N. Kühnert 1 , N.A. Krämer 1 , H.C.W. Donker 1 , J. Otto 1 , I. Slabu 1 , M. Baumann 1 , G.A. Krombach 2 , C. Kuhl 1 , U. Klinge 1 ; 1 Aachen/DE, 2 Giessen/ DE (nicolas.kuehnert@rwth-aachen.de) Purpose: Prosthetic mesh implants are widely used in hernia surgery. To reveal long-term mesh-related complications such as shrinkage, dislocation or deformation, a precise visualisation of meshes in vivo is important. Modification of mesh MRI properties can help to delineate them. The aim of this study was to demonstrate time-dependent mesh shrinkage in vivo by MRI. Methods and Materials: PVDF-meshes with incorporated super-paramagnetic iron oxides (SPIO) were implanted as an abdominal wall replacement in 30 rats. MRI was performed at day 1, 7, 14, or 21 using gradient-echo sequences TR/TE = 50/4.6, flip-angle 20°. Length and width were measured on axial, coronal and sagittal images and geometrical deformation was assessed. These data were compared to the post-mortem explanted meshes. Results: In all orientations, MRI visualised all meshes and showed a mean shrinkage of length and width (day 21, 13.1%, 3.1%; day 14, 9.2%, 4.1%; day 7, 5.9%, 0.5%). Post-mortem measurements validated these results. Purpose: The poor prognosis of gut ischaemia is due to the lack of specific findings, either clinical or radiological, that leads to delayed diagnosis and ineffective treatment. Early diagnosis assures a remarkable reduction of mortality rate. Therefore, the aim of this study is to value the experimental use of 7 T microMR in the detection of acute intestinal ischaemia (AMI) due to venous occlusion, relating the MR imaging patterns to the evolution of intestinal morphodynamism and histological analysis.

The study was conducted on 16 Sprague Dawley rats. After anaesthesia, animals in the 'control group' (n = 8) underwent the superior mesenteric vein (SMV) occlusion by tight ligation and, after macroscopical monitoring, were killed at different timing and the bowel removed for histological analysis; in reverse, rats in the 'experimental group' (n = 8) had a loop tied around SMV without occluding it. 3 days after surgery, basal MR scans were collected using a Bruker-Biospec-70/16-US; then SMV was occluded squeezing the loop and MR sessions were repeated after 5 min, 4 and 8 hrs. Results: Rats scanned using rare-T2 sequences showed no pathological pattern at the first time-point, while significant bowel wall thickening (> 1.5 mm) and mesenteric hyperintensity were detected at following time-points (4 and 8 hrs). Conclusion: Compared to histological analysis and macroscopical evidences, MR imaging correctly detected morpho-functional alterations of ischaemic gut and early identify (4 hrs) the specific pattern of lesions caused by SMV occlusion. Its future application in the early diagnosis of mesenteric venous ischaemia is highly reasonable.

A specific rat model of acute gut ischaemia due to arterial occlusion: a 7 T micro-MR imaging study F. Somma, D. Berritto, N. Landi, C. Cavaliere, S. Cappabianca, A. Rotondo, R. Grassi; Naples/IT (daniela.berritto@libero.it) Purpose: Mesenteric ischaemia is an uncommon but often underestimated cause of nontraumatic acute abdomen and its global prevalence is around 0.1% of all hospital recovers. Despite the great improvement in its diagnostic techniques, bowel infarction is still a fatal disease, with high mortality rate. Thanks to a simple animal model, intestinal ischaemia was induced occluding superior mesenteric artery (SMA) without laparotomy. The aim of the study was to identify MR imaging patterns of lesions due to SMA acute occlusion, avoiding pneumoperitoneum and ileus post surgery. Munich/DE (michael.scherr@med.uni-muenchen.de) Purpose: To evaluate, using a porcine model, LD CT strategies to examine suspected body packers. Methods and Materials: Nine samples of typical illicit drug compounds packed in standardised ovoid plastic containers were rectally inserted in a pig cadaver. Using a 64-row CT (CT750HD, GE Healthcare), images were obtained at 120 kVp and fixed tube currents of 80 mA (LD80), 30 mA (LD30), and 10 mA (LD10). Images were compared to standard dose abdominal CT (STD) and rated, first, for general diagnostic image quality in abdominal indications (general assessment score: 0 = impossible, 1 = limited, 2 = relevant diagnostics, 3 = excellent). Second, with respect to body pack detection, visibility of three main characteristics (wrapping, content and shape) was rated (visibility: 0 = definitely not, 1 = presumably not, 2 = presumably, 3 = definitely) and summarised to a detection score (DS) ranging from 0 to 9 with a score ≥6 representing sufficient detectability. Results: Mean detection scores (DS) for the different LD-protocols were 8.2 ± 0.7 (LD80), 6.8 ± 1.3 (LD30), and 4.6 ± 1.9 (LD10), compared to 8.8 ± 0.5 (STD), with p < 0.05, respectively. Effective doses were calculated at 2.3 mSv (LD80), 0.9 mSv (LD30), 0.3 mSv (LD10), and 9.4 mSv (STD). For all LD-protocols, mean general assessment scores were inferior to STD: 1.4 ± 0.7 (LD80), 1.1 ± 0.7 (LD30), and 0.3 ± 0.6 (LD10) compared to 2.3 ± 1.0 (STD), with p < 0.001, respectively. Conclusion: While general diagnostic image quality was limited for all LD protocols, LD80 provided an excellent level of body pack detectability at a tolerable effective dose while LD30 reached a still sufficient level of detectability at a dose comparable to x-ray.

Cumulated radiation dose and risk of developing radiation-induced cancer in IBD patients treated with biological therapy C. 

The incidence of errors was low. Nevertheless common patterns of error were detected with potentially significant consequences. The MDT offers an opportunity for double reporting and increasing accuracy. Education to illustrate common pitfalls may help to reduce the number of future reporting errors for both urological and non-urological malignant disease.

The incremental value of contrast-enhanced MR imaging in biopsy-proven local recurrence after radical prostatectomy C. Wassberg, O. Akin, A. Shukla-Dave, H. Vargas, J. Zhang, H. Hricak; New York, NY/US (cecilia.wassberg@gmail.com) Purpose: To assess the impact of contrast-enhanced magnetic resonance imaging (CE-MRI) on the accuracy of readers of different experience levels in the detection of post-surgical local recurrence of prostate cancer by MRI, using biopsy as a standard of reference. Methods and Materials: The Institutional Review Board approved and waived the informed consent requirement for this HIPAA-compliant study of 52 patients who, between March 2005 and December 2008, underwent 1.5-Tesla-MRI with multiphase post-contrast endorectal MRI following post-surgical biochemical recurrence of prostate cancer and had biopsy within 3 months of MRI. Two radiologists (reader 1, with one year and reader 2, with six years of experience) retrospectively and independently reviewed each MRI study and recorded their level of suspicion for recurrent cancer on a five-point scale. Areas under receiver operating characteristic curves (AZ) were calculated to assess diagnostic performance of T2W imaging with and without CE-MRI. Inter-reader agreement was assessed using kappa statistics. Results: Thirty-three patients (63%) had prostate cancer recurrence on biopsy. With the addition of CE-MRI, the AZ for cancer detection increased significantly for reader 1 (0.77 vs. 0.85; P = 0.04) but not for reader 2 (0.86 vs. 0.88; P = 0.72). Use of CE-MRI increased inter-reader agreement from fair (K = 0.39) to moderate (K = 0.58); it also allowed correct reclassification and greater certainty in image interpretation in 13/52 patients (25%) for reader 1 and 6 patients (12%) for reader 2. Conclusion: Adding CE-MRI to T2W endorectal MRI can improve post-operative detection of locally recurrent prostate cancer, especially by relatively inexperienced readers, while reducing interobserver variability.

Role of 1H-MRS and DCE-MR in identification of pattern changes after neoadjuvant hormonal therapy F. Galati, V. Panebianco, A. Sciarra, V. Buonocore, D. Lisi, R. Passariello; Rome/IT Purpose: To determine the time-dependent metabolic and angiogenic changes that occur in PC during neoadjuvant HT using multiparametric MR. Methods and Materials: Seventy-five patients were submitted combined endorectal MRI with MRSI and DCE-MRI using 3-T magnet (Verio, Siemens, Erlangen, Germany). Scan protocol included morphologic imaging with TSE T2-weighted sequences on three planes, spectroscopic imaging with 3D CSI and 3D FLASH T1-weigthed sequence. All patients underwent MRI examination before therapy, at 4, 12 and 24 weeks from therapy. The time course of loss of prostate metabolites was delayed in PCa tissue as compared to healthy tissue. Results: We noticed that at 4 weeks from therapy there was a citrate reduction with an apparent increase in choline peak in 90% of patients; during the last 24 weeks there was a significant time-dependent loss of the prostatic metabolites (citrate level decreased faster than creatine and choline levels, resulting in an increase in ratio value) till the complete loss of all metabolites (metabolic atrophy) in 30% of patients on long-term therapy. In patients with metabolic atrophy we observed neoangiogenic pattern reduction at 6 months. Conclusion: Metabolic and angiogenic changes during a neoadjuvant HT may provide important prognostic informations and may help to define the optimal time to start the definitive therapy in each patient.

Assessment of early response to external beam radiation therapy for prostate cancer: a prospective study with 3 T diffusion-weighted imaging J. Seo, C. Kim, S. Park, B. Park, W. Park; Seoul/KR (noir1130.seo@samsung.com) Purpose: To investigate the changes of apparent diffusion coefficient (ADC) in prostate cancers before, during and after external beam radiation therapy (EBRT), and to assess the reproducibility of ADC measurement in the prostate. Methods and Materials: Eight consecutive patients with biopsy-proven prostate cancer were examined by T2-weighted, dynamic contrast-enhanced and diffusionweighted imaging (DWI) at 3 T[b-values, 0 and 1000 s/mm 2 ]. All patients treated done with the linear probe. Hard lesions (encoloured in blue) were considered suspicious for testicular cancer and histopathologic findings were used as gold standard. Lesions with normal or decreased tissue stiffness (encoloured in red to green) have been thought to be benign changes, i.e. focal orchitis or ischaemia, which can mimic testicular cancer on gray scale ultrasound. When benign lesions were suspected, follow-ups were performed to exclude progression of the mass. Results: Overall, 19 (70%) hard lesions and 8 (30%) soft lesions were detected. Seventeen hard lesions (89.5%) histopathologically revealed testicular cancer. The 2 other hard lesions were thought to be scars and had no progression at follow-up. No soft lesion was found to be cancer (100%). Because of clinical findings and findings on B-mode ultrasound soft tissue changes were thought to be cysts or inflammation and all showed no growth at follow-up. Conclusion: Our preliminary results suggest the capability of RTE to distinguish between benign and malign testicular masses and may therefore refrain men scheduled for unnecessary surgery of the testis in future.

Voiding MR-cystourethrography: a new diagnostic imaging technique for the evaluation of male lower urinary tract M. Di Girolamo, C. De Cecco, E. Pandolfi, P. Venneri, S. Galassi, V. David; Rome/IT (digirolamomarco@hotmail.com) Purpose: To evaluate the diagnostic accuracy of a new diagnostic imaging technique called voiding MR-cystourethrography that allows the visualisation of the male urethra. Methods and Materials: 10 normal volunteers and 86 male patients with bladder outlet obstruction (evaluated with urine-flow velocity recording) underwent voiding MR-cystourethrography performed with an 1.5 T magnet with the patient placed in supine position. The filling of the urinary bladder with paramagnetic contrast agent was obtainded by the i.v. administration 20 mg of furosemide followed by the i.v. administration of ¾ of the normal dose of a paramagnetic contrast agent (Magnevist, Schering, Germany). When the bladder was filled with contrast-material-enhanced urine, the patient was asked to urinate. During the micturition T1-weighted spoiled 3D gradient-echo acquisitions on sagittal plane were performed. 25 patients performed conventional cystourethrography in the month preceding MRI. Results: Homogeneous opacification of the bladder lumen was always obtained. 10 patients were unable to perform the MR examination. In all the volunteers and in all the patients studied (76 pts) we obtained a perfect evaluation of the male urethra with voiding MR-cystourethrography. The visualisation of the urethra with MIP reconstructed images was considered comparable to that obtained with conventional cystourethrography. We detected 24 cases of bladder neck obstruction, 36 cases of urethral stricture and 2 urethral papillomatosis. The site, length and the number of urethral strictures were accurately determined by MRI. Conclusion: Voiding MR-cystourethrography demonstrates the morphology of the bladder neck and urethra during the micturition and can substitute standard retrograde and micturating cystourethrogram.

Radiological reporting errors identified in the prostate, bladder and penile cancer multidisciplinary team meeting -a prospective review to identify the patterns and severity of errors at second reporting: beware the common pitfalls S.J. Kennish, J.T. Smith; Leeds/UK (s_kennish@yahoo.co.uk) Purpose: The urology multidisciplinary meeting (MDT) offers an opportunity for second reporting of urological malignancy imaging. Errors can be detected, categorised and graded as to potential severity. The frequency, category and pattern of common errors are presented with case illustrations to help radiologists avoid common pitfalls and improve personal reporting quality. Methods and Materials: Continuous prospective collection and analysis of errors identified during independent second review of urological malignancy imaging at multi-disciplinary meetings over a 2-year period. Cases identified by a single radiologist and confirmed by consensus. Errors were categorised into: false negative, satisfaction of search, interpretation error, typographical error and false positive findings. Severity of error was graded from 1 to 4 with 4 representing the highest potential clinical significance. Results: Of 3347 MDT patients, 41 radiological errors were identified (1.2%). The average delay to identification was 47 days. The majority of errors were false negative (37%), interpretation (29%) or satisfaction of search (22%). Failing to identify significant lymph nodes (false negative) and missing significant foci of disease in the presence of disseminated malignancy (satisfaction of search) were amongst the commonest patterns. Mean severity of error was 3 (2-4), a high potential for significant consequences. 

A S283 C D E F G H Methods and Materials: Fifty-three female patients (mean age 48.1, range 28-81 years) with histologically proven BCa underwent conventional MRI (including T2weighted and T1 dynamic imaging post-contrast) and diffusion-weighted (DW) MRI (b-values 0, 250, 500 and 1000 s/mm 2 ). The ADC for the BCa tumour volume, as defined manually on the highest b-value images, was correlated with molecular and histological prognostic factors. ADC values were assessed for difference between gene profiling subtypes (Luminal A, Luminal B, HER-2, triple receptor negative), vascular invasion (present/absent), grading (G1/G2/G3), ER/PgR/HER-2 expression (positive/negative), Ki67 (< 14% or > 14%) and TNM staging as factors using analysis of variance (ANOVA). Spearman coefficient was used to assess correlation between ADC values and continuous variables (age, percentage of ER, PgR, HER-2, Ki-67). Results: The mean ADC of BCa lesions visible on the high b-value image was 1.14 ± 0.20 x 10 -3 mm 2 /sec. The ADC of HER-2 was lower than other subtypes, reaching significance when compared with triple receptor negative (p = 0.055). The ADC for the small T3 subgroup (n = 8) was lower than other T grades, but only significantly so relative to T1 (p = 0.03). There was a marginally significant correlation (Spearman r = -0.35, P < 0.05) between ADC and HER-2 expression. Conclusion: Our results showed little support for a correlation between ADC and the molecular and histological prognostic factors examined. A possible exception is HER-2 expression. The T3 subgroup may also bear a distinct ADC behaviour.

Diffusion-weighted imaging of breast cancer at 3 T: does the apparent diffusion coefficient value correlate with prognostic factors M. Telesca, F. Pediconi, V. Casali, M. Luciani, F. Vasselli, E. Miglio, C. Catalano, R. Passariello; Rome/IT (mariannatelesca@yahoo.it) Purpose: To evaluate the correlation of the apparent diffusion coefficient (ADC) value with prognostic factors in patients with biopsy-diagnosed breast cancer. Methods and Materials: A prospective study was conducted in 45 patients (mean age 46 years) with biopsy-proven malignant breast lesions measuring ≥6 mm. All patients were examined at 3 Tesla using dedicated bilateral breast coil. Diffusion weighted imaging (DWI) of the breast was performed using a single shot echo planar imaging with a b-factor of 0 and 800 sec/mm 2 . ADC maps were reconstructed and findings were compared with findings from histology. Lesion size, histotype, receptor status (ER, PgR) and Ki67 expression were determined. Results: Breast cancer was detected in 41 patients (91.1%). The mean ADC value was 1.41 ± 0.23 x 10 -5 mm 2 /s. There were no correlations between the ADC value and prognostic factors. In particular, there was no correlation between ADC and ER (p = 0.5) or between ADC and PRG (p = 0.6). A small correlation was noted with Ki67 expression, but it was not statistically significant. Conclusion: ADC values are potentially an additional tool to predict tumour malignancy, although no statistically significant correlation with prognosis factors was found in this study.

3.0 Tesla diffusion tensor magnetic resonance imaging of the breast: preliminary results A. Tagliafico, G. Rescinito, F. Monetti, C. Faedda, M. Calabrese, F. Chiesa, A. Villa; Genoa/IT (chiesa.fabio@gmail.com) Purpose: To evaluate whether 3.0-T diffusion tensor imaging (DTI) allows differentiation between breast tumours and normal breast tissue.

The study was IRB approved and written informed consent was obtained. The study included 54 women with 65 breast lesions (56 malignant, 9 benign). DTI was performed during 3.0-T breast MRI examinations, and fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values were measured for breast lesions and normal tissue in each subject. FA and ADC were compared between cancers, benign lesions, and normal tissue by univariate and multivariate analyses. Results: FA of carcinomas (mean ± SD: 0.32 ± 0.09) was significantly lower than normal breast tissue in the same subjects (0.42 ± 0.12; P < 0.001). Multiple logistic regression showed that FA and ADC were each independent discriminators of malignancy (P < 0.001). FA improved discrimination between cancer and normal tissue over ADC alone. There was no difference in FA between malignant and benign lesions. Conclusion: Diffusion anisotropy is significantly lower in breast cancers than normal tissue, which may reflect alterations in tissue organisation. Our preliminary results suggest that FA adds incremental value over ADC alone for discriminating malignant from normal tissue but does not help with distinguishing benign from malignant lesions.

Breast DTI is a completely non-invasive, fast (approx. 5 minutes) method that demonstrates a high potential for cancer detection and diagnosis as a stand alone method or in conjunction with DCE-MRI.

Diffusion-weighted magnetic resonance imaging (DWI) of the breast: high-resolution sagittally-oriented versus conventional axial echo-planar sequences C. Losio, P. Panizza, A. Salerno, G. Cristel, I. Fedele, V. De Iorgi, F. De Cobelli, A. Del Maschio; Milan/IT (losio.claudio@hsr.it) Purpose: DWI is known to improve the specificity of dynamic breast MRI. However, echo-planar sequences, usually acquired in the axial plane, suffer from low spatial resolution and artefacts that may significantly degrade image quality. Our purpose was to set-up a high-resolution sagittally-oriented DWI sequence and to test its performance during clinical breast MRI compared to conventional axial imaging. Methods and Materials: 84 women underwent breast DWI (b-value: 900) in the axial plane (both breasts acquired together) and sagittal plane (each breast acquired separately), followed by dynamic breast MRI. Standard post-processing of dynamic study was completed with qualitative and quantitative assessment of water diffusivity (ADC maps). ADC values were calculated for all enhancing lesions > 5 mm and compared for the two sequences. DW Images were also analysed in terms of lesion conspicuity and image quality. Diagnostic performance was assessed for both sequences, using cyto-histology or follow-up as reference standard. Results: DWI identified 44/52 (axial) and 48/52 (sagittal) enhancing lesions (range: 5-38 mm). ADC values were significantly lower in malignant than benign lesions (0.98 ± 0.28 vs 1.76 ± 0.29; p < 0.001), with no differences between the two sequences. High-resolution sagittal sequences better depicted small enhancing foci, with less artefacts and more accurate determination of ADC; however, the conspicuity of benign lesions was inferior compared to axial DWI. Conclusion: DWI with high-resolution sagittally-oriented sequences is superior to axial one in terms of image quality, reducing distorsions and artefacts particularly in the right breast, with benefits in both qualitative and quantitative assessment of malignant lesions, and without significant increase of examination time.

Correlation of apparent diffusion coefficient (ADC) value of invasive ductal carcinoma on diffusion-weighted MR imaging with pathologic prognostic factors S. Park, H. Choi, S. Hahn, H. Kim, H. Jung; Incheon/KR (shpark@gilhospital.com) Purpose: To measure the apparent diffusion coefficient (ADC) values of invasive ductal carcinoma (IDC) of the breast and to analyse correlations between ADC values and prognostic factors including tumour size, axillary lymph node status, histologic grade, estrogen receptor (ER), progesterone receptor (PR), and c-erb-2 (HER2) using diffusion-weighted MR imaging. Methods and Materials: A total of 110 patients (mean ± SD: 48.1 ± 9.2 yrs) with 110 pathologically confirmed IDC were examined on 3.0 T MRI with diffusion weighted MR imaging. The ADC values of breast cancer were calculated using two b factors (0 and 1000 s/mm 2 ). The mean and standard deviation (SD) of the ADC value for each prognostic factor (size, grade, lymph node metastasis) and hormonal receptor (ER, PR, HER2) was calculated using independent sampled Student-t test. The correlations between ADC values and prognostic factors were analysed and assessed by one-way ANOVA test and Spearman's correlation test. Results: The mean ADC value of IDC (n = 110) was 0.88 ± 0.15 x 10 -3 mm 2 /s. The mean ADC of HER2-positive IDC (n = 36) was significantly higher than that of HER2-negative IDC (n = 74) (0.92 ± 0.13 mm 2 /s vs. 0.86 ± 0.16 mm 2 /s) (p = 0.02). On the other hand, ER status (p = 0.16), PR status (p = 0.57), histologic grade (p = 0.12), axillary lymph node status (p = 0.27) showed no statistically significant correlation with ADC values. Conclusion: HER2-positive IDC showed higher ADC than HER2-negative IDC, respectively, ER status, PR status, histologic grade and axillary lymph node status showed no statistically significant correlation with ADC. Munich/DE (mike.notohamiprodjo@med.uni-muenchen.de) Purpose: Dermal backflow is associated with obstruction of the lymph collecting vessels. The aim of this study was to evaluate MR-lymphangiography for visualisation of dermal backflow in patients with lymphoedema of the lower extremity. Methods and Materials: 30 consecutive patients with lymphoedema of the lower extremity underwent MR-lymphangiography on a 3.0 T-scanner (Magnetom VERIO, Siemens) with a highly-resolved isotropic T1-weighted FLASH3D-sequence after intracutaneous injection of Gd-DTPA. Four levels were examined: lower leg, knee, upper leg and pelvis. One radiologist and one nuclear-physician evaluated depiction of lymph-collectors, diffusity of drainage and number of visualised levels in consensus. All examinations were correlated with corresponding lymphoscintigraphyexaminations performed with colloid-bound 99Tcm. Correlation of the two methods was assessed with weighted-kappa-coefficients. Results: All examinations were diagnostic and MR-lymphangiography and lymphoscintigraphy showed excellent correlation (r = 0.84). Normal and instant lymphatic drainage in unaffected extremities was observed in both modalities shortly after contrast application. Depiction of delayed drainage was concordant in both modalities. In 12 patients dermal backflow in terms of complete or partial diffuse lymphatic drainage was detected. Five of these patients showed localised dermal backflow with directed lymphatic drainage of the levels below. Compared to lymphoscintigraphy contrast media leakage was more clearly depicted in MRlymphangiography. The three-dimensional reconstruction capabilities proved beneficial for exact anatomical localisation of the lymph vessel obstruction. Conclusion: MR-lymphangiography and lymphoscintigraphy show excellent correlation. MR-lymphangiography allows for a more accurate anatomical localisation of dermal backflow in patients with lymphoedema of the lower extremity and is a valuable tool for planning of microsurgical therapy. (pascal.baltzer@med.uni-jena.de) Purpose: Diffusion weighted imaging (DWI) is increasingly recognised as a powerful quantitative tool for breast lesion differentiation. Changes of the extracellular space, i.e. increased cellularity in malignant lesions, leads to characteristic restriction of water diffusion. Beyond differential diagnosis, this method may also be applied to the assessment of breast cancer aggressiveness. Higher grading is associated with

B A S285 C D E F G H

Foot perfusion CT in patients with peripheral arterial occlusive disease (PAOD): a novel chance for treatment planning? R. Iezzi, R. Dattesi, F. Pirro, M. Santoro, G. Tinelli, F. Snider, L. Bonomo; Rome/IT (r.iezzi@rad.unich.it) Purpose: To prospectively assess the technical feasibility and reproducibility of a quantitative foot perfusion multidetector-row computed tomography (MDCT) technique in patients with PAOD. Methods and Materials: After institutional review board approval and informed patient consent was obtained, 10 patients with PAOD referred to our department to undergo a single-limb endovascular treatment were prospectively enrolled. All patients underwent dynamic foot 64-row-CT examinations before and after (within 72 hours) endovascular treatment, acquiring eight contiguous 5-mm reconstructed sections, with 60-second acquisition time, during injection of 50 mL of contrast medium (Iomeprol 400 mgI/mL, at 5 mL/sec). Data were analysed by two experienced blinded readers using a dedicated software to calculate perfusion parameters, such as blood flow (BF), blood volume (BV), mean transit time (MTT), and permeabilitysurface area product (PS). Interobserver and intraobserver agreement of perfusion CT analysis of untreated foot were statistically assessed using intraclass correlation coefficient (ICC) and Blant-Altman analyses. Results: A good interobserver and intraobserver agreement of perfusion CT analysis was obtained in all patients. All perfusion parameters obtained for untreated foot showed good agreement between the two repeated studies. Conclusion: Foot perfusion CT is a feasible and reproducible technique. By providing the functional foot microvasculature, it could be a promising method to select the best treatment available for patients with PAOD; furthermore, in case of endovascular treatment, it could be useful to determine which and how many arteries should be recanalised in order to obtain the best result. Purpose: To describe the different branching patterns and mean caliber of DIEA (deep inferior epigastric artery), a costant vessel is used by microvascular surgeons in autologus breast recostruction. To assess the incidence and anatomy of SIEA (superficial inferior epigastric artery) which can be used by surgeons as an alternative to DIEP flaps. To evaluate the incidence of SICA (superficial iliac circumflex rtery), a vessel can mimic SIEA. Methods and Materials: The volume datasets of 174 consecutive abdominal CT angiography studies (CTA) in female patients performed using a 64-detector-row scanner (rotation time 400 ms, beam collimation 64 x 0.5, 120 kVp, reconstruction interval 0.5 mm) were reviewed (mean age 72 years old). Branching pattern was classified following the Taylor and Moon classification which describe the DIEA arterial branches entity. Results: DIEAs were identified bilaterally in all but one patient with previous appendicectomy, in whom right DIEA had been ligated during the operation (n = 347). The Taylor branching pattern was: single branch (type I) 61% (n = 211), two branches (type II) 34% (n = 119), three branches (type III) 5% (n = 17). DIEA median calibre was 3.1 mm ± 0.5 DS. SIEA was present in 49/174 patients (28.1%), bilateral in 8 cases (16%) and monolateral in 40 (84%). Superficial iliac circumflex artery was present in 47/174 patients (27%), bilateral in 19 cases (40%) and monolateral in 28 (60%). Conclusion: Vascularisation of the inferior abdominal wall is variable, and detailed preoperative knowledge of the anatomical variations is necessary in patients undergoing autologous breast reconstruction. CTA allows detailed evaluation of all surgically relevant arteries.

Planning deep inferior epigastric artery flaps for breast reconstructions: a comparison between CT, MR and colour-Doppler US A. Cina, L. Barone-Adesi, P. Rinaldi, A. Cipriani, S. Santoro, M. Salgarello, L. Bonomo; Rome/IT (acina@sirm.org) Purpose: Deep inferior epigastric artery perforators free flaps (DIEP) surgery has become the state of the art for autologous breast reconstructions borrowing skin and fat tissue from the abdominal wall. The role of imaging is crucial during the DIEP planning to select a perforator pedicle adequate to supply the flap. Caliber of perforator and its intra-muscular course (MC) are important data for the surgeon. ence and grade of nodal varices was plotted against the presence of limb edema. Results: incidence of oedema was 35.5% in the operated group and 30.4% in the untreated one (p = ns).The incidence of lymph node varices, oedema and no oedema, was 64.8% and 12.2% in the operated group (p < 0.01) and, respectively, 35.3% and 11.5% (p < 0.05) in the untreated group. In the treated group, two patients without oedema presented grade 2 dilatation; this was encountered in 10 patients (28.5%) in the oedema group. For the untreated group, there was no grade 2 dilatation in the nonoedema subgroup and three patients (25%) showed grade 2 nodal varices in the oedema group. Transnodal veins were only observed in the operated group. Conclusion: The results infer a high prevalence of groin intranodal venous varices in patients with lower limb venous varices and oedema. Transnodal veins have, to the best of our knowledge, not been previously described. This finding may represent a compensatory mechanism or an indirect cause of oedema. MRA. The images were reviewed by two radiologists independently. The image quality was graded using a five-point scale: grade 5, good demonstration of interlobar artery; grade 4, well demonstration of segmental artery; grade 3, good demonstration of anterior or posterior renal artery; grade 2, fair demonstration of main renal artery only; grade 1, poor demonstration of main renal artery. The diagnostic confidence was graded by a four-point scale: excellent, good, fair and incapable to make diagnosis. The Wilcoxon signed rank test was used for comparison between the two techniques. A two-tailed P value of 0.05 or less was considered statistically significant. Results: IFIR MRA showed comparable imaging quality as compared with 3DCE MRA in 40 cases (46%), inferior in 33 cases (37.9%) and superior in 14 cases (16.1%) (P= 0.001). IFIR MRA showed equal diagnostic confidence as compared with 3DCE MRA in the majority of the cases (N=65, 74.7%) and even superior in 6 cases (6.9%). There was no statistical significance between the two techniques (P = 0.059). Conclusion: IFIR MRA is a non-contrast-enhanced MRA technique which has high diagnostic confidence for evaluation of renal artery stenosis.

Simultaneous acquisition of renal morphology and function using a CT perfusion protocol A. Helck, M. Notohamiprodjo, U. Schönermarck, M. Wessely, R. Morello, K. Nikolaou, D.-A. Clevert, M.F. Reiser, C. Becker; Munich/DE (andreas.helck@med.uni-muenchen.de) Purpose: To verify the accuracy of a renal CT-perfusion protocol for assessing renal function. Methods and Materials: Eleven patients with renal graft dysfunction received a CTperfusion protocol using a 128-slice CT-scanner with continuous bidirectional tablemovement, allowing to cover a scan range of 14 cm within 1.5 sec. Twelve scans of the entire kidney were acquired every 3.5 seconds to obtain renal perfusion data. Tube potential and current were lowered to 80 kV/120 mAs (BMI < 25) and 100 kV/120 mAs (BMI ≥ 25), respectively. Perfusion data were corrected for haematocrit and CT-derived glomerular filtration rate (GFR) was calculated by the modified Patlak method. GFR obtained from a two-point plasma disappearance curve of Iomeprol served as the standard of reference (blood samples 2h and 4h after CT/slope-intercept method). Maximal HU-values in the iliac, renal artery, and renal vein were quantified. Results: GFR obtained from CT-perfusion correlates well with the GFR derived by Iomeprol clearance with a correlation coefficient of r = 0.85 (p < 0.005). Using the optimal of the 12 dynamic phases, maximum mean attenuation in the iliac artery was 357 ± 129 HU, in the renal artery 341 ± 118 HU, and the renal vein 176 ± 28 HU. The mean amount of contrast medium required was 35 ml. Conclusion: A renal CT-perfusion protocol offers the possibility to determine the GFR and thus facilitates simultaneous assessment of morphology and function. In addition, the CT-perfusion protocol helps to reduce the amount of contrast medium, which is beneficial for patients with impaired renal function. Results: MRI of the foetal heart showed good anatomical correlation with the US images in the normal fetuses. MRI also showed good correlation in the pathologic cases, being superior to US in the conjoined twins case, as the US failed to demonstrate the correct anatomy of the two hearts. Conclusion: Foetal MRI with SSFP sequences in static and cine-resonance acquisition may help in the evaluation of the normal foetal heart and identification of the main morphological alterations related to CHD, when the US examination is limited by the foetal position, maternal obesity, and oligohydramnios. Additionally, both obstetric US and foetal MRI presented good correlation of findings and can be used as adjunct imaging techniques.

We compared CT, MR and CDS to assess accuracy in selecting and measure the "dominant" perforator, employed to supply the flap, and to estimate its MC. Methods and Materials: Ten women (mean age 49) underwent before surgery: 64-slices CT (90 cc iodine contrast 400 mgI/d)l; 1.5 T MR (LAVA sequences, 12 cc Gadobutrol) and CDS (12 MHz probe). Surgical findings were the standard of reference.

The dominant perforators was correctly identified in 9/10 cases (90%) by CT, 9/10 (90%) by MR and 10/10 by CDS (100%). Accuracy for evaluation of perforator's caliber (< 1 mm, 1-1.2 mm, 1.3-1.5 mm, > 1.5 mm) was 90% for CDS, 70% for CT and 60% for MR, with a systematic overestimation for CT (mean error 0.8 mm) and MR (1 mm). Accuracy for IMC evaluation (< 1 cm, 1. , and 11 reperfused PAVMs (due to recanalisation, opening of collaterals, insufficient packing) in 7 patients (interval 3 months-5 years). Interestingly, HAVM occurred predominantly in females (78.7%). Arterio-venous shunts lead to right-heart-insufficiency and biliary necrosis due to a left-to-right shunt. Arterio-portal sinusoidal shunts lead to pseudo-cirrhosis and hyperplastic nodular liver changes similar to FNH; arterio-portal pre-sinusoidal shunts lead to portal hypertension. 3 patients with symptomatic HAVM were treated symptomatically.

A S287 C D E F G H Methods and Materials: Six country pigs (46±4 kg) received stent placement (3.0 mm) of the proximal LAD. and balloon catheter placement within the stent and inflated to obtain luminal narrowing of the vessel diameter (50% and 75% diameter stenosis, respectively, by flow wire). All animal models underwent adenosine-mediated (280μkg/min) dynamic stress and rest CT imaging using a fast Dual-Source CT system with acquisitions every second for 30s. CT-based MBF was derived using a model-based parametric deconvolution method. Fluorescent microspheres were injected at each perfusion state. Results: All study procedures were performed without complications. At rest and stress, the average MBF was 93±6 and 125±11 ml/100 ml/min for rest and stress imaging, respectively. There was a strong positive association between CT-based and microsphere-based MBF (r = 0.86, p < 0.001) without overestimation by CT (average difference: 0.005±0.9 ml/g/min, p = 0.9). In addition, CT-based MBF within myocardial segments pertaining to 50% and 75% coronary stenoses was strongly correlated with microsphere MBF (r = 0.67, p = 0.005) with lowest MBF at 75% luminal narrowing (2.1±0.1 vs. 1.4±0.4 vs. 0.8±0.3 ml/g/min for 0%, 50%, and 75% luminal narrowing under stress, respectively; p < 0.01) similar to microsphere MBF. Conclusion: CT-derived MBF measurements correlate well with microsphere MBF. Perfusion CT may allow for assessment of haemodynamic significance of different degrees of coronary artery stenosis. Purpose: To quantify differences in regional myocardial perfusion at different levels of coronary stenosis using dynamic perfusion dual source CT (DSCT). Methods and Materials: In five pigs, an adjustable, remote controlled hydraulic occluder was placed around the LAD to induce various degrees of coronary stenosis. In addition, catheters were placed into the aortic arch and the coronary artery distally to the occluder for pressure measurements and fractional flow reserve (FFR) determination. Myocardial blood flow (MBF) was assessed during infusion of contrast using a second generation DSCT. Images were acquired at baseline and under maximal adenosine vasodilatation at FFR = 1 (no stenosis), FFR = 0.7 (moderate stenosis) and FFR = 0.5 (severe stenosis). MBF measured in the downstream LAD perfusion area was compared with the remote posterior myocardium. In addition, correlation between MBF and FFR was investigated. Results: Adenosine resulted in an increase in MBF from 110 ± 13 to 262 ± 12 ml/100 ml/min (P < 0.05). In the presence of adenosine, both moderate and severe stenosis reduced MBF in the LAD perfusion territory (204 ± 45 and 107 ± 29 ml/100 ml/min both P < 0.05 vs no stenosis) as compared to MBF in the remote myocardium (317 ± 29 and 254 ± 26 ml/100 ml/min). Moreover, percent MBF reduction correlated very well with the FFR measurements (R 2 = 0.90, P < 0.05). Conclusion: Dynamic perfusion DSCT is suitable for the assessment of myocardial hypoperfusion and can be used to discriminate between several degrees of clinically relevant coronary stenoses in a large animal model.

Dual energy CT for the detection of delayed enhancement in reperfused chronic infarction: a comparison with magnetic resonance imaging and histopathology in a porcine model J.M. Kerl, S. Deseive, R.W. Bauer, T. Lehnert, B. Schell, T.J. Vogl; Frankfurt a. Main/DE (matthias.kerl@kgu.de) Purpose: To evaluate the performance of delayed enhancement dual energy CT (LE-DECT) for the detection of irreversibly damaged myocardium compared with delayed-enhancement magnetic MRI and histology in a porcine model of reperfused chronic myocardial infarction. Methods and Materials: MI were induced by 30-min occlusion of the proximal left anterior descending coronary artery in 8 healthy minipigs. 61 days after LAD occlusion LE-DECT was performed 5, 10 and 15 minutes subsequent to contrast material injection using a DSCT-scanner in dual energy mode. 50 minutes after LE-DECT imaging, LE-MRI and histopathology were acquired. Two experienced radiologists interpreted in consensus all imaging series in a random fashion and blinded to the results of the other techniques for the presence of LE using the AHA 17-segment model. Results were compared with histopathology. Results: Of totally 136 myocardial segments in 8 minipigs, histopathology found scar tissue in 27 segments. The best results were observed in CT 10 minutes after contrast material injection. Here, LE-DECT was able to determine segments containing scar tissue with 76% sensitivity, 93% specificity, 70% PPV, 94% NPV and 89%

Measurements of the great vessels for the detection of aortic coarctation with fetal cardiovascular MRI: preliminary experience L. Manganaro, V. Vinci, A. Tomei, F. Fierro, M.E. Sergi, P. Sollazzo, S. Bernardo, F. Ventriglia, M. Marini; Rome/IT (lucia.manganaro@uniroma1.it) Purpose: Aortic coarctation (CoA) is the most common duct-dependent cardiac defect missed at routine physical screening of the neonate; indeed, despite new development in foetal echocardiography, prenatal detection rate for congenital outflow tract anomalies is relatively low. This study aimed to evaluate the ability of foetal cardiovascular MRI to characterise and measure anatomical details in order to offer new diagnostic possibilities in the detection of CoA. Methods and Materials: From November 2009 to July 2010, we enrolled 10 foetuses with the US suspect of CoA. Foetal cardiovascular MRI was performed with a 1.5-T Magnet System, in standard condition and without mother sedation. Retrospectively, we evaluated the mediastinal pulmonary artery to ascending aorta diameter ratio, according to the preliminary experiences in echocardiography which considered 1.60 as limit for normality. Results: We diagnosed CoA in 7/10 foetuses; we disconfirmed the US suspect of CoA in 2/10 cases identifying a ratio under 1.60; in one foetus the ratio was exactly 1.60, therefore classified as diagnostically inconclusive. The 7/10 pathological cases were confirmed either by after birth follow-up (5 cases) or autoptical findings (2 cases). After birth follow-up confirmed, in 2/10 cases, the normal aortic diameter findings of MRI. At last, the unclassified case at US and MRI find out to be normal in the after birth follow-up. (Harald.Quick@imp.uni-erlangen.de) Purpose: MR-guided transcatheter, transarterial aortic valve implantation (TAVI) has been performed in swine following device modifications towards MR compatibility of the comercially available stent valve delivery device.

The self-expandable Medtronic CoreValve® aortic bioprosthesis (Medtronic Inc). is composed of a nitinol stent frame with integrated trileaflet valve and is either implanted via femoral or subclavian artery. Its 18 Fr delivery catheter comprises the crimped prosthesis which can be released stepwise maintaining transaortic blood flow. The original catheter shaft revealed ferromagnetic braiding, considerably compromising MR imaging and safety. Device modifications obviating any metal braiding resulted in full MR compatibility. MR-guided TAVI was performed on 8 swine (75-85 kg) via subclavian access on a 1.5 T Avanto MRI system (Siemens) equipped with in-room monitor. Catheter placement and stent release were performed under TrueFISP real-time MR guidance providing 5 fps. Results: Device modifications resulted in artefact elimination and excellent real-time visualisation of catheter movement and valve deployment using rt-TrueFISP imaging. MR-guided TAVI was successful in 6/8 swine. Post-interventional therapeutic success could be confirmed using ECG-triggered cine-TrueFISP sequences and flow-sensitive phase contrast sequences revealing or excluding regurgitation. Final stent valve position was confirmed by ex vivo histology. Conclusion: The self-expandable CoreValve aortic stent-valve is potentially suited for real-time MRI-guided placement after suggested design modifications of the delivery-system. MR imaging provided excellent pre-interventional anatomic and functional evaluation of the native aortic valve, precise real-time instrument guidance allowing accurate placement of the stent-valve within the native aortic annulus, and finally, detailed post-interventional evaluation of therapeutic success.

Accuracy of dynamic CT-based adenosine stress myocardial perfusion imaging to estimate myocardial blood flow in a porcine animal model with various degrees of coronary artery stenosis F. Bamberg, F. Schwarz, T. Sandner, B. Wintersperger, R. Hinkel, E. Baloch, A. Becker, M.F. Reiser, K. Nikolaou; Munich/DE (fabian.bamberg@med.uni-muenchen.de) Purpose: To determine the accuracy of CT-based dynamic stress myocardial perfusion imaging to estimated myocardial blood flow (MBF) in a porcine animal model with variable degrees of coronary artery stenosis in comparison with microsphere-derived MBF.

A D E F G H units offer a tomography mode, using a half-rotation to reconstruct tomographic images. Here, we analysed the correlation of aortic valve diameters obtained from pre-procedural MDCT and intra-procedural C-arm-tomography acquisitions. Methods and Materials: 15 patients underwent ECG-triggered, high-pitch CTangiography of the thorax and abdomen 3-30 days prior to PAVR. During valve replacement, all patients underwent C-arm-tomography while contrast material was administered intraarterially at the level of the aortic valve. Rapid ventricular pacing was used to reduce motion artefacts. For this study, two readers independently determined valve diameters on a coronal and double-oblique plane. All datasets were anonymised, presented in random order and readers had to adjust planes manually. Results: Aortic valve diameters could successfully be measured in all datasets.

Although there was a tendency towards better interobserver agreement for MDCT, it was excellent in both modalities (ICC = 0.82 vs. 0.89). Overall, agreement between both modalities was high (ICC = 0.80). There was a small but significant difference in means of valve diameters in the coronal plane (23.9 mm vs. 22.9 mm, p = 0.008 Purpose: The CIRTACI is a prospective multicenter study exploring immediate reactions after contrast media. The goal was to differentiate between allergic and non-allergic hypersensitivity to iodinated and gadolinium contrasts. Final analysis of the 330 cases in 31 centers will be reported.

The following data were obtained: clinical severity (Ring and Messmer classification), plasmatic tryptase and histamin levels, allergological skin tests performed 6 weeks to 6 months after the reaction. 5 Gd chelates or 10 iodinated agents were tested. Positivity of the intradermal test was assessed by the increase in diameter of the injection papula together with inflammatory signs. The final diagnosis concerning the culprit contrast agent had to be made between allergic and non-allergic hypersensitivity reaction based on these clinical, biological and allergological criteria. Results: Among 319 available cases, there were 179 grade 1, 91 grade 2, 39, grade 3, 5 grade 4 and 5 undetermined. 275 patients could follow the entire protocol, among them 233 reactions to iodinated and 42 to gadolinium agents. All contrast agents could be responsible for a reaction. Allergic hypersensitivity reactions were almost always found in grade 3 and 4 reactions. In grade 1 (diffuse urticaria) allergy was found in about 30% of the cases. Detailed results of the study will be presented at the meeting. Metabolite and triglyceride signal strength in relation to water signal strength from the same voxel in spectra were estimated using General Electric's SAGE package.

The 100 participants completed all study procedures and 86 participants (46males and 40 females, age 29±11 years; BMI 22.1±2.6 kg/m 2 ) were included in the final analysis. 14 participants (14%) were excluded because of poor spectra. The spectral peaks were identified with following chemical shifts: water, choline, creatine and triglyeride. No statistical significant inter-gender difference of creatine and choline concentrations were observed (p > 0.05), while myocardial triglyceride concentration was 88% higher in myocardial tissue in females than in myocardial tissue in males (p < 0.05), despite the former having a lower mean body-mass index.

The concentration of myocardial triglyceride is related to gender, and higher in females, while the concentrations of myocardial creatin and choline are not related to gender.

Evaluation of aortic distensibility and cardiac function in normal wild type and diabetic mice with MRI at 9.4 T P. Fries, J. Reil, A. Müller, R. Seidel, K. Altmeyer, A. Massmann, A. Bücker, G. Schneider; Homburg/ DE (drpeterfries@googlemail.com) Purpose: The aim of this study was to compare aortic distensibility (AD) and cardiac function in wild type (C57BL/6) and type II diabetic (db-/db-) mice by means of cine MRI. Methods and Materials: C57BL/6 and db-/db-mice (n = 5 each) were evaluated using a horizontal bore 9.4 T MRI animal scanner (Bruker, Karlsruhe, Germany) with a dedicated 2 x 2 phased-array surface coil. A double-gated cine sequence (Intragate Flash, TR/TE = 8.9/2.1 ms, FA = 10°, FOV = 2 x 2 cm, matrix = 384 x 384) located perpendicular to the ascending aorta was acquired for imaging of AD. Cardiac function was assessed acquiring five consecutive short axis cine sequences covering the whole ventricles. AD was defined as the difference of the subintimal vessel area at systoly and diastoly. Enddiastolic (ED), endsystolic (ES), stroke volume (SV) and ejection fraction (EF) were analysed using an image processing software (Osirix). Student's T-Test was used for statistical analysis (P < 0.05). Results: Diabetic animals showed a significant lower AD (mean ± SD: 0.467 ± 0.05 mm 2 ) compared to C57BL/6 (0.668 ± 0.1 mm 2 , p = 0.004). ED volumes of diabetic animals were significantly smaller (db-/db-: 41.9 ± 7.1 ml, C57BL/6: 55.5 ± 11.8 ml, P = 0.03) while there were no significant differences in SV between the two groups (db-/db-: 32.7 ± 8.1 ml/C57BL/6: 30.4 ± 7.0 ml). EF was significantly larger in diabetic animals (78.1 ± 0.1%) compared to the C57BL/6-group (54.9 ± 0.2%, P < 0.001). Conclusion: Diabetic mice demonstrate a significantly lower AD. As there is no significant difference in SV these characteristics are most likely based on vascular changes. In addition, db-/db-animals show signs of restrictive cardiomyopathy with decreased LVEDV and increased EF as compared to the reference.

Rotating Nijmegen/NL (h.dekker@rad.umcn.nl) Purpose: To evaluate the application of a new CT request form and the simultaneous introduction of a special outpatient clinic for patients at high risk for contrastinduced nephropathy (CIN). Methods and Materials: During six months, request forms were collected from outpatients of the department of Urology, who were ≥18 years and undergoing elective ceCT examinations. Information to be indicated on the request form was: MDRD-GFR in ml/min/1.73 m 2 and the risk factors: diabetes, heart failure, peripheral vascular disease, age > 75, anaemia, symptomatic hypotension, large doses of contrast medium (> 150 ml), dehydration, and nephrotoxic drugs. Based on the risk factors a risk score was calculated to define patients at high risk for CIN, who were referred to the "outpatient CIN clinic". Results: Of the 402 patients included, 24% had a pre-CT MDRD-GFR < 60 ml min/1.73 m 2 . Of these, 50 patients were defined as high risk and referred to the "outpatient CIN clinic". Thusfar, 32 patients have completed the protocol. Five patients did not need hydration, 27 patients received pre-and post-hydration in combination with the ceCT. Five patient were hydrated overnight to avoid overhydration. In nearly all patients, data given on the request form were reliable, and the MDRD recently determined. In the patient group with a MDRD < 60 ml/min/1.73 m 2 pre-CT, there was one patient with CIN. Conclusion: The introduction of a new request form allows good identification of patients at risk for nephropathy, while the "outpatient CIN clinic" resulted in closer patient monitoring and optimal preparation, thereby improving patient safety.

Comparison of lesion detection and characterisation in patients with neuroendocrine tumours using DOTATOC-PET in correlation with contrastenhanced MRI and contrast-enhanced CT F.L. Giesel, K. Clemens, S. Wulfert, J. Weller, N. Schimpfke, U. Haberkorn, H.-U. Kauczor, S. Ley; Heidelberg/DE (f.giesel@dkfz.de) Purpose: Most of the gastroenteropancreatic neuroendocrine tumours (GEP-NETs) present an increased somatostatin receptor expression. Therefore, PET with 68Ga-DOTATOC has both high sensitivity and specificy in tumour staging. However, additional morphologic characterisation of tumour lesions is often essential, e.g. the relation to adjacent structures to determine weather the lesion is resectable. We evaluated the rate of successful lesion characterisation comparing contrast enhanced CT with MRI each in correlation with DOTATOC-PET. Methods and Materials: 8 patients with GEP-NET were investigated using late enhanced MRI (1.5-Tesla, T2w-BLADE and T1w-VIBE, before and 15 min after injection of Primovist®, BayerSchering/GER), contrast enhanced (ce) CT and DOTATOC-PET imaging. DOTATOC-PET and ce-CT were performed in a BIO-GRAPH-6 PET/CT (Siemens/GER). Results: Using DOTATOC-PET as standard reference for specific detection of metastases, 28 liver metastases were diagnosed. All 28 liver metastases were also detected by ce-MRI and extension could be adequately assessed using the late phase Primovist®-MRI in all lesions (100%). In contrast, not all liver metastases were detectable by ce-CT (79%) and neither relation to adjacent structures nor exact lesion extension was assessable in the occult (21%) liver metastases. Conclusion: Late ce-MRI was superior to ce-CT in providing additional morphologic characterisation and exact lesion extension of hepatic metastases from neuroendocrine tumour detected with DOTATOC-PET. Therefore, late enhanced Primovist®-MRI seems to be the adequate imaging modality for combination with DOTATOC-PET to provide complementary (macroscopic and molecular) tumour characterisation in hepatic metastasised GEP-NETs.

Iodine contrast media iso-attenuating with gadolinium doses proven diagnostic in CTA and angiography: a way to void both CIN and NSF in azotemia? U. Nyman 1 , B. Elmståhl 2 , H. Geijer 3 , P. Leander 2 , T. Almén 2 , M. Nilsson 2 ; 1 Trelleborg/SE, 2 Malmö/SE, 3 Örebro/SE (ulf.nyman@bredband.net) Purpose: To establish iodine (I) contrast medium (CM) doses iso-attenuating with gadolinium (Gd) CM doses regarded as diagnostic in computed tomography angiography (CTA) and percutaneous catheter-angiography/vascular interventions (PCA/PVI) in azotemic patients. Methods and Materials: CT Hounsfield units (HU) were measured in 20-mL syringes containing 0.01/0.02,/0.05/0.1 mmol/mL of iodine or gadolinium atoms and placed in phantoms. Relative contrast was measured in 20-mL syringes filled contrast medium (mean volume 85.6 ml). 21.8% of all patients had at least one risk factor (i. a., renal impairment), 7.3% were at-risk patients with allergies or who had previously reacted to contrast medium. Antiallergic pretreatment before contrast medium administration was given in 1144 patients (0.7%). Adverse events were documented and the image quality was assessed. Results: A diagnosis was possible in 99.5 % of all cases. The image quality was rated good or excellent in 92.2%. The adverse event rate (e.g., nausea, urticaria) observed was 0.6% in all patients, 1.6% in patients with allergies and 6.0% in patients with a previous reaction to contrast medium. Adverse events occurred more often in women than in men (p < 0.001). The rate of adverse events was not increased in higher doses of Iobitridol, even if administered to high-risk patients. Conclusion: Iobitridol was shown to be a well-tolerated contrast medium that allowed a diagnosis with very good or excellent imaging quality to be made in more than 90% of all cases. The incidence of adverse events was low even in patients with risk factors.

Contrast-enhanced ultrasound (CEUS) to assess the perfusion of patients with peripheral arterial disease (PAD) treated with endothelial progenitor cells (EPCs) F. Alessandrino, S. Alessi, C. Dellabianca, V. Arici, C. Perotti, F. Calliada; Pavia/ IT (pragia@hotmail.com) Purpose: Since contrast-enhanced ultrasound (CEUS) has been recognised as a valid method for the assessment of muscle perfusion deficits, the aim of this study is to evaluate CEUS as a diagnostic tool to assess the perfusion of the muscle of patients with PAD treated with endothelial progenitor cells (EPCs) and its correlation with other diagnostic tools as ankle brachial index (ABI) and clinical data. Methods and Materials: 12 Patients underwent intramuscular injection of EPCs from September 2008 to April 2010. CEUS was performed before therapy and scheduled at 3, 6, 9, 12 months after the therapy. Patient performed baseline and CEUS examination with Philips IU22 (Philips Bothell WA) and GE Logiq e9 (GE, Milwaukee, WI). The perfusion was assessed evaluating contrast agent timeintensity curves with QLab (Philips Bothell WA) and Qontraxt (Amid and R&D, Bracco, Milan, Italy) software. Region of interest (ROI) were placed on the calf muscle and on posterior tibial artery (PTA). Results: All the enrolled patients showed relevant clinical improvement, except one, who, later, underwent leg amputation. Mean AUC before treatment was 688.43 +-SD 172.34 dB s on PTA and 545.18 +-SD 188.39 dB s on calf muscle. Mean AUC at six months after treatment was 1054.62 +-SD 360.74 dB s on PTA and 1372.9 +-SD 478.21 dB s on calf muscle. Conclusion: CEUS is a valuable tool to display variation of the perfusion of the calf muscle and tibial artery in patients treated with EPCs with good correlation with clinical data.

Changes of water diffusion coefficient in the kidney after application of iodinated contrast media G. Jost, D. Lenhard, M.A. Sieber, P. Lengsfeld, J. Huetter, H. Pietsch; Berlin/DE (gregor.jost@bayerhealthcare.com) Purpose: The viscosity of x-ray contrast media (CM) may have an influence on the diffusion of water molecules within the kidney. The aim of our preclinical study was to investigate the changes in the apparent diffusion coefficient (ADC) of the kidney after application of monomeric and dimeric CM using diffusion weighted MRI. Methods and Materials: Han-Wistar rats (8 per group) received an intravenous injection of dimeric CM (Iodixanol) and monomeric CM (Iopromide) at a dosage of 4 g I/kg b.w. The control group received saline (0.9% NaCl) at the same volume. The renal ADC was monitored up to 40 min p.i. by diffusion weighted MRI using a 1.5 T clinical MR-scanner (Avanto, Siemens). The correlation between CM viscosity and the ADC of water was assed by phantom measurements. Results: A slight ADC increase was observed immediately after application of the monomeric agent (ΔADC = 80 μm²/s) and isotonic water (ΔADC = 89 μm²/s), which normalised nearly completely to the baseline level at 40 min p.i. In contrast, after application of the dimeric agent a strong decrease of the ADC was determined. This was most prominent 6-12 min p.i. and did not recover to baseline level during the investigation (ΔADC = -418 μm²/s). The in vitro measurements yield an agentspecific correlation between increasing viscosity and decreasing water diffusion. Conclusion: A clear difference in the renal ADC was observed between monomeric CM/ water (saline) and dimeric CM. The strong ADC decrease observed for the dimeric agent might be caused by the high viscosity of dimeric CMA. Nephrogenic systemic fibrosis (NSF) is a rare, but serious, acquired, idiopathic disorder. To date, the exact pathogenesis of NSF has not been identified, but an association has been postulated between the administration of gadoliniumbased contrast agents (GBCA) and the occurrence of NSF. The purpose of this study was to test the hypothesis that long-term gadolinium (Gd) deposits play a role in the induction of NSF-like lesions in an animal model.

Han-Wister rats were administered a triple intravenous injection of Gd-DTPA-BMA formulated without excess ligand at a dose of 2.5 mmol/ kg body weight at different time intervals (24h or 14-, 28-, and 56-days). Gd concentrations were measured by inductively coupled plasma mass spectrometry and occurring skin lesions were determined macroscopically and histologically. (anno.graser@med.uni-muenchen.de) Purpose: We sought to determine the impact of automated exposure control (AEC), organ-based tube current modulation (TCM) and iterative reconstruction (IR) technique on effective radiation dose and image quality in an experimental set-up and a consecutive cohort of patient undergoing CT scanning of the brain. Methods and Materials: An Alderson-Rando-phantom equipped with thermoluminescent dosimeters was used to estimate the radiation exposure of organs within the brain and neck scanned. We also measured the grey and white matter signal-to-noise ratios from CT brain images of 100 consecutive patients and images were subjectively evaluated for quality criteria in different territories of the brain. Moreover, measures of effective radiation exposure were collected. Results: In the phantom, AEC and TCM significantly reduced the radiation exposure of the lenses, cerebrum, cerebellum and thyroid gland by 41.9%, 34.5%, 30.5% and 34.9%, respectively. Head CT scan from patients investigated with AEC and TCM and IR had significantly higher image quality (p < 0.05) and identifiability of intracerebral structures (< 0.05). The CT dose index and dose length product was 24.1% and 20.2% lower, respectively. Conclusion: The combination of AEC, TCM and IR improves image quality while radiation exposure can be reduced, particularly in dose-sensitive organs, such as the lenses and thyroid gland.

Feasibility of cerebral CT angiography using lower tube voltage ( Purpose: To determine the radiation dose delivered during a one-stop scanning programme of the nervous system performed on a 320-detector row dynamic volume CT. Methods and Materials: Patients who were suspected to have suffered acute ischaemic stroke from September 2009 to January 2010 were evaluated using a one-stop scanning programme on a 320-detector CT. Scan plans included NCCT covering the whole brain as a mask image, CT arterial angiography, CT venous angiography and whole-brain perfusion imaging with 19 total exposures. DLP measurements were displayed after each scan on the scanner's console. Radiation exposure was quantified with a DLP conversion factor of 0.0021 mSv • mGy-1 • cm-1. Results: A total of 45 cases of suspected acute ischaemic stroke were successfully identified using the scanning procedure and used in the analysis of radiation dose. An average DLP covering the entire brain resulted in a dose of 2695.3 mGy•cm for an average effective radiation dose of 5.66 mSv. The effective dose of the NCCT was 0.45 mSv. The greatest contribution to the overall dose was the CTA of 3.27 mSv. The average dose generated from CTV was 1.96 mSv. Conclusion: Although the radiation dose was higher than that of brain NCCT scanning, the one-stop scan mode provided significantly more anatomical detail. As it is possible to carry out subtraction in any phase, pure arterial and venous phase images can be obtained. Furthermore, cerebral ischaemia in any part of the brain or functional changes in any region of the brain can be detected.

Evaluation of dose reduction in 64-row CT of the cervical spine with adaptive statistical iterative reconstruction (ASIR) using standard clinical protocols M. Körner, R. Hempel, S. Wirth, M.F. Reiser, U. Linsenmaier; Munich/ DE (markus.koerner@med.uni-muenchen.de) Purpose: To evaluate radiation exposure from 64-row CT examinations of the cervical spine using adaptive statistical iterative reconstruction (ASIR) compared to standard filtered back projection (FBP) raw data processing. Methods and Materials: 64-row C-spine CT with a standard clinical protocol of 67 studies (mean age 55 ± 20 y, 30 males, 37 females) with FBP (LightSpeed VCT XT, GE Healthcare, Waukesha, WI) was retrospectively compared to 80 studies (50 ± 20 y, 49 males, 31 females) with ASIR (Discovery 750HD, GE Healthcare). The scan parameters for both examinations were identical (120 kV, dose modulation, 0.625 mm collimation, pitch 0.531:1) but the noise index (NI) was increased from 5 to 25 with an ASIR level of 30%. From the dose reports generated automatically, the scan length, CTDI, and DLP were recorded. Statistical tests were performed with the Mann-Whitney U test. Results: In the FBP group, mean CTDI was 21.43 (±1.43 standard deviation), scan length was 186.30 mm (±23.60), and DLP was 441.15 mGy*cm (±51.73). In the ASIR group, CTDI was 9.57 (±3.36), scan length was 195.21 mm (±26.82), and DLP was 204.23 mGy*cm (±68.01). The differences were significant for CTDI and DLP (p < 0.001) and scan length (p = 0.013). Estimation of mean effective dose resulted in 2.38 mSv in the FBP group and 1.10 mSv in the ASIR group. All images were of diagnostic quality as routine clinical acquisition protocols were used. Conclusion: Dose reduction of about 50% can be achieved with the use of ASIR for 64-row CT of the cervical spine in standard clinical protocols.

Ultrasound-guided pain control in the cervical spine: a prospective randomised clinical trial A. Loizides, H. Gruber, S. Peer, K. Galiano, J. Obernauer, S. Ostermann; Innsbruck/AT (stefanie.ostermann@i-med.ac.at) Purpose: Pararadicular and facet-joint injections are widely used for alleviation of cervical pain. Injections are preferentially performed as fluoroscopy or computed tomography (CT)-controlled interventions. Ultrasound (US) provides real-time monitoring, does not produce ionising radiation, and is broadly available. Methods and Materials: We studied feasibility, accuracy, time-savings, radiation doses, and pain relief of pararadicular ultrasound-guided and CT-controlled facetjoint injections versus CT-controlled interventions in a prospective randomised clinical trial. Sixteen adult patients with chronic cervical pain were consecutively enrolled and evenly assigned to the ultrasound or a CT-group.

Eighty-two patients were referred for cerebral CT angiography because of screening for intracranial aneurysm. Forty-one patients were scanned with 120 kVp and 150 mAs eff after 70 mL of HC-CM (370 mgI/mL) administration and the other 41 patients were scanned with 80 kVp and 390 mAs eff after 70 mL MC-CM (300 mgI/mL) administration. Two groups of patients were compared with each other in terms of Hounsfield number of cerebral arteries, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), subjective degree of arterial enhancement, image noise, sharpness of arteries, and overall diagnostic image quality of both maximum-intensity projection and volume-rendered images. Results: Mean attenuations of cerebral arteries, SNR, and CNR were significantly higher in 80 kVp with MC-CM (390.0 ± 31.5 HU, 28.4 ± 3.7, 26.4 ± 3.7, respectively) than 120 kVp with HC-CM (280.9 ± 43.0 HU, 26.8 ± 4.5, 24.0 ± 4.4, respectively). 80-kVp protocol had significantly higher quality scores in arterial enhancement, sharpness of the artery, image noise, and overall diagnostic image quality compared to 120-kVp protocol. Effective dose of 80-kVp protocol (0.7 ± 0.0 mSv) was 26.5% lower than that of 120-kVp protocol (0.9 ± 0.1 mSv). Conclusion: The use of 80 kVp with MC-CM could improve arterial enhancement, SNR, and CNR, and provide superior image quality than 120 kVp with HC-CM, despite using smaller amount of iodine administration and decreased radiation dose.

Virtual non-contrast images of head CT angiography by dual-source CT dual-energy mode: initial experience D. Han, L. Wen, Y. Xu; Kunming/ CN (kmhandan@sina.com) Purpose: To investigate the clinical value of virtual non-contrast (VNC) images of head CT angiograpgy (CTA) by dual-source dual-energy mode (DECT). Methods and Materials: For 100 cases including 46 normal and 54 abnormal cases of which 15 were with haemorrhage, 13 infarction, 22 tumour and 4 calcification head CTA by DECT was performed. VNC images were generated from contrast images by DECT and were imported into DE VNC software after real non-contrast scan and DE-CTA. Mean CT values of brain parenchyma, signal-to-noise ratio, image quality, sensitivity of lesion detection and radiation dose were compared between real non-contrast scan and VNC. Results: Image quality was measured by a five-score grading criterion. That the score was greater than or equal to 3 was believed to meet the diagnosing requirement. The score of image quality was statistically significant difference between non-contrast images and VNC images (P < 0.05). The image quality in non-contrast scan was better than VNC. However, the scores of all VNC images were beyond 3. (2) Radiation dose of DE head CTA was lower about 9.2%, 47.6% than non-contrast scan and Neruo-DSA (radiation dose of Neuro-DSA for sum on non-contrast and enhanced dose), respectively.

(3) Sensitivity of lesion detection showed no significant difference between non-contrast scan and DE-CTA VNC. But tiny calcified lesions might not be demonstrated by VNC. Conclusion: DE head CTA could get head CTA images and cerebral VNC images in one scan and VNC images could meet the diagnosing requirement and significantly decrease the radiation dose.

Detection of aneurysms by 320-slice volume CT angiography in patients acutely suspected of having an intracranial aneurysm: feasibility of reducing tube voltage G. Sun, J. Ding, M. Li, Z.-H. Peng; Jinan/ CN (cjr.sungang@vip.163.com) Purpose: To prospectively assess the feasibility of lower tube voltage (80 KV)for detection of intracranial aneurysms by 320-slice volume CT angiogyaphy. Methods and Materials: Forty-eight patients with suspected subarachnoid hemorrhages underwent both CTA with 320-slice volume CT and DSA.Patients were randomly assiged to two groups (GA:80 KV, GB:120 KV). Contrast medium was injected automatically with bolus tracking. We compared vascular enhancement,image quality,sensitivity and specificity for detection of aneurysms, and radiation dose in two groups of CTA images. Results: The 80 kV protocol had a higher mean vessel attenuation (42% to 58% increase), correlated with a higher image noise and SNR than 120 kV, There were no significant differences in subjective image quality between patient group A and B regarding the scores for arterial enhancement, depiction of small arterial detail, interference of venous structures, and overall scores. Mean effective dose for the 80 kV was significantly lower than for the 120 kV (0.56 mSv±0.25 vs 1.84 mSv±0.002;-69.73%;p < 0.001), With DSA as the standard of reference ,the sensitivity and specificity of group A and group B were 94.12%,100%and 100%,100% respectively. There was excellent correlation between the maximum size of each aneurysm on CTA of both groups and 3D DSA.No statistical difference was found between mean maximum diameter measured on CTA of two groups and those measured on DSA images. Conclusion: Using 320-slice MDCT,dignostic image quality can be chieved with 80 kV for detection of intracranial aneurysms ,which resulted in significant reduction of effective dose without significant loss of dignostic image quality and accuracy of aneurysm. Figure IQF inv . Dose was expressed as average glandular dose (AGD). Performances have been evaluated in clinical setup using automatic exposure control (AEC) modality and in optimized setup. Optimization was based on contrast to noise ratio (CNR) measurements.

Results: DQE of all systems ranked between 55% and 80%, being higher in systems applying direct conversion technology. Generally, a higher DQE reflected in a higher IQF inv score. Target/filter combinations producing spectra of higher mean energy confirmed better performances than conventional Mo/Mo ones. Photoconductive switching readout exhibited a potential average 20% dose saving; also scanning acquisition showed improved image quality at equal dose compared to standard flat panel detectors.

The technological innovations allow better image quality with lower AGD, even though the AEC of some system could be further improved.

Results: Out of the 16 subjects 6 met after assignment the criteria for the randomised US procedure of the present study. All (US and CT) combined pararadicular and facet-joint injections were performed correctly. The duration of procedure and radiation dose was mean 4.6 (1.3-8.4) minutes and 22.7 (16-40) mGy*cm, respectively, for the ultrasound-guided CT-controlled group, and 10.4 (5.3-15.5) minutes and 144.8 (56-312) mGy*cm for the CT group. Both groups showed a comparable overall benefit from pararadicular and facet-joint injections.

The US-guided approach to the pararadicular compartment and the facet-joints in the cervical spine is feasible with minimal risks and the results indicate to date comparable pain reduction in the proposed study setting. However, zero radiation by US-guided performance in daily routine must be taken into account additionally.

Apple iPad and emergency CT brain interpretation: a phantom and clinical imaging study P. Purpose: The specifications of the Apple iPad meet ACR criteria for CT image interpretation; namely its display has a spatial resolution of 1024 x 768 pixels and it can display 16-bit DICOM images allowing radiologists adjust window width and level. To better quantify the clinical utility of the recently introduced iPad we compared its performance with a 3-megapixel diagnostic-monochrome display using CT-brain datasets and a contrast-detail phantom. Methods and Materials: 100 consecutive emergency CT-brain examinations were reported in consensus by two general radiologists using the Ipad. The datasets were also reported by one of two neuroradiologists using a PACs workstation. Reporting discrepancies were scored using the ACR Radpeer criteria. The Artinis CDRAD2.0 contrast-detail phantom was used to objectively evaluate the iPad and diagnostic display. Results: 75% of the CT datasets were reported in complete agreement. There were 3 clinically significant but understandable misses (Radpeer2b) with the iPad including a cervical fusion anomaly and two cases of early basal ganglia stroke. It is noteworthy that there were 4 clinically significant pickups when reporting with the iPad including parietal and supraorbital fractures (Radpeer3b). No significant difference in CDRAD score was demonstrated between the diagnostic monochrome (169points) and Ipad displays (163points). Conclusion: There were 3 clinically significant but understandable misses according to ACR criteria. The Ipad radiologists were blinded to previous imaging and clinical information which likely contributed to these discrepancies. No significant difference was demonstrated between displays using a contrast-detail phantom. iPad would appear to represent a very promising tool to facilitate remote review of imaging studies.

Automated slice positioning in MRI: application and reproducibility in brain examinations L. Gerigk 1 , M. Purpose: Different slice angulations often lead to limited comparability and inaccurate measurements of brain lesions in follow-up examinations. We evaluated a tool for automated slice positioning in brain MRI.

In 51 consecutive patients, a 3D localiser was acquired and retrospectively processed using the algorithm intended for automated positioning. Slices were reconstructed in predefined orientations and visually rated for accuracy, which was classified into optimal (2), acceptable (1) and need for manual intervention (0). For reproducibility, the categories optimal congruence (2), visible deviation but usable (1) and inacceptable (0) were used.

The accuracy of the alignment to the corpus callosum (CC) was optimal or acceptable in all but one patient (medium score 1.4), the resulting axial slice was rated with a medium score of 1.8. The best reproducibility was achieved parallel to the CC with a score of 1.7, followed by alignment to the temporal lobe with a score of 1.5 (right) to 1.6 (left). The reproducibility of slices through the internal auditory canal and the optic nerves was acceptable with scores of 1.0 to 1. Conclusion: Extra dose of doing a 3D compared to a 2D mammogram on this system is lower for dense breasts (BI-RADS 3 and 4) than for less dense breasts (BI-RADS 1 and 2). This should be taken into account when assessing the potential gain of using tomosynthesis in a screening setting.

The determination of dose-area-product trigger levels to predict the exceeding of thresholds for deterministic effect at the skin for interventional radiology: TRIR project L. Struelens 1 , K. Bacher 2 , F. Bleeser 3 , H. Bosmans 4 , E. D'Agostino 1 , M.-T. Hoornaert 3 , F. Malchair 5 ; 1 Mol/BE, 2 Gent/BE, 3 La Louvière/BE, 4 Leuven/BE, 5 Liège/BE (lara.struelens@sckcen.be) Purpose: The dose to the skin of the patient in interventional radiology/cardiology can sometimes be high, and even get to the dose limits for deterministic effects. In practice, the radiologist/cardiologist has no direct means to know the doses that are given to the skin of the patients during the procedures. The goal of the TRIR project is to verify the correlation between maximum skin dose and DAP values for different interventional procedures. The final goal is to determine specific trigger levels for reaching threshold values of deterministic effects where follow-up of the patient is advisable. Methods and Materials: Seven different interventional procedures are considered (cerebral embolisation (30 patients), chemo-embolisation (10 patients), radio-frequency ablations (30 patients), TIPSS (20 patients), embolisation of vena spermatica (40 patients), ERCP (30 patients) and biliary drainages (25 patients)) in 11 Belgium hospitals. For each procedure, the DAP value is registered. A sheet filled with up to 50 thermoluminescent dosemeters is wrapped around each patient to measure the procedure-specific maximum skin dose distribution. Results: Especially for TIPSS, chemo-embolisations of the liver and cerebral embolisations the dose value for deterministic skin damage (2Gy) was attained, with maxima up to 7.5 Gy for TIPSS (trigger value: 35500 μGy.m²). For the different procedures, a significant correlation was found between the DAP values and the measured maximum skin doses. However, this correlation is different for different procedures. Purpose: To quantify cumulative exposure to ionising radiation in patients with end stage kidney disease (ESKD) and to investigate factors which may be independently associated with risk of high cumulative effective dose (CED) exposure.

The study had local institutional review board ethical approval. We conducted a retrospective study of 394 period prevalent ESKD patients attending a single tertiary referral centre between 2004 and 2009. Patient demographics were obtained from case records. Details of radiological investigations were obtained from a central radiology database. CED was calculated using standard procedure specific radiation levels. High exposure was defined as CED> 50mSv, an exposure which has been reported to increase cancer mortality by 5%. Data were compared using Pearson Chi-squared and Mann Whitney U test or Kruskal-Wallis tests. Results: 394 patients were followed for a median of 4 years, (1518 patient years follow-up). Of these 63% were males. Seventeen per cent of patients had a cumulative effective dose (CED) of > 50mSv. Computed Tomography (CT) accounted for 9% of total study numbers while contributing 61.4% of total study dose. Median cumulative dose and median dose per patient year were significantly higher in the Haemodialysis (HD) group (15.13mSv and 5.79mSv respectively) compared with the transplant group (2.9mSv and 0.52mSv respectively), p < 0.001. Conclusion: Patients requiring renal replacement therapy are at risk of cumulative exposure to significant levels of diagnostic radiation at young age. The majority of this exposure is imparted via CT examinations to patients in the Haemodialysis group.

Proposed European protocol for the estimation of average glandular dose for digital breast tomosynthesis D.R. Dance 1 , K.C. Young 1 , R.E. van Engen 2 ; 1 Guildford/UK, 2 Nijmegen/NL (daviddance@nhs.net) Purpose: There is a need for a European protocol for the estimation of average glandular dose for digital breast tomosynthesis. A protocol is proposed which is an extension of the current European methodology for dosimetry in conventional mammography.

The formalism introduces t-factors for single projections and T-factors for a complete examination which are used in conjunction with the current methodology for estimating dose for conventional mammography. The factors were calculated for a wide range of x-ray spectra, breast thicknesses and breast glandularities using a Monte Carlo model. Two geometries were considered: one using a fixed detector and the other a scanning narrow beam. Results: t and T-factors show little dependence on breast glandularity (less than 2%). For the fixed geometry, tables of t-factors are provided as a function of breast thickness and projection angle, which may be used to calculate T for any selection of projection angles in the range -/+ 30 degrees. Dependence of T on x-ray spectrum and breast thickness is less than +/ 4%. Typical values of T are in the range 0.94-0.99. For the geometry using a scanned beam, only tables of T are provided. Conclusion: The protocol is easy to implement and is a natural extension of the methodology currently used for dosimetry in conventional mammography. T-factors for the geometry with a fixed detector are close to 1. For the same tube loading, the doses for conventional mammography and for breast tomosynthesis will be similar.

Contribution of compression paddle flexibility to estimation of breast glandularity N. Geeraert, R. Klausz, M. Lemuhot, D. Sundermann, S. Muller; Buc/FR (nausikaa.geeraert@ge.com) Purpose: Flexible compression paddles are used to improve patient comfort in mammography, but may reduce the accuracy on the compressed breast thickness (CBT). This study investigates if it is acceptable when CBT is used to determine breast glandularity for the automatic optimisation of parameters (AOP) at acquisition. Methods and Materials: The operation over one week of 1016 Senographe DS and essential mammography equipment (GE Healthcare, Chalfont, UK) was sampled using remote access. The parameters retrieved were paddle type, anode/filter, kV, mAs, detector dose and CBT. The detector dose is determined for the lowest level in the image compatible with a glandular area detected by the AOP from a pre-exposure. For each image the glandularity corresponding to these parameters (peak glandularity) was computed with the same algorithm as the AOP. Two populations were separated depending on the paddle: rigid or flexible, and the distribution of glandularities compared. Results: Data from 58974 images were collected, with 30227 rigid paddles and 28747 flexible paddles. The difference in the mean peak glandularity between rigid and flexible paddles was found statistically significant: respectively, 38.3% and 45.7%. For the rigid paddle the 25 th , 50 th and 75 th percentile of the peak glandularity were 14%, 33% and 57%, respectively, and 23%, 42% and 63% for the flexible paddle. Conclusion: Depending on the compression paddle a significant difference was observed for the peak glandularity distribution over a priori equivalent populations. When using the flexible paddle the bias on the peak glandularity is small enough to justify its full usage. Purpose: It is known that the output of commercially available tumour perfusion CT packages is not necessarily directly comparable and acquisition protocol recommendations differ. Our goal was to better understand model-dependent differences between two frequently used commercial implementations.

We compared one implementation (MP4) that uses an impulse residue function with 4 independent parameters (scan time recommendation 60s to 120s) with another (MP3) which assumes no backflow from interstitial space (3 independent parameters, scan time recommendation 45s). We used measured arterial input functions from an ongoing breast tumour perfusion study and generated a large number of enhancement curves from them by convolution with MP4. The independent parameters were varied such that blood flow (BF), blood volume (BV), flow-extraction product (E*F) and permeability-surface area product (PS) all covered the full range typically found for tumours. We then determined these parameters again by deconvolution with MP3 and analysed the differences for 45s and 60s scan time. Results: 45 s data showed excellent correlation for BF and BV (r 2 > 0.99), deviation from unity slope was less than 3%. Permeability measures E*F and PS correlated equally well, but absolute differences increased for higher values to 20%. These differences could be completely corrected for by a simple quadratic function. Absolute agreement for 45s data was better than for 60s. Purpose: PET/CT metabolic imaging using F-18 FDG is becoming increasingly used in oncology. Standardised uptake value (SUV) that reflects the metabolic activity measured within specific region of interest represents the most frequently used semi-quantitative parameter. Beside the biological variability of this measurement, intrinsic variability is related to performances of the PET/CT scanner and technical apparatus settings. Our aim was to validate a PET/CT cross-calibration protocol and to assess intra-and inter-apparatus reproducibility of SUV measurement. Methods and Materials: Dedicated phantom filled with a standardised FDG solution was used. Reference base-line was calculated. During 7 months, consecutive measurements were performed to monitor and evaluate the long-term intra-apparatus variability (CV) and longitudinal daily quality control (QC). Statistical CUSUM and

Manifestations of small bowel disease in paediatric Crohn's disease on magnetic resonance enterography M.R. Torkzad 1 , U. Ullberg 2 , N. Nyström 1 , L. Blomqvist 2 , P. Hellström 1 , U.L. Fagerberg 1 ; 1 Uppsala/SE, 2 Stockholm/SE (michael.torkzad@radiol.uu.se) Purpose: Magnetic resonance enterography (MRE) has rapidly gained popularity for imaging in patients with Crohn's disease (CD). We report the manifestations of CD as observed at MRE in a paediatric population at the time of diagnosis of CD. Methods and Materials: MRE of 95 consecutive paediatric patients with suspected or confirmed inflammatory bowel disease (IBD) examined between 2006 and 2009 were retrospectively analysed. The findings were documented based on location of the small bowel disease and additional findings. Results: 51 were boys and 44 girls (median age 13.7 years). 54 had CD, 31 other forms of IBD, and 10 no IBD. The most common site of small bowel involvement was the terminal ileum in 29 (53.7%) patients, ileum in 10 (18.5%) and jejunum in 9 (16.7%). Solitary jejunal inflammation was seen in two patients (3.7%). Two cases with small bowel stenoses leading to surgery were found. Perienteric lymphadenopathy could be seen in 30 (55.6%) patients and fatty proliferation in 9 (16.7%). Fistula formation and abscess/phlegmonous tissue were seen only in one patient each. The most common manifestation of small bowel inflammation was increased contrast enhancement of bowel wall (93.5%), thickening of the bowel wall (90.3%) and derangement of bowel shape with saccular formations (25.8%). Conclusion: MRE in the paediatric population often demonstrates acute inflammatory changes of the small bowel such as increased contrast uptake, bowel wall thickening and perienteral lymphadenopathy. More chronic small bowel changes, stenoses leading to surgery and jejunal involvements, are less commonly seen.

Safety of intravesical use of a second generation US contrast agent in children for vesicoureteric reflux detection F. Papadopoulou 1 , A. Ntoulia 1 , E. Siomou 1 , K. Sarantidou 1 , C. Tsampoulas 1 , K. Darge 2 ; 1 Ioannina/GR, 2 Philadelphia, PA/US (katerinantoulia@yahoo.com) Purpose: To evaluate the safety of intravesical administration of a second generation US contrast agent (CA) for the diagnosis of vesicoureteric reflux (VUR) in children. Methods and Materials: After institutional board review approval and written informed consent of parents 360 children (140 boys, 220 girls, mean age 33.8 months, range 1m-17y) underwent voiding urosonography (VUS) to rule out VUR. A second generation CA (SonoVue®, Bracco, Milan) was administered intravesically through a transurethral bladder catheter. Adverse events were recorded during the examination and also reported by phone call within 7 days thereafter. Urine analysis and culture were performed 3-5 days before and 24-48 hours after the VUS. Results: Minor adverse events were reported in 11 (3%) patients.These included transient abdominal (n = 1), urethral pain (n = 6), transient macroscopic haematuria (n = 2), frequency of micturition (n = 1) and one episode of vomiting after 3 hours (n = 1). No severe or directly to the CA SonoVue® related adverse event was recorded. Symptoms subsided in the next 24-48 hours. The adverse events did not differ from those reported in VCUG studies and are most likely to be attributed to the catheter placement rather than CA. VUR was detected in 133 patients (37%). Conclusion: Only a small percentage of patients experienced minor adverse events during VUS, most likely due to the catheter. The intravesical use of SonoVue® for possible VUR detection in children is a safe procedure.

MRI features and volumetric changes of suprarenal neuroblastoma following (wait and see) protocol of therapy N.-E.A. Nour-Eldin, A. Ola, N.N. Naguib, T.J. Vogl; Frankfurt a. Main/DE (nour410@hotmail.com) Purpose: To retrospectively evaluate the MRI features and volumetric changes of neuroblastoma following conservative therapy (wait and see protocol). Methods and Materials: This retrospective study was approved by the institutional review board; informed consent was waived. The study included 72 patients of suprarenal neuroblastoma of which 10 patients undergone conservative therapy (wait and see protocol) at age of less than 6 months (6 females and 4 males) in the time period between January 1997 and January 2010. Initial tumour stage were stage 1: in 40% (4 out of 10), Stage 2 in 30%(3 out of 10) and stage 4S in 30% (3 out of 10). CE-MRI was performed at 3 months interval. Mixed linear modelling and logistic regression were performed including correlation to tumour markers. Results: Mean tumour size at diagnosis was 5.2 cm in diameter (range 2.5-7.5). Mean tumour volume 55 cc (range: 4.5-220 cc, SD: 15.5).The median follow-up tions were the most frequent findings and 5 patients underwent balloon dilatation and/or percutaneous stent placement. Two patients benefited from a meso-Rex bypass to restore portal flow. Conclusion: 1-CDU plays the leading role for post-transplant assessment and MRI, MDCT and angiography should be reserved to confirm diagnosis before treatment. 2-In this series, a higher incidence of complications was observed after using trans-hilar split technique (SB).

Imaging the Rex recessus vein pre-operatively using wedged hepatic venous portography (WHVP) and Doppler ultrasonography and the role of post-operative Rex shunt surveillance A. Lawson 1 , P. Rischbieter 2 , A. Numanoglu 1 ; 1 Cape Town/ZA, 2 Nelspruit/ZA (drajlawson@gmail.com) Purpose: The objective of this study was to assess the efficacy of wedged hepatic venous portography (WHVP) at detecting a patent umbilical segment of the left portal vein (Rex vein) pre-operatively in paediatric patients suffering from extrahepatic portal vein obstruction (EHPVO) who were considered for mesentericoportal bypass. Methods and Materials: A retrospective chart review was done in patients with EHPVO who had been considered for a mesoportal bypass between January 2001 and January 2010. Results: Sixteen patients (13 boys, 3 girls, mean age 5, range 1-13 years) were considered for mesoportal bypass including four post reduced-size liver transplant patients. Ten patients (62%) underwent WHVP. The Rex vein was clearly identified in eight (80%) patients. One Rex vein was seen at surgery despite not being demonstrated at WHVP. Six mesoportal bypasses were performed without WHVP of which three (50%) were successful. The average shunt sizes on the first postoperative day as documented using Doppler ultrasonography was 7.8 mm with an average flow of 20.8 cm/s. The average flow rate increased to 27.1 cm/s after one month. All shunt occlusions (n = 1) or partial thromboses (n = 2) were detected by post-operative Doppler ultrasound. Conclusion: Our series demonstrates the use of WHVP as an effective tool with a sensitivity of 80% and specificity of 100% in the preoperative patency assessment of the Rex vein. Doppler ultrasonography is useful for evaluating the size and flow rates in the post-operative shunt and is sensitive (100%) to the detection of early shunt thrombosis. Purpose: EPVT is a condition in which the obstruction of the portal vein trunk results in non-cirrhotic portal hypertension (pre-hepatic type) with complications ranging from variceal haemorrhage, hypersplenism, biliopathy to growth/development deficiency or subclinical encephalopathy. MRB (a direct bypass from mesenteric to left portal vein) has been shown to restore the physiologic portal flow to the liver. This study intends to evaluate the role of CDU for assessing of patency of the shunt and intra-hepatic portal flow at the long-term. Methods and Materials: Between 2007 and 2010, 28 children affected by EPTV with recurrent bleeding from oesophageal varices and/or severe hypersplenism (age range: 9 months-18 years) underwent MRB. Intra-and extra-hepatic vascular anatomy was assessed preoperatively using CDU, CT, MRI and retrograde portography. CDU only was used for intra-operative checks, in the early (1 week) (EPO) and in the late post-operative period (LPO). Current follow-up ranges from 1 to 36 months. Results: Peri-operative flow problems diagnosed at CDU led to successful immediate revision of the shunt in two cases and conversion into a meso-caval shunt in another one. Purpose: The diagnosis of hip osteoarthritis is often complicated by co-existing pathology in the knee and spine, and mismatch between clinical and radiological signs. Temporary pain relief from a local anaesthetic injection into the hip joint has been reported to help localise symptoms, reducing the risk of unnecessary surgery being performed. We hypothesise that good surgical outcome is predicted by significant analgesia following diagnostic injection, and that alternative pathology can be demonstrated when there is no response to injection. Methods and Materials: Data were analysed from a prospective database of 164 consecutive patients who underwent diagnostic hip injection for possible osteoarthritis. We recorded demographics, symptomatology, result of injection, and whether hip replacement was performed. A good outcome to surgery was defined as subsequent pain relief without significant residual symptoms. Results: 56 patients had a total hip replacement following a good response to diagnostic injection. Of these 52 had a good outcome following surgery (93% Purpose: To assess pre-treatment apparent diffusion coefficient (ADC) for predicting treatment response in lymphoma. Methods and Materials: Thirty-two patients (< 18 years) with histological confirmed lymphoma underwent whole-body anatomical MRI and diffusion weighted imaging (DWI) prior to and following treatment between August 2007 and July 2010. For the initial seventeen patients (group 1), the anatomical site of greatest nodal disease was selected at pre-treatment staging for localised DWI (9.2 cm tissue block), and reassessed using DWI following 2 cycles of chemotherapy. DWI measurements from group 1 patients were used to derive a relationship between ADC (volume and single-slice) and percentage residual tissue volume following treatment. For the subsequent fifteen patients (group 2) whole-body DWI was performed as part of routine clinical care. The anatomical site post-treatment containing the maximum volume of residual tissue was identified for the assessment of ADC (volume and single-slice) and total volume of disease at pre-treatment, and residual volume of tissue at after treatment. The ability of pre-treatment ADC to predict disease response (based on the correlation established in group 1) was assessed for patients in group 2. Results: A strong positive linear correlation was present between pre-treatment ADC and percentage residual tissue volume following 2 cycles of chemotherapy (group 1: single-slice R2 = 0.84, volume R2 = 0.70). Sensitivity and specificity of pre-treatment ADC for predicting treatment response was 100%, 100% and 100%, 92% for volumetric and single-slice measurements, respectively (group 2). Conclusion: Pre-treatment ADC correlates positively with residual tissue volume and can predict response to chemotherapy.

A S297 C D E F G H in internal rotation). The presence of enough intra-articular contrast material for interpretation, extra-articular leakage of contrast material and duration of procedure were recorded and the intensity of the patient's pain was measured using a visual analogue scale. Results: Intra-articular contrast material was satisfactory for interpretation in 94.8% of all applications, without any significant difference between anterior and posterior application. Mean duration of procedure was 444 ± 128 (A490 ± 190, P372 ± 77) seconds and mean pain intensity was 4.6 ± 0.7 (A4.9 ± 1.5, P3.3 ± 0.4), contrast material leakage to extra-articular space was detected in seventeen (A:11, P:6) cases; differences between the two applications were statistically significant (p < 0.005). Conclusion: MRI arthrography of the shoulder is known to increase diagnostic accuracy in comparison with conventional MRI. Especially in patients with suspected anterior instability, the contrast injection using posterior approach has the advantage to avoid interpretation difficulties due to contrast extravasation which is more possible in anterior approach. In our clinic, the posterior approach is assumed as optimal injection site for MR arthrography, because it is less likely to cause interpretation difficulties due to contrast leakage. Moreover, the posterior approach is more tolerable to patients, easier to perform for the operator. 

In 20 patients with suspected soft tissue tumours, preoperative MR-guided biopsy by the use of functional sequences was prospectively performed at 3 Tesla with a flexible surface coil. In addition to the routine sequences, in 18 patients, a DCE sequence was conducted over 3 min, with a temporal resolution of 7.3 sec during the first 72 sec, followed by 15 sec image intervals. In 17 patients, DWI was applied with the use of DWI-MSh FH (b-value: 0-800), and the apparent diffusion coefficient (ADC) was calculated. Multivoxel 1 H-spectroscopy could be conducted in 13 patients with sufficient tumour size. Results: In 12/17 patients with heterogeneous tumours, DCE was available. The biopsy area was chosen at the site of the highest slope in the contrast enhancement curve, which corresponded with the maximal Cho/Cr ratio in 8/11 patients. The diffusion weighted sequence was of limited value for the selection of the biopsy area. In all patients, the biopsy was successful, in 19/20 patients, the highest mitotic rate was hit. The final histopathological results of the surgical specimen confirmed the diagnosis of 10 malignant tumours of different grades, 7 benign entities and 3 tumours of intermediate dignity.

Conclusion: Dynamic contrast-enhanced MR imaging and multivoxel 1 H-MR-spectroscopy are valuable tools for the selection of the biopsy site in MR-guided biopsy.

Which is the more effective for the treatment of low back pain: intra-facet or peri-facet joint injection? A blinded randomised controlled trial F.W. Smith, A. Nandakumar, D. Wardlaw; Aberdeen/UK (franciswsmith@hotmail.com) Purpose: There is controversy regarding the efficacy of facet joint steroid injections for the treatment of low back pain. This study is designed to identify any difference in clinical outcome between intra-facet (IF) and peri-facet (PF) injections in patients with low back pain (LBP). The difference in screening time and radiation dose, between the two techniques, are also recorded. Methods and Materials: 88 patients were randomised into IF and PF groups. For the IF group bilateral L4/5 and L5/S1 levels were injected with 40 mg of Depo-Medrone with Lignocaine (total 2 ml). The PF group received 80 mg of Depo-Medrone with 1 ml of 0.5% Marcaine (total 3 ml) at the same sites. Pain visual analogue score (VAS) and analgesic chart were completed weekly for six months. Results: 88 patients were entered into the study and 14 withdrew leaving 69. Sixty-three (91%) had significant pain relief a week following injection, 86% and 94% in IF & PF groups. The mean duration of pain relief was 12.4 weeks (IF-12.2, PF-12.5). The mean change in VAS before and after injection was 4 (IF-4, PF-4). Mean change in analgesic use before and after injection was two tablets. (IF-2.9, PF-2.3). The radiation dose area product (DAP) for IF injections was 603 mGy. cm2. compared with 45 mGy.cm2 for the PF group. Conclusion: Both techniques afford significant pain relief (91%), (PF 94% vs. IF 86%). PF is easier to perform at a lower radiation dose and is the preferred technique.

One-year survey of two different ultrasound (US)-guided percutaneous treatments of lateral epicondylitis: results of a randomised controlled trial G. Ferrero 1 , E. Fabbro 1 , A. Muda 1 , L. Sconfienza 2 , E. Silvestri 1 ; 1 Genoa/IT, 2 San Donato Milanese/ IT (giulio.fer@libero.it) Purpose: Lateral epicondylitis is a common cause of elbow pain in general population. We show the efficacy of a percutaneous treatment under US guidance in treating this condition. Methods and Materials: 32 patients (18 males,14 females,mean age 45±8.6) suffering from lateral epicondylitis underwent an US-guided percutaneous treatment. They were randomly subdivided into two groups. In the first group (16 patients), under local anaesthesia and US guidance,a needle was advanced into the enthesis of the common extensor tendon. There, we performed multiple punctures to obtain a scarification of the enthesis and of the preinsertional portion of the tendon. In a second group (16 patients), an US-guided steroid injection was performed. A visual analogue scale was used to evaluate the degree of pain pre-and post-treatment at 2, 12, 24 and 48 weeks. Results: In the first group, no significant improvement compared with baseline was found at 2 weeks but was present at 12, 24 and 48 weeks (p <.001 for all). In the second group, significant improvement compared with baseline was found at 2 weeks (p <.001) but not at 12, 24 and 48 weeks. Comparison between the groups showed significantly different outcome in favour of the second group at two weeks (p <.001) and in favour of the first group at 12, 24 and 48 weeks (p <.001). Conclusion: US-guided percutaneous dry needling alone is more effective than steroid injection. The efficacy of this treatment seems to be long-lasting. It can be considered as an effective and minimally invasive treatment for lateral epicondylitis. Purpose: In chronic supraspinatus clacific tendonitis refractory to conservative treatment, a number of non-surgical options are available. Most of these interventions incorporate needling of the calcific deposit as part of the procedure. However, an evaluation of needling without further intervention, e.g. attempting to aspitrate the deposits as in barbotage, has not been undertaken. This was recently highlighted by the National Institute of Clinical Excellence in England and Wales in its assessment of autologous blood use where dry needling often forms part of the procedure, and as such calls into question what is yielding the improvement in these procedures. We undertook a study to determine whether dry needling on its own would be of benefit to these patients. Methods and Materials: Patients who had ultrasound guided dry needling for supraspinatus calcific tendonitis between April 2008 and February 2010 at South Manchester University Hospital were assessed using the Oxford shoulder score. Results: 31 patients were identified: 6 non-contactable, 1 unable to understand questionnaire and excluded. There was a 79% improvement demonstrated, which compares with an average of 60% with other interventional procedures utilised for this condition, including those that incorporate dry needling. There was an average increase in Oxford shoulder score of 17 p < 0.001, 95% CI (12.0, 23.0). The average pre-procedure (16) score is regarded as requiring surgical intervention, whilst the average post procedure score (33) is regarded as requiring lifestyle changes such as exercise. Conclusion: The procedure was effective at relieving shoulder symptomatology from surpraspinatus calcific tendonitis and avoiding surgery.

Optimal injection site in shoulder MR arthrography: anterior or posterior approach? M.G. Yilmabasar, B. Pekar, K. Akan, A. Hayirlioglu; Istanbul/TR (yilmabasar@gmail.com) Purpose: The purpose of our study was to optimise shoulder MR arthrography by comparing the anterior with the posterior ultrasound-guided arthrographic injection approach. Methods and Materials: Thirty-nine patients were divided into two groups, according to the injection site selected: seventeen were injected using the anterior approach (just superior to subscapularis tendon, in external rotation), twenty-two were injected using the posterior approach (just superior to infraspinatus tendon, Purpose: Two papers in the New England Journal (August 2009) compared treatment of vertebral fractures in randomised controlled trials reporting improvement in the majority of patients but with no significant difference between the local anaesthetic and vertebroplasty groups. Possible explanations include placebo response or therapeutic response to the "control procedure". Pain associated with osteoporotic compression fractures may be due to factors other than the fracture. Methods and Materials: Suitable patients referred for treatment by vertebroplasty were first offered a local anaesthetic and steroid facet joint injection at the most painful level; a two-week pain diary was employed. Those who failed to respond were offered a vertebroplasty. Results: Of 91 patients referred, 16 went straight to vertebroplasty (13 successful). 61 out of the remaining 75 were offered facet joint injection as initial treatment. 21 were successful, 2 relapsed and had further facet injections with good results, 3 elected not to have treatment, 5 had temporary benefit and were offered rhizolysis and 1 died from unrelated causes. Of the 29 who failed to respond to facet injections, 24 underwent vertebroplasty (23 successful). Conclusion: Around a third of patients referred for treatment of osteoporotic fractures by vertebroplasty responded beneficially to facet joint injection. It is possible that in this group the pain mediator is one of instability and overload on the facet joints produced by adjacent wedge fracture. This protocol allows a reduction in the number of vertebroplasties undertaken with enhanced success rate from those vertebroplasties that are performed.

Are we using discography appropriately? C.L. Gan, D. Lloyd; Cardiff/UK (cheelingan@yahoo.com)

Purpose: Discography is often used in conjunction of MRI findings to aid diagnosis of symptomatic lumbar disc degeneration. Positive discography with degenerate discs on MRI can help to differentiate between symptomatic and asymptomatic lumbar disc degeneration. Although discography is relatively safe, it is invasive and painful. We aim to investigate the appropriate use of discography in a U.K. tertiary centre. Methods and Materials: All the discography requests within Cardiff & Vale Healthcare NHS Trust from 01/11/2008 to 30/04/2009 were studied retrospectively. Electronic letters and test results from Clinical Workstation were reviewed to ascertain the indication for request and management plan following the procedure. Results: 73 requests were received in total. 42 patients had positive discography finding, whereas 21 were negative, 1 was inconclusive and the remaining 9 were abandoned either because of technical difficulty or patients' intolerance to procedure. All patients in the positive discography group had positive MRI correlation. However, only 55% of these patients either had operation or on surgical waiting list. 36% either were not offered operations or declined operations. Patients' decision towards operation was still awaited in the remaining 9%. Conclusion: Not all patients with positive discography had surgery. Possible implications following either positive or negative discography finding should be discussed with patients prior to referring for discography. This can help to prevent the patients with no intention for operation from going through this invasive and potentially painful procedure. The findings and recommendations can also apply to other hospitals in order to improve practice. Conclusion: DECT has a slightly higher sensitivity than MRPA with no statistically significance; DEPI has higher accuracy than MRPP to visualise perfusion defect derived from pulmonary emboli in a rabbit model.

: Despite a mean decrease of 35.5% in the dose-length-product, there was no statistically significant difference between group 2a and group 1 in the objective noise, signal-to-noise (SNR) and contrast-to-noise (CNR) ratios and distribution of the overall image quality scores. Compared to group 1, objective image noise in group 2b was significantly reduced with increased SNR and CNR and a trend towards improved image quality. Conclusion: Iterative reconstructions using 3 iterations provide similar image quality compared with the conventionally used FBP reconstruction at 35% less dose, thus enabling dose reduction without loss of diagnostic information. According to our preliminary results, even higher dose reductions than 35% may be feasible when using more than 3 iterations. Purpose: To evaluate the frequency and severity of cardiogenic motion artefacts on dual-source computed tomography (CT) angiograms of the chest acquired with a high-pitch and high-temporal resolution scanning mode. Methods and Materials: 100 consecutive patients underwent nongated chest CT angiography with dual-source, single-energy CT without administration of betablockers (temporal resolution: 75 msec). The presence, location and severity of cardiogenic motion artefacts were evaluated at the level of mediastinal vessels, central and peripheral pulmonary arteries, heart borders and lung parenchyma.

The mean (±SD) duration of data acquisition and heart rate were 0.9 (± 0.07) s and 78.9 (±15.2) bpm, respectively. Cardiogenic motion artefacts were depicted at the level of the ascending aorta in 4 patients (4%) and pulmonary trunk in 2 patients (2%). Among the 7000 pulmonary arteries analysed (5 central, 5 lobar, 20 segmental, 40 subsegmental arteries/patient), pseudofilling defects were depicted at the level of 3 segmental arteries (0.042%). A "double contour" of the left ventricle was noted in 52 patients (52%) and in 7 patients (7%) at the level of the right atrium, with a mean (±SD) To evaluate image quality and diagnostic yield of a high-pitch dual source comprehensive cardiothoracic CT protocol in patients presenting with acute chest pain. Methods and Materials: Consecutive symptomatic subjects with undifferentiated acute chest pain underwent ECG-synchronised high-pitch dual-spiral CT angiography (Definition Flash, Siemens) of the thorax. Independent investigators determined image quality of each vascular territory, measured contrast-to-noise-ratio (CNR) and determined radiation exposure. In addition, the prevalence of CT findings (pulmonary embolism, (PE), aortic dissection (AD), and significant coronary stenosis (≥ 50%) was determined. Univariate and multivariate analyses were performed to determine the subpopulation with highest diagnostic yield. Results: Among 51 subjects (66% males, average age: 63±15.8), the prevalence of positive CT findings was moderate (overall: 11.7%). Overall, image quality of the pulmonary, aortic and coronary vasculature was good (1.26±0.43 and CNR: 2.52) with an average radiation dose of 3.82 and only 3.2% of segments rated nonevaluable. Among the different vascular territories, image quality was lowest in the coronary artery tree (p = 0.02) and strongly associated with heart rate (r = 0.52, p < 0.001). In subjects with a heart rate of ≤ 65 bpm, subjective image quality and CNR of the coronary arteries were higher (1.6±0.5 vs. 2.1±0.5, p = 0.03 and 1.21±0.3 vs. 1.02±0.3, p = 0.05) with only 1.5% segments classified as non-evaluable. Conclusion: High-pitch dual-spiral comprehensive cardiothoracic CT provides low radiation exposure with excellent image quality at heart rates ≤65 bpm. In subjects with higher heart rates, image quality of the aortic and pulmonary vasculature remains excellent, while the assessability of the coronary vasculature degrades substantially.

Uterine artery chemoembolisation: its application in treating C-section scar pregnancy G. Zhang, S. Gu, T. Wang, Z. Han; Shanghai/ CN (guofuzh@msn.com) Purpose: To discuss the efficacy and safety of uterine artery chemoembolisation in treatment of Cesarean-section scar pregnancy. Methods and Materials: During the period of Jun 2005-Jun 2010, uterine artery chemoembolisation was performed in 130 patients with C-section scar pregnancy in our hospital. Via bilateral uterine arteries,100-150 mg MTX and 80000 U gentamycin were infused separately, after that, the bilateral uterine arteries were embolised with gelfoam particles to slow down the blood flow and then bars of gelfoam to block blood flow. Among all the patients, 23 cases of them who were combined with leiomyoma were embolised with PVA instead of gelfoam. Curettage of uterus was carried out within 24h to 72h after the procedure. Results: The success rate of bilateral uterine artery chemoembolisation was 100%. And the success rate of curettage combined with the first time embolism operation is 96.92% (126/130), The blood loss was between 12 and 110 milliliters (mean 28.5 ml) during the procedure of curettage. The other 4 cases (4/130) whose embryonic month were large and combined with arteriovenous fistula were still rich in blood supply after the first embolism operation which were proved by ultrasound examination, need a second embolism 2 weeks later after the first time embolism operation. No serious complications occurred in the treatment of 130 cases and the uteruses of all cases were preserved in the treatment. Conclusion: The uterine artery chemoembolisation is an effective and safe treatment for C-section scar pregnancy, which can greatly reduce the risk of hysterectomy and preserve fertility of the pregnant women.

Immediate and long-term follow-up after uterine artery embolisation in post-partum haemorrhages: monocentric study of 88 patients L. Mezzetta, R. Loffroy, B. Guiu, P. Sagot, J.-P. Cercueil, D. Krausé; Dijon/FR (denis.krause@chu-dijon.fr) Purpose: To evaluate immediate efficacy and mid-long-term complications after uterine artery embolisations in post-partum haemorrhages (PPH). Methods and Materials: 88 consecutives patients (mean age 31.3 years, (17-45) were treated by embolisation for PPH. The median follow-up was 55 months (8-117). The PPH risks factors were: uterine scars (n = 21); curettage (n = 15); multiparous women (n = 13); obesity (n = 8); previous PPH (n = 7); coagulopathies (n = 5); caesarean surgical interventions (n = 7). Results: Immediate results: primary technical success (94.4%, 83/88); 5 haemostasis uterine surgical ablations were necessary (in 3 cases, for abnormal placental insertion). The complications were: haematomas on the arterial puncture site (n = 16); pelvic pain during 72 hours (n = 22); endometrites (n = 3), with one uterine necrosis; endometrium atrophy (n = 7). Mid-and long-term results: menses return (86.4%; 64/74), with a mean delay of 2.9 months without or with modifications of the regularity (43.7%) and/or of the abundance (56.3%). Oligo-amenorrheas (20%, n = 15), with uterine synechies (4%, n = 6) -sterilities (19%, 17/88): definitive menauposes (n = 5); voluntary tubal ligatures (n = 4). Pregnancy desires (28%, 25/88); 17 pregnancies for 14 patients, 14 at term, 4 with a new PPH; child birth deliveries at 38 amenorrheas weeks (mean weight of the foetus 3129 g). Conclusion: Embolisation of the uterine arteries for post-partum haemorrhages representa a very safe, efficient technique, which allows a secondary return of menses, and a satisfying fertility rate, especially in the long term.

Chorionic villus sampling and marked membrane separation S. Akhlaghpoor, A. Aziz Ahari; Tehran/IR (alireza_ahdr@yahoo.com)

The major concern about invasive prenatal diagnostic tests is the frequency of procedure induced pregnancy loss. Chorionic Villus Sampling (CVS) is the invasive test of choice in the first trimester after the 10 th gestational week. Our experience suggests after the 10 th gestational week the marked chorioamniotic separation is an uncommon finding. This study assesses the rate of marked membrane separation in 10-to 14-week gestational period and post CVS foetal loss in this group. Purpose: We retrospectively assessed the interventional parameters and outcomes of the oesophageal balloon dilatation (EBD) in children with benign oesophageal stricture due to corrosive material intake and surgical anastomosis. Methods and Materials: Between January 2006 and August 2010, 58 patients (age range 3-249 months) who underwent EBD due to corrosive and surgical strictures were included the study. Associations between fallowing parameters were analysed by Fisher exact test and Mann Whitney test with using SPSS 16 software: period between initial and last EBD, number of EBD, complications, location and type of strictures, severity of lesions, balloon size and radiologic-clinic success.

Results: In all cases technical success was 100%. Radiologic and clinic success rate was 92%. There was no significant correlation between gender and procedure parameters. The most significant factors in success rate were severity of lesions and balloon size. There were no statistical significance between number of EBD, complications, location and type of stricture, period between initial and last EBD, but number of EBD and complications were higher in patients with severe-segmental/ diffuse stricture and anastomotic stricture, respectively. Conclusion: EBD in children with surgical and corrosive oesophageal stricture is valuable and effective method. Especially severity of lesions and balloon size plays important role in radiologic and clinic success. , computer-controlled positioning system, radiofrequency amplifier system, operates in conjunction with a 3 T MR unit (GE). Pre-treatment imaging set (T2-w and T1-w images, pre-and post-Gd -BOPTA) were obtained to measure leiomyoma volume. Immediately after treatment, T1-w contrast-enhanced fat-suppressed MR images in three planes were used to measure nonperfused volume ratio. The average volume of treated fibroids was 175 ± 201 (SD) cm 3 . Follow-up images were obtained 4 and 6 months after treatment and served to determine leiomyoma shrinkage. Qualitative and quantitative relations between fibroid volume, nonperfused volume ratio at treatment, and 6-month shrinkage were measured. Results: The average nonperfused volume ratio was 65% ± 25% immediately after treatment. At follow-up treated fibroids showed volume decrease to 112 ± 141 cm 3 with an average volume reduction of 36% ± 25%. A linear regression model showed highly significant correlation between post-treatment nonperfused volume ratio and shrinkage at 4 months. Conclusion: MRgFUS therapy for leiomyoma results in substantial nonperfused volume ratio and shrinkage. Improved relief of symptoms can be achieved in a safe and noninvasive manner without affecting surrounding tissue. 

A S303 C D E F G H Methods and Material: Pixel CdTe detector 1 cm x 2 cm x 2 mm bonded to dedicated ASIC that is thinned down to 50 μm. 4 of such assemblies mounted on 50 μm thin kapton printed circuit board (PCB) form a unit detector with effective CdTe material of 4 cm x 2 cm x 2 mm. A stack of such unit-detectors will form a module-detector of a can be arranges to form a PET scanner with effective 4 cm CdTe absorption and highly segmented in 3D (1 mm x 1 mm x 2 mm) making it ideal for brain imaging. Results: Given the high detection efficiency of 4 cm CdTe for 511 keV (90%) and the excellent energy resolution (1% FWHM), simulation results followed by image reconstruction show, when compared with state-of-the-art PET scanner for head, the detection efficiency increases by a factor of 2.3 and the scattered photons are reduced from 98% to 3%, thus allowing to resolve the 1.2 mm rod in Derenzo-like phantom with peak-to-valley ratio of 3 to 1. Purpose: The accuracy of apoptosis imaging was analysed for assessing antiangiogenic therapy response. Treatment effects on tumour vascularisation were additionally addressed. Methods and Materials: AnnexinVivo 750 (apoptosis) and AngioSense 680 (vascularisation) concentrations were determined by fluorescence molecular tomography (FMT2500) in nude mice with subcutaneous epidermoid carcinoma xenografts after 4 days of anti-angiogenic treatment with SU11248. Tumour localisation was facilitated by parallel morphological μCT-scans. 3D FMT data were compared with corresponding 2D fluorescence intensities. Vascularisation was additionally analysed by contrast-enhanced ultrasound (2D MIOT). In vivo-data were validated on corresponding histological sections (CD31-staining: vascularisation; TUNELstaining: apoptosis). A Student´s t test was applied for statistics. Results: Strikingly lower AnnexinVivo concentrations were measured in treated tumours compared with controls (3D: control: 81.3 pmol/cm 3 , therapy: 27.5 pmol/ cm 3 ; 2D: control: 13 FI/cm 2 ; therapy: 11 FI/cm 2 ). In contrast, immunohistology demonstrated a significantly increased apoptosis in treated tumours (TUNEL+ area: control: 0.011%, therapy: 0.461%; p < 0.001). Lower AngioSense concentrations were detected in treated tumours (3D: control: 13.8 pmol/cm 3 , therapy: 6.3 pmol/ cm 3 ; 2D: control: 38 FI/cm 2 , therapy: 15 FI/cm 2 ). Ultrasound and immunohistology confirmed a significantly reduced vascularisation in treated tumours (ultrasound: control: 285.3 a.u./mm 2 , therapy: 87.3 a.u./mm 2 ; p < 0.01; histology: CD31+ area: control: 4.8%, therapy: 1.7%; p < 0.001). Conclusion: In vivo apoptosis imaging fails in assessing increased apoptosis after anti-angiogenic therapy. This can be explained by the significant breakdown of the vasculature in response to treatment, thereby impairing the delivery of the annexin probe to the tumour. This favours apoptosis imaging only in therapies that do not severely interfere with the vasculature.

Combination of vascular endothelial growth factor antisense oligonucleotide therapy and radiotherapy increases the curative effects against maxillofacial VX2 tumours in rabbits Purpose: To study effects of combination of vascular endothelial growth factor (VEGF) antisense oligonucleotide and radiotherapy on maxillofacial VX2 tumours in rabbits. Methods and Materials: 24 New Zealand rabbits with maxillofacial VX2 tumour were randomly divided into the following 4 groups: group A, treated with 16 Gy of radiotherapy; group B, treated with an injection of 150 μg of VEGF antisense oligonucleotide into tumour; group C, treated with an injection of 150 μg of VEGF antisense oligonucleotide into tumour immediately after 16 Gy of radiotherapy; and group D, treated with an injection of 300 μl 5% aqueous glucose solution into tumour. On days 3 and 14 after treatment, dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) was performed to calculate maximal enhancement ratio (MER), slope of enhancement (SLE), and tumour volume change. Tumour samples were obtained for pathological examination and immunohistochemical staining for VEGF. Purpose: Harnessing the full potential of complementary imaging modalities is the driving force in the development of multimodal imaging probes. Here, we investigate the use of radiopharmaceuticals as a dual modality probe (PET-optical) by imaging the inherent optical emissions (Cerenkov radiation) from their radioactive decay. Methods and Materials: Cerenkov luminescence imaging (CLI) of a range of radionuclides ( 18 F, 89 Zr, 64 Cu, 124 I, 131 I, 225 Ac) was investigated using in vitro phantom and a standard optical imaging device (Ivis 200, Xenogen). In vivo, the novel radiolabelled monoclonal antibodies (mAb) 89 Zr-DFO-J591 and 89 Zr-DFO-trastuzumab were used for the simultaneous PET and optical imaging of PSMA positive prostate cancer and Her-2 positive breast cancer xenograft, respectively. Biodistribution studies combined with ex vivo CLI of radiotracer uptake in organs was performed. Results: Intensity of the Cerenkov radiation in phantoms correlated linearly (R 2 = 0.98) with both the activity concentration and the measured PET signal (%ID/g). In vivo and ex vivo CLI provided accurate, quantitative information on mAb biodistribution and tumour uptake that increased over time from 24h (3.4 ± 0.35 p/s/cm 2 /sr) to 96h (6.0 ± 0.45 p/s/cm 2 /sr) correlating with the corresponding immuno-PET images. CLI was used to visualise tumour margins and successfully guide surgical resection. Conclusion: CLI of medical isotopes shows excellent promises for the use of radiopharmaceuticals as a dual-modal imaging modality allowing both PET and direct optical imaging suitable for intraoperative surgical application. Purpose: A whole-body hybrid PET-MR imaging unit was implemented and tested against PET-CT imaging for diagnostic and follow-up of oncology patients. Our goal was to evaluate the performance and clinical applicability of combined imaging protocols on hybrid PET-MR compared to PET-CT Methods and Materials: A whole-body scanner combining a 3 T MR and a time-of-flight PET scanner sharing a single bed allowing sequential acquisition of co-registered MR and PET images was evaluated clinically in 35 patients referred for diagnostic PET-CT study. PET-MR images were acquired following standard clinical PET-MR studies. Images were interpreted by multidisciplinary teams of radiologists and nuclear physicians in the same way it is routinely done for hybrid PET-CT studies. Oncological studies included, lymphomas, head and neck tumours, prostate and breast tumours as well as lung and colon cancers. Results: Optimised imaging protocols combining whole body MR attenuation correction data set with standard MR diagnostic protocols of both modalities while reducing the total time of the study were developed. Diagnostic quality of fused PET-MR images were comparable to corresponding PET-CT images and measured local SUV were comparable. None of the PET-MR studies had insufficient diagnostic quality and finding were comparable to PET-CT findings. Purpose: Tumours are not a homogenous compound of malignant cells but comprise a large variety of recruited host cells, inter alia a significant amount of macrophages. There is strong evidence that those tumour-associated macrophages (TAM) are promoters of tumour progression and metastasis, found, e.g. at sites of basement-membrane breakdown, enabling the tumour cells to locally egress into surrounding stroma. TAM density and activity correlates with poor prognosis in numerous malignant diseases, reflecting advanced tumour progression. This study should confirm the possibility to visualise TAM activity in tumour models using optical imaging (OI) and a macrophage-specific fluorescence contrast agent. Methods and Materials: An antibody addressing macrophage-specific S100 A9 was labelled with Cy5.5 for in vivo OI. IGG-Cy5.5 served as control for unspecific label distribution. Mice were inoculated with 4T1 murine breast cancer or MOS murine osteosarcoma, respectively. At the time, tumours reached a size of 4 mm, OI was performed after injection of the fluorescence contrast agents. For correlation of imaging results, immunohistochemistry of tumour sections was performed. Tumour cells were analysed for S100 A9 expression using ELISA. Results: Neither 4T1 nor MOS cells showed significant S100 A9 production in ELISA. In fast-growing 4T1 tumours, anti-S100 A9-Cy5.5 injection resulted in fluorescence signal which was regularly about two-fold higher compared to IGG-Cy5.5 (2356.8 vs. 1283.2; p < 0.005). At same size, slow-growing MOS accumulated less anti-S100 A9-Cy5.5 than 4T1 ( (clemens.cyran@med.uni-muenchen.de) Purpose: To investigate dynamic, contrast-enhanced computed tomography (DCE-CT) for monitoring the effects of the multikinase inhibitor sorafenib on experimental prostate carcinomas in rats with immunohistochemical validation. Methods and Materials: Copenhagen rats (n = 18) were implanted subcutaneously with prostate carcinoma allografts (MLLB-2) and underwent DCE-CT-enhanced with Iopromide (Ultravist370 ® , Bayer Schering, Berlin, Germany) on day 0 and after one week of therapy. The treatment group (n = 9) received daily applications of sorafenib (Nexavar ® , Bayer, Leverkusen, Germany, 10 mg/kg bodyweight) via gavage. Quantitative measurements of tumour perfusion (PF, ml/100 ml/min) and tumour vascularity (PV, %) were calculated based on a two-compartment kinetic model. On day 7 tumors were excised for immunohistochemical staining including tumour vascularity (RECA-1), tumour cell proliferation (Ki-67) and apoptosis (TUNEL).

Results: Under sorafenib therapy tumour perfusion declined significantly between days 0 and 7 (PF day 0 31.7 ± 21.7 vs. PF day 7 9.3 ± 5.7 ml/100 ml/min, p < 0.02). Tumour vascularity decreased significantly under sorafenib therapy (PV day 0 = 14.3 ± 8.0 to PV day 7 = 5.2 ± 4.0%; p < 0.05). In the control group, no significant alteration (p > 0.05) of tumour perfusion or tumour vascularity was observed. Immunohistochemical measurements revealed significantly lower tumour vascularity in the therapy than in the control group (RECA-1 1.415 ± 71 vs. 2.391 ± 120, p < 0.01). In Sorafenib-treated tumours, significantly more apoptotic cells (TUNEL 55.806 ± 2.790 vs. 22.332 ± 1.116, p < 0.01) and significantly less proliferating cells (Ki-67 77.026 ± 3851 vs. 140.457 ± 7.022, p < 0.01) were observed than in the control group. Conclusion: Tumour perfusion and tumour vascularity, quantified by DCE-CT, may be applicable as non-invasive surrogate parameters for monitoring the antiangiogenic, anti-proliferative and pro-apoptotic effects of sorafenib on experimental prostate carcinomas in rats as validated by immunohistochemistry.

In group C, tumour volume was significantly reduced on day 14 after treatment, and the difference was statistically different as compared to that before treatment, on day 3 after treatment and other groups. Values of both MER and SLE after treatment were significantly lower than the values before treatment. Pathological specimen revealed decreased VEGF expression. The immunohistochemical score (IHS) of group C was significantly different from groups A and D. Conclusion: Injection of VEGF antisense oligonucleotide after radiotherapy can enhance the curative effect on rabbit maxillofacial VX2 tumour, and DCE-MRI can serve as a reliable technique for in vivo monitoring.

The precise co-registration method of high-resolution SPECT and MR images of in vivo mouse brains K. Purpose: The precise co-registration of SPECT and MR images would be useful to evaluate detailed localisations of molecular probes in the small animal imaging. We investigated precise co-registration techniques of in vivo mouse brain images obtained from commercial small animal SPECT and MR scanners.

We developed a rodent holder with three external fiducial markers (FMs) to immobilize its head. To demonstrate the usefulness of these markers as guides, the resolution and the sensitivity at FMs on SPECT images were measured. A cylinder phantom and three mouse brains treated with I-123 iomazenil (IMZ) were scanned by both SPECT and MRI scanners. The obtained images were manually merged based on FMs. Co-registration errors were calculated using the mutual information and the rigid three-dimensional transformation algorithms.

The resolution and the sensitivity at FMs on SPECT images were 713 ± 42 μm FWHM and 425 cps/kBq, respectively. These results indicated that FMs can be used as guides for the image co-registration. On the merged images, the maximum translational and rotational co-registration errors were 171 ±71μm and 0.012 ± 0.010°, respectively. In the animal studies, merged images clearly depicted the distribution of IMZ restricted to the cortex.

We proposed a precise co-registration technique of high-resolution SPECT and MR images of small animals. The images merged by this method were good enough to evaluate anatomical localization of molecular probes in mouse brains in vivo.

In Purpose: To investigate whether in vivo micro-CT and micro-PET imagings enable an evaluation of different classes of tumour promoting agents in a c-myc transgenic mouse model of liver cancer. Methods and Materials: Sixty c-myc transgenic mice were treated with different genotoxic or non-genotoxic chemicals or vehicle at the age of 2 months. In vivo contrast-enhanced CT (DHOG, Fenestra) and 18 F-FDG-PET of the abdomen were performed at 4 different time points between the age of 4 to 8.5 months. Liver volume, tumour percentage, tumour multiplicity and tumour-to-non-tumour ratio were determined and compared with histopathology results. Results: MicroCT detected lesions of ≥1 mm with certainty. All animals that were positive for hepatocellular carcinoma in histopathology could be identified by in vivo imaging. There was a strong correlation between the determined tumour volume and the liver volume (r = 0.94). Significant differences (p < 0.01) in tumour percentage, tumour multiplicity and tumour-to-non-tumour ratio were observed between treated and control groups. Notably, in animals treated with the genotoxic carcinogen N-nitrodiethylamine hepatocellular carcinoma developed faster and tumour percentage was higher, therefore demonstrating its tumour promoting effect. Furthermore, metastatic spread to the lung as well as primary lung cancer could be observed in this group. Purpose: To determine the feasibility of the 3D-FIESTA MR curved planar reconstruction (CPR) for evaluating the anterior cruciate ligament of the knee joint. Methods and Materials: A magnetic resonance (MR) examination of the knee was performed on 30 subjects using a GE 3 T scanner and a phased-array extremity coil. MR 3D-FIESTA sequence with CPR and oblique sagittal proton density-weighted fast spin-echo (PD-FSE) sequence were performed. Signal-to-noise efficiency, and contrastto-noise ratio (CNR) measurements of the anterior cruciate ligament were calculated for all sequences and compared with the use of paired t-tests. The presence of the originating point, anteromedial (AM) and posterolateral (PL) bundle, inserting point of the anterior cruciate ligament on the two sequences were compared with the use of rank sum test. Results: The anterior cruciate ligament signal-to-noise efficiency of MR 3D-FIESTA sequence with CPR was not significantly different from that of the PD-FSE sequence (P > 0.05). MR 3D-FIESTA sequence with CPR produced CNR between anterior cruciate ligament and synovial fluid was not significantly different from that of the PD-FSE (P> 0.05). MR 3D-FIESTA sequence with CPR allowed excellent visualisation of the anterior cruciate ligament of the knee joint. The presence of the originating point, AM and PL bundle, inserting point of the anterior cruciate ligament on MR 3D-FIESTA sequence with CPR was significantly higher than oblique sagittal PD-FSE sequence (P < 0.05). Conclusion: MR 3D-FIESTA sequence with CPR significantly enhances our ability to identify the anterior cruciate ligament compared with oblique sagittal PD-FSE sequence.

Are all patients with anterior cruciate ligament tear able to undergo a double-bundle reconstruction? An MRI study A. Ntailiani, G. Tzoanos, N. Tsavalas, E. Vassalou, A.H. Karantanas; Iraklion/GR (kdailiani@gmail.com) Purpose: To estimate the incidence of anatomic restrictions which may prevent performing a double-bundle reconstruction of a torn anterior cruciate ligament (ACL).

We retrospectively studied 658 knee MRI examinations, which corresponded to 635 patients. All MRI examinations were performed on a 1.5 T scanner using a synergy coil. Both the coronal and sagittal images were evaluated for the measurements. The femoral notch width (FNW) and the ACL tibial insertion site (TIS) length were digitally measured using a computer workstation. We measured FNW on coronal T1-w spin echo images at the level of decussation of ACL and posterior cruciate ligament and the ACL TIS length on sagittal fat suppressed proton density turbo-spin-echo images at the point of maximum length. Exclusion criteria included patients with open physes, severe osteoarthritic changes (Grade III or greater) and multiligamentous injuries. Results: We found 8.51% of patients to have open physes, 9.27% severe osteoarthritic changes and 0.759% multiligamentous injuries. The above were excluded from the analysis. 14.741% of subjects in our study group proved to be non-proper candidates for double-bundle reconstruction technique depending on the results of measurements. Of them, 13.677% had short TIS length, 1.367% had narrow femoral notch and 0.3% had both. Conclusion: There is a significant rate of patients who do not meet the criteria for double-bundle ACL reconstruction technique. MRI is able to depict this subgroup of patients and thus is important to avoid failure of a demanding treatment.

Evaluation of external displacement of implanted menisci on MR imaging T. De Coninck 1 , W. Huysse 1 , P. Verdonk 1 , E.-L. Heinrichs 2 , K. Verstraete 1 ; 1 Gent/BE, 2 London/UK (tineke.deconinck@ugent.be) Purpose: To evaluate the external displacement of a polyurethane meniscal scaffold. The displacement is compared to that of 3 control populations: normal, torn, and transplanted menisci. Methods and Materials: Fifty-two patients, implanted with a polyurethane scaffold (34 medial and 18 lateral scaffolds) received MR-scans at 1 week, 3, 12 and 24 months post-surgery. The findings of these patients were compared to 3 control populations: 100 normal, 100 torn and 16 allograft transplant menisci. Coronal MR images were used to analyse the external displacement of the body of the meniscus, which was defined as the distance (mm) between the edge of the tibial plateau and the peripheral edge of the meniscus. Results: The external displacement of normal (mean = 1.36 mm) and torn (mean = 1.37 mm) menisci does not differ significantly. The lateral transplanted (mean = 4.04) and lateral scaffold (mean = 4.49 mm) menisci have a similar displacement; however, both are significantly more displaced than normal and torn menisci. Purpose: A non-invasive method to follow structural changes occurring during muscle damage/repair process was an essential need to investigate the role of immune system in muscle healing. Aim was to set up and validate a magnetic resonance (MR)-based follow-up of muscle repairing process.

The model of acute muscle damage was obtained through cardiotoxin injection (CTX) into tibialis-anterior and quadriceps muscles of C57BL/6N mice. MR imaging was performed before CTX injection and 1-3-5-7 and 10 days after injury on a 7T magnet (Bruker). T2w-MSME sequences for T2 relaxation-time (T2-rt) quantification and diffusion-tensor images (EPI-DTI) for fractional-anisotropy (FA) assessment were acquired. MR results were correlated with histological analysis performed by haematoxylin-eosin staining in the same animal model.

In healthy muscles T2-rt resulted 16,723 ± 0.8149 ms and FA 0.404 ± 0.0901. T2-rt strongly increased after acute damage with maximum values at days 1 and 3 and a slow decrease in the later time points (day 1: 34.879 ms, day 3: 44.84 ms, day 5: 34.579 ms, day 7: 36.875 ms, day 10: 31.521 ms). Modifications in T2-rt were apparently related to histological evidence of inflammatory infiltrate. Besides, we observed a marked reduction of FA at days 1 and 3 after damage, with a complete recover of FA values at day 7, according to histological evidence of fibres regeneration (day 1: 0.161, day 3: 0.113, day 5: 0.3, day 7: 0.336, day 10: 0.377). Conclusion: T2-rt and FA modifications seem sensitive and quantitative markers of two different phenomena occurring in the acute muscle damage, tissue oedema and muscular architecture disruption. MR imaging allows quantitative and non-invasive monitoring of muscle healing process with oedema reabsorption and fibres integrity recover. Purpose: To evaluate the MRI characteristics of suprapatellar fat pad (SFP), correlate the findings with mass effect on the suprapatellar joint recess and determine its clinical significance.

We retrospectively reviewed 698 consecutive knee MRI examinations from 671 patients for the presence of SFP oedema with subsequent mass effect on the posterior joint recess. We measured the maximum anteroposterior (A-P), craniocaudal (C-C) and oblique (OBL) diameter of the suprapatellar fat pad on sagittal fat-saturated intermediate-w TSE images. The ratio DΙ/SD was determined after assessing the difference of the signal intensities of the suprapatellar and prefemoral fat pad (D|) as well as the corresponding standard deviation of the background noise (SD). Findings of anterior knee pain according to history and physical examination were also recorded. Statistical analysis was performed with Mann-Whitney test and Spearman's correlation.

The presence of SFP oedema with mass effect on the posterior joint recess was detected in 12.89% (90/698) of MRI examinations. The A-P (p < 0.0001), C-C (p = 0.0158), and OBL (p < 0.0001) diameter of the pad as well the DΙ/SD ratio (p < 0.0001) were significantly larger in knees with mass effect. Significant correlation was found between the Δ|/SD ratio and the A-P (rho = 0.123, p = 0.0011), C-C (rho = 0.105, p = 0.0057) and OBL (rho = 0.140, p = 0.0002) diameters of the pad. Mass effect was significantly associated with anterior knee pain in 3 ball game players (3/90 = 3.33%). Conclusion: SFP oedema with mass effect is a common finding in MRI examinations of the knee but very rarely associated with anterior knee pain. Purpose: To evaluate the presence of vascular tissue in a degradable polyurethane meniscal device implanted after partial meniscectomy. The vascularisation was analysed by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Methods and Materials: Fifty-two patients were assessed by DCE-MRI at 1 week and 3 months and a representative sample of subjects at 12 and 24 months postsurgery. The change in T1-relaxation time in a region-of-interest (ROI), as depicted on a time-intensity-curve (TIC), is used to measure gadolinium influx during the first 3 minutes after contrast administration. An increase of gadolinium can only be explained by the presence of blood vessels, hence this influx corresponds to vascularisation in this area. ROIs were drawn over the inner and outer half of the scaffold. Results: In the first week after surgery, the outer half of the scaffold only shows enhancement in 6.8% (3/44) of the subjects. After 3 months signal enhancement, and thus evidence of tissue ingrowth into the peripheral half of the scaffold, is seen in 81.4% ( 18-78m) . Male to female ratio was 9:6. Radiological evaluation was performed with x-ray and 3 T-MRI. Results: 2/15 patients received full knee replacement due to insufficient cartilage repair through MFX during FU period. Evaluation of pre-and postoperative MRI showed good cartilage repair tissue in one (7.7%), moderate repair in two (15.4%) and poor fill in ten patients (76.9%). In these ten patients the defect size increased. Average defect size preoperatively was 187 mm2 (12-800 mm2) and postoperatively 294 mm2 (40-800 mm2). Ten patients showed a varus leg axis deviation (Ø 5.9°) and three had a neutral alignment. The alignment correlated positively with declining knee function. Conclusion: Our study demonstrated that MFX as a treatment option for cartilage defect in the knee did not show the anticipated radiological long-term results. In 12/15 patients the cartilage defect size had increased after MFX, in 2/15 indicating full-knee-replacement. Purpose: The objective of this study was to determine gender-related differences in the repair tissue after MACT as assessed by advanced morphological and biochemical MRI. Methods and Materials: 40 patients (20 females, 20 males) after MACT of one femoral condyle of the knee were included. Both groups were matched by age (females: 35.1 ± 13.7; males: 35.2 ± 8.7 years) and post-operative follow-up (females: 34.0 ± 17.4; males: 34.2 ± 16.8 months). At 3.0 Tesla MRI, morphological magnetic resonance observation of cartilage repair tissue (MOCART) scoring, based on high-resolution PD-TSE, dual-FSE and TIRM sequences, was performed as well as biochemical T2 mapping, based on a multi-echo spin-echo sequence. A zonal (deep and superficial) region-of-interest analysis was performed to assess the quantitative T2-values in native control cartilage and cartilage repair tissue.

medial compartment there is also no difference between the external displacement of the scaffold (mean = 4.81 mm) and transplanted allografts (mean = 4.71 mm). The displacement of the medial scaffold is significantly less 1 week post-surgery compared to 3 and 12 months. The lateral displacement remains unchanged. Evaluation of the 24-month images is in progress.

The position of the medial polyurethane scaffold shifts between 1 week and 3 months postoperatively, but remains stable thereafter. Its final position is similar to that of a meniscal allograft.

Locoregional deformation patterns of knee cartilage: a 3D volumetric highresolution in-vivo study at 3 T A. Horng 1 , R. José 2 , M. Zscharn 1 , M. Notohamiprodjo 1 , U. Hoehne-Hückstädt 3 , U. Glitsch 3 , R. Ellegast 3 , M.F. Reiser 1 , C. Glaser 2 ; 1 Munich/DE, 2 New York, NY/US, 3 St. Augustin/ DE (annie.horng@med.uni-muenchen.de) Purpose:To evaluate knee cartilage deformation as a surrogate parameter for joint contact areas under defined loading conditions. Methods and Materials: Sagittal 3D-T1-w-FLASH-WE-sequences (TR14.2 ms/ TE7.2 ms/FA15°/resolution 0.3 2 x 1.5 mm 3 /matrix 512²) were acquired of 10 healthy knees before and after loading (10 min kneeling at 90°/squatting/after 50 knee bends) and after further 90 min rest at 3 T. After segmentation 3D-reconstruction of the cartilage was used to generate thickness-and 2σ-significance-thicknessdifference-maps. Volumetric parameters (volume, cartilage-bone-interface-area, mean thickness) were calculated. Voxel-based reproducibility for calculation of significant changes was assessed by RMSA, statistics by t-tests. Results: Similar deformation pattern was found for patellar cartilage in the medial and caudolateral facet, most pronounced after squatting. Corresponding deformation was found in the posterior aspect of the femoral condyles (medial > lateral). Tibial cartilage revealed inner anterior and central deformation (lateral > medial), most pronounced after squatting. Dynamic revealed shallower changes than static loading. Voxel-based reproducibility depended on cartilage thickness (Th) ranging from 0.12 to 0.35 mm. For Th ≥1 mm reproducibility was < 0.31 mm (<voxel size). Significant global changes for volume and mean thickness ranged from 1.1% to 4.9%, while locoregional changes reached 14%. Conclusion: These first data of direct in-vivo locoregional cartilage deformation patterns indicate potential contact and loading zones after kneeling, squatting and knee bends. They are anatomically and functionally plausible and may represent areas where stress-induced degeneration and subsequent osteoarthritis might originate. Voxel-based quantitative analysis of regional deformation combining information from all knee compartments complement common global analysis of cartilage volumetry. The approach may facilitate understanding of individual knee loading and kinematic properties and contribute to improving and validating biomechanical knee models. Purpose: The aim of this study was to assess several MRI-features as baseline risk factors that may predict patello-femoral (PF) and tibio-femoral (TF) cartilage loss at 6-month follow-up. Methods and Materials: For 177 subjects with chronic knee pain, 3 T MRI of both knees was performed at baseline and follow-up. Knees were semiquantitatively assessed using a modified WORMS score. All MR joint morphologic features were dichotomised into present (≥ 1) versus absent (0). Logistic regression models were applied to predict cartilage loss controlling for clustering by person and by knee. 

A S307 C D E F G H Results: Mean attenuation of main renal artery was significantly higher in 80 kVp with MC-CM (370.0 ± 65.0 HU) than 120 kVp with HC-CM (269.9 ± 27.8 HU, p < 0.001) without significant differences in SNR and CNR. 80-kVp protocol had significantly higher quality scores in arterial enhancement, sharpness of the artery, and overall diagnostic image quality compared to 120-kVp protocol. Effective dose of 80-kVp protocol (4.5 ± 0.3 mSv) was 8.2% lower than that of 120-kVp protocol (4.9 ± 0.4 mSv, p < 0.001). Conclusion: The use of 80 kVp with MC-CM could improve arterial enhancement and provide superior image quality than 120 kVp with HC-CM, despite using smaller amount of iodine administration and decreased radiation dose. Purpose: To evaluate the accuracy of T3b-renal-cell-carcinoma (RCC) detection using dual energy-CT (DECT) and to evaluate the image guality. Methods and Materials: 48 pt. (26 males, 22 females) who underwent surgery due to the suspected RCC were enrolled in the study. DECT in the arterial phase of contrast material distribution was performed using dual-source CT system with simultaneous exposition of 140 and 80 kV, respectively. The data were reconstructed in 0.6 mm thickness with one-third overlapping in the order of isotropic submilimeter resolution. The presence of vascular invasion or massive venous contrast return in the early arterial phase was signed as the suspected vascular invasion. The image analysis included the assessment of averaged, low-kV and virtual nonenhanced images. Following image quality was estimated: 1-excellent, 2-good, 3-evaluable, 4-insufficient. After resection or nephrectomy, all specimens were assessed, typed and staged by experienced pathologist according to WHO criteria and the pathologic results were compared with CT findings. Results: The accuracy of the RCC detection reached 93/7% (45/48) -two benign tumours and one inflammatory process were confirmed. The sensitivity according to the detection of T3b RCC stage was 87%(7/8), specificity reached 92%(37/40). The average image quality vas marked 1.3, no one examination was marked as insufficient, the decreased quality of some patients were caused by obesity and expansion over the field of view of tube B. Conclusion: The vascular invasion of RCC could be depicted with DECT with the high specificity and sensitivity using high quality imaging.

Dual energy CT in patients with polycystic kidney disease A. Graser, N. Arndt, M. Staehler, M.F. Reiser, T.R.C. Johnson; Munich/ DE (anno.graser@med.uni-muenchen.de) Purpose: To evaluate the efficacy of dual source-dual energy CT (DECT) in the detection of neoplasia in patients with polycystic kidney disease. Methods and Materials: 20 patients with polycystic kidney disease (ADPKD, n = 11; multicystic kidney disease, n = 9) underwent DECT on a dual source scanner (Somatom Definition Flash, Siemens Healthcare), using kVp settings of Sn140/100 kVp, 32 x 0.6 mm collimation, and 232/310 mAs, online tube current modulation. A single phase examination 70s after iv injection of 1.3 mL/kg body weight of Imeron 400, Bracco Diagnostics was acquired. Colour-coded iodine maps and virtual nonenhanced images were used to determine enhancement within cysts and to differentiate haemorrhagic from simple cysts. A cut-off of 15 HU was used as threshold for malignancy. The number of cysts on each kidney was recorded. In patients with malignancy, histopathology was the gold standard; otherwise, patients underwent follow-up imaging for 120-381 days. Average radiation exposure was calculated. Results: Each patient had > 50 kidney cysts. Based on measured enhancement, malignant masses were seen in 8 patients, these were confirmed at histopathology as clear cell RCC (n = 5) and papillary RCC (n = 3); follow-up imaging showed no malignancy in 12 patients. Cysts did not enhance by more then 15 HU, while masses showed a mean enhancement of 45 (25-123) HU. Mean radiation exposure was 5.5 ± 1.3 mSv. Conclusion: In patients with polycystic kidney disease DECT greatly facilitates the detection of malignancy, at the same time reducing radiation exposure by omission of a true nonenhanced phase.

The MOCART score was comparable in female (69.4 ± 20.0) and male (69.3 ± 13.0) patients (p = 0.982). Seven of the nine variables did not differ significantly. Whereas the structure of the repair tissue was more homogeneous in female patients (p = 0.026), changes in the subchondral bone and bone marrow oedema were less often visible in male patients (p = 0.007). Quantitative T2-values (ms) showed comparable results for the native control cartilage (female: deep = 47.7 ± 9.6, superficial = 53.1 ± 9.2; male: deep = 49.8 ± 12.6, superficial = 53.7 ± 10.5; p = 0.445 (deep), p = 0.826 (superficial)); whereas the repair tissue showed significant shorter T2-values of the deep cartilage layer in female compared to male patients ((female: deep = 43.5 ± 9.8, superficial = 48.5 ± 9.8; male: deep = 48.2 ± 7.7, superficial = 52.6 ± 11.0; p = 0.009 (deep), p = 0.052 (superficial)). Conclusion: Our initial results suggest that gender-related differences in the ultrastructure of the repair tissue, mainly in the subchondral bone and the adjacent deep cartilage layer, can be assessed using advanced morphological and biochemical MRI. Purpose: To evaluate the potential dose reduction of a dual-energy (DE) singlephase split-bolus CT-urography protocol in comparison with standard single-energy (SE) dual-phase split-bolus CT-urography protocol. We also compared the quality of virtual nonenhanced (VNE) images derived from DE acquisition with that of true nonenhanced (TNE) data-set in the same patients. Methods and Materials: 64 patients underwent CT urography with the following study protocol: TNE images acquisition in SE-mode, 50 mL bolus, 6-minute delay, 80 mL bolus, 100 seconds delay, urinary tract acquisition in DE-mode. On a secondary work-station VNE images were obtained. Two blinded and independent readers rated the quality of VNE respect to TNE data-set with a five-point scale: for 1-3 scores were judged comparable, not comparable for 4-5 scores. Effective radiation doses for DE single-phase protocol and standard SE dual-phase protocol (obtained by SE-mode scanning dose x2-multiplying) were calculated. Cohen's k was used to assess the interobserver agreement for VNE images evaluation. Results: VNE images were considered comparable (scores 1-3) to TNE data-set in the same 57 out of 64 patients (k =.891). In the remaining 7 patients VNE images were judged not comparable (scores 4-5) for excessive noise, kidney exclusion or lower conspicuity of calcification. Mean effective dose of DE single-phase protocol was 6.37 mSv ± 1.54 and that for standard SE dual-phase protocol was 8.92 mSv ± 1.22. Mean calculated dose reduction was 28.59 %. Conclusion: Split-bolus CT-urography in DE-mode allows a radiation exposure lowering by 28%, with VNE images that could replace TNE data-set. Purpose: To investigate the feasibility of 80 kVp with moderate-concentration of contrast material (MC-CM) in CT angiography of renal arteries compared to conventional 120 kVp with high-concentration of contrast material (HC-CM). Methods and Materials: Fifty patients with a clinical suspicion of renovascular hypertension were referred for CT angiography of renal arteries. Twenty-five patients were scanned with 120 kVp and 200 mAs eff after 110 mL of HC-CM (370 mgI/mL) administration and the other 25 patients were scanned with 80 kVp and 585 mAs eff after 110 mL MC-CM (300 mgI/mL) administration. Two groups of patients were compared with each other in terms of Hounsfield number of renal arteries, signalto-noise ratio (SNR), contrast-to-noise ratio (CNR), subjective degree of arterial enhancement, image noise, sharpness of renal arteries, and overall diagnostic image quality of both maximum-intensity projection and volume-rendered images.

A D E F G H Results: At laparoscopy 97 lesions (bowel = 68; peritoneum other than Douglas pouch = 15; bladder = 7; ureters = 7) were identified in 75 patients, with 19 women presenting multiple implants. MR identified 77/97 lesions, missing 14 bowel lesions, 3 of the peritoneum, 2 of the ureters, 1 of the bladder, with 10 false positive findings. MRI diagnosis was based on hyperintense lesions on T1-WI (36), plaque-like lesions (31), indirect findings (10). Sensitivity, specificity, PPV, NPV of MRI were 81%, 92%, 94% and 77%, respectively. The correlation regarding lesion size at MRI (mean 33 mm) and laparoscopy (mean 34 mm) was good (p < 0.01). Interobserver variability was excellent (k = 0.92). Conclusion: Identifying extrapelvic endometriosis preoperatively, MRI could be useful for treatment planning in case of severe pelvic disease. Purpose: To evaluate the dose reduction potential of an adaptive statistical iterative reconstruction (ASIR) on a high-definition CT (HDCT) for imaging female pelvis by comparing dose and image quality between HDCT and a 64-slice standarddefinition CT (SDCT). Methods and Materials: Forty consecutive female patients for pelvic examination were included, with 23 scanned on SDCT and 17 on HDCT. Similar scan parameters were used on both scanners expect the tube currents of 310 mA and 200 mA with SDCT and HDCT, respectively. Images were reconstructed with a standard kernel on both scanners and ASIR (30%) was applied on HDCT. Radiation dose was recorded and image noises for the background, uterine myometrium, internal obturator muscle and fat were measured and compared. Results: All images were clinically acceptable. The image noises for SDCT group were 5.1 ± 0.7 HU, 11.5 ± 2.6 HU, 11.3 ± 2.6 HU, and 9.8 ± 1.6 HU from the background, uterus myometrium, internal obturator muscle, and fat regions, respectively. The image noises for HDCT with ASIR were 5.5 ± 0.9 HU, 10.4 ± 2.0 HU, 10.7 ± 1.6 HU, and 10.2 ± 1.7 HU from the corresponding four regions, respectively. There was no statistical difference for the image noises between the two groups (P > 0.05). On the other hand, the average x-ray dose with HDCT was only 16.9 ± 5.0 mGy, much lower than that of SDCT group (26.1 ± 2.0 mGy) (P < 0.01).

The use of 30% ASIR on HDCT maintained the same image quality for female pelvic scans as conventional SDCT while providing 35% dose reduction to patients. Purpose: This study aimed at investigating the use of the BOLD (blood oxygen level dependent) effect during an oxygen-challenge magnetic resonance imaging in an intra-uterine-growth-restriction (IUGR) rat model.

The IUGR rat model is based on a left uterine artery ligation. A small animal MRI 4.7 T unit was used, and a multiple gradient echo sequence (25 echo times [1.8 to 49.8 ms]; TR 3500 ms) was developed. T2* were measured in maternal livers, in the placentas, in foetal livers and brains, before and after the mother breathed oxygen (6 l O 2 /min). All foetuses and placentas were weighed afterwards to validate the IUGR model. Wilcoxon's signed rank sum tests were performed to compare the right and left foetoplacental units (FPU). Results: Sixteen rats were studied, with a significant-induced growth restriction of 21.89% in the weight of the foetuses. A BOLD effect was detected in the maternal liver, as well as in all the placentas (n = 61, 34 right, 27 left), the foetal livers (n = 62, 41 right, 21 left), and the foetal brains (n = 56, 36 right, 20 left) (p < 0.002). The BOLD effect was significantly different (p < 0.0036) between the left ligated horn (IUGR) and the right horn for the placentas and the foetal brains. No significant difference was noted for the foetal livers. Conclusion: BOLD MRI at 4.7 T is a good non-invasive tool to evaluate FPU oxygenation, and to detect IUGR. This technique may potentially be applied to the identification and understanding of placental dysfunction in IUGR in humans. Purpose: To describe MR imaging morphologic-and signal intensity abnormalities of endometriosis infiltrating the bowel and to assess the value of MR imaging in predicting depth of bowel wall infiltration. Methods and Materials: This, single-centre, retrospective study was performed in a tertiary referral centre for endometriosis (2004) (2005) (2006) (2007) (2008) (2009) (2010) . A total of 473 consecutive patients were analysed for diagnosis of endometriosis infiltrating the bowel. All MR images were analysed by one experienced reader (number of lesions, location, size, SI, and depth of bowel wall infiltration ("adhesions", serosal, muscular or suspicion of mucosal infiltration)). Inter-rater agreement for infiltration of the muscular layer of the bowel wall was calculated in a subset of patients. In 24 patients (27 lesions) MR imaging was correlated with histopathology. Results: MR images revealed endometriosis infiltrating the bowel in 192 (41%) patients (222 lesions). The inter-rater agreement for infiltration of the muscular layer of the bowel wall was 0.80. All lesions showed isointense or slightly hyperintense signal compared with muscle with foci of high signal intensity on T2-and T1-weighted imaging in 42 and 33 cases, respectively. Depth of infiltration was assessed as serosal and muscular 29 and 178 lesions, respectively, and suspicion of mucosal infiltration was raised in 15 lesions. Compared with histopathology, bowel wall infiltration on MR imaging was correctly scored in 22 out of 27 (81%) lesions. Conclusion: MR imaging is useful to predict serosal and muscular infiltration of the bowel wall, providing a roadmap to surgery.

Comparison of 3D and 2D FSE T2-weighted MR imaging in the diagnosis of pelvic endometriosis: preliminary results M. Bazot 1 , A. Stivalet 1 , I. Thomassin-Naggara 1 , E. Darai 1 , C. Coudray 2 , E. Poncelet 3 ; 1 Paris/FR, 2 Buc/FR, 3 Lille/FR (marc.bazot@tnn.aphp.fr)

Purpose: To prospectively evaluate overall image quality and diagnostic accuracy of 3D and 2D T2-weighted MR imaging for the evaluation of pelvic endometriosis. Methods and Materials: One hundred and ten consecutive patients with clinical suspicion of endometriosis were recruited at two institutions. Two readers with different degrees of experience were independently asked to interpret images using both 3D and 2D MRI. Each dataset was independently evaluated by readers for the presence or absence of pelvic endometriosis (ovaries, uterosacral ligaments, vagina, rectosigmoid, Douglas pouch and bladder). Descriptive analysis, Mc Nemar test and Chi-square test were used. Intra-and interobserver agreement was calculated using k coefficients. Results: Both readers found that 3D yielded significantly lower image quality than 2D MRI (p < 0.0001). Acquisition time for 3D was significantly less than that for 2D MRI (p < 0.01). Out of twenty-three women undergoing surgery, 23 had endometriosis at histology. 3D offered a similar accuracy to diagnose ovarian and deep infiltrating endometriosis compared with 2D MRI. For all locations of endometriosis, a high or variable intraobserver agreement was observed for reader 1 and 2, respectively. Conclusion: Despite a lower overall image quality, 3D provides significant time saving and similar accuracy than multiplanar 2D MRI in the diagnosis of specific locations of endometriosis.

Occurrence Purpose: To evaluate the accuracy of magnetic resonance (MRI) in the assessment of extrapelvic endometriosis. Methods and Materials: The MRI exams of 127 patients (mean age 29.8 years) suffering from severe pelvic endometriosis, performed within 1 week before laparoscopy, were retrospectively reviewed by two readers. Number, site, size of each implant were evaluated. At MRI, diagnosis was based on hyperintense lesions on T1-WI, plaque-like hypointense lesions on T1 and T2-WI, or indirect findings (tethering, thickening of peritoneal fat and visceral wall). Using laparoscopy as reference standard, MRI diagnostic accuracy was calculated. Pearson test was used to compare lesions size. Interobserver variability in determining the presence of extrapelvic disease was assessed with k statistics. A value of p < 0.05 was considered significant.

A S309 C D E F G H Methods and Materials: On 119 patients we performed the ultrasonic guided core cut biopsies (7.5 mHz-array, core cut biopsy systems BIP™, today BARD Angiomed™ and Pflugbeil™, coaxial-and core cut needle with a length of 10 cm, diameter 2.1 mm (14G), core cut depth 1.6 resp. 2.2 cm, 1-3 specimen) with a benign histology. Control intervals were 3.6,12 and 24 months sonographically, 12 and 24 months mammographically and a follow-up after 15 years. Results: For 103 of the 117 patients it was possible to perform a follow-up. On one patient a histologically verified fibroadenoma with the tendency to grow was operated and the benign histology was confirmed. A second patient was therapied at a lobular carcinoma 43 months after core cut biopsy and the histology "simple mastopathy", and the third patient was operated at a breast cancer 56 months after intervention in that area, in which the result of the core cut biopsy was "to less material". Sensitivity, specificity ppV and npV were in each case 98%, at a p < 0.0001. Conclusion: Precondition for the feasibility of ultrasonic guided core cut biopsies is the consequent use of the European guidelines. and apparent diffusion coefficient (ADC) values in the diagnosis of malignant thyroid nodules. Methods and Materials: After institutional review board approval, a prospective study was conducted in 14 patients with malignant thyroid nodules, H-MRS, and Diffusion weighted imaging (DWI) with b factors 200-400 s/mm2 were performed. ADC mapping was constructed and ADC values of the nodules were measured. Standard Cho peak, Cho /Cre ratio and ADC values of malignant nodules were correlated with the five benign nodules and four normal appearing thyroid lobe parenchyma. The gold standard reference was the fine needle aspiration biopsy (FNAB) and histopathology. Statistical analysis of this research was assessed by Fischer's exact t-test, Variance analysis test (ANOVA) and receiver operating characteristics (ROC) curve. Results: Two acquisitions at TE: 40-144 msec.were applied in the H-MRS. Average Cho/Cre ratio for the malignant nodules was 2.95+_1.54 at short TE, 5.30+_2.38 at long TE, had significant statistical differences from the benign nodules, and normal thyroid parenchyma (p 0.805 at TE 40 msec. and > 1.225 at TE 144 msec. were assumed to be malignant. Average ADC values of malignant ones were 0.06+_0.02, had significantly lower than the benign nodules and the normal thyroid lobe ADC values (p < 0.05). Conclusion: H-MRS acquisitions at short and long TEs, DWI and ADC mapping are new non-invasive imaging modalities, give diagnostic data about the nature of the nodules and can easily distinguish the malignant thyroid nodules from benign ones.

The efficacy of apparent diffusion coefficient value calculation in differentiation between malignant and benign thyroid nodules A. Dilli, U.Y. Ayaz, E. Cakir, E. Cakal, B. Hekimoglu; Çankaya/TR (umityasarayaz@yahoo.com) Purpose: Differentiation of malignant and benign thyroid nodules is crucial for determination of therapeutic approach. Our purpose is to evaluate the efficacy of apparent diffusion coefficient (ADC) calculation in differentiation between malignant and benign thyroid nodules. Methods and Materials: A prospective study was conducted in 52 consecutive patients (8 males, 44 females) with 59 nodules. Diffusion-weighted single-shot echo-planar magnetic resonance imaging was performed. The b factors were 0 and 400 s/mm 2 . ADC values of the thyroid nodules were calculated. Nineteen nodules were proved to be malignant by performing fine needle aspiration biopsy (FNAB) first and by histopathological examination during and after surgery. Forty nodules were accepted benign, as cytopathological examinations after FNAB were taken as Methods and Materials: 161 thyroid nodules were studied (157 patients) using conventional US examination (MyLab Gold, Esaote. Italy) and SWE elastography (Aixplorer, Supersonic Imagine, France) with a high frequency linear transducer. Nodule stiffness and ratio were calculated. Cytological correlations were obtained in all cases using fine needle aspiration (FNAB). In addition, histological confirmation of cancer was obtained in case of suspicious cytological findings. Results: SWE measurements from 32 nodules were finally excluded during the review process, corresponding to the learning curve. Statistical calculations were performed from the 129 nodules included. The stiffness of the embedding tissue was not significantly different between all the categories. Cancers exhibited a significant increased stiffness (114 ± 61 kPa) compared to normal parenchyma (31 ± 12 kPa) and other nodules (34 ± 17 kPa). The ratio between nodule and parenchyma was significantly higher (p < 0.001). Conclusion: Shear wave elastography can provide additional information to the clinician to diagnose thyroid malignancy, and direct FNAB to the most suspicious lesions. It is an easy and promising imaging technique for characterisation of thyroid nodules. Multicentric study remains necessary to confirm these data.

Prospective evaluation of multiparametric ultrasound and quantitative elastosonography in the differential diagnosis of benign and malignant thyroid nodules V. Cantisani, P. Ricci, O. Medvedeieva, M. Olive, M. Di Segni, E. Mancuso, E. Maggini, R. Passariello; Rome/ IT (vito.cantisani@uniroma1.it) Purpose: To assess the clinical value of quantitative sono-elastography in differentiating hystologically proven thyroid nodules as compared with multiparametric ultrasound evaluation. Methods and Materials: 150 patients (range 21-83 years, mean age: 44 yrs) with thyroid nodules previously detected at colour-Doppler ultrasonound (CDUS), were prospectively examined with sono-elastography with dedicated quantitative software (Elasto-Q, Toshiba) before surgery (complete or partial thiroydectomy). Ultrasound examination and sono-elastography were evaluated by two readers in consensus. US features, colour-Doppler pattern, polychromatic map elastography and strain ratio value were evaluated. Sensitivity and specificity of CDUS and sono-elastography were compared using X 2 test and ROC curves. Results: Sensitivity and specificity considering hypoechogenicity, halo sign, colour-Doppler pattern, polychromatic map elastography and strain ratio were 65, 67, 70, 85 and 96% and 66, 65, 45, 74, 92%, respectively. Sono-elastography was more sensitive and specific than all ultrasonographic in predicting malignancy of the thyroid nodules (p < 0.01). According to the different elastosonographic features the lesions characterised by score 3 or 4 as evidenced at polichromatic map or with a strain ratio > 2 may be considered malignant. Conclusion: Sonoelastography is highly sensitive in the characterisation of thyroid nodules, more accurate than colour-Doppler ultrasound; its application in the current daily clinical practice may limit unnecessary FNAB. Purpose: To investigate the clinical utility of quantifying iodine concentration in spectral CT imaging for differentiating thyroid solitary nodes. Methods and Materials: Thirty patients (ten solitary adenomas, ten solitary papillary carcinomas and ten normal) underwent thyroid gland CT scanning using spectral imaging mode with fast tube voltage switching. Material decomposition images were reconstructed. Regions of interest (ROIs) from unenhanced CT images of two thyroid gland lobes (including lesions) and carotid artery were chosen on iodine-basis images to measure the effective iodine concentrations (eIC). Kruskal-Wallis test was used to analyse eIC differences among the adenoma, carcinoma, and control groups. Results: There was no difference for eIC in the thyroid of the control (2.01 ± 0.79 mg/ cc), the uninjured lobes of the adenoma (1.98 ± 0.6 mg/cc) and carcinoma (2.05 ± 0.72 mg/cc) groups (p = 0.85). The eIC for the carcinomas (0.75 ± 0.2 mg/ cc) was significantly lower than that for the adenomas (1.77 ± 0.48 mg/cc) (p < 0.01), but no statistical difference was detected between the adenomas and the control groups (p = 0.51). There were statistical differences for the carotid arterial eIC of the control, adenoma and carcinoma groups (0.39 ± 0.09 mg/cc, 0.83 ± 0.12 mg/ cc, 1.89 ± 0.26 mg/cc, respectively, p < 0.01). The carcinoma group had the highest carotid arterial eIC. Conclusion: Spectral CT imaging allows for quantitative measurement of effective iodine concentration in solitary benign and malignant thyroid nodes. This may enable the identification of different thyroid diseases in the future.

A S313 C D E F G H was found between DVCT and TTE (r = 0.89, limits ± 0.31 cm 2 ) and between DVCT and TEE (r = 0.84, limits ± 0.39 cm 2 ). Absolute exposure time was 1.28 ± 0.16 s ranging from 0.89 to 1.53 s depending on the heart rate (71.9 ± 10.2 bpm ranging from 51 to 89 bpm). The resulting mean effective radiation dose was 0.61 ± 0.07 mSv ranging from 0.44 to 0.73 mSv. Conclusion: Quantification of the AVA is feasible with a reasonable low radiation dose.

Evaluation of aortic valve morphology using multi-imaging modalities: comparison with intraoperative finding J. Son, S. Ko, J. Choi, Y. Kim, M. Song, J. Shin, H. Hwang, S. Lee; Seoul/ KR (20070035@kuh.ac.kr) Purpose: To retrospectively compare the diagnostic accuracy of dual-source computed tomography (DSCT), cardiac magnetic resonance imaging (cMRI) and transthoracic echocardiography (TTE) for the assessments of the aortic valve (AV) morphology with intraoperative findings as reference standard Methods and Materials: A total of 251 consecutive patients underwent DSCT, cMRI and TTE before undergoing aortic valvuloplasty or replacement. Morphology of AVs evaluated visually by using short-axis DSCT, cine-cMRI and TTE images were compared with each other. Interobserver agreement was measured using Pearson's correlation analysis. The accuracy of diagnosis for multi-imaging modalities was obtained from the comparison with intraoperative findings using the chi-square test. Results: One hundred seventy-seven patients had tricuspid AV, 72 patients had bicuspid AV, and three patients had quadricuspid AV. There was excellent correlation between cMRI and DSCT for AV morphology (r = 0.93, p < 0.0001). There was good correlation between DSCT and TTE (r = 0.78, p < 0.0001), MR and TTE (r = 0.75, p < 0.0001). DSCT, cine-cMRI and TTE had an excellent interobserver agreement (r = 0.93, r = 0.92 and r = 0.82, p < 0.0001). DSCT (n=248) had a sensitivity of 90.3% and a specificity of 92.3% for assessment of AV morphology (bicuspid vs tricuspid valve), and cardiac MRI (n=248) was 87.5% and 88.9%, TTE (n=243) was 90.9% and 73.1%, respectively. Conclusion: DSCT and cMRI enable accurate noninvasive assessment of the preoperative AV morphology.

Diagnostic accuracy of dual-source computed tomography coronary angiography in patients undergoing evaluation for transcatheter aortic valve implantation P. Blanke, G. Pache, S. Bulla, C. Lang, M. Langer; Freiburg/ DE (philipp.blanke@uniklinik-freiburg.de) Purpose: To prospectively investigate diagnostic accuracy for dual-source computed tomography coronary angiography (CTCA) to diagnose significant coronary artery stenosis compared with catheter coronary angiography (CCA) in patients with severe aortic stenosis undergoing assessment of aortic root anatomy for transcatheter aortic valve implantation (TAVI). Methods and Materials: One-hundred and twenty consecutive patients (64/56 female/male, mean age 79.0 ± 10.9 years) with severe aortic stenosis (0.7 ± 0.1 aortic valve area cm 2 ) underwent a specific dual-source computed tomography protocol for evaluation of aortic root and coronay anatomy prior to TAVI. CCA was considered the standard of reference. Results: Mean heart rate was 73.0 ± 15.9 beats/min. Thirty-four (28%) patients had non-sinus rhythm. Prevalence of coronary calcifications was 98% (117 of 120). Mean Agatston score was 873 (median 475; range 0-7412). CCA identified 112 significantly stenosed segments in 46 patients. One-hundred and thirty-nine of 1699 segments (8%) in 43 patients were considered non-diagnostic with computed tomography because of calcium (n = 116), motion artefacts (n = 17) or stents (n = 6). On a patientbased analysis, considering non-diagnostic segments as false positive, sensitivity, specificity and positive and negative predictive values for the detection of significant lesions (≥50% diameter stenosis) were 100%, 68%, 66% and 100%, respectively. Conclusion: In patients undergoing DSCT for assessment of aortic root anatomy prior to TAVI, ruling out of significant coronary artery disease is impaired in a considerable portion of patients due to high calcium burden or heart rhythm irregularities.

reference. ADC values were correlated with histopathological results in malignant nodules and FNAB results in benign nodules. Statistical analysis was performed. Results: The mean ADC value of malignant thyroid nodules was 0.829 ± 0.1786 x 10 -3 mm 2 /s and mean ADC value of benign thyroid nodules was 1.984 ± 0.4819 x 10 -3 mm 2 /s. The mean ADC value of malignant nodules was significantly lower than the mean ADC value of benign nodules (P = 0.0001, < 0.05 = α). There was no significant difference between the mean ADC values of various malignant thyroid nodules. The 95% confidence interval for the mean ADC of malignant nodules was between 0.7429 and 0.9151, whereas it was between 1.830 and 2.138 for the benign nodules. Conclusion: ADC value calculation is a noninvasive, effective method in differentiation of malignant thyroid nodules from benign ones.

Vanishing thyroid in patients with renal cell carcinoma treated with tyrosine kinase inhibitor S. Takahashi, K. Kitajima, T. Maeda, T. Yoshikawa, M. Fujii, Y. Ohno, H. Miyake, M. Fujisawa, K. Sugimura; Kobe/JP (kitajima@med.kobe-u.ac.jp) Purpose: Hypothyroidism is a common complication in patients treated with tyrosine kinase inhibitors. We evaluated the relationship between the thyroid size shown on CT and thyroid function in patients with advanced RCC treated with tyrosine kinase inhibitors. Methods and Materials: Thirty-two metastatic RCC patients treated with tyrosine kinase inhibitors (sorafenib; n = 22, sunitinib; n = 10),with follow-up period of > 12 months were evaluated. The patients who ever showed elevated TSH of > 10 mU/l were defined as "hypothyroidism". CT scans were performed before and 3, 6, 9, and 12 months after the initiation of the treatment. The area of the thyroid at the maximum section in each examination was measured and compared to those before the treatment. Differences in the thyroid size were also compared between the patients groups with or without "hypothyroidism", as well as the type of the drug. Results: 16 patients (8 sorafenib, 8 sunitinib) presented "hypothyroidism" 97.1 ± 91.0 days (range: 12-315 days) after the beginning of the treatment. In patients with "hypothyroidism", the thyroid reduced in size of 87 ± 18% after 3 months, 78 ± 22% at 6 months, 67 ± 22% at 9 months, and 64 ± 23% at 12 months, whereas the patients without "hypothyroidism" maintained the size of 91 ± 13% even 12 months after (p = 0.0027). The patients treated with sunitinib showed more prominent size reduction of the thyroid than sorafenib (50 ± 22% and 78 ± 13% at 12 months; p = 0.0039). Conclusion: After tyrosine kinase inhibitor treatment, the thyroid showed apparent size reduction in the patients with hypothyroidism, especially treated with sunitinib. Purpose: To evaluate a low radiation scan protocol using dynamic volume computed tomography (DVCT) for aortic valve opening area (AVA) quantification. Methods and Materials: A total of 21 patients (body mass index 25.9 ± 3.4 kg/m 2 ) with aortic valve stenosis underwent a contrast-enhanced electrocardiogram-gated scan using a DVCT equipped with a 320-row detector. A targeted dynamic data set of the aortic valve was acquired during a single heart beat with 160 x 0.5 mm collimation, 350 ms gantry rotation time and fixed x-ray tube settings of 100 kV and 100 mA. Aortic valve visibility was judged on a 5-point scale (from 0 = insufficient to 4 = excellent) and AVA was determined by planimetry. Transthoracic and transoesophageal echocardiography (TTE and TEE, respectively) served as references. The effective radiation dose was estimated from the displayed dose-length-product. Results: The aortic valve was visualised with invariably diagnostic quality rated as 2.1 ± 0.8 (range 1-4). The mean AVA on DVCT (0.91 ± 0.36 cm 2 ) was not significantly different from TEE (0.94 ± 0.39 cm 2 ; p > 0.05) but slightly larger compared to TTE (0.83 ± 0.28 cm 2 ; p < 0.01). A good correlation with acceptable limits of agreement S314 B C A D E F G H mode (pitch 3.2; 80 kV tube voltage, 0.28 seconds gantry rotation time). Overall image quality was evaluated using a 5-grade scoring system and was assessed by looking at cardiac and vascular structures. The image quality for the coronary arteries was also evaluated using a 5-grade scale. Quantitative image quality parameters (noise, contrast-to-noise ratio) were measured. Effective radiation doses were calculated. Results: Images of diagnostic quality (grade ≥ 3) were obtained in all 26 children with a mean image quality grade of 4.5 ± 0.5 (range 3-5). Interobserver agreement in grading image quality of graft segments was good (kappa = 0.75).Mean coronary artery visibility grade was 4.1 ± 0.6 (range 3-5) for the right coronary artery and 4.6 ± 0.5 (range 3-5) for the left coronary artery. Image noise in the aorta was 28.9 ± 7.5, contrast-to-noise ratio was 11.9 ± 5.2. Effective radiation dose was 0.21 ± 0.08 mSv. Conclusion: Low dose high-pitch prospectively ECG-triggered DSCT angiography provides adequate thoracic and coronary artery image quality in children with CHD.

Role of low-dose 64-slice multi detector CT in evaluation of children with cyanotic congenital heart disease, and its correlation with echocardiography and cardiac catheterization angiography C. Guruprasadh, K.S. Sodhi, A.K. Saxena, R.M. Kumar, N. Khandelwal; Chandigarh/IN (cguruprasadh@gmail.com) Purpose: To study the role of low-dose 64-slice multi detector CT (MDCT) in evaluation of children with cyanotic congenital heart disease, and its correlation with echocardiography and cardiac catheterization angiography (CCA).

This prospective study included 53 patients in the age group of 1 month to 24 years, with cyanotic congenital heart disease who underwent low-dose MDCT examination. The results were compared with echocardiography and CCA. 23 patients who did not undergo CCA were excluded from the study group.

Results: Out of total study group of 30 patients, patent pulmonary arterial confluence was identified by low-dose MDCT in 23 out of the 29 patients, whereas echocardiography and cardiac catheterization angiography could identify it in only 19 patients. Low-dose MDCT was superior in detecting pulmonary artery and its stenosis, associated cardiovascular and extracardiac thoracic anomalies, whereas CCA was better in detection of major aortopulmonary collaterals. Echocardiography and cardiac catheterization angiography were also better in detection of intracardiac valvular and septal pathologies. The effective radiation dose to the patients less than 2 years of age was in the range of 1.7 to 2.5 mSv, while in patients more than 2 years of age was in the range of 2.1 -4.2 mSv, which is much less than radiation dose entailed in CCA. Conclusion: Low-dose MDCT is a useful tool and can replace invasive cardiac catheterization angiography for diagnosing cardiac and extracardiac thoracic anomalies with markedly reduced radiation dosage to the patient.

Diagnostic value of cardiac magnetic resonance (CMR) in estimation of pulmonary pressure gradient before and 4 months after transcatheter valve implantation F. Secchi, A. Giardino, V. Nardella, G. Di Leo, M. Carminati, F. Sardanelli; Milan/IT (francescosecchimd@gmail.com)

Purpose: To evaluate the diagnostic value of CMR in estimation of pulmonary pressure gradient before and after transcatheter valve (Melody, Medtronic) implantation.

After IRB approval and informed consent, patients with pulmonary valve disease underwent transcatheter valve implantation during which peak-to-peak gradient between right ventricle and pulmonary artery was measured (CAT). Before and 4-months after implantation, they were scheduled for echocardiography (US) and 1.5-T CMR. A turbo fast low-angle shot phase-velocity mapping sequence (TR/TE = 41/3.2 ms, slice thickness 5 mm; velocity of encoding 250 ms) for pulmonary flow estimation was acquired. Pressure gradient (ΔP) was estimated with both US and CMR from peak flow velocity using Bernoulli's equation. Wilcoxon test were used for comparisons. Results: we enrolled 19 patients, all of them studied within one week before valve implantation and 4 months later. All CMR examinations were diagnostic, with metallic artifacts limited to the inner space of the valve and not impairing the flow sequence. At CMR, ΔP before valve implantation (38 mmHg) was larger (p = 0.001) than that after valve implantation (19 mmHg). Before valve implantation no significant differences were observed in terms of ΔP among the three techniques (p > 0.382); after valve implantation no significant difference was observed between CMR and US (p = 1.000). Conclusion: the Melody valve implantation does not impair the pulmonary flow evaluation. It allows for a significant reduction of the pulmonary pressure gradient as confirmed at CMR. Finally, CMR is accurate and comparable with echocardiography. We assessed the ability to discriminate pledgets from contrast-enhanced solutions on MDCT images based on the difference in attenuation. Methods and Materials: PTFE felt pledgets were scanned in 4 different contrastenhanced (Ultravist® -300 mg jopromide/ml) saline concentrations (10.0, 12.0, 13.6 and 15.0 mg/ml) in various experimental conditions; (1) pledget only, (2) sutured to the sewing ring of a mechanical PHV, and (3) sutured to the sewing ring of a mechanical PHV and porcine aortic annulus. Scanning was performed on a 256-slice scanner with 8 different scan protocols with various kV (100,120) and mAs (400, 600, 800, 1000) settings. Attenuation of the pledgets and surrounding contrast-enhanced saline were measured. In addition, the attenuation of pledgets and contrast-enhanced blood was measured on scans of 20 patients with PHVs. Results: The attenuation of pledgets was consistently higher than that of the surrounding contrast-enhanced solution in all in vitro conditions. Attenuation values in the pledgets and solutions were significantly higher with 100 kV tube voltage and increasing contrast concentration. No effect of tube current was present. Attenuation values of the pledgets and contrast-enhanced blood in patients were 420 ± 26 HU (range 383-494), and 288 ± 41 HU (range 202-367), respectively. Conclusion: PTFE felt pledgets have an higher CT attenuation than surrounding contrast-enhanced blood. This helps to differentiate the pledgets from paravalvular leakage.

Percutaneous aortic valve replacement: predictive value of cusp calcium quantities for peri-procedural complications F. Schwarz, P. Lange, F. Bamberg, K. Nikolaou, M.F. Reiser, P. Boekstegers, C. Becker; Munich/ DE (florian.schwarz@med.lmu.de) Purpose: Percutaneous aortic valve replacement (PAVR) has been established to treat patients with severe aortic stenosis who are not amenable to surgical intervention. Although MDCT plays a central role in workup, little is known about CT-derivable parameters to predict procedural risk. Here, we analysed the predictive value of calcium on individual cusps for the occurrence of av-blocks and aortic regurgitation (AR).

After IRB-approval, 63 consecutive patients with severe aortic stenosis underwent ECG-synchronised high-pitch MDCT 3-30 days prior to intervention. Valve calcification was assessed quantitatively for each cusp using a semi-automated approach. Postinterventional aortography and diastolic aortic pressure were used to quantify AR. Patients were followed up clinically for 3 months. Results: PAVR was successfully performed in all patients. 30-day mortality was 1/63. 10 patients developed AR > 1°. Pacemaker implantation was required in 15 patients. Mean calcium-quantity was 572 ± 454, distributed evenly among the cusps. In the subgroup of patients with AR > 1°, overall calcium-quantity was significantly higher than in those with AR <= 1° (919 vs. 500, p < 0.05). This relation was preserved on the cusp level. Patients requiring pacemaker implantation had lower amounts of overall calcium; however, this relation became significant for the non-coronary cusp only (145 vs. 266, p < 0.05). Calcium-quantity on the non-coronary cusp showed highest predictive value among all cusps and higher predictive value than overall calcium (ROC). Conclusion: Analysis of the correlation between CT-derived parameters and clinical outcomes in PAVR can lead to more individual risk prediction models and thus to a further increase in safety of this emerging technique.

Low dose high-pitch 128-slice dual-source computed tomography for the evaluation of congenital heart disease in infants and children Z. Cheng, X. Wang; Ji Nan/ CN (chengzhaoping110@sina.com) Purpose: To assess ECG-triggered dual-source computed tomography (CT) in high-pitch spiral acquisition mode for the evaluation of congenital heart disease (CHD) in infants and children regarding image quality and radiation dose. Methods and Materials: 26 consecutive patients (16 men, age 2.5 years) underwent clinically indicated 128-slice dual-source CT angiography of the entire thorax for the evaluation of CHD using a prospectively ECG-triggered high-pitch spiral acquisition Purpose: To characterise the brain volume abnormalities associated with the diagnosis of first-episode schizophrenia in young patients using voxel-based morphometry (VBM). This issue has never been investigated for the Russian population until now. Methods and Materials: Group of 12 right-handed young patients with first-episode schizophrenia (F20.09, ICD-10; 19-25 years) and group of matched mentally healthy subjects were compared utilising 3 T MRI. The data were acquired using isotropic T1WI based on 3D Turbo Field Echo. For VBM analysis we used Statistical Parametric Mapping program (SPM5) developed in the Wellcome Institute (London, UK). This technique allows the evaluation of grey matter volumes in subjects with schizophrenia compared with healthy controls in an automated fashion, across the whole brain. Original MR-images were imported into the SPM, preprocessed and statistically compared using two-sample t-test with false discovery rate (FDR) of 0.05. Age and total intracranial volume were used as covariates. Results: Patient group had significantly reduced grey matter in the left hemisphere Brodmann's area 9 and 10. Height threshold: T = 3.39, p = 0.002 FDR p < 0.05 Extent threshold: k = 15 voxels, p = 0.664 Conclusion: Left hemisphere prefrontal cortex grey matter reduction in schizophrenia patients may be crucial to the pathogenesis of schizophrenia. This finding could become useful prognostic marker. Finally, we emphasise future research directions: study of the relationship between VBM findings of multifocal grey matter deficits and changes in the interconnecting white matter tracts; the need for populationbased VBM studies; and study for correlation between VBM measurements and other important schizophrenia markers.

Concentration changes of whole brain grey and white matter in child patients with Tourette's syndrome: evidence from voxel-based morphometry Y. Liu, Y. Peng, P. Gao; Beijing/ CN (liuyue20036@yahoo.com.cn) Purpose: Pathophysiological evidence suggests involvement of fronto-striatal circuits in Tourette syndrome (TS). To identify TS-related abnormalities in grey and white matter,we used optimised voxel-based morphometry (VBM) which provides a quantitative measure of concentration. Methods and Materials: Three-dimensional T1-weighted (3DT1) MRI were acquired in 31 TS children (28 boys, 3 girls, mean age 8.4 years, range 2-15 years) and 50 age-and sex-matched controls on a 1.5 Tesla Philips scanner. Images were pre-processed and analysed using a version of VBM2 in SPM2. Results: Using VBM, TS patients showed significant increases in grey matter concentrations in prefrontal areas, sensorimotor areas, hippocampus, the left posterior cingulate gyrus, right caudate nucleus, right cerebellar hemisphere, left pons and parietal-occipital association cortex bilaterally. Decreases in grey matter concentrations were seen in medulla oblongata and left pons. Decreases in white matter concentrations were found in right thalamus, right caudate nucleus, right orbitofrontal structures, left anterior cingulate gyrus, left precentral and postcentral gyrus. Increases in white matter concentrations were detected in pons and anterior midbrain (P < 0.05, K E ≥ 10voxels). Tic severity was not correlated with any con-

In utero tractography initially allowed the consistent 3D visualisation of trajectories corresponding to the anatomical course of the bundles of Probst in 2/10 CCA cases uni-and in 8/10 CCA cases bilaterally from 20 GW onwards. In PCA an atypical "sigmoid" bundle connecting the opposing frontal and occipital lobes, which could not be visualised by conventional (T2-and T1-weighted) foetal MR sequences, was depicted by DTI in 3/3 cases. Conclusion: Prenatal DTI provides further insights in the abnormal connectivity in cases of commissural abnormalities and thus helps to establish a more specific imaging diagnosis in a heterogeneous group of developmental brain pathologies. Results: 100 foetuses were included. Foetal MRI was performed at a mean gestational age of 33 weeks (min: 24 weeks, max: 39 weeks) and postnatal MRI, at a mean age of 3 months 18 days. The results were subdivided as follows: group 1 (n = 52), group 2 (n = 32), group 3 (n = 4), group 4 (n = 11) and group 5 (n = 1). In 16 cases (16%), the postnatal diagnosis was different from the prenatal one and this had a prognostic impact in 10/16 cases. Most discrepancies involved the corpus callosum anatomy (partial vs complete agenesis) and cortical and migration disorders. Conclusion: On a large cohort of 100 foetuses, pre-and postnatal MRI data showed good agreement in 84% of cases. Disagreement had a prognostic impact in 10/16 cases. Purpose: Pilomyxoid astrocytoma (PMA) is a recently described astrocytic tumour that has been previously diagnosed as pilocytic astrocytoma (PA). The purpose of this study was to describe the imaging features of PMAs in comparison with PAs. Methods and Materials: We retrospectively reviewed CT/MR images and medical records of 10 patients with PMA and 38 patients with PA. The mean ages of patients with PMA and PA were 10 and 15 years, respectively. Imaging features including location, composition, enhancement pattern, presence of calcification, haemorrhage, and leptomeningeal dissemination were compared in patients with two tumoru types. Results: Six PMAs (60%) occurred at the suprasellar area and the cerebellum was the most common (45%) site of PA. Solid component was dominant in eight PMAs (80%) and in 19 PAs (50%). All of the PMAs containing solid mass (n = 8) included non-enhancing portion while 12/37 (32%) PAs included non-enhancing solid portion (p < 0.05). Leptomeningeal dissemination was noted in five PMAs (50%) and one PA (3%) (p < 0.05). Other imaging findings were not significantly different. Conclusion: A younger age, more frequent occurrence at the suprasellar area, mainly solid mass containing non-enhancing portion, and more frequent leptomeningeal dissemination are helpful differential features of PMAs as compared to PAs.

Tumour/normal brain apparent diffusion coefficient value ratios allow for distinguishing common paediatric cerebellar tumours K. Koral 1 , L. Gargan 1 , K. Cederberg 1 , D.C. Bowers 1 , B. 

A S317 C D E F G H -0.62 to 0.04) and intracellular uptake of Gd-EOB-DTPA (r = -0.20; p = 0.29; 95% CI: -0.53 to 0.17) were not significantly correlated to SUV max . Conclusion: MRI-perfusion parameters from the DCE-MRI provide functional information for liver metastases of neuroendocrine tumours. Especially arterial plasma flow shows a good correlation with SUV-values derived from the specific PET tracer Ga-Dotatate.

Gadoxetic acid-enhanced MRI with MR cholangiography for the preoperative evaluation of bile duct cancer H. Sun, J. Lee, H. Park, B. Choi, J. Han; Seoul/ KR (hysunsun@hanmail.net) Purpose: To assess the diagnostic performance of gadoxetic acid-enhanced MRI with MR cholangiography (MRC) in the preoperative evaluation of bile duct cancer staging and resectability. Methods and Materials: 73 patients, with suspected of having bile duct cancer detected on CT, were included in this study. Two radiologists evaluated the MRI, including 3D-MRC and gadoxetic acid-enhanced dynamic images regarding the tumor respectability. Of the 73 patients, 69 patients underwent surgical exploration, otherwise 4 patients had palliative drainage because of unresectability. 69 patients were included in the assessment of tumour extent, vascular involvement, and lymph node metastasis. The results were compared with the surgical and pathology findings used as standard reference. Diagnostic performance was evaluated using receiver operating characteristics. Results: The overall accuracy of each reviewer for determining the tumour respectability was 0.62 and 0.94, respectively. The Az values were 0.802 for reviewer 1 and 0.892 for reviewer 2 for assessment of the both secondary biliary confluence tumour involvement, and 0.773 for reviewer 1and 0.846 for reviewer 2 for assessment of the intrapancreatic duct involvement. Reargding the vascular involvement, the Az values were 0.718 and 0.906, respectively, for the hepatic artery evaluation and 0.55 and 0.88, respectively, for the portal vein evaluation. For assessment of lymph node metastasis, the overall accuracy was 0.69 and 0.79, respectively. Conclusion: Gadoxetic acid-enhanced MRI with MRC is accurate and feasible for assessing the tumour extent of bile duct cancer; however, there was limitation in the evaluation of vascular and lymph node involvement, which influence diagnostic accuracy for assessing the tumour resectability.

Correlation between histologic diagnosis and MR signal intensity after Gd-EOB-DTPA administration of nodules detected within cirrhotic explanted livers: retrospective analysis V. Battaglia, E. Bozzi, D. Campani, C. Bartolozzi; Pisa/IT (novetrequarti@yahoo.it) Purpose: To retrospectively evaluate MR signal intensity on baseline, dynamic and hepatobiliary phases of nodules proven at histology on explanted cirrhotic livers. Methods and Materials: Histology detected 71 native nodules on 25 explanted livers (33 hepatocellular carcinomas-HCC, 20 high grade-HGDN, 18 low grade-LGDN dysplastic nodules). MR studies were retrospectively analysed to evaluate signal intensities of nodules. Nodules signal intensities were analysed on baseline T2w.i and T1w.i, dynamic arterial (AP) and late (DP) phases and hepatobiliary (HBP) phases (20 minutes after injection of Gd-EOB-DTPA). Dynamic enhancement patterns of nodules were classified as: A-isointense on AP and DP; B-iso/hypointense on AP and hypointense on DP; C-hyperintense on AP and hypointense on DP; D-hyperintense on AP and isointense on DP. Nodules on HBP were classified as iso, hyper or hypointense. Results: On baseline, a significant difference (p < 0.00001) among signal intensity of HGDN/LGDN and HCC was found. On dynamic study, pattern C showed a positive predictive value (PPV) in assessing the diagnosis of HCC of 100%.Ten out of 33 HCC showed pattern B; no significant difference was observed between neoplasms' dynamic patterns and histological grade (p < 0.5). No diagnostic predictive dynamic pattern was found for HGDN, while a pattern A resulted to be typical of LGDN (p < 0.00001). On HBP, 48/71 lesions resulted to be hypointense, of whom 32 were HCC. The remnant 16 hypointense lesions on HBP were HGDN.

LGDN were all iso or hyperintense. Conclusion: Patterns B and C associated with nodular hypointensity on HBP are diagnostic for nodular premalignancy/malignancy. centration changes of grey and white matter in brain. Tic course was negatively correlated with right caudate nucleus (Beta = -3.233, P = 0.003) and the left posterior cingulate gyrus (Beta = -0.382,P = 0.044). Conclusion: Our MRI in vivo neuropathological findings support the hypothesis that alterations in frontostriatal circuitries underlie TS pathology. We also suggest that abnormalities of frontal-parietal-occipital regions, limbic system, cerebellar hemisphere and brain stem are associated with TS.

Preliminary fMRI study on the active dependence of visual cortex in children with anisometropic amblyopia L. Jiang, X.M. Wang, B.N. Luo; Guangzhou/CN (lxjiang427@gmail.com) Purpose: To investigate the active dependence of visual cortex in the visual cortex with blood oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI) in children ametropic amblyopia and the relation to the spatial frequencies of visual signal. Methods and Materials: Eight children with anisometropic amblyopia on single eye were performed BOLD-fMRI examination using Siemens Magnetom 3 T trio equipment. The brain cortex activated by amblyopic eye and contralateral normal eye were evoked and detected by visual stimuli of reversal black-and-white checkerboard pattern with different spatial frequencies of 0.2, 0.4, 0.7 cpd individually. The data were processed with SPM software and the visual cortex activated by amblyopic eyes and contralateral normal eyes were compared. Results: BA17, 18, 19 were activated by both eyes. Compared to the contralateral normal eyes, the size and the intensity of the visual cortex activated by amblyopic eyes decreased (p < 0.05) mainly in the areas of BA17, 18. The spatial frequency of black-and-white checkerboard pattern would affect the activation of visual cortex on the active size and intensity. With higher spatial frequency, the size and intensity of the activation visual cortex activated by amblyopic eyes decreased in the area of BA18, 19 (p < 0.05), while there was no difference between the size and intensity of the activation visual cortex activated by contralateral normal eyes. Conclusion: The active dependence of visual cortex activated by amblyopic eye decreases in size and intensity. And the higher spatial frequency will inhibit the excitability of visual cortex of amblyopic eye. Purpose: To correlate perfusion parameters from dynamic-contrast-enhanced MRI (DCE-MRI) with SUV derived from PET-CT imaging in liver metastases from neuroendocrine tumours (NET). Methods and Materials: 24 patients with proven metastases of NET underwent DCE-MRI at 3 T (Siemens Verio) using a 3D gradient-echo sequence (TWIST; 2.1 sec temporal-resolution, 5 min acquisition-time) and liver-specific contrast-agent Gd-EOB-DTPA (Primovist, Bayer). All patients underwent PET-CT with the specific tracer Ga-DOTATATE within 1 week after DCE-MRI. ROIs were placed in hepatic metastases (> 3 cm) visible both on PET-CT and DCE-MRI-images. In addition, one ROI was placed in non-metastatic liver tissue. DCE-MRI parameters (arterial and venous plasma flow, extracellular MTT, extracellular volume and intracellular uptake of Gd-EOB-DTPA) and corrected SUVmax (SUVmax metastasis / SUVmax normallivertissue ) from PET-CT datasets were analysed for all ROIs using the software PMI 0.4. Correlation of DCE-MRI parameters and SUV max values was analysed using Pearson's correlation coefficient (r). Results: A total of 62 ROIs were placed in hepatic metastases of NET. Arterial plasma flow showed a high correlation value with Ga-Dotatate uptake in PET-CT (r = 0.77; p < 0.0001; 95% CI: 0.56 to 0.89) followed by extracellular volume (r = 0.43; p = 0.020; 95% CI: 0.08 to 0.69 A S321 C D E F G H resolution of standard dual energy CT is 140 ms. Therefore, additional images were generated using a dedicated, mixed single source/dual source reconstruction technique available at the scanner that should result in a resolution of 75 ms. In addition, scans were reconstructed at 140 and 280 ms. All coronary segments, the aortic valve and the epicardial border of the myocardium were evaluated with regard to motion artefacts on a four-point scale (1: non-diagnostic to 4: excellent). Results: Heart rate did not significantly differ between the single energy and dual energy acquisitions (49.8 ± 9.6 and 52.5 ± 11.0 beats/minute). Image quality was similar for single energy and dual energy CT at a temporal resolution of 75 ms. However, moderate to severe artefacts were found more frequently in dual energy images. Significant lower image quality was scored at 140 and 280 ms temporal resolution. Conclusion: Temporal resolution of dual energy DSCT with a mixed reconstruction technique appears to be similar to the resolution of single energy DSCT. Only at high heart rates image quality of dual energy DSCT seems to be lower than single energy DSCT.

14:00 -15:30 Room Q

Managing patient dose, quality assurance Purpose: Current radiological reports are individual and created by conventional free-text. Gathering clinical key parameters is therefore often time consuming and difficult to establish in clinical routine. To enhance diagnosis and therapy monitoring structured standardised reporting (SSR) might be beneficial.

To assess communication between the University departments of Vascular Surgery and Radiology, a technology acceptance model was adopted to evaluate the acceptance of current reports in the field of cardiovascular radiology. Based on a SWOT (strengths, weaknesses opportunities, threats) analysis and a detailed literature review of relevant clinical parameters, a new concept for SSR for abdominal aortic aneurysms (AAA) was developed. Qualitative data were analysed from field observations and interviews were conducted with experienced clinicians within the related departments. Results: An innovative web-based reporting application standardised for AAA was implemented to enhance effectiveness of radiological reports. Clinical parameters of highest interest were subdivided into three main categories: characteristics of the pathology and adjacent anatomy, measurements, as well as additional findings. Integrated graphical widgets (e.g. drop-down-menus) provide the selection of predefined data entries with high usability. Figures for optional use to guide and standardise the reporting are embedded. The new system is open to communicate also with radiology and hospital information systems (RIS, HIS). Conclusion: Structured standardised reporting has strong potential to improve clinical communication, to accelerate the reporting process and to enable a more comprehensive overview of the patient's diagnosis and monitoring. In summary, the presented SSR application may facilitate a faster extraction of relevant clinical information.

Development and long-term validation of web-based CAD server for synergic performance improvement of CAD software and radiologists Y. Nomura, N. Hayashi, Y. Masutani, T. Yoshikawa, M. Nemoto, S. Hanaoka, S. Miki, K. Ohtomo; Tokyo/JP (nomuray-tky@umin.ac.jp) Purpose: In this study, we aimed at development and validation of our computerassisted diagnosis/detection (CAD) server with web interface for improving performance of both CAD software and radiologists.

We have been building a clinical CAD server system for processing DICOM data transferred from PACS systems, and for displaying the There was a significant increase in the mean scan length of the abdomen from 519.7 ± 50.6 to 681.8 ± 263.9 cm with ASIR (p < 0.001). Total mean dose estimation was 24.1 mSv (FBP) and 17.1 (ASIR), resulting in a reduction of 30%. Conclusion: 64-row whole-body trauma CT with ASIR helps to significantly reduce exposure dose to the patients by 30% in comparison to CT with FBP.

Abdominal multidetector-row computed tomography ( We considered additional findings which could be related to CS as CS-related additional findings (CSrAF). We considered non-CS-related additional findings requested to be actioned as RIFs: RIF1 (further investigations/follow-up); RIF2 (nonsurgical consultation); RIF3 (biopsy/surgical consultation). Incidental findings not requested to be actioned were excluded. Examinations were classified: R1 = confirmed-CS, no CSrAFs/RIFs; R2 = non-confirmed CS, no CSrAFs/RIFs; R3 = confirmed-CS, CSrAFs; R4 = non-confirmed-CS, CSrAFs; R5 = confirmed-CS, RIF; R6 = non-confirmed-CS, RIFs. Chi-square, 95% confidence intervals (95% IC) with binomial distribution were used. Results: Rate of males and patients' age were not significantly different for inpatients and outpatients; 53.5% versus 54.0%; 67.0 ± 14.6y (mean ± standard deviation) versus 67.0 ± 12.3 y, respectively. Inpatients: R1 = 61%; R2 = 11%; R3 = 10%; R4 = 3%; R5 = 12.5%; R6 = 2.5%; outpatients: R1 = 66%; R2 = 13.5%; R3 = 5%; R4 = 1.5%; R5 = 10%; R6 = 4%, respectively (p = n.s). RIFs rate was 58/400 overall (14.5%, 95% IC 11.2%, 18.3 %); 15.0% for inpatients,14.0% for outpatients (p = n.s). Of 58 RIFs, 28 were RIF1 (48.3%), 25 RIF2 (43.1%), 5 RIF3 (8.6%); inpatients 33.7%, 60.0%, 6.7%; outpatients 64.3%, 25.0%, 10.7%, respectively (P =.001). Conclusion: About 1/7 patients was diagnosed for a RIF prompting other medical actions. Request of further investigations/follow-up was more probable than medical consultation for outpatients. aMDCT exams work as an unwanted screening for undefined diseases. Consequences in terms of patient outcome and rate of overdiagnosis and overtreatment are not known.

Results of a new quality assurance visual test for medical display calibration S. Busoni 1 , C. Fulcheri 2 , G. Belli 1 , D. Deroo 3 , T. Kimpe 3 , C. Gori 1 ; 1 Florence/IT, 2 Sesto Fiorentino/IT, 3 Kortrijk/BE (busonis@aou-careggi.toscana.it) Purpose: A new generation QA visual test, based on dynamically generated disks, was developed to test if displays are properly calibrated with respect to reading room illuminance and DICOM GSDF.

The visual test is based on detectability of discs, with variable size and luminance, displayed randomly over a uniform background. The user is asked to click on detected discs and the software registers the percentage of correct answers. Two types of visual tests were optimised to assess if the display is properly calibrated with respect to the ambient light settings and the GSDF. Optimal ranges for relevant parameters (disc size/luminance, background luminance) and the effect of a time constraint were investigated. A dataset, large enough to meet statistical significance, was assembled. Tests were performed on a BARCO Coronis medical display. Display calibration and illuminance were measured with a calibrated measurement device. Results: Optimum disc size (between 8 and 20 mm), background and disc contrast values (±2, ±5, ±7JNDs) were defined to maximise test effectiveness. The 95% confidence level that the ambient illuminance is not correct, for 3 lx reading room calibrated display, requires at least 10 and 50 readings, if 75 lx and 25 lx are used, respectively. The 95% confidence level that the display is calibrated with a GSDF curve instead of a gamma 1.8 or 2.2 curve, requires at least 10 and 20 readings.

The developed visual test provides a quick and reliable way to check if the ambient illuminance and the display calibration curve are set properly.

A S323 C D E F G H results as web pages. The data processing function is implemented as a "plug-in" of the server, for visualisation, automated lesion detection, measurement, etc. Another important feature of the server is a function for radiologist's feedback. For example, each radiologist personally classifies lesion candidates of CAD result to four categories; true positive (TP) recognised before CAD reference, TP missed before CAD reference, false positive (FP), and pending. This feedback scheme is purposed not only for collecting additional learning samples for CAD software but also for explicit display of radiologist's interpretation characteristics by pointing out the number and the locations of lesions missed by the radiologist. That is, it realises synergic performance improvement of CAD software and radiologists. Results: In the clinical use for 20 months, feedback data of over 2,300 cases have been collected. Using additional learning samples, the sensitivity of the cerebral aneurysms detection in MRA data was improved from 83.6% to 90.9% at 6.0 FPs per case. The display of the individual radiologist's interpretation characteristics has been useful for calling attention to the radiologist's weak points in lesion detection.

We have developed a novel CAD server system, and validated its effectiveness through long-term clinical use.

Cloud computing and radiology: applications, challenges and future developments C.J. McCarthy, D.P. Brophy, C.P. Cantwell; Dublin/IE (colin.mccarthy@iol.ie) Purpose: We review the advent of cloud computing, its potential for use in radiology and outline how we used cloud computing to create an original web-based radiology portal.

The modern radiology department is heavily dependent on often expensive computing infrastructure. Cloud computing offers the possibility of online databases and customised application delivery over the internet, with the required infrastructure located off-site. Benefits such as savings on capital expenditure and scalability need to be balanced against issues such as data security and patient confidentiality. We developed a web-based resource using open source software and obtained seamless integration with a cloud-based application delivered securely from a remote server. Results: The portal incorporates a teaching file, reporting tools, group calendar, content-management system and a forum with private-messaging capabilities. In addition, dedicated modules allow tracking of discrepancy cases and monitoring of adverse patient events. The portal can be accessed at any radiology workstation on the campus or using a mobile device, and has lead to an exponential increase in data collection for quality assurance and teaching purposes.

Conclusion: Using open source software and cloud computing together has improved education, discrepancy reporting and quality assurance in our multi-site PACS environment. Ensuring compliance with data protection directives, such as HIPAA (US), the EU Data Protection Directive (95/46/EC) and local regulations poses an important challenge to the use of cloud computing in radiology.

Ratios to thalamus were consistently lower in all cases and again measurements performed by the paediatric neuroradiologist and fellow were concordant. ROC analysis distinguished between PAs and ependymomas (threshold = 1.8033: sensitivity = 93.55, specificity = 78.57) and also between ependymomas and MBs (threshold = 1.2461: sensitivity = 100, specificity = 92.59). MBs and AT/ RTs were not significantly distinguishable. Conclusion: Tumour/normal brain ADC ratio is a useful clinical tool in distinguishing common paediatric cerebellar tumours. and portal vein-to-liver contrast-to-noise ratio (CNR) were calculated from each image set. These calculated values and the results of clinical and laboratory tests were statistically analysed. Results: Serum bilirubin level showed relatively good correlation with imaging results: liver SNR at 20 minutes (r = -0.706, p <.01), portal vein-to-liver CNR at 20 minutes (r = -0.7248, p <.01), portal vein-to-liver CNR at 60 minutes (r = -0.7124, p < 0.01), and the ratio of liver SNR at 20 minutes to non-enhanced phase (r = -0.784, p <.01). The liver SNR and portal vein-to-liver CNR were higher in patients with Child class A than with Child class B or C (p <.003), and patients with normal serum bilirubin level than those with elevated serum bilirubin level (p <.003). In both groups, portal vein-to-liver CNR were higher in hepatobiliary phase images obtained at 60 minutes than those

Lesion conspicuity on 100% radiation-dose FBP-images and 50% radiation-dose MBiR-images were rated equally. Decreasing radiation-dose from 100% to 10% increased mean image noise from 30.0 HU to 87.7 HU (FBP), and from 12.1 HU to 17.2 HU (MBiR)

Methods and Materials: 24 patients with abdominal 64-row CT (LightSpeed VCT XT, GE) as reference examination and a follow-up study on a CT with ASIR option (Discovery 750HD, GE) were included. The noise index was increased from 29 (reference) to 43 (follow-up's) with other relevant scan parameters held constant. Images (5 mm ax, 3 mm sag and cor; standard kernel) were reconstructed using different ASIR levels (0, 30, 50, 70, and 100%, slice and volume mode). 2 experienced radiologists performed consensus reading with respect to image quality regarding soft tissue evaluation in comparison to the reference (-2:clearly inferior, -1:inferior, 0:equal, +1:superior, +2:clearly superior). The Mann-Whitney-U-test was used for statistical testing. Results: ASIR 30 and 50 performed equal to the reference (mean -0.1; p> 0.05), ASIR levels of 0 or 100% were rated inferior (-0.3 and -0.7; p < 0.01) without significant differences for slice and volume mode (p> 0.05). ASIR 70 volume mode performed equal to ASIR 30 and 50 slice mode (0.0; p> 0.05) while ASIR 70 slice mode was inferior (-0.2; p < 0.05). Mean CTDI of the follow-up studies decreased significantly from mean 19

Noise reduction and image quality improvement of low-dose and ultra-lowdose brain perfusion CT by HYPR-LR processing R

Methods and Materials: Simultaneous BPCTs were acquired in 8 patients over 60 seconds on a dual-source CT system by applying LD (80 kV, 200 mAs) on tube A and ULD (80 kV, 30 mAs) on tube B. Both datasets were post-processed using HYPR-LR. Correlation coefficients between mean attenuation values within corresponding ROIs, area under attenuation curve (AUC), maximum attenuation (MAX), and signal to noise ratio (SNR) of brain parenchyma were assessed. Subjective image quality was assessed on a 5-point scale by two blinded observers. Results: Radiation dose of ULD was more than six times lower compared to LD

compared to groups 1-3 (p < 0.02) with no difference between groups 1-3. BN was significantly lower (p < 0.01) in group 4 (12 ± 3 HU) and group 1 (13 ± 6 HU) compared to groups 3 and 2 (16 ± 6 HU and 23 ± 9). Conclusion: On CTPA with a DECT scanner of the second generation noise levels are comparable to a 120 kV protocol, while dose can be significantly reduced with a 80 kV/140 kV+Sn filter configuration. Purpose: To evaluate the image quality of an iterative reconstruction algorithm (IRIS) in low-dose chest CT in comparison with standard-dose filtered back projection (FBP) CT. Methods and Materials: 80 consecutive patients referred for a follow-up CT examination of the chest, underwent a low-dose CT examination (group 2) in similar technical conditions to those of the initial examination (group 1) except for the milliamperage selection and the replacement of regular FBP reconstruction by iterative reconstructions using 3 (group 2a) and 5 iterations (group 2b).

No difference was found between the CT protocols in sensitivity, specificity, positive and negative diagnostic likelihood ratios at all ramification levels of the pulmonary arteries (P between.343 and 1). The overall sensitivity and specificity with the normal and simulated low-dose protocol were 79.9% vs 81.3% and 98.0% vs 98.2% (P =.444 and.702), respectively. The diagnostic confidence (2.81 ±.39 vs 2.77 ±.47; P =.297) and overall image quality (3.92 ±.52 vs 3.83 ± 0.54; P =.216) were similar at 120 kV and 80 kV. Purpose: To evaluate the assessibility of coronary arteries from standard chest CT angiographic examinations acquired with a high-pitch mode. Methods and Materials: 242 patients underwent a nongated CT angiographic examination of the chest with dual-source, single-energy CT obtained with a pitch of 3 and a temporal resolution of 83 ms (prototype version; group 1; n = 123) or 75 ms (commercial version; group 2; n = 126). Results: The mean duration of data acquisition and mean dose-length-product were 1.49 ± 0.42 s and 136.9 ± 49.8 mGy.cm, respectively. The median value of the heart rate was 81.7 ± 15 bpm. The percentage of assessable segments was 88% at the proximal level (i.e., 4 segments), 75% at the proximal and mid-segment level (i.e., 7 segments), 61% and 48% when considering 10 and 15 segments, respectively. The mean (±SD) number of assessable segments per patient was 3.5 ± 0.78 and 5.2 ± 1.50 when considering 4 and 7 segments per patient, respectively. The percentage of patients with 4 segments assessable was 67% (126/242), decreasing to 23% (55/242) with 7 segments assessable and 3% (7/242) with 10 segments assessable while no patient's entire coronary artery tree was assessable. No statistically significant difference was observed when comparing group 1 and group 2 patients. Conclusion: Diagnostic image quality is attainable at the level of proximal and mid-coronary segments in 23% of patients with a dual-source, high-pitch scanning mode despite the absence of ECG gating and elevated heart rates.

Triphasic contrast injection improves evaluation of dual energy lung perfusion in pulmonary CT angiography J.M. Kerl, R.W. Bauer, M. Renker, P. Weisser, B. Schell, T.J. Vogl; Frankfurt a. Main/DE (matthias.kerl@kgu.de) Purpose: Lung perfusion analysis in DECT is sensitive to beam hardening artefacts from dense contrast material. We compared two scan and four CM injection protocols in terms of severity of artefacts and attenuation levels in the thoracic vessels. Methods and Materials: Data of 120 patients who had undergone DECT pulmonary angiography were evaluated. Group 1 (n = 30) was scanned in a craniocaudal scan direction versus groups 2 -4 (n = 30 each) were scanned in a caudocranial scan direction. In group 1 -3 biphasic injection protocols with different amounts of CM and NaCl were investigated. In group 4 a triphasic protocol with an initial CM bolus followed by a NaCl/CM mixture and NaCl chaser bolus was used. Attenuation in the subclavian vein, superior vena cava and pulmonary artery tree was acquired. Artefacts on DE iodine distribution map were rated on a scale from 1 to 5 by two blinded readers in consensus. Results: In protocol 4 mean attenuation in the SV (645±158 HU) and SVC (389±114 HU) was significantly lower compared with groups 1-3 (p < 0.002). Artefacts in group 4 (1.1±0.4 and 1.5±0.7) were rated significantly less severe compared with all other groups for the subclavian vein and SVC (p < 0.01). Attenuation in the PA was maintained on a diagnostic level in all 4 groups. Conclusion: Split-bolus injection provides sufficient attenuation for pulmonary DECT angiography while beam hardening artefacts arising from high-density contrast material in the thoracic vessels can generally be avoided.

The effect of experience and skill on volumetric helical perfusion CT quantification of primary breast cancer A. Gogbashian, S. Li, V. Shah, I. Simcock, J. Stirling, A. Makris, V. Goh; Middlesex/UK (andrew@cardiacforum.com) Purpose: To assess the effect of experience and skill on inter-and intra-observer reliability of quantitative volumetric helical perfusion CT (vPCT) in primary breast cancer. Methods and Materials: Following ethical approval, 20 prospective patients with primary breast cancer underwent vPCT (4D adaptive spiral; Dual Source CT, Siemens) encompassing the whole tumour prior to surgery. Five observers participated (attending radiologist, oncology fellow, radiology fellow, radiology resident, CT technologist). Following dedicated training, each observer evaluated 20 datasets (deconvolution analysis; vPCT Body, Siemens). Maximum tumour size, volume, whole tumour blood flow (BF), blood volume (BV) and extraction fraction (ES) were recorded for the volume of interest. Each dataset was re-evaluated with > 4 weeks between reads. Analysis of variance was performed to assess inter-observer differences in parameters. Bland-Altman statistics were applied to assess intra-observer reliability. Significance was at 5%. Results: Mean (SD) tumour size was 4.1 (2.1) cm and volume 42 (90.9) cm 3 . Within observers, the intra-class correlation coefficient ranged from 0.83 to 0.99 demonstrating excellent intra-observer agreement. There were no significant differences in quantitative parameters between observers; BF (F = 0.56; P = 0.69), BV (F = 0.84; P = 0.50) and ES (F = 1.13; P = 0.34) indicating that quantification using a volumetric region of interest and vPCT is highly reproducible irrespective of experience and skill. Conclusion: Quantitative vascular parameters are being applied increasingly to assess tumour biology and therapy response in clinical practice. With vPCT and volume of interest analysis, data can be reliably and reproducibly obtained by observers with differing experience.

Impact of real-time virtual sonography, a coordinated sonography and MRI system with magnetic navigation, on the preoperative staging of breast cancer S. Nakano 1 , K. Yorozuya 1 , K. Fujii 1 , M. Yoshida 1 , T. Fukutomi 1 , J. Kimura 1 , Y. Ohshima 1 , T. Ishiguchi 1 , O. Arai 2 ; 1 Aichi-gun/JP, 2 Kashiwa/JP (snakano1@aichi-med-u.ac.jp) Purpose: We recently developed a real-time virtual sonography (RVS) that can simultaneously display images of the same obtained by sonography and magnetic resonance imaging (MRI) using magnetic navigation. The purpose of this study was to evaluate the role of RVS in the management of enhancing lesions visualised with MRI. Methods and Materials: Between June 2006 and April 2007, 65 patients underwent MRI for staging of known breast cancer at our hospital. All patients were examined using mammography, sonography, MRI and RVS before surgical resection. Results were correlated with histopathologic findings. MRI was obtained on a 1.5 T imager, with the patient in the supine position using a flexible body surface coil. Detection rate was determined for main tumours and incidental enhancing lesions (IELs), with or without RVS. Results: Overall sensitivity for detecting main tumours was 85% (55/65) for mammography,91% (59/65) for sonography, 97% (63/65) for MRI and 98% (64/65) for RVS. IELs were found in 26% (17/65) of the patients. Of 23 IELs that were detected by MRI, 30% (7/23) of IELs could be identified on repeated sonography alone, but 83% (19/23) of them were identified using the RVS system (P = 0.001). The

Sexual development in utero: testicular descent on prenatal MRI U. Nemec, S.F. Nemec, M. Weber, P.C. Brugger, G. Kasprian, D. Prayer; Vienna/AT Purpose: Foetal gender determination including the evaluation of testicular descent is an important tool to assess normal male sexual development. In view of the increasing role of foetal magnetic resonance imaging (MRI) as an adjunct to US in prenatal diagnosis, we aim to demonstrate testicular descent in correlation to the gestational age on MRI in utero. Methods and Materials: After chart review, this retrospective study included foetal MRI scans of 202 male foetuses [17 to 39 gestational weeks (GW)] with normal anatomy or minor anomalies. On a 1.5 Tesla unit, multiplanar T2-weighted turbo-spin-echo sequences were applied to image the scrotal content. Relative frequencies of unilateral and bilateral testicular descent were calculated and correlated to the gestational age. Purpose: The ability to objectively critique and assess the image quality of mammograms obtained by each department is vital to assess the efficacy of the services provided by breast imaging units. As evidenced in the analysis of symptomatic breast images for a large national optimisation of mammographic examinations study, current methods of classification for symptomatic mammograms are largely subjective resulting in poor inter-and intra-departmental agreement on image quality. The subjectivity within current image quality criteria allows mammogra-RVS system was able to correctly project enhanced MRI information onto a body surface, as we checked sonography form images. Purpose: To determine the type and frequency of medicolegal claims at a Dutch breast cancer screening programme.

The study population consisted of all 80019 women who underwent screening mammography at a southern breast screening region of the Netherlands between January 1997 and July 2007 (301139 screens). We included all medicolegal claims that had been recorded at the central screening department within 3 years following screening mammography. During 2-year follow-up, we collected the biopsy results and surgery reports of all referred women. Two screening radiologists reviewed the screening mammograms of all screen-detected cancers (SDC) and interval cancers (IC) and determined whether the cancer had been missed at the previous screen (in case of SDC) or latest screen (in case of IC). The radiologists were blinded to each other's review; discrepant readings were followed by consensus reading. Results: Just 3 medicolegal claims had been reported, all of them related to financial compensation following a diagnosis of IC. The verdicts of these cases still have to be finalised. Excision biopsy had been performed in 10.7% (234/2183) of false-positive referrals. Review showed that 20.8% (261/1254) of SDCs had been missed at the previous screen and 23.6% (139/588) of ICs should have been detected at the latest screen. Conclusion: Medicolegal claims were very rare, although a substantial proportion of referred women with benign follow-up had been confronted with excision biopsy and more than 20% of SDCs and ICs had been missed at the previous screen or latest screen, respectively. Purpose: The influence of previous benign breast surgery on screening mammography accuracy is unknown. We determined whether the sensitivity of screening mammography and tumour characteristics are different for women with and without previous benign breast surgery.

We included a consecutive series of 317398 screening mammograms performed between 1997-2008. During 2-year follow-up, clinical data, breast imaging reports, biopsy results and breast surgery reports were collected of women with screen-detected or interval breast cancers. Screening sensitivity, tumour biology and tumour stages were compared between 168 women with breast cancer in the ipsilateral breast after prior benign surgery and 2039 women with breast cancer, but without previous ipsilateral, benign breast surgery.

The sensitivity of screening mammography was significantly lower for women with prior surgery (64.3% (108/168) versus 73.4% (1496/2039), p = 0.01). The concomitant increased interval cancer risk remained significant after logistic regression adjustment for age and breast density (OR = 1.5, 95%CI: 1.1-2.1).Comparing screen-detected cancers in women with and without prior breast surgery, no significant differences in oestrogen-receptor status (p = 0.56), mitotic activity (p = 0.17), proportions of large (T2+) tumors (p = 0.6) or lymph node positive tumors (p = 0.4) were found. For interval cancers, oestrogen-receptor status (p = 0.41), mitotic activity (p = 0.39) and proportions of large tumours (p = 0.9) and lymph node positive tumours (p = 0.5) were neither statistically significant. Conclusion: Sensitivity of screening mammography is significantly lower in women with previous benign breast surgery than without, but tumour characteristics are comparable. At screening women should be informed about the decreased sensitivity of screening mammography in post-surgical breasts.Conclusion: SPECT-CT improves localisation of parathyroid adenomas compared to planar imaging or US alone. Although US is not as sensitive as SPECT-CT, a combination of the two modalities reported by a single radiologist is the most accurate technique in the localisation of parathyroid adenomas. Ultrasound, sestamibi and MDCTA were reviewed and the results were compared with surgical and histopathologic findings. The operations were declared successful if the calcium and PTH levels were normalised after surgery. Results: MDCTA was performed in 18 patients. Scan range included the neck and upper mediastinum. In 9 of 18 (50%) patients ultrasound and sestamibi were positive, but there was discrepancy in accurate localising of adenoma (n = 7) or sestamibi revealed an adenoma in the upper mediastinum (n = 2). MDCTA identified and localised an adenoma in all cases. Surgical resection, which was performed in 7 patients, confirmed the imaging findings. In 7 of 18 patients (39%) adenoma was revealed by either ultrasound or sestamibi. MDCTA confirmed the presence of an adenoma in all patients, which was surgically confirmed. In 2 patients (11%) both ultrasound and sestamibi were negative. MDCTA revealed an adenoma in one patient, which was surgically confirmed. Conclusion: MDCTA is a useful imaging technique in identifying and accurate localising parathyroid adenoma in patients with primary hyperparathyroidism in which ultrasound and sestamibi are discrepant or negative.

Thyroid cysts: their ultrasound worrisome appearance after evacuation C. Crespo, M. Navarro, J. Calbo, L. Humanes, M. Castro, L. Pérez;

Purpose: To show the worrisome ultrasound characteristics that thyroid cyst could became after their liquid content will be evacuated. The lesion wrinkled and appeared with characteristics suspicious of malignancy.

We reviewed the ultrasound examination of ten patients who had thyroid lesions with ultrasound criteria of malignancy. Only in two cases, there was medical information about previous puncture of a thyroid cyst. In five cases, it was the patient who told the radiologist about this. Results: The studied thyroid nodules were solid and hypoechoic in ultrasound with irregular shape and ill defined margins. Three presented hyperechogenic foci and three showed a mixed pattern of vascularization. We performed follow-up ultrasound exam in all patients and biopsy in four of them. The pathologist found a benign lesion in all four cases. The review of the images of previous ultrasound exam (only disposable in six patients) showed characteristic features of thyroid cyst. Conclusion: All radiologists working in head and neck section knows the ultrasound appearance of a thyroid cyst. However, it is a habitual practice in clinical setting to evacuate a symptomatic thyroid cyst. In more instances the lesion is palpable and the procedure is done by the clinician or by the pathologist. The radiologist did not know this data and found a suspicious lesion in the gland. The knowledge of the lesion appearance at the initial ultrasound exam together with the information about the previous procedure could avoid an unnecessary biopsy and prevent the anxiety of the patient. Purpose: The correlation between a nodule stiffness and malignancy has been found by clinical palpation and demonstrated in in vitro studies. Strain elastography is able to provide stiffness information but the clinical routine use is suffering from limitations. Shear wave elastography (SWE) is becoming available and provides true quantitative measurement of stiffness with reduced variability between operators. The purpose of our study was to evaluate SWE in routine clinical practice.

phy departments to appear to achieve the high scoring requirements of only 3% inadequate images.

The European Quality Criteria for film-screen mammographic image quality, the proposed amended European criteria for digital mammography images and the breast screening quality criteria classification of images as inadequate, moderate, good and perfect were modified to remove all subjective criteria. These objective classifications of image quality were tested for inter-and intra-rater reliability by a panel of experts.Results: Inter-rater reliability (K > 0.701; p < 0.001) and agreement (Pearson's correlation ρ > 0.884; p < 0.01) by the evaluation panel was higher than using the original two methods. The intra-rater reliability was equally high (K > 0.7; p < 0.001) and agreement via Pearson's correlation: ρ > 0.844; p < 0.01. Conclusion: A method of scoring images by combining the most objective components of two major European scoring systems for images is suggested as tested in a large national study within the symptomatic breast services of Ireland. The removal of subjectivity from the scoring systems will remove all doubt regarding the achievement of these image quality goals.

Artefacts in direct digital full-field mammography Z.C. Milosevic, M. Nadrljanski; Belgrade/ RS (dr.m.nadrljanski@gmail.com) Purpose: The prospective analysis of artefacts in direct digital full-field mammography.

Between January 2009 and June 2010 screening or diagnostic mammography was performed in 4715 patients (Selenia, Hologic). The analysis of artefacts was based on following classification: patient-related artefacts (motion artefact, antiperspirant artefact, thin breast artifact), hardware-related artefacts (field inhomogeneity, detector-associated artefacts, collimator misalignment, underexposure, grid lines, grid misplacement, vibration artefact), and software processing artefacts ("breast-within-a-breast" artifact, vertical processing bars, loss of edge, high-density artifacts).

The most common artefacts were patient-related artefacts (antiperspirant artefact) followed by intermittent field inhomogeneity artefact and high-density artefact (more than 10% of the mammograms). Motion artefact, thin breast artefact, underexposure artefact, breast-within a-breast artefact and loss of edge artefact were noted in less than 1% of the mammograms. Conclusion: Artefacts can mask true abnormalities or create pseudolesions. Some artefacts in digital mammography are specific and, consecutively, should be identified and collected in order to avoid misinterpretation. Purpose: A comprehensive quantitative study of the three-dimensional foetal cortical development, its natural variability, and the correlates with specific pathologies, can aid early and accurate diagnosis of disease. It can foster the understanding of the development process, and can ultimately lead to effective treatment strategies.In this work, we evaluate a quantitative model of the development process derived from high resolution in utero magnetic resonance imaging.

We model the foetal cortical development from a data set containing 29 foetal brains with gestational ages from weeks 18 to 30. After a coarse alignment and an expert annotation of the cortical surface, the temporal development is modeled by a group-wise spatio-temporal registration. We evaluated the accuracy of the model in a leave on out cross validation. We measured the reconstruction error of individual cases, and the automatic estimate of the gestational age derived from a cortical shape.

The proposed spatio-temporal group-wise registration improves the error to 1.48. The correlation coefficient between nominal gestational age, and automatically estimated age increases during the iterative spatio-temporal modelling process and reaches r = 0.957. Conclusion: Results on 29 foetal brains with gestational ages ranging from 18 to 30 gestational weeks indicate that the group-wise spatio-temporal registration has advantages over registration that does not take time into account. The accuracy of the gestational age estimate indicates that the time specificity of the resulting model allows for a precise comparison of individual cases to the population characteristics.

Functional resting-state measurements of the foetal brain V. Schöpf, G. Kasprian, C. Mitter, P.C. Brugger, D. Prayer; Vienna/AT (veronika.schoepf@meduniwien.ac.at) Purpose: FMRI has already been used to study foetal brain activity in utero using different kinds of stimulation methods. The underlying neuronal mechanisms strongly connected to gestational weeks of these activity networks have not been described yet. In this study, we used functional connectivity analyses allowing the characterisation of interregional neural interactions during spontaneous activity during rest to study those underlying mechanisms, aiming at specifying the differences of functionally connected cortical networks over foetal brain development stages. Furthermore, we aimed at describing subplate specific activity patterns and interactions with subcortical structures for different developmental stages. Methods and Materials: Functional images of 10 foetuses with morphologically normal brain development, aged from gestational week 18-39, were acquired on a 1.5 T Philips Intera MR scanner using single-shot gradient-recalled echo-planar imaging (10-15 axial slices (5 mm thickness), matrix size 144 x 144, FOV 250 x 250 mm, TE/TR 50/1000 ms, flip angle 90°). Data sets were preprocessed including motion correction and brain extraction. Postprocessing was performed using functional connectivity analysis which was implemented in Matlab. Results: Bilateral resting-state activity could be proved in cortical regions and in the subplate of the frontal, occipital, and temporal. Furthermore, functional correlations of thalamic structures to temporal and frontal subplate regions were obtained. Conclusion: Resting-state measurements in utero provide consistent results and can be analysed by the means of functional connectivity analysis. Our data reflect the high metabolic and functional activity of the subplate between 20 and 30 GW. Thus, the regional distribution of foetal brain function can be mapped in normal conditions and monitored in developmental abnormalities. Purpose:Arrhythmogenic right ventricular dysplasia/cardiomyopathy (ARVD/C) is an important cause of sudden cardiac death in young athletes. Cardiac magnetic resonance (CMR) evaluation of athletes to detect possible ARVD/C is complicated by overlapping features such as right ventricular (RV) volume increase. The recently revised ARVD/C diagnostic Task Force Criteria (TFC) now incorporate cut-off values for RV ejection fraction (EF) and RV end-diastolic volume (EDV) on CMR. We compared CMR ventricular volume, function and cut-off values to distinguish ARVD/C patients from athletes. Methods and Materials: CMR was performed in 33 ARVD/C patients diagnosed by the revised TFC (68% male), sex-and age-matched to 66 healthy non-athletes and 66 healthy athletes. Quantitative dimensions and volumes were compared and evaluated using existing (revised TFC) and new cut-off values. Results: RV EDV in ARVD/C patients overlaps considerably with athletes. However, the ratio of left ventricular (LV)/RV EDV, end-systolic volume (ESV) and EF of athletes is similar to that of healthy non-athletes, but significantly different from ARVD/C patients. Areas under the ROC-curve are 0.683 (RV EDV index), 0.837 (LV/RV EDV ratio) and 0.932 (RV EF), demonstrating significantly (p < 0.001) better performance of RV EF and LV/RV EDV ratio. Conclusion: If a wall motion abnormality is present, RV EF is suitable to distinguish ARVD/C from physiologic cardiac adaptation in athletes, whereas RV EDV index is not. The latter results in numerous false ARVD/C diagnoses in athletes. As an alternative we suggest using the LV/RV EDV ratio in athletes, which best represents the balance between left and right ventricles. Purpose: To evaluate normal morphological development of the hippocampal region during the foetal period using magnetic resonance imaging (MRI) Methods and Materials: MRIs of 63 foetuses without intracranial pathology were reviewed independently by two radiologists who did not know the gestation ages. Three MRIs were performed post mortem and 60 in utero. Progress of the hippocampal inversion was analysed in the coronal slices and the left and right sides were compared in every case.

The foetuses in the post mortem examinations were at gestation weeks (GW) 17-18 and in the in utero examinations at GW 19-36. The hippocampal sulcus was open, bi-or unilaterally, in 39 foetuses. The oldest of them was at GW 32. The sulcus was closed at GW 21 at the earliest, unilaterally, and always from GW 33 onwards bilaterally. In 26/63 foetuses (41%), the hippocampal development was asymmetric and in 22 foetuses, the right side had developed faster. The non-ovoid hippocampal shape was interpreted to be an earlier form than the ovoid shape. A shallow collateral sulcus was found at GW 17. A deep collateral sulcus was visible earliest at GW 26 unilaterally but in all foetuses from GW 31 onwards it was seen bilaterally. Orientation of the collateral sulcus was not related to the gestational age. Purpose: The degree of liver enhancement in the hepatobiliary (HB) phase is known to be lower in patients with chronic liver dysfunction. The aim of our study was to evaluate the value of HB phase of gadoxetic acid-enhanced MRI in the diagnosis of hepatocellular carcinoma (HCC) in patients with impaired liver function.

Fifty-eight patients with 75 HCCs underwent gadoxetic acid-enhanced MR examination. Thirty-eight patients (group 1) had normal liver function, and the other 20 patients (group 2) had impaired liver function. Two radiologists reviewed two sets of MR images of two groups: set 1, unenhanced and gadoxetic acid-enhanced dynamic images; set 2, additional HB phase and unenhanced and gadoxetic acid-enhanced dynamic images. Diagnostic accuracy was compared using the area under the ROC curve (Az). Lesion-to-liver contrastto-noise ratio (CNR) was calculated in equilibrium phase (5-minute delayed) and HB phase. Results: Az value increased with the addition of HB phase in both group 1 (0.962 to 0.998, p = 0.365) and group 2 (0.875 to 0.929, p = 0.202), but did not reach statistical significance. CNR on HB phase (-126.8 ± 66.1) was significantly lower than that on equilibrium phase (-87.4 ± 51.4, p < 0.001) regardless of patients' liver function. However, CNR drop on HB phase of group 2 (26.3 ± 22.0) was significantly lesser than that of group 1 (39.6 ± 27.5, p = 0.026). Conclusion: Hepatobiliary phase images of gadoxetic acid-enhanced MR imaging may improve diagnosis of HCC although patient had impaired liver function.

Optimisation of the dynamic, Gd-EOB-DTPA-enhanced MRI of the liver: the influence of the injection rate C. Schmid-Tannwald, K.A. Herrmann, M.F. Reiser, C.J. Zech; Munich/ DE (christine.schmid-tannwald@med.uni-muenchen.de) Purpose: To evaluate the influence of the injection rate on the enhancement of liver vasculature in Gd-EOB-DTPA-enhanced liver MRI. Methods and Materials: Two groups of patients who were referred to a Gd-EOB-DTPA-enhanced liver MRI (1.5 T) were retrospectively evaluated. In both groups a Care Bolus technique to determine the start of the arterial-dominant phase and a power injector were used. The injection rate of the contrast agent was 2 ml/sec (group 1: n = 40) and 1 ml/sec (group 2: n = 40), respectively. Signal intensities (SI) in the unenhanced, early dynamic and hepato-biliary phase of T1w 3D GRE sequences were measured in normal liver parenchyma, in the aorta and in the intrahepatic portal vein (PV). Image noise was measured. Signal-to-noise-ratio (SNR), percentage enhancement (PE) and SNR increase were calculated. (wieland.sommer@med.uni-muenchen.de) Purpose: Dynamic contrast-enhanced MRI (DCE-MRI) of the liver is able to quantify perfusion parameters. The aim of this study was to develop and validate a perfusion model for liver-specific contrast agent (Gd-EOB-DTPA) which is able to quantify the uptake into the hepatocytes.

In order to determine the intracellular uptake rate (min -1 ), a dual-inlet two-compartment uptake model was designed, by modifying a singleinlet uptake model. DCE-MRI data were acquired at 3 T (Siemens Verio) in 25 patients using the 3D gradient-echo sequence TWIST (2.1sec temporal-resolution, 5 min acquisition). Gd-EOB-DTPA (Primovist, Bayer) was injected at 2 ml/s. Data were postprocessed using the software PMI 0.4. ROIs were drawn on parametric maps of descriptive indices in the abdominal aorta and the portal vein to measure arterial and venous input functions. In each patient, 5 circular ROIs were drawn in normal liver tissue.

The one-compartment model did not provide a good fit to any of the data, and produced unphysical values for the total blood flow and the arterial flow fraction. The uptake model fitted all data accurately, and provided values in the expected range for all known parameters (bloodflow 110±35 ml/min; arterial flow fraction: 29±14%; MTT 12±3.8sec). After correction for the difference in relaxivity between intra-and extracellular spaces, the average value for the uptake rate was 1.7±0.7 x 10 -2 min -1 .

These data indicate that the proposed model is able to determine Gd-EOB-DTPA kinetics in the liver. The method may present a new and practical paradigm in functional liver MRI, producing quantitative measures of both perfusion and hepatobiliary function. Purpose: To correlate commonly used liver function tests and Gd-EOB-DTPA enhanced MR imaging and to define laboratory tests that have potential to predict hepatobiliary phase imaging characteristics. Methods and Materials: 121 patients who underwent liver MR imaging with Gd-EOB-DTPA were included in this study. Non-enhanced, dynamic phases, and two phases of hepatobiliary image sets (20 and 60 minutes) were obtained using a T1-weighed 3D gradient-echo sequence, The liver signal-to-noise ratio (SNR) Purpose: To quantify the accuracy of material attenuation data from synthesised monochromatic images generated by dual energy CT (DECT).

A DECT scan (Discovery HD, GE Healthcare) was performed on two patients and a technical phantom with relevant material inserts. From the monochromatic reconstruction images, attenuation curves were obtained for different materials of interest, i.e. plastics (polyethylene, nylon, polycarbonate, acryl and polystyrene) and biological tissues (blood, adipose tissue and cortical bone). The attenuation curves were decomposed into a Compton (C) and photo-electric (PE) component by least squares fitting of the Klein-Nishina function and the inverse cubic function of the energy. Compton and PE decomposition coefficients and their ratio (PE/C), reflecting the spectral material properties, were used to investigate the accuracy of the obtained attenuation curves. Results: A sound congruence (variation < 1%) was found between Compton coefficients computed from measured and tabulated attenuation curves (NIST) of the various plastics. For biological materials, the congruence of the Compton coefficients was observed to be lower (3% to 4%). However, for these materials, the measured PE/C ratios agree well (< 1.5%) with the PE/C ratios obtained from the tabulated attenuation curves, indicating that the difference in Compton coefficients is due to a difference in density rather than a variation in spectral attenuation properties. Conclusion: Based on comparison of Compton and PE decomposition coefficients, we conclude that obtained attenuation data from DECT accurately match the tabulated attenuation data. Our results indicate that the decomposition coefficients are valuable metrics for tissue differentiation and characterisation based on DECT data.

Comparisons between two tube voltage combinations (80/Sn140 kV, 100/ Sn140 kV) in dual energy CT imaging Y. Ho 1 , S. Yu 1 , G. Lo 1 , W. Lam 1 , S. Liu 1 , W. Wong 1 , L. Johnson 2 ; 1 Hong Kong/CN, 2 Shanghai/CN (yickwingho@gmail.com)Purpose: Two kV settings, 80/Sn140 kV (80/140) and 100/Sn140 kV (100/140), available for dual energy imaging in dual source CT (DSCT), were compared in the aspects of image quality and radiation dose.

A solid water phantom with 11 inserts of physical densities (0.30-1.82 g/cm 3 ) was scanned using a DSCT at 80/140 and 100/140 kV settings, utilising current modulation and keeping other scanning parameters the same and CTDI was recorded. Monochromatic images (40-190 keV) and basis-material iodine images were reconstructed. Signal-to-noise ratio (SNR) of inserts in monochromatic images at different energies were measured as the ratio of the mean CT number to standard deviation. Contrast ratio (CR) was measured as the ratio of the mean CT number in the insert to that of background. The measured iodine concentration was compared statistically between the two settings using paired t-test. Results: Using 80/140 setting, the average SNR of inserts in monochromatic images were higher (p < 0.001) with CTDI being 18.9 % lower than those of 100/140. It demonstrated a higher CR in 40-90 keV (largest difference = 13%), and lower CR in 100-190 keV (largest difference = 6%) than the 100/140 setting in monochromatic images. No statistically significant difference was found in iodine concentration measurement (p = 0.126) between the two settings. Conclusion: 80/140 was found to have better imaging quality in terms of better SNR and higher CR in low keV range with lower radiation dose than 100/140, without significant difference in quantitative measurements such as iodine concentration.

Free-breathing dynamic contrast-enhanced (DCE-MRI) of the abdomen at 3.0 T using radial k-space sampling and k-space-weighted image contrast (KWIC) reconstruction: preliminary study K. Kim 1 , J. Lee 1 , T. Block 2 , J. Han 1 , B. Choi 1 ; 1 Seoul/KR, 2 Erlangen/ DE (medimash@radiol.snu.ac.kr) Purpose: To evaluate the feasibility of free-breathing DCE-MRI of the abdomen and thorax at 3.0 T using radial k-space sampling and KWIC reconstruction. Methods and Materials: Sixteen patients with malignant tumours (hepatocellular carcinoma, n = 5; liver metastasis, n = 3; pancreas cancer, n = 4; lung cancer, n = 4) underwent DCE-MRI at 3.0 T machine (Tim Trio, Siemens). Among them, 6 patients were scanned at two different time points. T1 map was generated from a triple flip-angle (α = 2°, 8°, 15°)-based free-breathing radial T1 mapping procedure. Free-breathing dynamic scan was conducted using a modified 3D FLASH-VIBE sequence (TR/TE, 3.3/1.3 ms; flip angle, 11°; matrix size, 256 x 256) with Gd-DTPA bolus administration. The pixel-wise perfusion maps of the volume transfer coefficient (K-trans) and extraellular extravascular volme fraction (Ve) were generated using a software (4D Perfusion, Siemens). The overall image quality regarding the lesion conspicuity, artefact, and the sharpness of abdominal solid organ of the DCE-MRI were rated on a 4-point scale (1, poor, to 4, excellent quality) comparing the breath-hold T1-weighting gradient-echo sequence scanned immediately after the dynamic scan.

The proposed free-breathing DCE-MRI technique provided excellent motion-correction sum-up images with the reasonably high spatial resolution (1.48 x 1.48 mm) and temporal resolution (4.1 sec) that allowed for high frame-rate arterial input function sampling and excellent pixel-wise perfusion maps. Overall image quality score of DCE-MRI was 3.68 ± 0.47 with negligible radial artefact and motion artefact. Conclusion: Free-breathing DCE-MRI with radial k-space sampling and KWIC reconstruction is feasible for perfusion analysis of tumour in the abdomen and thorax. It can overcome respiratory motion while providing high spatial and temporal resolutions.

Value of MRI for the diagnosis of benign liver tumours: correlation of updated MRI criteria with histo-molecular analysis of resected specimens L. Baranes 1 , J. Purpose: To investigate the radiation dose and image quality in high-pitch chest and abdominal CT compared to standard-pitch CT. Methods and Materials: 131 patients (chest n = 100, abdomen n = 85) scanned on a DSCT with high-pitch mode (pitch 3; group 1) were retrospectively included and compared to 148 patients (chest n = 111, abdomen n = 98) scanned on a DSCT with standard-pitch mode (pitch 1; group 2). The groups matched for age, gender and BMI. The patients were subdivided into 4 subgroups (BMI ≤ 25 and > 25, chest and abdomen scan). Dose length product (DLP), image quality (image noise and scoring) were compared (Kruskal-Wallis test). DLP was correlated to BMI and noise (Pearson correlation). Interobserver agreement was analysed (Kappa-analysis). Results: DLP of group 1 was significant lower compared to group 2 (chest and abdomen: group 1 mean = 367 ± 92 mGy cm; group 2, 577 ± 286 mGy cm; p < 0.0001).In group 1 image noise was significantly higher for abdomen scans when compared to group 2 (group 1: mean = 26.2 ± 5.7 HU; group 2: 18.2 ± 4.3 HU; p = 0.0001), but not for chest scans (group 1: mean = 17.3 ± 3 HU; group 2: 17.7 ± 3.6 HU; p = 0.37). The highest correlation (r = 0.64) comparing noise to BMI was found for abdomen scans of group 1, reflecting stronger dependency of noise on BMI in exams with reduced dose. The overall interobserver agreement was good for group 1 (k = 0.75) and excellent for group 2 (k = 0.92). Conclusion: A significant dose reduction is possible using hp-DSCT. Diagnostic image quality is preserved with slightly higher image noise for abdominal CT scans. were included in the study. All examinations were performed on a 256-slice multidetector CT (iCT, Philips Healthcare) from chest to pelvis. All patients had a previous examination several months before, using a standard dose protocol (protocol A) adapted to body mass index (BMI). Patients were imaged with a low dose protocol (protocol B) and an iterative reconstruction algorithm (iDOSE, Philips Healthcare) was applied to the images. Image noise was measured in different body regions of standard dose, low dose and low dose images reconstructed with the iDOSE algorithm. Subjective image quality was addressed. Statistical analysis was performed. Results: CT examinations with low dose protocol (protocol B) showed an average dose reduction of approximately 50% compared to the standard protocol (protocol A). Image noise was 21-55% higher in low dose images. By applying the iDOSE reconstruction algorithm image noise was lowered to the level of standard dose CT images (p < 0.01). Subjective image quality was equal in the low dose images reconstructed with iDOSE compared to the standard dose CT images.

The iDOSE reconstruction algorithm allows dose reduction of approximately 50% in multi-detector computer tomography examinations retaining an image quality equivalent of standard dose CT images without loss of information. To evaluate radiation exposure of 64-row whole-body CT in patients with major trauma using ASIR. Methods and Materials: ASIR is a CT raw data processing method that compensates image noise more effectively than conventional filtered back projection (FBP), consequently allowing for reduction of exposure. Dose reports of 64-row whole-body trauma CT (head to symphysis) of 105 studies (72 males, 33 females) with FBP (LightSpeed VCT XT, GE Healthcare, Waukesha, WI) were retrospectively compared to 81 studies (52 males, 29 females) with ASIR (Discovery 750HD, GE Healthcare). The scan parameters for both examinations were identical but the noise index (NI) was increased from 5.2 to 6.0 (head) and 29.0 to 46.0 (chest, abdomen and pelvis) with an ASIR level of 30% (head) and 50% (chest, abdomen and pelvis). Scan length, CTDI, and DLP were tested for statistical difference with the Mann-Whitney-U test.

Conclusion: SNR and image quality of ULD BPCT can be improved to a level similar to LD when using HYPR-LR algorithm without distorting attenuation measurements. This can be used to substantially reduce radiation dose.

CT perfusion reproducibility assessment with a moving perfusion robot A. Hirschenbein 1 , S.Y. Mahgerefteh 2 , J. Sosna 2 ; 1 Ra'anana/IL, 2 Jerusalem/IL (avivhir@gmail.com)Purpose: Dynamic contrast-enhanced CT is a potentially useful noninvasive technique for assessing tissue perfusion. Our aim was to assess the reproducibility of CT perfusion parameters derived with a perfusion robot.

We developed a unique moving CT perfusion robot containing 12 tubes placed 30 degrees apart, with varying concentrations of iodine (1.5 mgI/ml-60 mgI/ml corresponding to 15-1300 HU, respectively). The main robot components are motor, reduction gearbox, test-tube wheel, x-ray sensor, position feedback, construction frame, and controller. To simulate repeated time density curves performed at different times, 30 experiments were performed on 5 different days on a 64-slice MDCT using parameters: 140 kVp, 250 mAs, and 5 mm slice thickness. Cycle times ranged from 1.5 to 5 seconds for 3 minutes. ROI data were collected, plotted, and analysed for trends and reproducibility using a regression fit model.

The CT perfusion robot produced constant movements for predetermined locations. Time density curves for different experiments had a perfect fit (r2 = 0.995). However, the changes in HU from the anticipated values ranged from 5% for the 60 mgI/ml iodine concentrations (1283 HU to 1343 HU for anticipated 1300 HU) to 200% for the 1.5 mgI/ml iodine (16 HU to 50.7 HU for the anticipated 15 HU). Changes were more prominent for physiologic density ranges up to 400 HU and decreased for the higher concentrations. Conclusion: CT perfusion exam data simulated with a moving robot are not stable, with especially marked differences at the physiological range of iodine density in blood vessels.

Reducing abdominal CT radiation dose with iterative reconstruction in image space X. Hu, W. Lin, G. Shen; Hangzhou/CN (lakewest@gmail.com)Purpose: To assess radiation dose reduction and image quality for abdominal computed tomography (CT) examination with iterative reconstruction in image space (IRIS). Methods and Materials: Low-dose abdominal CT examinations with IRIS were performed in 40 patients who had previously undergone routine-dose CT. Phantom also scanned with the same parameters. The image noise, high-contrast resolution and CT radiation dose were compared with and without IRIS. The images quality was graded on a scale of 0 (non-visualisation), 1 (poor), 2 (average), 3 (good), 4 (excellent). Signal-to-noise ratio (SNR) was compared with IRIS and without IRIS. Student t-test was used to determine differences of abdominal CT images with and without IRIS in numerical scores of images quality, SNR and radiation dose.