key: cord-0010769-ieededv4 authors: nan title: ECR 2020 Book of Abstracts: Vienna, Austria. 15 March 2020 date: 2020-05-05 journal: Insights Imaging DOI: 10.1186/s13244-020-00851-0 sha: 0b375c189c5935072929c76fb1cba68edd629d77 doc_id: 10769 cord_uid: ieededv4 nan Influence of the minimum b-value on prostate cancer assessment using conventional DWI and DKI models Radiomic analysis in renal cell carcinoma: the impact of computed tomography vascular phase on parameter quantification Acoustic radiation force impulse elastography as an alternative to diffusion-weighted imaging for the characterisation of breast lesions Aichi/JP, 2 Nagakute/JP (snakano1@aichi-med-u.ac.jp) Reduction of pressure sensor artifacts in the simultaneous acquisition of wide-angle breast tomosynthesis and mechanical imaging Classification of lung opacities in supine chest radiographs: artificial intelligence approach (external algorithm evaluation) versus a radiological assessment The eye-tracking system as an assessment tool for mammography positioning In-vivo measurements of ADC of invasive ductal breast cancer: a multicentric study to assess the quality of DWI protocols in breast MRI at 1.5T Comparison of prone digital breast tomosynthesis-guided vacuumassisted biopsy (DBT-guided VAB) and prone atereotactic-guided vacuum-assisted biopsy (S-guided VAB) Myocardial perfusion imaging and infarct characterisation: diagnosis and prognosis RPS 203-1 Improved sensitivity in diagnosing myocardial infarction using frequency selective nonlinear blending in computed tomography ; Tübingen/ DE (eva.stock@med.uni-tuebingen.de) Purpose: Sequential strategy including FFRCT plus stress-CTP impacts on the management of patients with stable chest pain: the stress-CTP RIPCORD study Left atrial dysfunction in children and adolescents with severe obesity: a cardiac magnetic resonance imaging myocardial strain study High signal intensity in the globus pallidus (GP) and dentate nucleus (DN) on unenhanced T1-weighted magnetic resonance images: an assessment of two macrocyclic gadolinium-based agents The absence of T1 hyperintensity in the brain of high-risk iron-loaded thalassemia patients after multiple administrations of high-dose gadobutrol No changes in T1 relaxometry after a mean of eleven administrations of gadobutrol Impact of the novel contrast agent gadopiclenol on decision making in patients with brain metastases The value of 4D-MR angiography at 3T compared to DSA for the follow-up of treated dural arteriovenous fistulas Comparison of first-order radiomic parameters to the mean ADC for the prediction of clinically significant cancer from prostate MRI Multi-parametric magnetic resonance imaging of prostate cancer: correlation between Ktrans, a Gleason score, and a PI-RADS score Correcting versus resolving respiratory motion in free-breathing wholeheart MRA: a comparison in patients with thoracic aortic disease Value of T2 mapping in patients with chronic thromboembolic pulmonary hypertension (CTEPH) before and after balloon pulmonary angioplasty (BPA) A prospective randomised non-inferiority study of low iodine dual-energy coronary CT-angiography CTiR 3-9 Dose reference levels during fluoroscopically-guided procedures performed using mobile X-ray systems in operating rooms The neurodevelopmental effect of intracranial haemorrhage observed in hypoxic-ischaemic brain injury in hypothermia-treated asphyxiated neonates: an MRI study Quantitative paediatric MRI going clinical: comprehensive brain assessment from a single sequence based on quantitative brain maturation atlases The role of imaging in hepatocellular carcinoma (HCC) management MyT3 4-4 Evaluation of different keV-settings in dual-energy CT angiography of the siphon of internal carotid artery using noise-optimised virtual monoenergetic imaging ; Shanghai/ CN (kdidjs_fu@163.com) Purpose: Results: Application of FLAIR vascular hyperintensity-DWI mismatch in ischaemic stroke, depending on semi-quantitative DWI-Alberta stroke programme early CT score An apparent diffusion coefficient of different areas of the brain in growthrestricted foetuses MyT3 5-8 Hand extensor compartments: how to study them and is it always their fault? MyT3 5-10 Accuracy of volumetric trabecular bone mineral density assessment using dual-source dual-energy CT: a prospective phantom study and comparison with quantitative CT Use of DISCOGEL® to treat cervical and lumbar discal bulging: results and consideration in our monocentric experience The early effects of a mammography screening program on advanced breast cancer incidence in the Friuli Venezia Giulia Italian region Delayed breast cancer diagnoses in women recalled at screening mammography: trends in the proportions and lengths of delay over two decades of screening A radiomics-based model to identify the aetiology of liver cirrhosis using gadoxetic acid-enhanced MRI Using myocardial extracellular volume fraction to differentiate healthy from cardiomyopathic myocardium using dual-source dual-energy CT Interobserver variability in the evaluation of magnetic resonance elastography in patients with fibrotic liver disease Can T1-mapping serve as an alternative to MR-elastography in the staging of liver fibrosis? Diagnostic performance of US-guided core-needle biopsy versus fineneedle aspiration for diagnosing thyroid neoplasm as a first-line biopsy method: a propensity score matching study The role of multi-detector computed tomography angiography for preprocedural planning and radiation dose reduction in bronchial artery embolisation A preclinical endogenous rat HCC model system for the prospective evaluation of imaging-derived biomarkers in interventional tumour therapy Non-invasive assessment of liver cirrhosis with multiphasic dual energy CT using iodine maps: correlation with model of end-stage liver disease score The influence of AI-based computer-aided diagnosis systems on diagnosis confidence in medical experts with different levels of experience Utility of dynamic oesophagogram in the grading of achalasic patients: comparison with static x-ray barium swallow and correlation with clinical subtypes The predictive power of MRI in identifying the pathological grading of meningiomas The role of sonosalpingography (SSG) using normal saline in female subfertility: diagnostic or therapeutic? ; Moradabad/IN (rajulrst@yahoo.co.in) Purpose: Results: Limitations: Ethics committee approval: Funding: Author Disclosures: The accuracy of the characterisation of adnexal masses with noncontrast pelvic MRI: an 11-year experience from a tertiary referral centre in the United Kingdom Ş Contrast-enhanced spectral mammography with a compact synchrotron source Orientation-resolved means of diffusivities and transverse relaxation times in heterogeneous brain tissue The presence of globally-decreased perfusion on post-reduction MRI does not reliably predict proximal femoral growth disturbance at midterm follow-up in developmental dysplasia of the hip Single fast high-resolution 3D T1 VIBE FS MR sequence versus CT in the evaluation of tarsal coalition in children and adolescents: a reliable radiation-free alternative? Relative enhancement of the liver during hepatobiliary phase after Gd-BOPTA administration: correlation with bilirubin levels and renal function Role of contrast-enhanced imaging in the management of lesions of uncertain malignant potential (B3) The predictive role of the lumbar disc FA (fractional anisotropy) map in diffusion tensor imaging (DTI) to select patients with low back pain who may benefit from intradiscal oxygen-ozone injections Author Disclosures: Radiomics analysis of enhancing residual tumours better predicts survival in post-surgery MRI patients with brain glioblastomas Between and within rater agreement in white matter hyperintensity segmentation from manual rating and a supervised automated classifier (FSL-BIANCA) The added diagnostic value of a model-based reconstruction algorithm in detecting acute trauma-related lesions in brain CT examinations in an emergency setting Distinguishing benign from malignant focal liver lesions (FLL) using the apparent diffusion coefficient (ADC): impact of post-processing on the diagnostic accuracy? Improved performance of orthopaedic metal artefact reduction in virtual mono-energetic spectral images compared to conventional polyenergetic images: a total hip arthroplasty phantom study Quantification of metal artefact reduction of various MRI techniques in 10 common total hip arthroplasty implants: when less is more! Machine learning-based assistance to context-sensitively suggest ASPECT score during the reporting of neuroradiological emergencies Validation of a high precision semantic search tool using a curated dataset containing related and unrelated reports of clinically relevant search terms Natural language processing enables a correlation of clinical information with positive findings in low-dose computed tomography in patients with suspected urolithiasis Imaging of the left atrial appendage prior to occluder device placement: introduction of a new single-contrast bolus dual-phase protocol adapted to spectral-detector CT The small-size detail detection performance of digital breast tomosynthesis, synthetic 2D, and conventional full-field digital mammography images for different mammography systems: a multicentre study Image quality evaluation of a new high-resolution 3-D digital breast tomographic (DBT) imaging system The incidence of penetrating aortic ulcer (PAU) as a cause for nonaneurysmal rupture of the abdominal aorta: a single-centre experience Vascular hyperintensities on a post-contrast 3D fast-spin-echo T1weighted sequence: a sign of poor collateral pathways in sickle-cell disease cerebral vasculopathy? The arterial remodelling ratio and normalised wall index of the middle cerebral artery are associated with leptomeningeal collateral status: a study of high-resolution vessel wall imaging Quantification of carotid intraplaque haemorrhage: a comparison with manual segmentation and semi-automatic segmentation on magnetisation-prepared rapid acquisition with gradient-echo sequence Iodine accumulation in the liver in patients treated with amiodarone can be unmasked using material decomposition from multiphase spectraldetector CT Association of retrospective peer review and positive predictive value of magnetic resonance imaging-guided vacuum-assisted needle biopsies of breast Calculating the cut-off value of the damping index using the ROC curve to identify Child-Pugh C patients First national teleradiology pilot project in the Saudi Ministry of Health The safety of a new stent design during magnetic particle imaging and magnetic resonance imaging The effectiveness of dual-layer spectral detector CT for reducing the amounts of contrast agents and injection flow rate The validation of iodine contrast flow velocity quantification from timeresolved CT-angiography in a flow phantom A comparison between routine and targeted CT-based radiomics classification models for predicting malignant pulmonary nodules Is there any news in CT protocols for renal masses and stones evaluation? RPS 1407a-1 Substantial radiation dose reduction with consistent image quality using a novel low-dose stone composition protocol The classification of pulmonary nodules by Lung-RADS 1.1: a randomised prospective analysis of four ultra-low dose CT protocols in a lung cancer screening trial The fight between CAD-RADS and the calcium score: who is the winner? Non-contrast-enhanced magnetic resonance imaging for the visualisation and quantification of an endovascular aortic prosthesis, their endoleaks, and aneurysm sacs at 1.5T The effect of different thresholds for CT-perfusion volumetric analysis on estimated ischaemic core and penumbral volumes Virtual monoenergetic dual-energy CT reconstructions at 80 keV are optimal for early stroke detection The diagnostic efficacy of whole-body low-dose CT in the staging of patients with multiple myeloma compared to whole-body magnetic resonance imaging A novel computed tomography-based scoring system of intra-articular mineralisation of the knee: BUCKS (Boston University calcium knee score) Application of the dual-layer spectral detector CT on CTPA with a low dose of iodine contrast agent and low injection rate Purpose: It is known from the literature that anxiety correlates with reduced interhemispheric connectivity on resting-state functional connectivity. The aim of this study is to verify this hypothesis with the connectometry technique. We recruited 20 healthy subjects (13 males, 7 female; age 25-80). Each patient on the same day performed the Italian version of the trait scale of the state-trait anxiety inventory (STAI-I) for the assessment of a situation independent general condition of anxiety and then underwent an MR examination on a 3T scanner (Vantage Titan 3T -Canon) with a 32 channel head coil. The MR protocol included a 12 direction DW sequence for the connectometry analysis. The connectometry analysis was performed using a linear regression model to consider the effects of the STAI-I score, choosing three different T-score thresholds (T-threshold) values (1, 2, and 3). Results were considered statistically valid for p-value adjusted for a false discovery rate (p-FDR)<0.05. The network topology analysis measured the influence of the STAI-I score on brain networks. The connectometry analysis identified several white matter bundles negatively correlated to STAI-I score for T-threshold=3, in particular at the level of corpus callosum (48% of the total amount of tracts identified). Several properties of the network influenced the network topology analysis. Purpose: Magnetic resonance imaging has great relevance in CNS pathology, mainly thanks to the multiple, excellent contrasts between tissues. Among its limitations stand long scan times and sensitivity to patients' motion. The present work compared conventional MR sequences with three-dimensional SSFP MRfingerprinting (MRF) with spiral projection k-space trajectories. Through such technique, acquisition of quantitative data of T1, T2, and proton density allows for the post-hoc synthesis of conventional qualitative contrasts with reduced motion sensitivity. The aim of this study was to compare MRF with conventional MRI in examining normal and pathological brain structures. The signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated on one phantom and two healthy volunteers. Assessment of absolute and relative contrast between grey and white matter, and qualitative evaluation of both normal encephalic structures and pathological findings representation, was conducted in 10 patients with different diseases, part of a protocol including 250 patients. Results: SNR and CNR were comparable between MRF and conventional imaging (p>0.05 for all comparisons, except higher SNR of conventional FLAIR sequences in white matter, p=0.016). In MRF, absolute contrast was higher in T1-weighted (p=0.01) but lower in T2-weighted and FLAIR images. Relative contrast was higher in T1-weighted (p=0.0003) and comparable in T2-weighted and FLAIR images. Although not superior to conventional imaging in anatomical detail, MRF alone provided sequences with high diagnostic confidence in 7/10 patients, while the others needed further MR scans to reach conclusive diagnoses. Conclusion: MRF is a promising tool for MRI in that it may preserve a good diagnostic accuracy while reducing the detriment of a long scan time. Limitations: The small sample size. Purpose: To evaluate the role of multi b-value synthetic diffusion-weighted imaging in the evaluation of stroke. Methods and materials: 52 patients who presented with symptoms suggestive of stroke with ages of 39-81 years (mean age 49.5+/-10 years; M: F ratio 2:3) were evaluated by a dedicated stroke protocol and supplemented by an additional 2-minute synthetic magic DWI multi b-value protocol after a waiver from IRB. The parameters of synthetic DWI sequence are as follows: multi b-value DWI acquisition with b-values of 0-1,000 and synthesis of b-values of 1,500 and 2,000, with the maximum b-value not exceeding 2,500. The sensitivity of the detection of infarcts on conventional DWI was compared to MAGIC DWI. Results: Of 52 patients, 31 had multiple acute lacunar infarcts in the MCA territory and MCA and PCA watershed zones, 21 in the infratentorial region including the brain stem and cerebellum, and 6 with a sensitivity of 92% and specificity of 97% with conventional DWI. Using the MAGIC DWI with the synthesis of multiple b-values between 1,000 to 2,500, the sensitivity and specificity significantly improved to 97% and 99%, with an NPV of 99%. Conclusion: Multi B-value ultra-fast synthetic MAGIC DWI, which is a synthetic diffusion technique, is a powerful adjunct tool to the conventional stroke protocol offering the user a retrospective synthetic slider to generate multi b-value data from 100-2,500, increasing the sensitivity of stroke detection with no time or SNR penalties. Purpose: Quantitative assessment of myelination is an important clinical biomarker in treatment and follow-up. Myelin can be measured using synthetic MRI; the measurement of the R1 and R2 relaxation and proton density PD, in conjunction with a myelin model, can provide myelin partial volume maps for the entire brain. Recently, a 3D acquisition method was developed for highresolution, isotropic synthetic MRI. The purpose of this work was to compare myelin detection using 2D and 3D. Methods and materials: The 3D QALAS sequence is a segmented spoiled gradient-echo sequence with 5 parallel acquisitions, interleaved with a T2 preparation and inversion pulse. The 2D MDME sequence (MAGiC) is a saturation recovery multi-slice TSE sequence with multi-echo read-out. A group of 12 volunteers was acquired two times with 3D QALAS and two times with MDME, both at 1.5T and 3T. Results: The mean myelin volume for the entire group was 183 mL and the mean brain volume was 1,300 mL (14.1%). A high correlation was found between volumes determined by QALAS and MDME. The Pearson correlation coefficient was 0.94 and the mean difference was 0±13 mL. The difference between measurement 1 and 2 was -2±10 mL at 1.5T and 1±13 mL at 3T for QALAS, whereas it was 0±4 mL at 1.5T and -3±4 mL at 3T for MDME. Conclusion: Myelin measurements using 3D QALAS provides very similar values of myelin and brain volumes in comparison to 2D MDME. The advantage of 3D QALAS is the ability to view the data in all orientations. Limitations: A small number of subjects, all of which were healthy. Ethics committee approval: The study was approved by the local ethical board. All participants provided written informed consent. Purpose: To explore the potential of superb microvascular imaging (SMI) technology in visualising brain microvessels in preterm neonates of different gestational ages (GA). A retrospective, observational, single-centre study including 15 newborns (8 females) was conducted. Preterm patients were equally divided into GA groups according to the WHO sub-categories: extremely (GA< 28 weeks, EP), very (28-31 weeks and 6 days, VP), and moderate to late preterm (32-37 weeks, MLP). All patients underwent transcranial ultrasound (US) using a Toshiba Aplio500 scanner (Canon Medical Systems Corporation) and a linear 14 MHz transducer. Exams were performed during the first day of life via the anterior fontanel including superficial and deep scans (each on coronal and parasagittal planes). Based on their SMI morphology and location, microvessels were classified as extrastriatal (subdivided into cortical and medullary), striatal, or thalamic. Two examiners independently classified the vessels as visible (echogenic linear structures) or invisible. To assess the association between vessel visibility and GA, a binomial logistic regression analysis (separate for each microvessel group) was performed, taking visibility as a dependent variable and both examiners and GA as predictor variables. Results: A statistically significant difference among GA groups was found in sex (p=0.030), GA by definition (p=0.002), weight at birth (p=0.007), and the Apgar score within 1 minute (p=0.024). GA significantly contributed to the visibility of superficial vessels (p<0.05 for both cortical and medullary ones) but not striatal and thalamic vessels. Conclusion: SMI-US is a feasible tool for visualising and assessing intraparenchymal brain microvasculature in pre-term neonates. The morphology and distribution of the brain microvessels seen in the current study are consistent with those described in previous anatomical and radiological research. Purpose: To assess the diagnostic accuracy of T1-and T2-weighted contrasts generated by the MR data post-processing software SyMRI for neonatal brain imaging. A total of 32 neonates were included in this retrospective study. 4 independent rating neuroradiologists assessed the neonatal brain on the basis of conventional and SyMRI generated T1-and T2weighted contrasts. The sensitivity and specificity of both methods were calculated and compared with each other. Results: Compared to conventionally acquired T1-and T2-weighted images, SyMRI-generated contrasts showed a lower sensitivity but a higher specificity Purpose: Metabolic derangements are known to occur after traumatic brain injury (TBI). Animal models point to initial Na + influx causing membrane depolarisation. White matter, particularly the corpus callosum (CC), is susceptible to damage. The purpose of this study is to compare the distribution of the total sodium concentration (TSC) in the CC between patients of mild TBI (mTBI) and controls using sodium MRI. Methods and materials: 11 patients (6 M, 5 F; 20-52 years) with a history of mTBI and 10 healthy controls were studied. Sodium MRI scans were performed on a Siemens Prisma3T scanner using a custom 8-channel dual-tuned (1H-23Na) Tx/Rx head coil. Twisted projection imaging was used (FOV=220 mm, matrix size=64, TE/TR=0.3/100 ms, FA=90, and TA=10.3 min). Sodium MR images were visually inspected in conjunction with proton MPRAGE. TSC was measured in genu, body, and splenium with 5 mm regions of interest (ROI). To assess for reproducibility, 5 measurements were independently taken within each area. A comparison was made between cohorts with a significance level of 0.05. To assess the distribution of TSC along the CC in the anteriorposterior dimension, the TSC genu/splenium ratio was calculated. Results: TSC was higher in the genu (51.59 vs 45.6 mmol, p=0.049) and lower in the splenium (50.80 vs 41.88 mmol, p=0.01) in mTBI patients compared to controls. The genu/splenium ratio was also higher in patients (1.2 vs 0.9, p=0.001). On visual inspection, mTBI subjects demonstrated a reversal of the normal TSC anterior-posterior gradient in the CC compared with controls. Conclusion: TSC distribution in the CC is altered after mTBI. Since changes in extracellular Na + concentration are transient after brain injury, detected alterations in TSC are believed to be attributable to changes in the intracellular Na + concentration. Our work supports the notion of ongoing Na + channelopathy after injury affecting callosal white matter such as perturbed expression of sodium channels. Purpose: Magnetic resonance-guided focused ultrasound (MRgFUS) is a noninvasive treatment for essential tremor (ET) that allows targeted thermal ablation of brain tissue. Previous studies demonstrate targeting the thalamic ventral intermedius nucleus (Vim) as an effective treatment in ET. This paper describes the world's first trial using MRgFUS to target both the Vim and the inferior zona incerta (ZI). Methods and materials: This prospective study enrolled 13 patients with medication refractory ET for a unilateral MRgFUS procedure. Tremor severity and functional impairment were assessed at baseline and regular intervals posttreatment for 24 months, using the clinical rating scale for tremor (CRST). BFS spirals were also collated intraoperatively, immediately pre-procedure and after targeting the Vim and ZI. All spirals were scored by 3 blinded neurologists. The percentage improvement in the spiral scores after Vim and ZI ablation were compared and analysed. Results: In all patients, there was successful thermal ablation of the target tissue at both Vim and ZI with improvement in all parameters stable over 24 months. The CRST tremor score improved by 73.5% (treated arm) and 24% (non-treated arm Results: 8 3D-features discriminated lesions with a low or high mitotic-count (best AUC 0.81, best sensitivity 86%, and best specificity 93%). 6 3Dparameters detected GISTs based on the mutational group (best AUC 0.77, best sensitivity 75%, and best specificity 79%) and 3 parameters correlated with the risk class (best AUC 0.76, best sensitivity 72%, and best specificity 85%). In the differentiation between GIST at a lower or higher risk of recurrence, the regression model used 6 different features and obtained AUC 0.78, sensitivity 65%, specificity 79%, VPN 71%, and VPP 73% on the training set, and AUC 0.83, sensitivity 88%, and specificity 75% on the test set. Our results show a good correlation between radiomics features, disease aggressiveness, mutational profile, and the risk of recurrence. Results are promising but validation on external datasets will be necessary to confirm the role as an imaging biomarker. Purpose: To accurately identify the relationship between the portal radiomics score (rad-score) and the pathologic superior mesenteric vein (SMV) resection margin in patients with cancer in the head of the pancreas. Methods and materials: 181 patients with pathologically confirmed cancer in the head of the pancreas who underwent multislice computed tomography within one month of resection from January 2016-December 2018 were retrospectively investigated. For each patient, 1,029 radiomics features of the portal phase were extracted, which were reduced using the least absolute shrinkage and selection operator (LASSO) logistic regression algorithm. Multivariate logistic regression models were used to analyse the association between the portal rad-score and SMV resection margin. Results: Patients with negative (R0) and positive (R1) margins accounted for 70.17% (127) and 29.83% (54) of the cohort, respectively. The rad-score, which is based on 11 selected portal phase features, was significantly associated with the SMV resection margin status (P<0.05). Multivariate analyses confirmed a significant and independent association between the portal rad-score and SMV resection margin (P<0.0001); a higher portal rad-score was associated with R1 resection (P for trend <0.0001). The portal rad-score is independently and positively associated with the risk of the SMV R1 resection in pancreatic head cancer. Limitations: Retrospective in nature, the results were established based on data obtained from a single centre, and the available samples were limited. Purpose: Attenuation characteristics of the pancreatic ductal adenocarcinoma (PDAC) and contrast material uptake vary depending on the histological grade and pancreatic parenchyma condition. The purpose of this study was to compare dynamic contrast enhancement (DCE) and radiomics features in predicting pancreatic adenocarcinoma grades. The data obtained from an analysis of the radiomics of pancreatic cancer can subsequently be used to predict the prognosis of patients with PDAC and for the inclusion of big data analysis in the diagnosis process. Methods and materials: 62 consecutive patients with histologically-confirmed PDAC who underwent CE-MDCT were enrolled in the study. We compared the mean lesion attenuation and relative tumour enhancement ratio (RTE) of different types of PA and radiomics features using arterial, venous, and delayedphase MDСT scans. Purpose: To investigate the CT imaging appearance of pancreatic solid pseudopapillary neoplasm (pSPN) in predicting pathological invasive behaviours. The clinical data and CT data of 103 patients with solid pseudopapillary neoplasm confirmed by surgical resection and pathology were retrospectively analysed. According to the pathological results, all cases were divided into two groups (the invasive group and non-invasive group). Image analysis included tumour number, location, size, shape, calcification, haemorrhage, fibrous pseudo capsule, and the proportion of cystic or solid component. The density of the solid components in all lesions was measured in plain scan, the arterial phase, and venous phase of CT images. The differences between the two groups were statistically analysed. Results: All 103 cases of pSPN had a single lesion. 38 cases were in the invasive group and 65 cases were in the non-invasive group. There were statistically significant differences in the fibrous pseudo capsule and the proportion of cystic or solid component between the two groups (P<0.05). There were no significant differences in neoplasm maximum diameter, location, morphology, calcification, haemorrhage, pancreaticobiliary dilatation, and pancreatic atrophy between the two groups (P>0.05). The density in the venous phase (76.65±16.71) in the invasive group was higher than that in the noninvasive group (70.09±12.02) and the difference was statistically significant (P<0.05). Saturday Scientific Programme Purpose: It was intended to compare interventional super-selective embolisation regarding renal parameters, clinical course and outcome to other therapeutical options and by that improving the approach to the treatment of acute renal bleedings of arterial origin. Methods and materials: Patients with renal injuries due to bleeding of arterial origin, pseudoaneurysms or arteriovenous fistulas underwent transcatheter arterial embolisation. The analysis included data of the technical and clinical success of embolotherapy, differences performing endovascular treatment during the clinical course and whether the post-interventional modified parenchymal renal volume amplifies the risk of developing renovascular hypertonia and chronic kidney disease. Results: A total of 167 patients was included. The technical success rate was 100 % and a clinical success rate of 92.5 % was achieved. Recurrent bleeding occurred in 14 patients (8.4 %). 6 (42.9 %) were re-embolised and 5 (35.7 %) underwent total nephrectomy. Recurrent bleedings tended to occur in correlation with subcapsular haematomas (p < 0.001). Renal function was measured by glomerular filtration rate and creatinine several times. It revealed no significant changes (p > 0.05). Patients with discontinuity of renal fascia were at a higher risk of parenchymal loss (p = 0.0237). The post-interventional renal volume loss was insignificantly low and over a period of 15 years, neither contributed to the progression of renovascular hypertonia nor to deterioration of renal parameters. Purpose: To evaluate the role of colour Doppler sonography in the evaluation of pregnancy with intrauterine growth restriction. The uteroplacental circulation, which are the two maternal uterine arteries. The foeto-placental circulation, which is the umbilical artery. The foetal circulation, which includes the foetal middle cerebral artery. The study included 30 antenatal women who were diagnosed as having a foetus with intrauterine growth restriction based on greyscale ultrasound findings and referred for obstetric Doppler ultrasound if the following inclusion criteria and exclusion criteria were met. The inclusion criteria were: singleton pregnancy with a gestational age of more than 28 weeks. Women with reliable dating of pregnancy confirmed by an early first-trimester ultrasound examination using CRL or BPD or with known LMP will be selected. Multiple pregnancies and foetuses with congenital anomalies. Doppler US evaluation was performed following a detailed clinical history and US biometry. The scanners and transducers used: the greyscale real-time ultrasonographic examinations were performed using ALOKA PROSOUND SSD -3500SX and PHILIPS HD 11XE. The transducers used for the study were 3.5MHz convex array transducer. The sensitivity of uterine artery Doppler study to detect adverse perinatal outcome was 83% when two Doppler parameters were considered. Conclusion: Assessment of both the uteroplacental circulation and the fetoplacental circulations together is more sensitive to predict a perinatal outcome, than the assessment of each alone. In suspected IUGR, cerebroumbilical ratio (MCA/UA PI) is a better predictor of adverse perinatal outcome than an abnormal MCA PI or umbilical artery PI. Best results are obtained when we use MCA/UA PI ratio, rather than PIs of a middle cerebral artery and umbilical artery separately. Limitations: This study is limited by the lack of patient follow-up. Ethics committee approval: Not applicable. symptomatic women in correlation with digital palpation and to define cut-offs for Hiatal dimensions that can predict muscle dysfunction. Methods and materials: This is a prospective study included 73 women with symptoms suggesting PFD. 3D-TPUS, MRI, and digital palpation of levator ani muscle were performed for all patients. Levator hiatal antro-posterior (LHap) diameter and area (LH area) were measured at rest and at maximum muscle contraction. The reduction in LHap diameter and LH area during contraction was significantly less in women with underactive pelvic floor muscle contraction (UpfmC) than those who have a normal pelvic floor muscle contraction (NpfmC) by digital palpation (p <0.001). Statistically significant positive correlations (p <0.001) were found between Modified Oxford Score (MOS) and 3D-TPUS and MRI regarding the reduction in the LHap diameter (r= 0.8, 0.82, respectively) and LH area (r= 0.6, 0.7, respectively). A reduction in LHap <6.5% on 3D-TPUS and <7.6% on MRI, predicted UpfmC with sensitivities of 46% and 83%, respectively. A reduction in LH area <3.4% on 3D-TPUS and <3.8% on MRI predicted UpfmC with sensitivities of 75% and 88.5%, respectively. MRI was more sensitive in detecting levator avulsion (63%) than 3D-TPUS (27%). show a progression disease hard to predict with common prognosis factors. We know the gradual expansion of the cysts increases the renal volume. A prognostic classification based on the renal volume measured with magnetic resonance imaging (MRI) corrected for height and age (HtTKV, height-adjusted total kidney volume) was developed. The proposed classification defines 5 classes (from A to E) of patients with a different risk of GFR decline based on renal growth rate and annual GFR reduction. Methods and materials: 25 patients (age 24-57) with ADPKD (14F / 11M) were followed clinically and radiologically with abdominal MRI (GE 1.5T, T1, T2, DWI). The MRI was used to calculate HtTKV. The renal volume was calculated by the Mayo clinic's web-based programme and MRI images by an algorithm based on the ellipsoid equation (π / 6xLxWxD), 4 major measurements (L,W,D,coronal, sagittal ,axial sequences), then correlating these values to the patient's age and height, GFR, we obtained the patient class. Results: The GFR range was (12-128) and the HtTKV (230 and 2078 mL/m). The patients were divided in 5 classes (4 class A, 7 B, 6 C, 5 D, 3 E). Patients in class A had normal value, 1/7 class B (GFR <60 ml / min / 1.73m2), class C: 4/6 (GFR <60 ml / min / 1.73m2). Our study shows the utility of renal MRI-based volume in patients with ADPKD to analyse prediction factors. Limitations: This study was limited by the number of patients due to a rare genetic disease. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: The purpose is to investigate texture analysis findings', the structure analysis application in radiological imaging, efficacy in the differential diagnosis of kidney masses renal cell carcinoma (RCC) and oncocytoma, contribution to the classification of RCC subtypes and relationship with the nuclear stage of RCC. Purpose: To investigate monoexponential, biexponential, and stretchedexponential models of diffusion-weighted imaging (DWI) for the assessment of prognostic features in endometrial cancer (EC). Methods and materials: 61 consecutive EC patients confirmed by pathology underwent multiple-b value DWI (0-1200s/mm 2 ) preoperatively at 3.0T MR scanner. The apparent diffusion-weighted (ADC), biexponential model-derived parameters (D, D* and f) and stretched-exponential model-derived parameters (DDC and α) were measured by two radiologists independently. All parameters were compared between the deep and superficial of myometrial invasion, with and without cervical stromal infiltration (CSI), presence and absence of lymphovascular invasion(LVSI). ROC curve and intraclass correlation coefficient analysis were used for statistical evaluation in predicting prognostic factors of EC. Results: Except for D* values, all of the parameters were significantly lower in tumours with deep myometrial invasion(DMI) than those invading superficial myometrial wall and the AUC was largest for ADC values(AUC:0.794). There was a significant difference between groups with and without CSI for ADC and f values and the f values exhibited best diagnostic performance in the prediction of CSI(AUC:0.825). The presence of LVSI had significant lower ADC, f and DDC values than those absences of LVSI and the DDC values showed the best diagnostic performance in the diagnosis of LVSI (AUC:0.759). The current study demonstrated that the prediction of prognostic factors for EC is feasible by multiple mathematical DWI models, which could be conducive to the pre-treatment risk classification to facilitate the selection of the optimal therapeutic approach and meet the demand for greater personalisation of cancer care. Limitations: It was a single-centre study with relatively small sample size and 13 b-values for the multiple b-value of DWI protocol prolonged the scan time. Ethics committee approval: Information about review board approval and written informed consent. Purpose: The aim of this study was to compare the diagnostic quality of acquired and computed high b-value images in diffusion-weighted magnetic resonance imaging (DWI) of the prostate in a setting, where the signal-to-noise ratio (SNR) is increased at high b-values by means of larger voxel sizes. Methods and materials: 34 patients with a PI-RADS 4/5 lesion in the peripheral zone were included. DWI was acquired (aDWI) with four b-values (0, 800, 1500, and 3000 s/mm²) at 3T. The nominal resolution was adapted to roughly keep the SNR constant. Computed-DWI (cDWI) were generated with the respective unweighted image and the acquired b-values, e.g. b3000 from b0 and b800 (cDWI3000-800). This resulted in two acquired and two computed images for each b-value (including b0) summarising to ten images. The three b0 images were pooled. Four readers performed a forced-choice pairwise comparison for representative lesions of all ten images and assessed lesion conspicuity and the ability to locate the lesion. Results: Concerning lesion conspicuity, cDWI3000-1500 scored highest (6.71) followed by cDWI1500-800 (5.71), and cDWI1500-3000 (5.14). For the ability to localise the lesion, cDWI800-1500 scored highest (5.82), followed by aDWI800 (5.41), and cDWI1500-3000 (5.40). In this study diagnostic performance of DWI in prostate magnetic resonance imaging was achieved for the combination of acquired b1500 and computed b3000. Therefore, we recommend to acquire images as high as b1500 and to compute the remaining images. Limitations: Potentially the most important limitation is that whole-mount prostate histology data were not available because prostatectomy is rarely performed at our hospital. We examined 36 patients (n = 100%) using dynamic perfusion computed tomography. 27 of them (n = 75%) diagnosed with transsexualism, underwent a female-to-male sex change operation using phalloplasty and consistent urethroplasty after 6 months. 9 (n = 25%) patients had different urogenital diseases such as: posttraumatic urethral stricture, microphallia etc. A free revascularised microsurgical allograft -skin-muscular thoracodorsal was used for the reconstruction of the penis, and for the reconstruction of the urethra was used skin-fascial radial allograft. Results: Processing of the data showed that in the early postoperative period there was a decrease in the level of arterial blood flow in the allograft tissue compared with preoperative indicators due to congestion and oedema. Then, an gradually increase in the level of arterial blood flow is observed due to the formation of anastomoses between the flap and the recipient's tissues and the further improvement of their function, as well as a decrease in the severity of oedema and congestion in the flap. The blood flow recovery in the flap to normal preoperative values was registered by the end of the first month after surgery. Purpose: To review bladder pathology with a multimodality approach and to define criteria when bladder pathology was found. Methods and materials: Review 300 cases with bladder pathology. The bladder can be approached by many modalities. Ultrasound provides anatomic detail, vascularity and nodules identification. CT is the next step and the gold standard with uro-CT the most commonly used protocol. The six points were defined: (1) imaging findings, (2) ureterovesical junction (UVJ) permeability, (3) bladder wall thickness, (4) presence of nodules/masses, (5) neurogenic bladder, (6) pelvic masses and other findings. The imaging findings included: ultrasound a wall focal (nodule) or diffuse parietal thickening, vascularity; CT -diffuse or nodular parietal thickening with enhancement. 53% of patients had pathology of the UVJ including lithiasis, masses and compression. Bladder wall thickness was defined as pathologic superior to 3mm when distended and superior to 5. Nodules/ masses from the bladder where found in 13 %, from the ureter in 4 % and form the prostate in 3 % of cases, adnexal masses and uterine masses in 12 %. Signs of the neurogenic bladder were present in 64 % of cases and graded from mild to moderate and advanced. Other findings included lithiasis, other urothelial cancers, and intestinal pathology. Conclusion: Bladder pathology is a common problem for radiologists and careful attention and awareness are a demand. Limitations: This study is limited by one-radiologist read and its limited sample. Purpose: Acute glomerulonephritis is an entity of renal disease in which immunologic mechanism cause damage to the basement membrane. The current study evaluates the renal vascular resistance in patients with acute glomerular diseases by measuring main renal and intrarenal arterial resistance (upper, mid and lower polar) (RI) of both kidneys and to correlate with laboratory data (creatinine, proteinuria) and histopathology. The study was conducted on twenty (20) patients with the age ranging from 3-13 years old. Cases with a non-glomerular cause of acute kidney injury, obstructive uropathy and known reno-vascular lesion were excluded from the current study. The mean RI of either main renal or parenchymal arteries is elevated in tubulointerstitial affection (mean value in main renal = 0.65, mean value in parenchymal arteries = 0.62 ) when compared to mean RI in cases without tubulointerstitial affection (mean value in main renal arteries = 0.59, mean value in parenchymal arteries = 0.57). The presence of crescent is not correlated with RI values in main renal and parenchymal arteries, creatinine level is positively correlated (p-value 0.1) with RI of the main renal arteries, no significant statistical correlation between RI indices and proteinuria (albumin /creatinine ratio). Purpose: Vesicoureteral reflux (RVU) is an abnormal condition, in which urine flows backwards from the bladder to the ureters and renal pelvis. If untreated, it may lead to recurrent urinary tract infections (UTIs), reflux nephropathy chronic renal failure. Currently, micturating-cystourethrogram (MCUG) and radionuclidecystogram (RNC) are the most commonly performed techniques for the diagnosis of RVU; both involve exposure to ionising radiations. However, in the last years, cysto-sonography (CSG) has also been introduced in clinical practice. The aim of this study is to prove the non-inferiority of CSG compared to MCUG, through a comparative analysis, in order to start introducing it as first-line technique, to reduce radiation dose. Methods and materials: 97 paediatric patients (49 boys and 48 girls) with UTIs were enrolled in the study. As first, after positioning of a urinary catheter, each patient underwent a CSG by the intravesical introduction of contrast agent (1ml-SonoVue), immediately followed by MCUG with intravesical iodinated-contrastmedium. Passive and active RVU was sought after spontaneous urination or abdominal press operation. The comparison of the results demonstrated a consistent concordance of the findings, further confirming the sensitivity of CSG in the diagnosis of RVU in the paediatric patient. During the ultrasound examination, it was also possible to assess the renal parenchyma and the possible dilation of the calicopyelic system as further findings useful for the purposes of the diagnostic classification of the patient. Conclusion: Cystosonography is a sensitive diagnostic technique that allows to identify RVU and to evaluate possible morphological alterations of the renal parenchyma avoiding the use of ionising radiation. Limitations: The limitation of this study are that it is a single centre study and that it refers to a short period of time. Kruskall-Wallis tests showed statistically significant differences between transverse and sagittal tumour diameter, tumour volume, cava-and aorta-tumour contact angle and contact length (in cm) (p<0.005). Outstanding discrimination (>0.9) was found for cavatumour contact angle. The optimal cut-off of 5 cm for transverse tumour diameter was confirmed (SENS 0.75 and 1-SPEC 0.18, CI95% 0.8-1). Cava-tumour contact angle had SENS 0.93 and 1-SPEC 0.17 with a cut-off of 98° (CI95% 0.9-1). At MVA aorta-(cut-off 64°) and cava-tumour contact angle (cut-off 98°) and poor IGCCCG score represented the three most important predictors of MVS (p<0.05). The model constructed has a PPV 100%, NPV 87% and an accuracy of 88%. Conclusion: Presence of aorta-tumour contact angle >64°, cava-tumour contact angle >98° on CT scans and poor IGCCCG score identify correctly 9 out of 10 patients requiring MVS at the time of PC-RPLND. These patients may be referred to specialised centres capable to adequately perform this multidisciplinary surgery. Purpose: To compare 3D PD fat-suppressed space turbo spin-echo (TSE) and conventional 2D PD sequences at 3T MRI in the detection of meniscal tears. Methods and materials: Dedicated MR knee imaging of 100 sequential patients with suspected meniscal injuries was retrospectively analysed by two independent radiologists in a novel study design using both a 2D PD standard multiplanar TSE MR technique and 3D PD TSE technique. The result was compared to arthroscopic results (arthroscopy was performed after MRI study) which were considered as a gold standard test. We determined the sensitivity, specificity, and interobserver agreement for each sequence. Results: The 3D TSE shows statistically higher sensitivity and specificity (100% and 100%) for the detection of meniscal radial tear compared to the 2D sequence (67% and 96%). No significant difference was found between the sensitivity and specificity of both techniques regarding the detection of horizontal, longitudinal, root, flap, and bucket handle tears. There was higher interobserver agreements on 3D TSE SPACE than 2D sequences for radial, root, and bucket handle tears. Conclusion: At 3T MRI, the 3D PD fat-suppressed space turbo spin-echo (TSE) has higher diagnostic performance compared to the routine 2D TSE protocol for the detection of meniscal radial tears and similar results as regarding the other types of tears with the advantage of faster acquisition. Limitations: n/a Ethics committee approval: Written and informed consent was obtained. Purpose: To evaluate the added advantage of shear-wave elastography (SWE) over grey-scale sonography in the diagnosis of plantar fasciitis and to obtain a reproducible and reliable sonoelastographic shear stress cut-off value. In this prospective case-control study, 30 consenting patients having unilateral heel pain for >6 months and clinically diagnosed as unilateral heel pain were included. Patients were examined by the lead author and a faculty radiologist by B-mode sonography (for plantar fascia thickness), followed by SWE, obtaining a colour map for shear characteristics and quantitative data (shear velocity in cm/s and shear stress in kPa) using a 1x1 cm ROI placed at the position of the highest shear on the colour map. Sensitivity and false-positive rates with extrapolated specificity, PPV, and NPV were calculated, and shear stress cut-off values were obtained using the Chi-square test and ROC curve analysis. The unaffected foot from respective cases was taken as control (normal) for the study. Results: Of 30 examined cases, mean thickness of plantar fascia was significantly higher (5.4 mm) in the affected foot compared to the unaffected (2.9 mm), with a standard ROC cut-off of 3.7 mm (sensitivity 93%, false-positive rate 10%). The mean shear stress in the plantar fascia on the affected side (60.1kPa) differed significantly from the unaffected side (150.7kPa). Taking 97.5kPa as the shear stress cut-off, considering affected patients having values below this cutoff, sensitivity was 81.8% and the false-positive rate was 10.6%. Extrapolated specificity, PPV, and NPV were 89.4%, 88.5%, and 83.1%. AUC was 0.91. Conclusion: SWE shows superior diagnostic accuracy compared to grey-scale sonography in patients with plantar fasciitis and promises the potential for routine use in addition to B-mode US for the assessment of plantar fasciitis cases. Purpose: The increasing use of cone-beam CT to diagnose ankle and foot problems urges the development of quantitative 3D analysis tools that deliver geometrical parameters corresponding to well-known 2D counterparts. We compared estimates of geometrical parameters from both cone-beam CT and radiographs as differences can be expected between 2D and 3D analyses. Methods and materials: WB and non-WB cone-beam CT-images of the left and right hindfoot and forefoot were acquired on a Planmed verity cone-beam CTscanner. Geometric analyses were performed on images of the foot in 2D using simulated radiography images and 3D using custom analysis software. Measurements included the calcaneal pitch, Meary's angle, the angle between 1 st and 2 nd metatarsal (MTT), the talocalcaneal angle, and cuboid height. Results: 5 patients were included with a mean age of 39.6 years. All measurement results were statistically different in 2D and 3D (p<0.001). The average calcaneal pitch decreased in 2D and in both 3D WB compared to non-WB results. Meary's angle decreased in 2D measurements, however, increased using 3D measurements. The average angle between MTT1-2 increased in both approaches. During WB, the talocalcaneal angle decreased in both lateral-lateral and anterior-posterior 2D views and increased in 3D. Cuboid height decreased between non-WB and WB images by 1.3 and 6.5 mm in 2D and 3D images respectively (p<0.001). The geometric parameters evaluated in 2D were different compared to 3D, which is likely caused by over-projection and the specific point of view. We expect that 3D measurement tools are more univocal and therefore more suitable for future use. Also, switching to 3D enables advanced analysis such as obtaining joint space maps. Purpose: Normal saline infusion radiofrequency ablation (NSRFA) is a widely accepted treatment for tumours up to 3 cm in diameter. Microwave ablation (MWA) and hydrochloric acid infusion radiofrequency ablation (HRFA) may be better alternatives for larger tumours, but few studies comparing these 2 modalities have been published. Thus, this study is aiming to compare ablation zone sizes and shapes resulting from HRFA and MWA, using NSRFA as a control in an ex vivo bovine liver model. A total of 90 ablation procedures were performed prospectively, each using 3 modalities: NSRFA, HRFA, and MWA. For each modality, 5 ablation procedures were performed for each combination of power (80W, 100W, or 120W) and duration (5, 10, 20, 30, 45, or 60 minutes) . Ablation zone outcome parameter includes transverse diameter (TD), longitudinal diameter (LD), volume (V), front distance (FD), and spherical ratio (SR Purpose: Bias field correction deals with the intra-volume intensity inhomogeneities in magnetic resonance imaging (MRI) data. It has little significance for visual diagnosis yet it is a crucial step for the automation of radiotherapy solutions. Extensive research and evaluation show that the often used and well-established correction methods typically require parameter tuning for the given dataset or are expensive to compute. A novel method has been introduced, allowing for the training of a convolutional neural network (CNN) on non-medical images, making the model general and evaluating it for bias field correction on medical MRI data. The proposed method has been compared to an optimised N4ITK in three experiments. It produces nearly identical results to N4ITK for a benchmark test and similar results in a coefficient of variation, mean absolute error, mean error, and structural similarity for a BrainWeb-based experiment, with an average speedup factor of 137. Different methods performed best for the different investigated tissue types. It was also tested for real cases, often producing more radio realistic bias fields in the cases where N4ITK struggles. Also, no datasets were encountered where the method would not improve image quality. Conclusion: Training a convolutional neural network implicitly, through the relation of the outputs of two similar training samples, is possible, omitting the need for true target data. The achieved model is not only generalised but also achieves comparable accuracy to N4ITK with a significant speedup. In 22 patients, a small local haematoma with a median size of 6 ml without further treatment was observed. 20 patients reported pain within the first hours after the intervention, which was easily treated with metamizole. 1 of those 15 patients with a transhepatic pathway showed an incidental arterioportal fistula in a follow-up CT, which did not require further action. The intervention duration from local anaesthesia to the control scan was a median 16 minutes, with a range of 7-35 minutes, and was well tolerated by the patients. Malignancy was confirmed in 92%. In 2 cases, an adenoma was detected; one lesion showed high FDG-uptake and the other high density on CT. In 2 cases, a definite determination was not possible. Conclusion: CT-guided biopsy of the adrenal gland is a safe, fast, and reliable method to verify and characterise metastases in patients suspected of lung cancer. Pathways, patient positioning, investigator experience, and procedural manoeuvres are essential and will be communicated. Turin/ IT (20027467@studenti.uniupo.it) Purpose: To evaluate the effectiveness and safety of UFE as an alternative to surgery in the treatment of uterine fibromatosis. Methods and materials: 255 patients (aged 26-55) with symptomatic UF, indicated for surgery, followed in our centre (2000) (2001) (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) (2011) (2012) (2013) (2014) , with single or multiple fibroids, and pain and/or functional/compressive disorders underwent embolisation; an injection of PVA particles (150-900 µm) from the distal portion of uterine arteries (ascending section). The primary end-point was a flow-stop distally to the injection site, disappearance of the lesion design, and the preservation of flow in the main trunk of UA. The secondary end-point was the control of pain and functional/compressive disorders during follow-up (2-7 years). The procedure was performed bilaterally in 250 patients (98%). The mean duration was 47 min (average fluoroscopy: 10-50 min). Post-embolisation pelvic pain (according to a VAS score) was on average 2.2 at discharge (24h). Follow-up at 2 years showed a resolution of menstrual disorders in 78% of patients and improvement in 14%, pain disappeared in 66%, significant improvement of menstrual flow and HCT/HB levels, and adecrease in total uterine (57.7%)/dominant fibroid (76.1%) volume. There was recurrence in 18 patients. Purpose: To analyse pre-UAE MR images with a view to predicting migration of intramural fibroids post-UAE to better advise patients of the procedural risks. We retrospectively reviewed all patients referred for UAE for symptomatic fibroids at our institution over one year. Patients without a pre-procedural MRI and a six-month post-procedural MRI were excluded from the study. Both pre-and post-UAE images were independently reviewed by two radiologists with subsequent consensus review. In each, dominant intramural fibroids were identified and the following measurements taken: a) 3-dimensional maximal measurements, b) the shortest distance between the fibroid and the endometrial wall, and c) the shortest distance between the fibroid and the serosal wall. Paired sample t-tests and two-sample t-tests were used to compare between pre-and post-UAE variations and between migrated and non-migrated intramural fibroids, respectively. Results: 35 dominant intramural fibroids were identified, of which, 8 migrated to become submucosal fibroids, while 5 were either partially or completely expelled. All migrating fibroids had a pre-UAE minimum distance from the endometrium of 1-2.4 mm and a maximum fibroid diameter greater than 5.1 cm, and none of the non-migrating fibroids had both these characteristics together. Conclusion: Not only is it distressing to patients to have post-procedural pain and bleeding, but fibroid migration could also increase infection risks and affect fertility. Our study found that intramural fibroids with a minimum endometrial distance less than 2.4 mm and a maximum fibroid diameter greater than 5.1 cm have a high likelihood of migrating after UAE. Limitations: Small sample size and low statistical power. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: A major challenge in prostatic artery embolisation (PAE) is an adequate and distinct visualisation of the feeding prostatic vessels and other branches of the internal iliac arteries. Novel colour-coded parametric images calculated from raw angiographic image data provide additional information compared to common digital subtraction angiography (DSA). Our objective was to investigate the clinical value of parametric images in a PAE setting. We retrospectively examined the selective internal iliac artery image series of 26 patients (mean age: 67.47; SD: 9.76; range: 42-82) undergoing PAE. We generated 230, parametric, colour-coded-delayangiograms (CDAs) from raw image data obtained from a latest-generation angiography system. Three experienced interventional radiologists evaluated the image quality using a 5-point-Likert scale based on clinically relevant criteria (e.g. the identification of dominant feeding arteries, stenoses, and non-target collaterals). A C 600 Saturday Results: Subjective image quality was evaluated with mean scores of 4.16 (range 3-5; std.deviat.+-0.73), significantly higher for DA compared to the mean scores of 3.43 (range 2-5; std.deviat.+-1.80) for standard DSA. Evaluators agreed CDAs provided more information than the simple BW angiograms (87%), the image would be useful to select the dominant feeding artery (93%), and the functional stenoses of the branches could also be determined (84%). The interrater agreement, Fleiss' kappa, and significance were 81%, 0.19, p<0.05, 85%, 0.32, p<0.01, and 79%, 0.11, p<0.05, respectively. Conclusion: Parametric delay images calculated from x-ray angiography series provide more information than standard angiograms and hold additional information about the branches of the internal iliac artery. Decision-making based solely on CDAs will require an appropriate prospective safety assessment. Limitations: Colour-coded-parametric-angiography has to be tested in upcoming prospective trials to prove our results for real-time image guidance. Ethics committee approval: University hospital ethics committee approval obtained. Purpose: To evaluate fibroid size using ultrasonography (US) and magnetic resonance imaging (MRI) after uterine artery embolisation (UAE) of leiomyoma. Methods and materials: 1,023 women with symptomatic leiomyoma were treated by bilateral transcatheter UAE. Angiography and embolisation were performed using 5F Cobra and Roberts uterine catheters by 355-1,000µm of PVA particles. The size of particles depended on the uterine artery diameter and anastomosis between the uterine and ovary arteries. The mean age was 29.5 years (range: 19-47). The moderate reduction of leiomyoma size was observed one month after UAE. The reduction of fibroid size after 1 month was 18.4% by US and 17.3% by MRI, after 3 months was 40.7% by US and 42.9% by MRI, after 6 months was 60.4% by US and 61.8% by MRI, after 12 months was 72.5% by US and 74.7% by MRI, and after 24 months was 70.8% by US and 72.2% by MRI. Doppler US-control showed a blood flow decrease during the first month after UAE; in 3 months, blood flow data was restored in uterine arteries. We also noticed a significant reduction or disappearance of clinical symptoms such as pain, menorrhagia, and pressure symptoms. The partial vascularisation of leiomyoma was only detected in one case 24 months after UAE. Conclusion: UAE is an effective, minimal-invasive form of leiomyoma treatment. MRI and US are the modalities that can accurately estimate fibroid changes and blood flow in uterine arteries. The tendency of fibroid shrinkage after UAE is well-defined in the period from 1-24 months. After UAE, the blood supply of myometrium is restored according to US Doppler data in the period from 1-3 months. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Cryoprobe removal before sufficient thawing may lead to tissue injury and bleeding. We analysed the melting time of radiographic ice-balls during thawing in renal cryoablation by using ultrasound (US) examination. A consecutive 27 patients who underwent percutaneous cryoablation with cryoprobes of 2-4 for a renal tumour (median size, 1.9 cm; range, 1.1-4.1 cm) were evaluated. Reconstructed CT images obtained during freezing were used to measure the radiographic ice-ball volume. After completing a last freezing, the melting time of the radiographic ice-ball was analysed with US. The melting time was defined as the time of complete disappearance of intrarenal posterior acoustic shadowing generated by a radiographic ice-ball, which was analysed by two independent radiologists. After removing the cryoprobes, the degree of perirenal haematoma was assessed (i.e. <1 cm versus ≥1 cm in thickness). Serum haemoglobin (Hb) was also investigated before and after cryoablation. The median radiographic ice-ball volume was 30.3 cm 3 (range, 21.0-62.8 cm 3 ). The median melting time of radiographic ice-balls on US was 13 and 14 minutes for radiologist 1 and 2 (range, 9-15 minutes for both radiologists) (p> 0.05). A positive correlation was found between melting time and radiographic ice-ball volume (Spearman's rho, 0.647 and 0.484 for radiologist 1 and 2). Perirenal haematomas of ≥1cm in thickness were seen in two patients (7.4%). No patient required treatment for the bleeding. The median decrease of serum Hb was 0.3 g/dl. Conclusion: Ice-balls of renal cryoablation may require 15 minutes for complete melting, which may be considered as the optimal time for cryoprobe removal. Limitations: A small, single-institution study. Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: M. J. Kim: nothing to disclose S. Y. Park: nothing to disclose Purpose: To evaluate the therapeutic success, complications, and economic impact of transperineal US-guided focal laser ablation by US/MRI fusion software as a primary treatment for focal unilateral prostate cancer. Methods and materials: 24 patients with newly diagnosed, histopathologicallyproven, focal unilateral prostate cancer were treated with US-guided transperineal focal laser ablation as a primary treatment. The inclusion criteria were no previous prostate treatment, a PSA level ≤20, a Gleason score ≤7, and stage ≤ T2b N0M0 with 3T multi-parametric MRI-visible index lesion (PIRADS ≥ 4). After the ablation, a 3T mpMRI of the prostate was obtained. Follow-up consisted of mpMRI at 1, 3, 6, and 12 months and a US/MRI fusion-guided biopsy at 6 and 12 months. Results: 24 patients were successfully treated with transperineal US-guided focal laser ablation. No complications occurred. The IPSS and SHIM did not significantly change after treatment. The mean operation time was 38.2 minutes (range 32.6-42.5), the mean ablation time was 21.7 minutes (range 18.3-26.8), the mean energy deployed was 3606J (range 3212-3804), the mean hospital stay was 113 minutes (range 55-178), and the mean catheterisation time was 261 minutes (range 95-412). At the 6-and 12-month follow-up, prostate mpMRI and US/MRI fusion-guided biopsy showed neither evidence of local residual disease nor recurrence. Conclusion: Transperineal US-guided focal laser ablation as a primary treatment for prostate cancer has shown encouraging results. Ten-year followup with an international registry is intended to confirm oncological long-term prostate cancer with index lesion control. Limitations: A small number of treated patients, short follow-up, and our first experience with this technique. Ethics committee approval: Protocol approved by the ethical committee of Policlinico Tor Vergata (Rome). Purpose: To evaluate our decision-making between RFA and CA based on tumour size, central position, and nearness to the collecting system on technical efficacy and oncologic outcomes for cT1 RCC. Methods and materials: All cT1 RCC from 2005-2019 were included. Patients treated before 2013 were included in the RFA group and patients treated after 2013 were included in the CA or RFA group, where CA was performed for large (≧35 mm) central tumours or RCC with a nearness to the collecting system, while RFA was performed for smaller (<35 mm) and non-central RCC. The IPTW based on clinical and tumour features was applied. The primary success was compared using logistic regression, while local recurrence-free survival was compared using Cox proportional hazards regression. Clavien complications and the change in renal function (MDRD) were also evaluated. Results: Data from 408 patients were analysed. The primary success rate was higher in the CA or RFA group compared to the RFA group (97.1% vs 87.1%, p<0.001). Local recurrence-free survival was better in the CA or RFA group compared to the RFA group (HR=0.49; 95%CI: 0.28-0.84; p=0.01). Survival without metastases was not different (p=0.65). The incidence of Clavien grade ≥2 adverse events (p=0.30) and the change of renal function (p=0.35) was similar. Conclusion: Our decision-making between CA and RFA achieved a better primary success rate and local recurrence-free survival compared to RFA alone, with excellent oncologic outcomes for RFA in selected patients and also for large and central RCC mainly treated with CA. Limitations: The retrospective design. Ethics committee approval: Institutional review board approval was obtained to review the medical records of patients who underwent ablation of RCC at Necker University Hospital. Methods and materials: In this retrospective analysis from 2010 to date, 17 patients (12 men and 5 women, mean age: 59.8 years) presenting with adrenal bleedings were treated by endovascular embolisation. The aetiology was traumatic in 7 cases, ruptured neoplasm in 8 cases, and spontaneous in 2 patients assuming oral anticoagulant therapy. After thin slice contrast-enhanced CT, a super-selective embolisation was conducted with different embolising agents according to the type of vessel lesion and operator preference. Results: Technical success rate, considered as an interruption of adrenal bleeding detectable at angiography, was 94.1%. Clinical success rate, considered as haemodynamic stability restoration within 24 hours from the procedure, was 82.3%. The vessels involved were the superior adrenal artery in 5 patients, the middle adrenal artery in 8 patients, the inferior adrenal artery in one patient, and more than one adrenal artery in 3 patients. No procedure-related major complications occurred and no patients had infarctions, necrosis, abscess formation, or required long-term steroid supplementation. Conclusion: Acute adrenal haemorrhages can be safely and effectively managed by catheter-directed embolisations. The source of bleeding has to be carefully investigated at CT and angiography as adrenal glands present with a wide and complex vascular arterial network. Limitations: This is a retrospective analysis and therefore bias selections should be considered. The number of patients is small, however, this is due to the low frequency of the scenario. Ethics committee approval: n/ Resting-state functional magnetic resonance imaging in predicting the long-term functional outcome after acute ischaemic stroke E. Camacho-Ramos 1 , A. Jimenez-Pastor 1 , C. Biarnés 2 , A. Alberich-Bayarri 1 , S. Pedraza 2 , J. Serena 2 , M. Terceño 2 , Y. Silva 2 , J. Puig 2 ; 1 Valencia/ES, 2 Purpose: Stroke is one of the main causes of disability in adults. Resting-state functional magnetic resonance imaging (rs-fMRI) is able to map functionalanatomic networks by analysing spontaneously correlated low-frequency activity fluctuations across the brain. However, the role of rs-fMRI to predict the functional outcome after acute stroke remains unclear. We aimed to evaluate the impact of rs-fMRI at 72h after stroke symptoms onset to predict the functional outcome at 1-year follow-up. Methods and materials: 35 reperfused stroke patients (12 female, age: 68±14 years, 3-day national institutes of health stroke scale (NIHSS) score: 6±5) were consecutively scanned 72 hours after symptom onset. A modified Rankin scale (mRS) was used to evaluate functional outcome at 90 days (patients with good functional outcome, mRS ≤2; poor outcome, mRS>2). Image pre-processing included motion and slice-timing correction, outliers' detection, brain tissue segmentation, MNI space normalisation, smoothing, and band-pass filtering to remove undesired components. Region-of-interest analyses were performed to calculate the correlation coefficients for every pairwise region and a weighted general linear model was used to discover significative differences (p<0.05 FWE) between groups. Results: Patients with good outcome presented higher intra-network connectivity, while patients with poor outcome showed lower inter-network connectivity. To predict the 1-year functional outcome after a stroke, the 3-day NIHSS score was the most relevant predictor. An accuracy of 74.3% was obtained through a discriminant model using only clinical scores. Adding anatomical networks increased accuracy to 88.6%, 97.1% (including functional networks), and 100% (with clinical and imaging variables). Conclusion: These results show the relevance of rs-fMRI in the acute phase to predict the long-term functional outcome in stroke patients. Purpose: This study applied brain resting and task state fMRI to detect the changes of brain activity and explore the mechanism of neurological damage in adult patients with common exotropia. Methods and materials: 20 patients (M:10, F:10) and 20 age-and sex-matched controls with normal corrected visual acuity from 18 to 26 years of age were enrolled in this study. Two independent samples t-tests were employed to measure the differences of regional homogeneity (ReHo) and amplitude of lowfrequency fluctuation (ALFF) in the resting-state fMRI and the brain-activity difference in the visual stimuli task fMRI of two groups. Results: Compared with healthy controls, patients had significantly higher ReHo in the right middle occipital gyrus and the left supramarginal gyrus, and lower values in the right cingulate gyrus, the right precentral gyrus, the right superior temporal gyrus, and the left frontal sub-gyral lobe (P<0.05). Patients showed a reduced ALFF in the left superior parietal lobe, the right middle cingulate gyrus, the left anterior cingulate gyrus, and the right inferior temporal gyrus (P<0.05). Moreover, under the condition of binocular 2D visual task stimulus, the brain activity increased in the left superior frontal gyrus, the right superior orbital frontal gyrus, the bilateral inferior orbital frontal gyrus, and the left rectus gyrus in the patient group. Meanwhile, the brain activity of the left postcentral gyrus and the left precuneus gyrus increased under binocular stereo-stimulation. Conclusion: Common exotropia may lead to disorders in the processing of visual-spatial information and the control of emotion and attention. Purpose: Task-based fMRI is the method of choice for determining the hemispheric dominance of language function. One important limitation is that the technique depends on the patient's ability to perform the task. The aim of our study was to test the feasibility of using resting-state fMRI instead of task-based approaches. Methods and materials: 15 healthy native English speakers underwent restingstate fMRI and two language fMRI paradigms: a verbal fluency (VF) and one verb generation (VG). Task-based fMRI was analysed using both a general linear model (GLM) in the SPM software and independent component analysis (ICA). Resting-state fMRI data was analysed by using both ICA and principal component analysis (PCA) using the SPM CONN toolbox, assuming a fixed number of 20 components. Resting-state networks were extracted by using both an automated template procedure and visually by two experts (VL and LM). The VG paradigm was more accurate than the VF paradigm or combination of the two and was therefore selected for further analyses. Some of the networks labelled as language networks by the software were considered inappropriate and discarded by the experimenters. There was a poor overlap between resting-state and task-related areas of activation as assessed by calculating the Dice similarity coefficients (mean: Saturday Scientific Programme 0.1815) using the masks from the VG task GLM analysis, and the selected resting-state components. The laterality index (LI) values of resting-state networks were significantly lower (mean: -0.0165) than the LI of the VG task map (0.4846). Our results do not support the use of resting-state fMRI instead of task-based fMRI for hemispheric dominance assessment. Limitations: Our sample is relatively small and heterogenous. Ethics committee approval: Written informed consent was obtained from all subjects. Funding: This work was part of the UCLH Project ID number: 09/H0716/18. Author Disclosures: V. Lolli: nothing to disclose S. Wastling: nothing to disclose L. Mancini: nothing to disclose Machine learning analysis in the diagnosis of medication overuse headache: a resting-state functional magnetic resonance imaging study C. Yang, L. Yao, S. Lui; Chengdu/CN Purpose: Medication-overuse headache (MOH) is a psychiatric comorbidity associated with chronic migraine. As machine learning is capable of identifying subtle disease patterns on a single subject level, we sought to determine the differences between MOH patients from healthy controls by using a machine learning classifier on resting-state functional MRIs (rs-fMRI). Methods and materials: 34 patients with MOH and 41 age and sex-matched healthy controls were recruited. Smooth mean regional homogeneity (smReHo) was obtained via DPARSFA. Support vector machine (SVMs) algorithms based on PRoNTo toolbox were used to explore the utility for smReHo in the differentiation of patients and controls individually. Finally, Pearson's correlation between the MOH patients and clinical measurement was investigated. The balanced accuracy of the correct classification of MOH and controls was 72.03% for smReHo (p<0.001 during 1,000 permutation testing). Areas contributing to classification accuracy mainly included in right Heschl's gyrus (3.23%), left temporal pole, and right inferior temporal gyrus. The direct comparison showed that MOH patients had significantly decreased smReHo activation in the right Heschl's gyrus (p<0.05, FDR corrected), which was corresponding to the highest weight area by machine learning classifier. A negative correlation was observed between the duration of medication intake and smReHo activation in the right Heschl's gyrus (r=-0.433, p<0.05). Conclusion: With machine learning computational analyses, our study was able to diagnose MOH patients on a single subject level by demonstrating decreased regional function in MOH patients, notably involving the right Heschl's gyrus, suggesting the right Heschl's gyrus might play an important role in the pathogenesis of MOH. Limitations: Studies with larger samples may be required for further verification. Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: C. Yang: nothing to disclose L. Yao: nothing to disclose S. Lui: nothing to disclose The reorganisation of language networks after temporal lobe epilepsy surgery: a clinical fMRI study O. Purpose: The reorganisation of language is common in epilepsy patients and a relevant factor in surgical therapy. The non-invasive gold standard in presurgical evaluation is language fMRI. fMRI network analysis could reveal subtle changes in language network architecture so far invisible with standard analyses and thus provide relevant additional information. We aimed to investigate the reorganisation processes of language networks in surgical patients with temporal lobe epilepsy (TLE) using standard fMRI activation and recent network analysis. Methods and materials: Task-based language fMRIs were performed on 28 TLE patients (16 right/12 left) before and after surgery, and on 22 healthy controls. FMRI activation analysis, a derived laterality index, and functional connectivity of global and language networks were used to investigate language architecture. Naming, semantic, and phonematic verbal fluency were correlated with imaging parameters. Results: Both left and right TLE were associated with widespread language network alterations, particularly ipsilesional, before, and more pronounced after surgery. Preoperatively, mesiotemporal connectivity was decreased in left TLE, while postoperatively bihemispheric changes with a shift to more atypical language laterality (especially frontal) occurred. In right TLE, right intrahemispheric frontotemporal functional connectivity was attenuated before and further decreased after surgery, while those patients with stronger connectivity scored worse in semantic verbal fluency. Better postoperative naming ability was generally accompanied by stronger interhemispheric frontal connectivity. In all cases, network analysis revealed more widespread alterations than activation analysis. The reorganisation of language function is a visualisable phenomenon in left and right TLE, both before and after surgery. Functional connectivity analysis reveals more widespread network changes than fMRI activation analysis and may offer a more comprehensive view on language architecture. Some connectivity measures may serve as biomarkers for language performance. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Functional connectivity abnormalities, such as default mode network (DMN), have been reported previously in trigeminal neuralgia (TN) of which chronic pain is a common symptom. However, little is known about the dynamic functional connectivity (dFC) in patients with TN. Methods and materials: 29 patients with TN were enrolled in this study and underwent rest-functional MRI scanning along with 24 age-, sex-, and educationmatched healthy control (HC) subjects. A sliding-window approach was used to study the dFC based on the GIFT (group ICA of fMRI toolbox). The functional networks were arranged into DMN, auditory network (AUD), visual network (VIS), and ventral attention network (VAN). Dynamic analysis suggested 4 distinct connectivity 'states' across the entire group (state I for 24%, state II for 16%, state III for 35%, and state IV for 25%) and a larger proportion of state III and a smaller proportion of state IV were observed compared with HC group (p=0.0000 and p=0.0058, respectively). Furthermore, the TN group showed decreased connections between networks (DMN-VIS, AUD-VIS, and VIS-VAN). Our analyses showed that increased dwell time in the state IV was associated with the presence of pain in TN. Conclusion: Our study indicates that chronic pain in TN is characterised by altered temporal properties in dynamic connectivity. Further studies on dynamic functional connectivity could help to better understand the progressive dysfunction of networks in TN. Limitations: Correlation analysis is in progress. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Diffuse low-grade gliomas (DLGG) have a slow infiltrating character, allowing for simultaneous plasticity and maintaining the patients' functional capacity. Here we addressed whether and how anatomical disconnection following left-hemispheric DLGG growth and resection interferes with restingstate connectivity, specifically in relation to language. Methods and materials: 39 native French persons with a left DLGG were included and underwent awake surgical resection of the tumour. The anatomical disconnectivity risk following the DLGG volume and the resection, and the functional connectivity of resting-state fMRI images in relation to language, were evaluated prior to and three months after surgery. Resting-state connectivity B A C 604 Saturday Scientific Programme patterns were compared with 19 healthy controls. Language was assessed with a picture-naming task. Results: Adaptive plasticity was observed in regions around the tumour/cavitas area, with increased connectivity of the left and right inferior parietal lobule with the left inferior temporal gyrus. Picture naming was surprisingly not dependent on the connectivity of the language network, but of the connectivity of regions know to be involved in semantics, in which their specific role could be explained in the light of the broader resting-state network they took part in. Also, the recruitment of the frontoparietal attention network was observed. Conclusion: A whole-brain approach with specific clinical data input is required for meaningful resting-state analysis in case of lesions. Pre-surgical local plasticity and functional compensation by increased task-directed attention allow a relative high level of performance maintenance after DLGG resection. Limitations: All participants underwent intensive speech-language rehabilitation post-surgery, which might have trained and strengthened attentional resources. Purpose: There have been accumulating studies demonstrating thalamic related structural, functional, and metabolic abnormalities in minimal hepatic encephalopathy (MHE). We conducted the first study to investigate thalamicrelated structural connectivity alterations in MHE. Methods and materials: Diffusion based probabilistic tractography was used to determine the structural connection between the thalamus and the cortical/subcortical regions in 22 cirrhotic patients with MHE, 30 cirrhotic patients without MHE (NHE), and 30 healthy controls. DTI (diffusion tensor imaging) measurements of these thalamic connections, including connectivity strength (CS), fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD), were calculated and compared across three groups. Neuropsychological assessment was performed as well. The correlation analysis was conducted to investigate the relationship between neuropsychological performance and the measurements about thalamic connections. Machine-learning classification analysis was performed to estimate whether diffusion measurements can distinguish MHE from NHE. Results: The probabilistic tractography revealed thalamic structural connection, which was disturbed in cirrhotic patients (reflected by decreased CS/FA and increased MD/AD/RD). Abnormal thalamic connection primarily involved the prefrontal cortex, sensorimotor cortex, parietal cortex, medial temporal cortex, hippocampus, and striatum. Thalamic connectivity abnormalities deteriorated from NHE to MHE and correlated with patients' impaired neuropsychological performance. The relatively high classification accuracy was obtained using CS as a discriminating index. Conclusion: Our results demonstrated the altered thalamic structural connectivity with both cortical and subcortical regions in MHE. The disturbed thalamic connectivity may underlie the mechanism about cognitive deficits in MHE and can serve as biomarker for the diagnosis of MHE and monitoring disease progression. Purpose: Constant stimulation of the nervous system is one of the most popular ways to activate neural networks in order to activate the brain and initiate neuroplasticity processes. The purpose of the study was to evaluate the impact of translingual neurostimulation of the brain to balance, coordination of movement, and the ability to form new motor skills in children with cerebral palsy in the late residual stage. This study involved 6 children with cerebral palsy, a form of spastic diplegia. Patients had intact intellect and no seizures in anamnesis. All children received standard treatment, including massage, medical-gymnastics with simulators, 10 daily sessions of physical therapy and neurostimulation of the brain (using a device for electrotactile stimulation called PoNS (portable neurostimulator)). Patients underwent functional MRI of the brain before the start of, and at the end of, the course of treatment. The research was carried out using movement functional paradigms for each extremity (both feet and hands) and an active "count" paradigm for patients who could perform the task for the count and depending on clinical status, and passive movement or sensory paradigms were performed for those who could not. Postprocessing was carried out using the software package SPM 12. Results: All patients noted decreased muscle tone and improved balance and coordination functions after neurostimulation. According to obtained fMRI data, patients who underwent translingual neurostimulation showed increased activation in the right-hand motor area and activation in the foot motor area in response to the stimuli. Conclusion: Translingual neurostimulation is innovative in the field of neurostimulation, non-invasive, and safe and easy to use. fMRI active paradigms, with proper and high-quality implementation, is an auxiliary method for the objective control of treatment efficiency. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: To investigate the alterations of brain structural and functional connectome in unilateral sudden sensorineural hearing loss (USSHL) patients. We constructed functional-structural connectomes for USSHL patients (41 for the left side and 44 for the right side) and 85 controls. Graph theoretical analysis was employed to evaluate the network properties of functional-structural brain connectome. Moreover, we quantified coupling between functional-structural connectome by the correlation between functionalstructural connectome edges within non-zero structural connectivity areas. The coupling of functional-structural brain connectome was compared by using permutation tests. To investigate the clinical relevance of altered brain network topologies in USSHL, Pearson's correlation analysis was performed between the clinical variables and the topological properties. Results: Compared with the control groups, both groups of USSHL patients exhibited a significantly increased clustering coefficient, global efficiency, and local efficiency, but a significantly decreased characteristic path length in functional brain connectome, while a significant decrease clustering coefficient and local efficiency in structural brain connectome. In addition, the primary increased nodal strength was observed in limbic and paralimbic systems primarily as well as in the auditory network brain areas. More importantly, the coupling strength of structural-functional connectome was decreased and exhibited a negative correlation with some USSHL clinical variables in patients. Conclusion: Detectable alteration of network organisation already occurred in USSHL patients within the acute period at the global and regional level. The functional connectome is characterised by a shift toward small-worldization while the structural connectome is toward randomisation, which may indicate a plastic reorganisation procedure of the brain to compensate for the loss of hearing in USSHL. Moreover, the degree of coupling between structural-functional connectome was decreased, which may reflect the pathophysiological mechanisms of USSHL. Purpose: To investigate whether multimodal network graph analyses are relevant in the discussion of the combined fMRI-DTI pre-surgical evaluation of language networks. Methods and materials: A retrospective cross-sectional evaluation of our 39 consecutive patients who underwent pre-surgical BOLD fMRI with 4 language tasks (TR=3s, 3x3x3 mm voxels) was conducted. DTI (32 encoding directions, b=800 s/mm 2 , 2 mm isotropic resolution) was done with a standard SPM12 (https://www.fil.ion.ucl.ac.uk/spm) pipeline for fMRI, with the resulting activation maps further analysed using ICN_Atlas (Kozák et al., NeuroImage, 2017) with AAL atlas ROIs (Tzourio-Mazoyer et al., Neuroimage, 2002) . DTI data was analysed using ExploreDTI with whole-brain tractography (Leemans et al., ISMRM, 2009 ) resulting in AAL-based connectivity matrices. Structural connectiome characteristics calculated with the brain connectivity toolbox (Rubinov and Sporns, NeuroImage, 2010) were compared against fMRI-based language network parameters using conventional parametric statistics in Matlab, focusing on Broca's and Wernicke's areas bilaterally. Results: fMRI lateralisation correlated significantly with structural connectome graph parameters in a task-dependent fashion. The temporal lateralisation of speech comprehension and auditory decision and the frontal lateralisation of picture naming and synonym decision showed significant positive correlations with network metrics-based lateralisation parameters of the temporoparietal areas. The lateralisation of frontal language areas was also reflected in the lateralisation of arcuate fasciculus (AF). The temporal lobe language networks seem to have a defining role in whole-brain language organisation, shown by the observed functionalstructural correlations that are in line with previous literature (Wylie and Regner Purpose: Magnetic resonance imaging (MRI) is the gold-standard imaging tool for compressive spinal cord myelopathy (CSCM). Diffusion tensor imaging (DTI), especially fibre-tracking or tractography (DTT), has been considered superior to routine MRI in predicting prognosis and urgency of decompression. This study aims to evaluate DTT as an advanced dimension of MRI in patients with CSCM in predicting the necessity of decompression management. Methods and materials: 12 patients with clinical signs of CSCM underwent conventional MRI and DTI examination with a high-resolution matrix on a 1.5T magnet system. 3D-colour-coded maps were obtained in the sagittal and coronal planes. The presence or absence of parenchymal signal alteration and change in colour-code of the neural fibre tracts in the spinal cord, in addition to the integrity and homogeneous thickness of the fibre tracts on DTT, was evaluated. A near-homogeneous blue colour in the spinal cord and cauda equina was considered normal in our study, with the blue colour representing diffusion in the craniocaudal direction. Alteration in colour, thickness, and integrity, especially at the level of compression on T2W images, was considered significant. Results: Out of 12 patients, 9 had signs of CSCM secondary to infective spondylodiskitis, 1 secondary to a disk protrusion, 1 secondary to a disk extrusion, and 1 secondary to the metastatic collapse of the vertebra. In 8/12 patients with no spinal cord parenchymal signal alteration on T2W images, DTT revealed significant alteration in the colour-code both at and above the level of compression, signifying altered diffusion. Conclusion: DTT is a more sensitive indicator of spinal cord compression than conventional MRI by revealing colour alterations, thinning and loss of integrity of spinal nerve fibre tracts, and providing an easy and fast tool to predict the need for surgical decompression. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: R. Rastogi: nothing to disclose N. Jain: nothing to disclose RPS 1611-13 17:12 The role of the waist-to-hip ratio in brain networks: a connectometry study M. Porcu, E. Scapin, P. Garofalo, A. Caneglias, F. Destro, A. Operamolla, L. Saba; Cagliari/IT (micheleporcu87@gmail.com) Purpose: It is known from the literature that BMI influences cerebral networks. The purpose of this study is to evaluate the influence of the waist-to-hip ratio (W-H ratio) on brain networks by using the connectometry technique. We recruited 30 healthy subjects (12 males, 18 female; age 25-80). The waist and hip circumferences of every patient were measured in order to calculate the W-H ratio and on the same day an MR examination was performed on a 3T scanner (Vantage Titan 3T, Canon) with a 32 channels head coil. The MR protocol included a 40 directions DW sequence for the connectometry analysis. The connectometry analysis was performed using a linear regression model to consider the effects of the W-H ratio, choosing three different T-score thresholds (T-threshold) values (1, 2, and 3). Results were considered statistically valid for a p-value adjusted for the false discovery rate (p-FDR) <0.05. The network topology analysis measured the influence of the W-H ratio on brain networks. The connectometry analysis identified several white matter bundles negatively correlated to the W-H score for T-threshold=2 and 3. Several properties of the network were influenced in the network topology analysis. The W-H ratio influences brain networks. More studies are needed in order to fully understand these results. Limitations: The small number of patients recruited and the DW parameters adopted, in particular, the low b-values. Ethics committee approval: The study was approved by the local ethical committee. Informed consent was obtained. Purpose: Insight is the sudden and unpredictable appearance of a problem's solution. The role of various brain structures in the generation of the insight phenomenon remains poorly understood. We aimed to study the neural correlates of the insight process during fMRI. We performed fMRIs in 32 healthy subjects with a paradigm that involved solving anagrams. After giving an answer, the participants labelled their strategy as 'insight' or 'analytical'. The data was analysed using a GLM with predictors corresponding to different stimulus types and button presses. A random-effects analysis for the population was performed using a t-test over the individual contrast images. Purpose: Our retrospective study investigates whether a correlation exists between functional language lateralisation using fMRI and structural lateralisation using diffusion data of language tracts in controls and tumour patients. Methods and materials: 15 controls and 61 patients were included and underwent structural, functional, and diffusion imaging on 3T MRI. Subjects performed verbal fluency and verb generation language fMRI tasks, and a laterality index (LI) was calculated using data from the cerebral hemisphere and the contralateral cerebellar hemisphere. The diffusion tracts included the long fibres of the arcuate fasciculus, anterior and posterior short fibres of the arcuate fasciculus, uncinate fasciculus, inferior longitudinal fasciculus, inferior frontooccipital fasciculus, and frontal aslant tract. An asymmetry index (AI) for each fibre pair was calculated using the tract volume derived from single tensor (ST) and spherical deconvolution (SD) diffusion data, as well as using the hindrance modulated orientational anisotropy (HMOA) derived from the SD method. A linear regression analysis between fMRI LI and DTI AI was performed. Results: In all subjects, there was no significant correlation between fMRI LI and DTI AI for any of the tracts. Conclusion: Our results do not justify replacing fMRI with tractography in the assessment of language lateralisation. The discordant results between ST-and SD-based asymmetry indices indicate that the structural lateralisation is less robust than the functional one. The predominately left-sided tumour distribution is due to the retrospective design, as included patients had all been selected clinically for presurgical assessment. Results: DSCT and TTE all have high sensitivity and specificity in the diagnosis of intracardiac anomalies and anomalies of great vessels，the difference is not statistically significant (p > 0.05), while the sensitivity of DSCT in diagnosing PA-VSD patients with major aortopulmonary collateral arteries (MAPCAs) was higher than that of TTE (100% vs 68.75%). The incidence of MAPCAs in the group of the large ventricular septal defect (57.7%) is higher than non-large ventricular septal defect group (11.1%) diagnosed by DSCT, the difference is statistically significant(p=0.018, < 0.05. Based on DSCT images, McGoon ratio, Nakata index and TNPAI index of patients were measured. There were significant differences in McGoon ratio and TNPAI index between different treatment groups (p < 0.05). Conclusion: DSCT can accurately diagnose PA-VSD and associated cardiovascular malformations，evaluate the development of pulmonary artery and anatomical characteristics of MAPCAs. It has great reference value for diagnosis and the choice of treatment methods of PA-VSD. Firstly, the sample size of this study is relatively small, which limits the application of the research results. Secondly, because this study is a retrospective study, the surgical methods are affected by the personal habits of the operator partly, which makes the results of the study have certain limitations and need further verification of randomised clinical trials. The feasibility and radiation dose of single-beat coronary CT angiography in patients with atrial fibrillation C. Zhao, Z. Yang; Beijing/ CN (chenglin.zhao@163.com) Purpose: To evaluate the feasibility and radiation dose of single-beat coronary CTA in patients with atrial fibrillation (AF) compared sinus rhythm (SR). From 2016 to 2017, 20 patients with AF (group A) and 20 patients with SR (group B) were referred for CCTA with120 kVp tube voltage and auto mAs. The objective IQ including image noise, signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were compared respectively. The subjective IQ was evaluated by two radiologists in consensus (1 excellent, 2 good, 3 adequate, 4 poor). The effective radiation dose was calculated for each patient. Results: Group A (8 male, 12 female, mean age 59.1±15.1years, BMI 28.4±5.1kg/cm 2 ), group B (8 male, 12 female, mean age 54.8±18.6 years, BMI 29.0±6.5 kg/cm 2 ) (p>0.05) were recruited. The image noise, CNR and SNR of the 2 groups were 28.7±5. Conclusion: Single-beat CCTA using wide-detector CT could achieve high rate of success and good IQ for AF patients and with acceptable ED compared with SR patients. Limitations: A systematic comparison of diagnostic accuracy between CCTA and the gold standard of invasive coronary angiography has not been performed. Ethics committee approval: We performed this study with the approval of the local institutional review board. Funding: No funding was received for this work. Author Disclosures: C. Zhao: nothing to disclose Z. Yang: nothing to disclose Purpose: Coronary artery calcinosis is a specific biomarker of coronary atherosclerosis. It is often detected on chest CT scans. Today the popularity of lung cancer screening with low-dose CT is growing. The aim of the study is to compare the diagnostic value of low-dose chest CT for the assessment of calcium score in comparison with the standard gated calcium scoring. Methods and materials: Non-gated low-dose chest CT and gated calcium scoring were performed in 251 asymptomatic patients. The Agatston calcium score values obtained from low-dose CT and gated scans were analysed independently by two radiologists. Inter-observer and inter-technique agreement were evaluated (data with "zero" calcium scores was excluded). Inter-technique differences in a stratification of patients into five risk categories (scores 0, 1-100, 101-400, 401-1000 and> 1000) using the kappa coefficient (k) were analysed. The sensitivity of non-gated calcium scoring compared to the standard technique was 95%, specificity -99%. The proportion of "zero" calcium score values in our study was 31% (79 patients). The inter-technique concordance was quite high, both with the inclusion of "zero" values (r=0.981, p<0.05) and without them (r=0.978, p<0.05). The inter-observer agreement was 0.998. The intertechnique agreement in the stratification of patients into the five risk groups according to the calcium score values was also high: k = 0.846. Effective radiation dose in low-dose chest CT was significantly lower than in the case of gated calcium scoring (0.96 ± 0.26 vs 1.51 ± 0.22 mSv, p<0.01). The study showed that non-gated low-dose chest CT can be effectively used both for lung cancer screening and coronary calcium scoring with lower radiation exposure to patients than in gated calcium score. Purpose: Minimally invasive multivessel coronary artery bypass grafting (MICS CABG) has shown its safety, effectiveness and high rates of reproducibility, but graft patency data is lacking. In this study we retrospectively analyse CTangiographic graft patency and mid-to early long-term outcomes in patients, subjected to MICS CABG in our institution. Methods and materials: 245 patients were subjected to MICS CABG by small left side thoracotomy approach between July 2014 and December 2018. LITA harvesting, proximal and distal anastomoses were performed under direct vision. SVG/radial artery conduits were harvested endoscopically. Patients were examined with 128-slice CT coronary angiography. Angiographic results achieved for 127 (51.8%) patients (mean follow-up 31.1±7.8 months, range 15 -45 months). Results: All patients achieved complete revascularisation. The mean number of grafts was 2.6±0.5. Perioperative mortality was 0.4% (1 patient). There were 2 conversions to sternotomy (0.8%), 4 reopenings for bleeding (1.6%), 3 perioperative MI (1.2%) and 1 CVA (0.4%). 22 (9.0%) patients received transfusions. 83.2% of patients were followed-up clinically (range 12 -56 months). Long-term mortality was 4.4% (9 patients Purpose: To detect survived myocardium by low-dose dobutamine stress with tissue tracking using cardiac magnetic resonance (CMR) imaging and evaluate the prognosis of survived myocardium. In 24 adult SD rats, the left anterior descending coronary artery was occluded for half hours and released. Low-dose dobutamine stress cardiac cine-MR and late gadolinium enhancement were both performed to analyse wall motion of left ventricle. A three-dimensional and two-dimensional strain of anterior wall and anterior septum, include global peak longitudinal strain (GPLS), radial strain (GPRS), circumferential strain (GPCS), peak radial strain (PRS) and circumferential strain (PCS), were measured at baseline, after reperfusion and dobutamine stress. Triphenyl tetrazolium chloride (TTC) and Evans blue staining were applied to assess viable or nonviable myocardium. PCS of viable myocardium group increased significantly compared with that detected after reperfusion (P<0.05). Significant variances were presented between viable and nonviable myocardial group during stress. In addition, there was a significant difference between two-dimensional strain measured from cine-MR at rest and with stress. Strain variance of nonviable myocardium group during dobutamine stress dose not observe compared that detected after reperfusion (P>0.05). The application of low-dose dobutamine stress cardiac MR with tissue tracking may provide an objective and accurate method to evaluate survived myocardium. Limitations: The data still has to be interpreted with caution given the small sample size. Ethics committee approval: This study was approved by Institutional Animal Research Committee of our local institute. (daniel.preuss@mailbox.org) Purpose: To investigate whether perfusion defect (PD) interpretation of myocardial CT perfusion (CTP) should be different in patients with suspected or known coronary artery disease (CAD) for the detection of obstructive CAD. Methods and materials: A total of 134 consecutive patients with known or suspected CAD prospectively underwent CTP. Rest and stress PDs were retrospectively interpreted in four reading categories (RC): RC 1 (stress volume positive), RC 2 (rest or stress volume positive), RC 3 (stress-induced PDs positive), RC 4 (stress-induced and partially reversible PDs positive). Detection of ≥50% diameter stenosis in the supplying artery on quantitative coronary angiography (reference 1) and intervention (reference 2) served as the reference standards. Results: When using reference 1, per-patient analysis revealed AUCs of RC 1-4 of 0.74, 0.65, 0.62, and 0.72, respectively, in patients with known CAD. In these patients, AUC of RC 1 was higher than RC 3 (P= .03) and, respectively, AUC of RC 4 was higher than RC 3 (P= .006). In patients with suspected CAD, AUCs of RC 1-4 were the same (0.83, 0.86, 0.85, 0.88; all P> .05). When using reference 2, AUCs of RC 1-4 in patients with known CAD (0.61, 0.60, 0.67, 0.69) and, respectively, in patients with suspected CAD (0.75, 0.80, 0.76, 0.80) were the same at per-patient level analysis (all P> .05). Conclusion: In patients with known or suspected CAD, every stress PD should be judged positive to detect obstructive CAD in myocardial CTP regardless of its presentation in rest. Limitations: Study design is retrospective. Ethics committee approval: IRB approval and written informed consent were obtained. IT (filippo.vaccher0@gmail.com) Purpose: To assess the potential of texture analysis (TA) of native images of the interventricular septum to detect structural myocardial changes in patients with aortic valve stenosis (AVS), candidates to transcatheter aortic valve replacement (TAVI). Retrospective single-centre study on 210 CTangiography exams performed for TAVI planning. Scans were obtained with dual-source CT scanner (Somatom Definition Flash, Siemens, Erlangen) with retrospective ECG-gated acquisition and bolus-tracking technique. Twodimensional TA was performed using commercial software (TexRAD Ltd, Cambridge, UK), with a filtration histogram method, on polygonal regions-ofinterest (ROI) manually drawn by a single operator on native axial sections reconstructed in diastole and systole; ROIs encompassed the interventricular septum below aortic valve level. Septal thickness (ST), ejection fraction (EF), and aortic valve area (AVA) were recorded in each patient, in order to correlate such parameters with TA features. Statistical analysis was carried out using the least absolute shrinkage and selection operator (LASSO) linear regression model. Significance was fixed at P=0,05. Results: After LASSO analysis 10 TA features were significantly correlated with ST; a radiomic score was calculated (mean value 14,57) with linear regression coefficient 1,3389 (p=2,22x10 -11 ; R 2 =0,2299). 12 TA features were significantly correlated with EF; a radiomic score was calculated (mean value 52,33) with linear regression coefficient 1,2790 (p=5,39x10 -16 ; R 2 =0,2804). 8 TA features were significantly correlated with AVA; a radiomic score was calculated (mean value 0,6670) with linear regression coefficient 1,9824 (p=2,55x10 -5 ; R 2 =0,08807). Conclusion: TA features extracted for myocardial septum correlate with morphologic and functional parameters in patients affect by SA. The correlation with ST suggests a possible correspondence between TA features and myocardial fibrosis or disarray. Limitations: The single-centre study, study design retrospective, small sample. Ethics committee approval: All patients gave written consent to CT. Purpose: To estimate the clinical value of MR tissue tracking in the assessment of left ventricular remodeling in patients with type 2 diabetes mellitus. Methods and materials: 29 patients with T2DM as well as 21 healthy controls were enrolled. Cardiac magnetic resonance(CMR)was used to obtain left ventricular function parameters while MR tissue tracking to get peak strains (PS. Radial PS， circumferential PS and longitudinal PS were calculated with global and 16 regional segments in DM and healthy controls. Together with LVEF， LVEDV and LV Mass, peak strains were compared between DM and controls. Results: LVEF and global longitudinal PS, basal longitudinal PS were statistical significantly smaller for DM (-16.51±2.76 vs -18.00±2.15，P＜0.05; and -11.26±3.69 vs -13.35±3.16 P＜0.05; respectively) . Global and segmental PS had a statistical significant correlation with LVEF (r=0.63, 0.407, 0.571, 0.517, 0.627, 0.243, 0.615, 0.655 , P＜0 .05 separately). Global and basal radial PS，mid longitudinal PS had a significant-good correlation with HaB1C (r=-0.375, -0.367, 0.381，P＜0.05); separate PS had a poor diagnostic value while the union of apical radial PS，basal circumferential PS, global/basal/apical longitudinal PS had a relatively good sensitivity and specify to distinguish DM from healthy controls. Conclusion: There is a statistically significant difference between DM and healthy controls on peak strain，which serves an early measurement of left ventricular remodelling which could potentially be included as a supplementary diagnostic procedure in the evaluation of DM. Limitations: We didn't discuss the left ventricular systolic function of DM. Ethics committee approval: The study was approved by our institutional review board and written informed consent was obtained from all participants. Purpose: Heart failure is a clinical diagnosis characterised by non-specific symptoms such as dyspnoea, fatigue and oedema. The European Society of Cardiology recommends obtaining objective measures for heart failure in order to confirm a diagnosis. However, it can be challenging to obtain objective measures such as echocardiography in the acute setting. CTPA is conventionally used to outrule pulmonary embolus, however, CTPA provides great detail of the heart and lungs as a whole. The aim of this study was to investigate what role CTPA could play in diagnosing heart failure. Methods and materials: We reviewed 230 CTPA results, of these we confirmed which of these patients had heart failure by confirming a heart failure diagnosis with BNP and echocardiogram criteria. We divided the groups into those who had heart failure and those that did not. We then analysed which features found on CTPA were most specific for a diagnosis of heart failure. Purpose: To evaluate the role of the left atrium and left atrial appendage in the recurrence of atrial fibrillation (AF) after radiofrequency ablation. Methods and materials: 63 patients with AF who underwent radiofrequency ablation for the first time were enrolled. According to the recurrence of AF after radiofrequency ablation, the patients were divided into the recurrence group (n = 20 cases) and non-recurrence group (n = 43 cases). All patients underwent a 256-slice spiral CT examination before the operation. The maximum volume of LAA (LAAVmax), minimum volume of LAA (LAAVmin), LAA emptying fraction (LAAEF), and LAA ejection volume (LAAEV), LAA volume strain (LAA-VS), maximum volume of LA (LAVmax), minimum volume of LA (LAVmin), LA emptying fraction (LAEF), LA ejection volume (LAEV), and LA volume strain (LAVS) were measured. Results: The LAAVmax, LAAVmin, LAVmax, and LAVmin in the recurrence group were higher than those in the non-recurrence group (P < 0.05), while LAAEF, LAEF and LAA-VS in the recurrence group were lower than those in the non-recurrence group (P < 0.05). There was no difference in LA-VS between the two groups. LAAEF was an independent predictor of recurrence after radiofrequency ablation of AF. LAAEF < 44.68% had the highest predictive value for recurrence after radiofrequency ablation. Purpose: To reproduce the feasibility of estimating myocardial extracellular volume fraction (ECV) through routine contrast-enhanced computed tomography (CT) in a cohort of oesophagal cancer patients and to investigate the correlation between an increase in ECV and exposure to radiation therapy. Methods and materials: After ethics committee approval, patients with oesophagal cancer who had undergone routine CT examinations before and after radiation therapy were retrospectively analysed. For each patient, we considered pre-treatment CT and the post-treatment CT with the longest followup interval available. Patients with pre-existing cardiovascular conditions, who previously had undergone cardiotoxic chemotherapy, or with heavily artefacted CT images were excluded. ECV was measured using round regions of interest (ROI) in the septum at mid-level, accounting for movement by excluding borders, and in the left ventricular blood pool at the same level. Septal ECV was calculated using the formula proposed by Bandula et al. T-test for paired data was used to compare distributions. Results: Twenty-one subjects were analysed, with a mean age of 64±18 years, four of which were women. Pre-treatment mean ECV was 27.89±3.54% and post-treatment mean ECV at follow-up was 32.08±4.51% with a median interval of 197 days. Post-treatment ECV was significantly higher than pre-treatment ECV (p<0.001). Conclusion: ECV in oesophageal cancer patients showed to be significantly higher after treatment. Its rise after radiation therapy could play a role in the screening of myocardial condition in oesophageal cancer patients undergoing such treatment. Limitations: This study is limited by its small sample size and its retrospective, monocentric nature. Ethics committee approval: Ethics committee approval was obtained. Written informed consent was waived. The effect of non-optimal tube voltage on radiation dose in lumbar spine radiography E. Alukić, N. Mekis; Ljubljana/SI Purpose: To investigate how exposure parameters or, more precisely, the tube voltage affects radiation dose to the patient in lumbar spine imaging. The data for patient weight, height, image field size, and DAP were collated for 100 patients that were referred for lumbar spine radiography on 2 different x-ray units. The first unit used the tube voltage in line with the European guidelines (79kV) while the other unit's tube voltage was not in line with the EU guidelines (63 kV). The data was collected only for the AP projection. After the data collection, the calculations of BMI, effective dose, and dose to selected organs were calculated using a PCXMC2.0 Monte Carlo simulation program. Results: No statistically significant difference was found between the BMIs of the compared groups (p=0.671). When the optimal exposure parameters were used, the DAP value was lower by 79% (p<0.001), the effective dose by 62% (p<0.001), and the average dose to the organs by 73% (p<0.001). The image field was significantly larger when the optimal parameters were used. Therefore, the dose difference should be even higher when the optimal exposure parameters were used if the field size would be the same. Conclusion: Based on the results, we can conclude that the DAP, effective dose, and dose to organs are significantly higher when the exposure parameters used are not in line with the European guidelines. Limitations: The data was obtained for AP projection only and the image quality was not assessed in the study. Ethics committee approval: The national medical ethics committee and patient consent form were obtained prior to the study. The optimisation of the lateral lumbar spine projection using an air-gap technique A. Bellizzi, F. Zarb; Msida/MT (andrea.bellizzi.15@um.edu.mt) Purpose: To investigate the feasibility of replacing an anti-scatter grid with an air-gap technique to achieve a dose reduction for lateral lumbar spine radiography while maintaining image quality on a direct digital radiography (DDR) system. The study comprised of 2 phases. Phase 1 was an experimental study using an anthropomorphic phantom to identify the optimal air-gap technique. Phase 2 was performing lateral projections of the lumbar spine on patients (n=50). Patients were randomly assigned into a control group (n=25, imaged using the anti-scatter grid) and an experimental group (n=25, imaged using the air-gap technique). The dose area product (DAP) of all examinations was recorded, keeping all other variables constant. Image quality evaluation was performed by 5 radiologists performing absolute visual grading analysis (VGA) using an image quality score tool, with the resultant scores analysed using visual grading characteristics (VGC). Results: A 10 cm air-gap in conjunction with a source-to-image distance (SID) of 121 cm was found as the optimal air-gap technique. The clinical application of this technique resulted in a statistically significant (p<0.05) reduction in the radiation dose of 72% in terms of DAP. Image quality scores were higher for the anti-scatter grid but the variation between the image quality of the two techniques was not significant (p>0.05). The results imply that replacing the anti-scatter grid with an optimal air-gap technique in lateral lumbar spine digital radiography provides a significant dose reduction whilst still maintaining diagnostic image quality. Limitations: Differences in lateral thickness was a variable, limited by matching mean thickness for both groups. More air-gap/SID combinations were available, but particular distances were due to the physical room limitations. Purpose: To determine the change of radiation dose to the patient and impact on image quality when the collimation referred to in the professional literature is used for thoracic spine radiography. The study was performed on 84 patients who were referred for thoracic spine radiography in a Croatian hospital using a CR imaging system. Patients were randomly divided into 2 equal groups. The first group was imaged with the current collimation protocol and the second group with the collimation protocol mentioned in the professional literature. For each patient, weight and height, image-field size, exposure conditions, and DAP were measured while the effective dose and absorbed organ doses were calculated using the PCXMC Monte-Carlo simulation method. Image quality was assessed by 2 radiologists and 1 radiographer using the image software ViewDEX. Results: There was no statistically significant difference in BMIs between the groups. With the optimal usage of collimator, it was found that the size of the primary field in the AP projection was reduced by 45% (p<0.001) and in the LAT projection by 41% (p<0.001). The study also showed reduced values of DAP for AP projection by 34% (p=0.007) and for LAT projection by 23% (p=0.040). The effective dose was reduced by 54% (p<0.001) for AP projection and 29% (p<0.001) for LAT projection. The mean absorbed dose to selected organs decreased by 26% in the AP projection and 28% in the LAT projection. Image quality evaluation showed improvement in AP projection by 13% (p=0.001) and in LAT projection by 15% (p<0.001). Conclusion: Optimal collimation in thoracic spine imaging has a strong influence on the reduction of patient exposure to radiation and the improvement of image quality. Limitations: n/a Ethics committee approval: Ethics committee approval obtained. Purpose: To produce a transparent algorithm with the capacity to quickly and cheaply measure the relative performance of plain radiography imaging systems (IS). We used an ETR-1 test plate in combination with several plain radiography IS based on both CR and DR detector technology. We collected 20 radiographs for each focal spot (FS) setting using every IS. The test plate was placed at 20 mm intervals along the z-axis in order to reveal the potential impact of the anode heel effect and focal effect on image quality. We used exposures of 0.5 mAs at 81 kVp in order to avoid noise elimination and saturation. We upgraded our pilot FIJI macro to include affine image transformation functionality and DICOM header interaction. The macro was used to carry out the consistent characterisation of ETR-1's line-pair segments. This data was exported to a series of spreadsheets which were subjected to B A C 617 Sunday Scientific Programme percentile-based formatting as well as an absolute threshold. We also compared equivalent data points belonging to different FS sizes using subtraction. Results: Our results take the form of 1 spreadsheet for each combination of IS and FS setting, containing contrast maintenance values for 20 different spatial frequencies belonging to each exposure. Percentile-based formatting revealed a deterioration of image quality at the margin of every FOV. This effect was more pronounced on the anode side. Threshold implementation revealed minimum levels of contrast at spatial frequencies consistent with each datasets' respective IS limiting spatial resolution. Subtraction revealed that larger FS produce lower quality radiographs. Conclusion: Our upgraded FIJI macro appears to possess the capacity to produce useful and accurate assessments of image quality in plain radiography. Scientific Programme substitute the portosystemic shunting by means of a new histopathological grading system for zone 3 congestion in BCS. In a retrograde study, we reviewed 34 cases of BCS, diagnosed on radiology/pathology basis, who underwent a biopsy of the explanted liver prior to transplantation. Alongside the extent and amount of irreversible fibrosis, we emphasised other pathologic changes including sinusoidal and central venular dilatation. The pathologic slides were rechecked and, according to a new congestion scoring system, patients were categorised considering the extent of central venular and sinusoidal dilatation. Total scores of 3-8 were applied for each patient. Results: According to radiologic findings, 11 patients were in the proximal type subgroup, all representing a histopathologic score of 4 or less (score 3 in 6 patients, score 4 in the others). The remaining 23 patients were in the distal type category, including 7 patients with score 5 and the rest demonstrating score 6 or more. After pathological/radiological correlation, the study revealed that the more proximal the location of the venous obstruction was, the less sinusoidal and central venular dilatation. Results: TIPS indications were refractory ascites (n=89), variceal bleeding (n=41), and other (n=9). The mean MELD score was 12.0 ± 4.6 (6-33). The mean age was 59.1 ± 10.4 (10-78) and the mean follow-up was 12.2 months (± 8.0, range 1-33). In 9 patients (cavernoma n=4, massive thrombosis n=4, and Budd-Chiari n=1), TIPS was directly dilated to 10 mm in diameter. In 122 patients, TIPS was dilated to 8 mm with a final PSG<12 mmHg or a PSG reduction≥40%, compared to the baseline PSG. In 8 patients, not reaching the haemodynamic target, the stent was further dilated to 10 mm in diameter during the same session, reaching the haemodynamic target. Overall clinical success was achieved in 118/139 (85%) patients (80% in refractory ascites, 95% variceal bleeding, and 80% other). Overt hepatic encephalopathy was observed in 21 patients (15% Purpose: We quantitatively evaluated the impact of inter-rater bias in manual segmentations of medical images on the output of artificial neural networks (ANNs). Consistent differences in manual image annotations influence supervised ANNs' training processes. Thus, automatic segmentations of ANNs trained on different sources will be consistently different as well. MRIs of MS patients annotated by two radiologists with different levels of expertise were collected. CT scans of intracranial haemorrhage (ICH) patients were annotated twice: manually and semimanually. We trained an ANN (U-Net) on annotations from one source and tested its output segmentations on annotations from another source, using Dice scores as matching criteria. We used classifier ANNs to test (using hit-rate) if two sets of automatic segmentations produced by two identical U-Nets that were trained on different source segmentations could be distinguished. We calculated MS-lesion loads and ICH volumes based on the raters' and the U-Nets' segmentations and compared the differences. Results: Lower Dice scores were obtained between the outputs of ANNs trained on annotations of one rater and tested on another (cross-evaluation experiment). Classification hit-rates were higher for U-Nets that were trained on different sources than for the sources' Purpose: To investigate whether the accuracy of machine learning models used in radiology degrades with time. Methods and materials: Machine learning (ML) gains an increasing role in radiology. Currently, most ML models use the same static "train once, deploy everywhere" approach. However, the environments and data used by these models evolve all the time. As a result, static models run into the risk of becoming outdated. To verify this hypothesis, we used a full year of operational patient records from Massachusetts General Hospital. We trained several ML models to predict operational events, such as patient wait times or imaging delay times. We varied the age of the training data between 0 (current) and 12 months, and we varied the volume of the training set from 1 to 12 months. The performance of the trained models was then assessed based on the prediction accuracy for one month of test data. The goal was to see whether the size or the age of the training dataset affects the accuracy of the model results. Purpose: To showcase a fully automated workflow combining deep reinforcement learning (DRL) for whole-body volumetric analyses and cloudbased post-processing and storing of data on 10,508 whole-body organ volumes. Methods and materials: 431 retrospectively acquired multiphasic CT datasets with 10,508 volumes were included in the analysis (10,344 abdominal organ volumes, 164 lung volumes). AI-based whole-body organ volumes were determined using a multi-scale DRL for 3D body markers detection and 3D structure segmentation. The algorithm was trained for whole-body organ volumetry on 5,000 datasets. The data was uploaded to a cloud-based application with integrated DRL software, allowing to group data by diseases. Total processing time for all volumes and mean calculation time per case was recorded. Repeated measures analysis of variance (ANOVA) were conducted to test for robustness, considering the contrast phase and slice thickness. Final whole-body organ metrics were automatically outputted in comma-separated values format. The algorithm calculated organ volumes for the liver, spleen, and right and left kidney (mean volumes in mL: 1868.6, 350.19, 186.30, and 181.91, respectively) , and for the right and left lung (2363.1 and 1950.9) . We found no significant effects of the variable contrast phase or the variable slice thickness on the organ volumes. The mean computational time per case was 10 seconds. The total computational time for all volumes was 1 hour and 11 minutes. We were able to show that DRL in combination with cloud computing enables a fast processing of substantial amounts of data, allowing to build up organ-specific databases. Purpose: The annotating process is tedious and time-consuming. Effectively selecting a minimum number of representative unannotated training datasets out of a large-scale heterogeneous dataset is challenging. Our purpose was to identify training datasets iteratively and retrain current model to segment thigh muscle and produce state-of the-art segmentation performance. Methods and materials: IRB-approved 3,000 clinical CT images were used for this study. A 3D-UNet base model, which was previously trained to segment thigh volume into 11 muscle classes by a small cohort of reference datasets, was used. An iterative active learning framework was developed which identified the datasets out of all datasets where base model segmentation prediction had a higher uncertainty in prediction and could produce a positive effect on segmentation accuracy if used for re-training. Identified datasets were further clustered by their similarity features and 10 representative datasets were selected for annotation in each iteration. All re-trained models performance was evaluated by randomly selected 18 test thigh volume datasets. We compared the average surface distance (ASD) as a segmentation performance metric when the base model was re-trained by the same number of randomly chosen datasets. Results: After the second iteration, the evaluation indicates a 53% improvement in median ASD over all muscle class as compared to a 22% increase when retrained with randomly chosen datasets. Paired t-test with a statistical significance score (p < 0.005) shows clear improvement of the segmentation performance. Conclusion: Retraining a base model with the identified datasets by the active learning framework achieved higher segmentation accuracy in comparison to when trained with a randomly selected dataset. To keep the segmentation performance at the expected range for growing datasets, the framework finds the effective training datasets for annotation and reduces unnecessary annotation workload. Purpose: The Research Consortium for Medical Image Analysis (RECOMIA) is a not-for-profit organisation with the objective of promoting research in the fields of artificial intelligence (AI) and medical imaging. RECOMIA aims at minimising the time and effort researchers need to spend on technical aspects such as display and transfer of DICOM images, legal aspects such as de-identification, GDPR, and HIPAA compliance, and quality aspects such as display tools so that medical experts can review and correct AI-segmentations in a cloud-based DICOM-viewer. Methods and materials: Radiologists and nuclear medicine specialists have used the RECOMIA platform to manually annotate up to 100 organs in more than 300 CT studies. Not all organs were annotated in all images; in total, more than 10,000 were annotated. The manual annotations were used to train convolutional neural networks (CNN) to perform automated organ segmentation in CT images. The CT scans were split into training (80%) and validation (20%) groups. The CNN-based segmentations were compared to the manual segmentations of the validation group. There was agreement on 93.7% of the manually-labelled pixels. Conversely, there was agreement on 94.7% of the automatically-labelled pixels. The pixels considered background both manually and by the CNN are not included in these numbers. The run-time of the CNNs was about 2 minutes on a desktop computer. Conclusion: AI-based tools can provide highly accurate and reproducible organ segmentation, similar to those obtained manually by radiologists, but much faster. We continue to train new CNNs to continuously improve performance. Purpose: To create a general methodology of the generation of databases for machine learning algorithms Methods and materials: 450 anonymised chest CTs were included in the dataset. The RadiAnt DICOM Viewer has been chosen for modifying and dealing with the dataset. The prepared dataset was archived on the server of the oncological centre before the morphological diagnostics. Then the data was anonymised with the DicomCleaner™, renaming them with a special code. We used nodule's shape, internal, and external structure as a feature and represented them into histograms for radiomix. After the morphologic confirmation, we assigned some class label to each case. Results: We created the database LIRA (Lung Image Resource Annotated) for the development and testing of CAD. All cases were confirmed morphologically. In our study, 65% of LCs had a typical CT, 26% of CT images corresponded to different diseases which require an additional differential diagnostic criteria, and 9% of LC cases were extremely difficult to recognise by means of CT due to an atypical visualisation image. Class labels for subsets were "typical LC", "atypical", and "not cancer". Conclusion: One of the conditions for successfully using the medical CAD systems is a correctly created database which corresponds with clinical and radiological interpretation. We would like to point out that the main proposed idea of the methodology of creating the medical databases can be formulated as follows: structurisation of the data, their homogenisation and verification of diseases, and the inclusion of the "atypical" cases and cases which look similar to studying disease. The dataset LIRA contains class labels for each nodule to develop the differential diagnostic intellectual algorithm. CT lung (open dataset). Methods and materials: 95 patients (65 training and 30 validation cohorts) who had undergone pre-treatment 18F-FDG-PET studies were included and classified as those who achieved complete pathological response (pCR) and those who did not. The primary tumours were segmented using a fixed threshold approach, radiomic features were extracted, features were annotated according to the international biomarker standardisation initiative. Feature reduction with the minimum redundancy maximum relevance method was performed. Logistical regression and random forest models for clinical outcomes were constructed and compared using ROC curves, decision curves, and McNemar's test. Of the 5 feature sets tested, AUC values for the two models ranged between 0.6 and 0.87. Reduced feature sets did not have much overlap (10-25%) and showed substantial differences in their ability to predict outcomes with our models and datasets. A combined superset showed the best performance (AUC 0.87) but did not show much reduction in feature correlation. Conclusion: High-dimensional data requires extensive feature reduction and increased scrutiny for overfitting, but if a proper methodology is applied, the results of combining multiple feature sets may be beneficial for modelling with radiomics in cancer. The PET dataset is not very large (low incidence and expensive modality) and the event frequency is slightly below 40%. The fixed threshold segmentation may be less accurate for lesion size, but it was chosen to maximise radiomics stability. Ethics committee approval: The study was approved by the local ethics committee in accordance with the Helsinki Declaration and all patient information was anonymised prior to analysis. Purpose: The value proposition of artificial intelligence (AI) solutions in healthcare have been well described and it is apparent that 'narrow AI' will have a role in every stage of the clinical workflow. The deployment of AI solutions in the clinical workflow of a multi-country healthcare organisation has many challenges, but establishing a systematic and unified framework can lead to successful outcomes. By defining measurable performance indicators, it is possible to track in real-time whether actions give benefits to all stakeholders; patients, referrers, and business partners. In order to establish the framework of AI deployment in the clinical workflow, working groups were formed at a group and country level between clinical, legal, data protection, digital, operational, commercial, and marketing to identify all areas of relevance in a project management approach. Purpose: The gold standard is critical for clinical validation studies but usually time-consuming for researchers. This project aimed to explore whether a deep learning (DL)-based artificial intelligence (AI) diagnostic system would benefit the researchers in establishing the gold standard by reducing the discrepancies between them. In this study, we utilised 196 CXR scans which were diagnosed as either pneumonia or effusion and employed the DL AI system (InferRead DR Chest Research, Infervision) to study its effects in eliminating the discrepancies when generating the gold standard. Two senior radiologists participated in the reader study. They reviewed the CXR images with and without the aid of an AI diagnostic system at an interval of 4 weeks. Discrepancies of their reviewing results were analysed and evaluated by Cohen's Kappa index. Results: There were 18 cases that were diagnosed differently by the two senior radiologists when they reviewed the CXR images alone. With the assistance of the AI diagnostic system, the number of differentially diagnosed cases dropped to 7 cases, leading to a 61.1% reduction in discrepancies. Of note, the Kappa scores increased from 0.806 to 0.928 after utilising the AI diagnostic system. In particular, according to the ground truth of these CXR images, utilisation of the AI diagnostic system eliminated 17 discrepancy cases in which 8 false-positive cases and 7 false-negative cases were corrected. Meanwhile, among the 6 new introduced discrepancy cases, one was a false-positive case that was corrected by one radiologist with the help of the AI system. Purpose: BOLD-MRI can provide regional measurements of oxygen content using deoxyhaemoglobin paramagnetic characteristics. Chronic kidney disease (CKD) can affect oxygenation levels in renal parenchyma, which influences the clinical course of the disease. Furosemide may make Na+ -K+ -ATP pump work reduction and oxygen consumption decrease. The goal of this study was to assess renal oxygenation levels in CKD using BOLD-MRI when injecting furosemide. Methods and materials: 18 healthy subjects and 39 patients with CKD underwent a renal scan using a multi-gradient-recalled-echo (mGRE) sequence with 16 echoes, and 28 of 39 patients underwent the same sequence 7 minutes after injecting furosemide. R2* values of three regions of interest (ROI) of the renal cortex and medulla were measured and their average values were calculated. The difference of R2* was compared between the healthy subjects and all patients. According to eGFR, 28 patients with injected furosemide were divided into a normal eGFR group (group 1,≥90 mL/min/1.73 m 2 ) and an abnormal eGFR group (group 2,<90 mL/min/1.73 m 2 ), and the changes of R2* in the two groups after injection were compared. 28 of the 39 patients were confirmed by ultrasound-guided puncture and the other 11 patients had a similar diagnosis based on clinical symptoms and investigations. Purpose: To retrospectively determine the diagnostic values of a VI-RAD score for the detection of muscle invasion in patients with bladder tumours. Methods and materials: This study included 347 consecutive patients with 339 tumour lesions who were previously diagnosed by CT imaging and subsequently underwent bladder multiparametric MR imaging between January 2015 and March 2019. Lesions level analyses were performed by using an overall score of VI-RADS. Histologic examination of TURBT was used as the reference standard. Two radiologists assessed the diagnostic scores of tumours with muscle invasion by using cutoff values of ≥ 4 and ≥ 3. Cutoff values for VI-RADS scores were estimated from operating points of AUC analyses. Sensitivity, specificity, PPV, NPV, and accuracy were calculated to assess the utility of VI-RADS Results: The interobserver agreement was excellent for three different MR imaging at the lesion level. T2W and DW imaging had a similar diagnostic accuracy of 79.3% for tumour lesions with muscle invasion compared to the overall score of 80.2%, with a cutoff value of 4. The overall VI-RAD score showed an accuracy of 80.2%, with a cutoff value of ≥ 4, granting a sensitivity of 91.3%, specificity of 76.0%, PPV of 83.3%, and NPV of 78.9%. When we applied an arbitrary overall score of ≥ 3, accuracy was 63.7%, sensitivity was 94.6%, specificity was 43.9%, PPV was 51.6%, and NPV was 63.7%. Conclusion: VI-RADS has an overall good performance for the diagnosis of tumours with muscle invasion. Limitations: The current study was a single-centre retrospective study with a small number of participants. As a consequence, a statistical significance was not observed in comparing each result. For these results to be more definitive, prospective studies with larger numbers of patients are required. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Cochlear implantation is the recommended treatment for patients presenting with severe to profound pre-or post-lingual sensorineural hearing loss. Imaging plays a pivotal role in the selection of candidates. A pre-operative knowledge of variant anatomy helps in the planning of the surgical approach and aids in identifying potential complications. Methods and materials: Patients suffering from profound to severe sensorineural hearing loss with the candidacy for cochlear implantation were evaluated by high-resolution CT of the temporal bone (HRCT). A total of 100 cases were referred from July 2018 to February 2019 and were included in the study after taking patient consent. Many anatomical variants were observed including mastoid pneumatisation, level of middle cranial fossa dura, Korner's septum, mastoid emissary vein, sigmoid sinus and jugular bulb position, bony facial canal, position of mastoid segment of facial nerve canal with respect to round window for visualisation of round window during surgery and alignment of cochlear basal turn with respect to axis of internal carotid artery canal for round window accessibility. The pre-cochlear implant imaging had a sensitivity of 97.22% and diagnostic accuracy of 90% in identifying accessibility of round window when correlating with intra-operative findings. Surgeon acceptability was a highlight of this study. Conclusion: Pre-cochlear implantation imaging in addition to performing the accepted role of imaging of the middle and internal ear anatomy is also crucial in the identification of various non-otologic landmarks which affect procedure planning and possible complications. Attention to these small but significant details can greatly improve surgical outcome. Limitations: There is possibility that changes could have happened in the period between imaging and surgery which was appreciated in surgery and was absent during imaging. Ethics committee approval: n/ Purpose: Artefacts of contrast agent in the subclavian vein often affect the display of cervical arteries at the corresponding level. We aimed to evaluate the effects of noise-optimised virtual monoenergetic imaging (VMI+) reconstructions on reducing these artefacts, compared to traditional virtual monoenergetic imaging (VMI) and linearly blended (F_0.6) reconstructions in patients underwent carotid CTA and recommend a better energy level to show calcified plaque. Methods and materials: Ten patients who underwent carotid CTA were evaluated retrospectively. Images were reconstructed with F_0.6, VMI and VMI+ at 20-keV intervals from 40 keV to 140 keV. Attenuation and noise were measured in the cervical arteries at the worst level obstructed by the contrast agent artefacts of each patient. The artefact index (AI) and iodine SNR were calculated. The AI values for the cervical arteries were lowest in the 140keV VMI+ series, which were lower compared to all the VMI and F_0.6 images (all p<0.029), but showed no significant differences compared to the 120keV VMI+ series (p=0.684). The SNR was highest in the 40keV VMI+ series, and there were no significant differences compared to the highest SNR in the VMI images and F_0.6 (P≥0.529), even compared to the 60-120keV VMI+ (p≥0.052). Purpose: ONSD (optic nerve sheath diameter) measured by ultrasonography would be different in patients with migraine with aura than in patients with migraine without aura. Methods and materials: 49 patients were enrolled. Patients were excluded from the study if they were haemodynamically unstable, needed emergency intervention, had an anatomical eye abnormality, had an eye disease that increased the ONSD or had a headache secondary to organic disease. Patients were assessed for headache type according to the criteria determined by the International Headache Society (IHS). The Siemens Acuson P500 portable ultrasound device in ocular imaging mode with a 10-MHz surface tissue probe was used bedside for the evaluation of the ONSD and eye segment diameter. The ONSD was scanned in the sagittal and transverse planes. Additionally, the eye segment diameter measurements were performed on both eyes. All measurements were repeated three times and then the mean value was recorded. Results: There was no significant difference in the eye segment mean diameters between patients with and without pulsating pain (p > 0.05). There was also no significant difference in the eye segment mean diameters between patients with and without aura using the Mann-Whitney U-test (p > 0.05). In this study, we found that there was a statistically insignificant increase in the ONSD in patients with migraine with aura compared to the patients without aura and also in patients with pulsating pain compared with patients without pulsating pain. We could not find any differences in the transverse diameter, anteroposterior diameter, anterior camera depth, and eye globe volume in patients with and without aura. Limitations: This study is limited by the following: the small number of participants, no control group, no follow-up measurements, and the inclusion of just migraine patients. Purpose: Our goal has been to determine if PEG treatment determined a reduction of pneumonia cases in a population with neurologic functional dysphagia. Methods and materials: 20 patients with neurogenic dysphagia have been retrospectively examined. In 13 PEG-treated patients were evaluate the thorax x-rays and CT to find signs of pneumonia. The observation period is between 500 days before and after the endoscopic surgical procedure. Other seven patients were treated only with rehabilitation; the same evaluation was referred to the radiographic dynamic study of the swallowing (DSS -also called a videofluoroscopic study of the swallowing, VFSS). All 20 patients performed a DSS to confirm or exclude the presence of aspiration, this latter only determining the aspiration nature of pneumonia. The exact Fisher test was applied to each group and subgroup. Results: In the first group analysed of the 13 PEG-treated patients, 7 presented pneumonia before the procedure and the prevalence increased of 8/13 after it, six of whom did not present before. Regarding those ones without aspiration at DSS(low-risk for pneumonia) presented an increase of the cases from 3/8 to 5/8. In the remaining five patients with aspiration (high-risk pneumonia with the burden of aspiration pneumonia defined at DSS) the ratio decreased from 4/5 to 3/5. No significant p-value has merged at the Fisher tests. In the second group of seven patients, who were not treated with PEG the Fisher test on pneumonia cases was not significant as well, but the prevalence of pneumonia was higher than in the other subgroups (6/7). The DSS helped to distinguish patients with aspiration, and therefore with aspiration pneumonia, where the PEG treatment determined a reduction of pneumonia cases. In the other subgroups the prevalence of pneumonia increased. Purpose: To establish the advanced diagnostic protocol of MSCT data postprocessing in patients with orbital trauma in order to select the correct treatment tactics and provide optimal preoperative planning. Methods and materials: A total of 107 patients with orbital trauma (100%) were admitted to the hospital on the 1-2 day after the injury. MSCT was performed using Canon Aquilion One 640, data postprocessing was performed using workstation Vitrea Core. The analysis of MSCT data included standard visualisation of bone and soft tissue injury of orbital structures as well as developed techniques including orbital volume measurements, evaluation of inferior orbital wall defects, analysis of globe position and assessment of periorbital soft tissue density. Results: Postprocessing of MSCT data according to the developed techniques compared to the standard visualisation of bone and soft tissue injury allowed to reveal additional increased traumatic orbital volume in 21 patients (19%); additional enophthalmos in 9 patients (8%); determination of 4 types of inferior orbital wall defects: small inferior orbital defects (n=18; 17%), moderate defects (n=31; 29%), severe defects (n=37; 35%) and total defects (n=20; 19%) as well as detailed assessment of periorbital soft tissue density: oedema (n=60; 56%), haematoma (n=10; 9%), atrophy (n=28; 27%), no changes (n=9; 8%), р<.001. The advanced techniques of orbital volume measurements, inferior orbital wall defects evaluation, analysis of globe position and assessment of periorbital soft tissue density statistically significantly increased the efficiency of MSCT data analysis in patients with orbital trauma in order to define the correct treatment tactics. The clinical data of 80 patients with progressive intracranial haemorrhage were analysed retrospectively, mobile CT and conventional CT examination were performed in each of 40 cases. On two groups of a bleeding area of image quality was evaluated by two senior radiologists and records of two patients completed the inspection time, radiation dose and statistical analysis of the results. Results: All patients were successfully completed by a CT head scan. The time required for a mobile CT group examination was (9.21±2.13 min), significantly lower than that of the normal group (47.43±7.10 min), the difference was statistically significant. The effective radiation dose was (1.52±0.00 mSv), lower than the normal group (1.90±0.40 mSv), the difference was statistically B A C 634 Sunday Scientific Programme significant. Two groups of bleeding area image quality by which difference was not statistically significant. The examining flow of mobile CT group was reduced three links and three medical staff than the normal group. Conclusion: Mobile CT examination has higher timeliness and reduces the risk of mobile CT scanning in intensive care patients, saved labor cost and reduced radiation dose, could be better served for the ICU patients who had head CT examining needs. Limitations: In this study, the body mass index, age and patients were not quantitatively analysed before or after surgery, and further studies will be conducted in the follow-up work. Ethics committee approval: n/a Funding: This study was funded by the 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University. Purpose: To test the hypothesis that Jaccoud arthropathy (JA) in systemic lupus erythematosus (SLE) patients is associated with abnormal extensor digitorum (ED) tendons displacement during flexion of the metacarpophalangeal (MCP) joints. Methods and materials: 16 SLE patients with JA (JA+), 12 SLE patients without JA (JA-), and 24 control subjects were included in the study. Transverse spinecho T1-weighted MR sequences of the MCP joints in flexion and in extension were obtained in each hand of SLE patients and in one randomly-selected hand of control subjects. Two radiologists separately measured the amplitude and the direction of the displacement of the ED tendons with respect to the midline at the level of the MCP joints. Statistical analysis included two-way ANOVA with random effects to assess differences in amplitude (p<0.0083), Fisher-Freeman-Halton's exact test to assess differences in direction (p<0.0063), and Gwet's AC1 score to determine interobserver agreement. The amplitude of the displacement of the ED tendons was statistically significantly higher in JA+ patients than JA-patients and controls in flexion for both readers (p<0.0001) and in extension for one reader (p<0.0048). Ulnar deviation of the ED tendons was statistically significantly more frequent in JA+ patients than in JA-patients and controls in flexion and in extension for both readers (p<0.0001). Conclusion: JA is associated with the abnormal displacement of the ED tendons in flexion and extension. Abnormal displacement of the ED tendons is absent in patients without JA. Limitations: In the absence of anatomical correlation, we are unable to precisely determine the anatomic lesions associated with the instability of the ED tendons. Ethics committee approval: This study was approved by our institutional ethics committee. All study participants provided written informed consent. (valentina.mori@med.uni-duesseldorf.de) Purpose: To assess radiocarpal cartilage after distal radius fracture with intraarticular and extra-articular extension using multiparametric, non-enhanced magnetic resonance imaging (MRI). In this prospective study, multiparametric MRIs of the radiocarpal cartilage were performed in 26 participants (16 male, 10 female; mean age 39.5±14.7 years, range 20-70 years) using a 3-Tesla MRI. The cohort consisted of 14 patients with distal radial fractures and 12 healthy volunteers. The radiocarpal cartilage was assessed using morphological (DESS, TrueFISP) and biochemical (T2*) MRI sequences without the application of an intravenous contrast agent. The modified Outerbridge classification system for morphological analyses and region-of-interest biochemical analysis were applied to assess the degree of cartilage damage in each patient. Results: Morphological cartilage assessment showed no significant difference between the DESS sequence and the reference standard, TrueFISP (p= 0.75). In the morphological (DESS, TrueFISP) and biochemical (T2*) assessments, patients with intra-articular fractures did not show greater cartilage damage than those with extra-articular fractures (DESS, p=0.62, TrueFISP, p=0.32, and T2*, p=0.97). Significantly greater cartilage degradation was observed after distal radius fracture compared to controls (DESS, p=0.0001, TrueFISP, p=0.0001, and T2*, p=0.009). Conclusion: Magnetic resonance imaging using advanced, multiparametric sequences facilitates accurate, non-invasive assessment of cartilage changes after distal radius fracture without the need for a contrast agent. Post-traumatic radiocarpal cartilage damage did not differ between fractures with intra-and extra-articular extension, but patients with fractures showed significantly higher cartilage degradation compared to healthy controls. Limitations: A small sample size. No arthroscopic or histological confirmation. Missing inter-and intra-reader reliability for biochemical imaging. Ethics committee approval: 5087R. Approved by the local review board. Funding: No funding was received for this work. Author Disclosures: V. Mori: nothing to disclose C. Schleich: nothing to disclose D. B. Abrar: nothing to disclose M. Frenken: nothing to disclose G. Antoch: nothing to disclose L. Oezel: nothing to disclose S. Gehrmann: nothing to disclose L. Wollschläger: nothing to disclose RPS 1710-9 09:28 Defining the normal reference range for the cross-sectional area of the median nerve at the wrist and forearm using high-resolution ultrasonography in asymptomatic Asian adults M. Rauf, F. Raza, R. Nazir, B. Yawar Faiz, R. Aqeel; Islamabad/PK (mari23392@gmail.com) Purpose: Ultrasonography (USG) is a reliable, inexpensive, readily available, and painless modality for the diagnosis of carpal tunnel syndrome. However, the main sonographic criteria of the cross-sectional area (CSA) of the median nerve shows a wide normal variation depending on regional and ethnical differences, which warrants the establishment of the normal range of variability in the dimensions of the median nerve in the local population. In our experience, the normal average cross-sectional area of the median nerve in the asymptomatic adult Asian population is relatively small compared to other populations worldwide. Methods and materials: 500 asymptomatic patients/1,000 median nerves were evaluated by high-resolution USG using Toshiba Aplio 500 at the distal wrist crease and mid-forearm 12 cm above the distal wrist crease by 3 expert radiologists independently over a period of 2 years (July 2017-July 2019). The median nerve CSA in our population at the right wrist was 6.8±1.9 mm 2 while at the left wrist was 6.6±1.9 mm 2 . The CSA at the right forearm was 5.3 ± 1.4 mm 2 while at the left forearm was 5.2±1.5 mm 2 (P<0.001). Also, the CSA was relatively larger in males than females. The CSA of the median nerve at the wrist and forearm in the Asian population is relatively smaller compared to other populations worldwide. This further enforces the idea to define normal reference range for CSA of the median nerve by obtaining normative data for our population. Limitations: Only Pakistani asymptomatic adults included. Results should be correlated with other available Asian population data as well as data available worldwide for better reference range determination. Ethics committee approval: Approved by our institutional review board. Purpose: Computed tomography (CT) features are the reference standard in evaluating the ossification process of the medial clavicular epiphysis for forensic age diagnostics in adolescents and young adults. Consequently, the highest efforts on radiation reduction are warranted. Therefore, the aim of this study was to investigate the feasibility of low-dose CT of the clavicula for age estimation. Methods and materials: 207 non-contrast chest CT of 144 patients born between 1988-2012, performed in 2018 due to various clinical indications, were included in this retrospective study. The mean patient age was 16.9±6.6 years. Patients were divided into a low-dose (LD; n=146) and high-dose (HD; n=61) group. Image quality and ossification stages (using the 5-stage classification including the subgroups 2a-3c) were assessed by two radiologists independently. Confidence levels were evaluated for subgroups 2a-3c. Radiation dose was determined via dosimetry software. Results: A dose simulation with a z-axis reduction to depict the clavicula only resulted in a median exposure of 0.1 mSv (IQR: 0.0) in LD compared to 0.9 mSv (IQR: 0.6) in HD (p<0.001). The median image quality was rated significantly worse in LD compared to HD with a median of 3 (IQR: 1) versus 4 (IQR: 0) by both readers on a Likert-scale ranging from 1-4 (p<0.001 for both readers). There was an almost perfect agreement for the ossification stages between both readers with a Cohen's kappa of 0.83 (p<0.001). Median confidence levels of both readers were not significantly different between LD and HD (reader 1: p=0.186; reader 2: p=0.074). Conclusion: Low-dose CT of the clavicula for age estimation is possible without a loss of confidence. Purpose: To investigate differences in perfusion-MRI parameters between healthy and pathologic regions in patients with a histologically confirmed medication-related osteonecrosis of the jaw (MRONJ). Methods and materials: in a retrospective analysis, standard parameters of dynamic contrast-enhanced (DCE) perfusion MRI of 22 patients, who received the examinations as part of a dedicated clinical MRI protocol for MRONJ assessment, were evaluated. Same sized regions of interest were placed into a representative bone of defined regions. All regions were assessed qualitatively by a blinded expert radiologist as ground truth (4-point Likert-scale; normal, mildsevere changes). Furthermore, histologic confirmation was present for the resected regions. The mean patient age was 75.2 years (+/-9.9; 12 female patients). Wash-in (0.15 vs 0.05) and positive enhancement integral (PEI; 0.17 vs 0.11) values were significantly higher in MRONJ-affected regions than in healthy jaw areas (p<0.05 for both). Furthermore, DCE-parameters partially correlated with qualitative expert ratings (wash-in, PEI, initial area under the curve, all p<0.05). Conclusion: DCE-MRI reveals significantly different bone perfusion in MRONJaffected regions of the mandible and maxilla compared to the healthy jaw. Disease extent according to MRI data was larger than visible necrotic areas during clinical examination and may help to evaluate the severity of this adverse drug effect. Limitations: This is a retrospective single-centre and single-scanner investigation. All perfusion data was retrieved from MRONJ-diagnosed patients. While clinical and radiological assessment did not indicate pathologic alterations in regions defined as healthy, possible subvisual alterations of the healthy jaw may represent a limitation. However, a control-cohort of healthy volunteers was not used for ethical reasons. Ethics committee approval: Written consent by all patients and IRB approval was given. Purpose: Specific personal identification (fingerprint or DNA matching) is strong evidence in forensic examinations but is mostly lost with decomposition. Sutures on skulls are commonly preserved. The morphology of frontonasal sutures varies among individuals and can be clearly demonstrated by three-dimensional CT (3DCT) images using a new rendering technique. Cinematic volume rendering is a new technique which improves the clarity of 3DCT images and can display skull sutures. We asked whether 3DCT images of the frontonasal suture created by the cinematic volume rendering technique can be used as specific personal identification. Methods and materials: CT images of 50 dry skulls were reformatted to 3DCT images by using a cinematic volume rendering technique on a commercially available workstation (syngo.via, version VB20). Photographs of the dry skulls were adjusted by a blurring effect on Photoshop. Four readers (an anatomist, forensic anthropologist, forensic doctor, and radiologist) responded to three questionnaires, firstly matching a 3DCT image with the photographs of dry skulls, secondly determining minimum levels of clarity of the suture on dry skull photographs that they can use for matching, and thirdly determining the level of clarity which 3DCT displayed using photographs of dry skulls as a reference. Results: Correct matching rates were high for all readers (98-100%). The mean values of visual grading analysis scores for clarity of the sutures on 3DCT (63-77%) are higher than that of the minimum level that each reader required for matching (37-50%) (p<0.001, p=0.0039). Conclusion: 3DCT images of the frontonasal suture created by cinematic volume rending can be used for personal identification. If a clinical CT of the skull of a missing person is available, 3DCT images can be later reformatted and used as pre-mortem data. Purpose: To evaluate the accuracy of image-guided localisation and removal of lymph nodes containing known metastases in breast cancer patients treated with neoadjuvant chemotherapy (NACT). Methods and materials: 64 patients with breast cancer and nodal metastases who underwent NACT between 2017 and 2019 were prospectively enrolled. A clip was placed in the sampled node. After NACT, patients underwent sentinel lymph node (SLN) dissection, removal of the clipped node (CN), and subsequent axillary lymphadenectomy (ALND). Pathological results of SLN and CN were compared to ALND to assess the false-negative rate (FNR). Results: Axillary pathological complete response at ALND was 43,8%. FNR of SLN was 25%, with 9 false-negative (FN) events in 36 patients with residual disease. In 6/9 patients with FN SLNs, the CN contained metastases, resulting in an FNR of 8.3%. Basing on the number of abnormal nodes on initial staging, in patients with 1-3 abnormal nodes (21 patients) residual disease was identified in 10 patients at ALND (47,6%). The SLN didn't reveal metastases in 4/10 patients (FNR=40%) while the CN didn't reveal metastases in 1 patient (FNR=10%). In >4 abnormal nodes (43 patients), the FNRs of SLN and CN were 19,2% and 7,7%, respectively. According to radiological response to NACT, in patients with complete radiological response (cRR) (38 patients), the FNR of SLND was 35.3% (6 FN events in 17 patients with residual disease). In 4/6 patients with FN SLNs, the CN contained metastases, with an FNR of 11,7%. In patients without cRR (26 patients), the FNRs of SLN and CN were 15.8% and 5.3%, respectively. Conclusion: CN could improve accuracy of axillary staging in node-positive patients who received NACT, especially in N1-stage and in patients with cRR. Limitations: Nodal-stage on initial staging. Ethics committee approval: The study was approved by an ethics committee. Purpose: A major role in the diagnosis of myocarditis is played by new CMRbased biomarkers. The aim of the study was to investigate the diagnostic capabilities of tissue-tracking (TT) strain analysis in patients with a CMR-based diagnosis of myocarditis. Methods and materials: 43 patients with a CMR-based diagnosis of myocarditis according to the standard Lake Louise criteria were retrospectively analysed. 27 healthy participants were selected as a control group. Cine-RM data was used for the analysis. Dedicated TT-software was used to perform radial, circumferential, and longitudinal strain for global, per-plane (basal, middle, and apical), and segmental (AHA 16-segments standard) evaluation in short and long-axis left ventricle images. (AUC up to 0.824) . No significant correlations were found with oedema, while a significant correlation was found between segmental radial and LE (p-value:0.019), and an interesting trend between segmental circumferential and LE (p-value:0.087). Conclusion: Tissue-tracking strain analysis has proved accurate in the evaluation of patients diagnosed with myocarditis based on standard Lake Louise criteria. Tissue-tracking strain analysis could improve diagnostic accuracy in clinical practice, adding useful information over the standard findings. Limitations: A lack of mapping sequence, especially in elderly examination, does not permit an adequate comparison with revisited Lake Louis criteria. Ethics committee approval: The study was conducted in accordance with the declaration of Helsinki. Purpose: The pericoronary fat attenuation index (pFAI) emerged as a marker of coronary artery inflammation (CAI), measurable by standard coronary CT angiography (CCTA). It compares well with the gold standard for the CAI assessment and can predict future cardiovascular events. pFAI could prove invaluable in differentiating CAI from a noninflammatory status, helping to unravel the mechanisms subtending an event classified as myocardial infarction with nonobstructive CA (MINOCA) or Tako-Tsubo syndrome (TTS). Methods and materials: Patients admitted with MINOCA and TTS between 2011-2018, who had both CCTA and cardiac MR during/shortly after the acute phase, were selected and pFAI measured in their CCTA. pFAI was also measured in control subjects who had CCTA for atypical chest pain workup, no obstructive CAD found in their CCTA, and no cardiac events at a 2-year followup. Results: In the n=106 MINOCA/TTS, mean pFAI was 68.378.29 vs. 78.036.20 in the n=106 controls (P<0.0001), and the difference confirmed also when comparing mean pFAI in each CA between MINOCA/TTS and controls (P<0.0001). Nonobstructive coronary plaques at CCTA, high-risk plaques in particular, were more frequently found (P<0.01) in the MINOCA/TTS group versus controls. Conclusion: In MINOCA and TTS pts, CCTA is not only able to detect angiographically invisible atherosclerotic plaques, but its diagnostic yield can be expanded using the simple measurement of pFAI to characterise pericoronary fat tissue. In MINOCA/TTS, mean pFAI demonstrates higher values compared with controls, a finding that has been associated with CAI. Limitations: The drug therapy was not recorded. We cannot exclude an effect of drugs used in acute coronary syndromes on pFAI in our MINOCA/TTS group. To determine the T1-relaxation time differences between chronic myocarditis (CM) and non-ischemic dilated cardiomyopathy (DCM) using an overall and segmental approach on native T1 mapping. We performed a retrospective analysis of 62 cases, including 52 consecutive patients who had a presentation at >14 days from initial symptoms and underwent CMR as part of a complete diagnostic work-up for clinically suspected myocarditis in accordance to current guidelines. 26 patients were included in each of the CM and DCM groups and 10 individuals served as normal controls. The CMR protocol at 1.5T included non-contrast T1-mapping using a 5(3)3 MOLLI acquisition. We analysed a mid-myocardial slice for overall and segmental myocardial T1-relaxation. A Wilcoxon rank sum test was used to assess the differences between the groups. Purpose: The diagnoses of acute myocarditis (AM) and hypertensive heart disease (HHD) remain difficult and are commonly based on a combination of clinical expertise and a multi-parameter diagnostic workup including contrastenhanced cardiac magnetic resonance imaging (CMR). It was our purpose to evaluate a multiparametric set of native imaging parameters for their diagnostic accuracy. Methods and materials: A total of 33 AM and 21 HHD patients who had been referred to our department between 09/2014 and 09/2017, as well as 50 carefully selected healthy volunteers, (HV) underwent CMR at 3T. Subsequent featuretracking strain analysis and native T1/T2 mapping were performed and results were processed in the form of binary logistic regressions. Cut-off values, areas under the curve (AUC), and corresponding sensitivities and specificities were derived from receiver operator characteristic curves. Purpose: Hypertensive heart disease (HHD) and acute myocarditis (AM) can both present with increased myocardial mass despite different pathophysiology. The differentiation between them remains difficult. The accompanying distinct patterns of altered myocardial deformation, however, can be investigated using a feature tracking (FT)-CMR strain. Hence, the aim was to evaluate whether new strain ratios including the myocardial mass per body surface area (MyoMass/BSA) yield additional value in the differentiation between both diseases and healthy volunteers. Methods and materials: Patients with AM (n=43) and HHD (n=28) underwent CMR at 3T between 09/2014 and 09/2017. A group of 61 healthy volunteers (HV) served as normal controls. FT-strain analysis was performed and natural strain values were evaluated for gender and age-specific differences. Subsequently, gender-neutral strain parameters were calculated and indexed to the MyoMass/BSA, leading to ratio strains. These were then evaluated for their discriminatory accuracy by means of areas under the curve (AUC), sensitivity, and specificity. Results: There were statistically significant differences in strains between genders (p<0.05) but not between age groups. For the differentiation between HV and AM, the global circumferential strain ratio performed best (AUC 0.86, 79% sensitivity, and 82% specificity). In discriminating between HV and HHD, as well as AM and HHD, the global longitudinal strain ratio outperformed all other parameters (AUCs 0.96/0.79, 92%/89% sensitivity, and 86%/66% specificity, respectively). The calculated ratios provide additional value in the differentiation of diseases with increased myocardial mass. As there is no need for additional sequences, time, or even contrast agents, strain ratios have the potential to be a powerful addition into currently developing multiparametric native diagnostic approaches. with myocarditis was observed. Also, a positive correlation between increased native T1 mapping values and decreased values of peak circumferential strain (p=0.01), peak short-axis radial strain (p = 0.009), and peak long axis radial strain (p=0.02) was observed for the medium slice in these patients. Our study suggests the utility of systematic use of T1 and T2 mapping sequences and feature tracking analysis in CMR as an integral part of the diagnosis and management of myocarditis identifies potentially high-risk patients with latent ventricular dysfunction. Limitations: In our institution, reference values of T1 and T2 mapping aren't yet available. A higher prevalence of susceptibility artefacts in the inferolateral region was observed. Endomyocardial biopsy was performed only in one case of myocarditis. The small sample of patients enrolled may have influenced the results we obtained. Ethics committee approval: n/a Funding: No funding was received for this work. ). TRO-CT identified 19 (32%) obstructive CAD, 1 (1%) acute aortic syndrome, 5 (8%) PE, 1 (1%) CAD+PE, and 1 (1%) CAD+acute aortic syndrome. LIE acquisition was performed in the remaining 33 (55%) TRO-negative patients. 2 (3%) had LIE with ischemic pattern, 16 (27%) LIE with non-ischemic pattern [13 (22%) myocarditis, 3 (5%) idiopathic dilated cardiomyopathy], 3 (5%) increased ECV suggestive for amyloidosis, 2 (3%) cardiac metastasis, 2 (3%) tako-tsubo cardiomyopathy, 1 (1%) pericarditis, and 1 (1%) basal hyperdensity of myocardium suggestive for haemosiderosis. 6 (10%) patients were negative to TRO and LIE and did not report major cardiovascular adverse events after a mean follow-up of 419 days. Conclusion: LIE acquisition increases the diagnostic value of TRO-CT, finding a diagnosis in 82% of TRO-negative patients and allowing a safe discharge of TRO-and LIE-negative patients. Limitations: A single-centre study, small sample size, and short follow-up. Ethics committee approval: The study was approved by the institutional review board. Purpose: This is the first prospective study evaluating if the presence of pericardial effusion (PE) is associated with increased cardiac mortality in thalassemia major (TM). Methods and materials: 1,259 patients (648 females, 31.02±8.64 years) enrolled in the MIOT were prospectively followed from their first CMR scan. CMR was used to quantify myocardial iron overload (MIO) by a multislice T2* approach and to assess biventricular function parameters and PE by cine sequences. Purpose: To evaluate the efficacy of high-resolution ultrasonography (HRUS) in the diagnosis of periapical lesions. Conventionally, the diagnosis of periapical lesions of the teeth has been based on clinical, radiological, and histopathological examinations. However, conventional radiological procedures do not allow for the differentiation between cystic and non-cystic periapical lesions. Recently, USG has been used to provide more detailed information. Limited studies have been done to evaluate periapical lesions using USG with colour Doppler and power Doppler. Methods and materials: 60 patients having inflammatory periapical lesions of maxillary or mandibular anterior teeth and requiring endodontic surgery were included in this study. We used conventional periapical radiographs as well as USG with colour Doppler and power Doppler for the diagnosis of these lesions. Their diagnostic performances were compared against histopathologic examination. A prospective study was done between October 2017-October 2018 in our tertiary centre. DECT was performed using a 128-slice dualsource scanner (Siemens Somatom Definition Flash) for indications including staging, loco-regional extent, nodal status, and metastatic workup. Non-contrast (NC) and post-contrast (PC) sequences were evaluated for nodule size, attenuation value (Av), and iodine concentration (Ic). FNA/histopathology was used as a gold standard. Results: A total of 30 patients (21 females) having a mean age of 47.4 years were included. 15 patients had BTD and 15 patients had MTD including papillary carcinoma (n=9, 60%) and non-papillary malignancies (n=6, 40%). Benign nodules were larger than malignant nodules (mean size 42.9 vs 24.7mm respectively, p=0.013). MTD had higher PC Av (mean HU 116.4) than BTD (82.6), while they had lower NC Av (48.7 vs 62.1, p=0.006). MTD had higher PC Ic than BTD (2.83 mg/ml vs 1.32 mg/ml, p=0.001). The difference between PC Ic and NC Ic was defined as ∆Ic and that between PC Av and NC Av was defined as ∆Av. MTD had higher ∆Ic and ∆Av than BTD (2.45mg/ml vs 0.93mg/ml, p=0.001 and 67.7 vs 20.5, p=0.002 Purpose: To review the preoperative MRI of patients with primary hyperparathyroidism who underwent surgery for parathyroid adenoma (PtA) and to investigate the morphological features that help differential diagnosis. Methods and materials: Parathyroid MRIs of 184 patients from 2015-2018 were reviewed. Only 69 patients were operated on. 5 patients were excluded and 5 patients had bilateral lesions. In summary, 69 lesions from a total of 64 patients were analysed. Axial, sagittal, and coronal fat-saturated T2W TSE and axial pre/post-contrast T1W TSE sequences were obtained in a 1.5T MR system. T2W hyperintensity, oblonged morphology, "reverse-D" appearance, strong enhancement, and the extension to paraesophageal region were investigated. Sensitivity, specificity, and positive and negative predictive value were calculated for each finding (p<0.05). Results: 56 were classified as parathyroid adenoma and 8 as non-PtA, pathologically. MR was negative in 5 lesions. The mean lesion size was 15 mm (4-44.8 mm). Of the 56 adenomas, 51 were T2A hyperintense while 38 showed strong enhancement. Oblonged morphology was present in 40 adenomas, "reverse D" appearance in 36, and paraesophageal extension in 16 adenomas. None of non-PtA lesions showed extension. There was a statistical significance for the strong enhancement (p=0.019) and sensitivity, specificity, and accuracy values were 70%, 75%, and 71%, respectively. The "reverse D" appearance described for the first time in the literature has been shown to be a common finding in adenomas and sensitivity, specificity, and accuracy values were 64%, 63%, and 64%, respectively. Conclusion: Images on the sagittal plane may contribute to a morphological analysis for PtA in parathyroid MRI and oblonged morphology with reverse D appearance may facilitate the diagnosis. Purpose: Visualisation of peroneal vessels prior to autologous transplantation of osteomyocutaneous fibular flap for mandibular reconstruction is essential for surgical success. Our aim was to improve and simplify the pre-surgical diagnostics of peroneal perforators using a dedicated dual-energy CTA protocol and semiautomatic vessel unfolding algorithm. Methods and materials: CTA of the lower limbs was performed in 20 patients using dual-energy acquisitions from a third-generation dual-source CT and a high iodine flux (7mL/s, 350 mg/mL). Monoenergetic reconstructions (40 keV) were automatically reconstructed from the scanner and used for the semiautomatic centreline labelling of the peroneal arteries and their lateral perforators on a post-processing console using a dedicated vascular workflow. Curved MIP reconstruction was regarded as a gold standard. Vessel unfolding reconstruction was performed using a prototype software application. Number and length of perforator arteries were compared between curved MIP reconstruction, thin slice MIP in coronal orientation, posterior VRT, and coronal vessel unfolding of each lower limb. Results: The vessel unfolding algorithm was applicable in all patients and significantly more perforator arteries could be detected compared to thin slice MIP and VRT reconstructions. The mean perforator length was only 7% shorter than in the gold standard, whereas values from thin slice MIP and VRT were significantly shorter. Bone superposition was less frequent in vessel unfolding compared to thin slice MIP. Conclusion: The combination of low monoenergetic reconstructions and vessel unfolding allows for a comprehensive and precise presentation of small peroneal perforator vessels prior to autologous transplantation. Purpose: Non-static metal implants like pacemakers frequently lead to heavy streak-shaped artefacts in reconstructed CT image volumes. The reliable evaluation of neighbouring anatomy, for instance with regard to inflammation or calcification, might thereby be limited. Furthermore, motion precludes application of standard second pass metal artefact reduction (MAR) methods, which implicitly assume a static object during CT acquisition. We propose a MAR pipeline which is robust regarding motion and applicable on a wide range of scanner types, acquisition modes, and contrast protocols. The MAR pipeline uses raw projection data and is therefore independent of 3D motion blur. It is comprised of three convolutional neural network ensembles which are trained from scratch. First, SegmentationNets identify metal-affected line integrals in the input raw projection data. Second, values within the predicted metal shadow are treated as missing data and refilled based on surrounding line integrals by means of the InpaintingNets. The CT volume without metal is obtained by filtered backprojection of the inpainted sinogram. Finally, the ReinsertionNets determine metal positions in the image domain based on the segmented metal shadow. The data for supervised learning is generated by introducing synthetic metal implants into the projection data of 14 metal-free clinical cases with desired acquisition settings. A pacemaker lead model ensures sensible insertion positions, pathways, and motion trajectories by taking the cardiac anatomy and concomitant ECG-data into account. The fully automatic pipeline is tested on 9 clinical cases with real pacemakers, whereby ECG-gated, as well as ungated contrast-enhanced CT, scan types are included. Significant metal artefact reduction is achieved. Limitations: While empirical evaluation was restricted to pacemaker leads, applicability to other metal implants is very likely. Ethics committee approval: n/a Funding: No funding was received for this work. Methods and materials: Breast cancer patients who underwent dynamic contrast-enhanced (DCE) MRI and received NST followed by surgery between 2014 and 2019 (n=116) were analysed. MRI protocols were as follows: T1WI, T2WI, DWI (TR/TE=7000/62 ms; b=0, 1000 sec/mm2), and DCE-MRI performed on 3.0 Tesla scanner (Siemens AG, Erlangen, Germany) with 18 or 16 channel dedicated breast coils. The MRIs were retrospectively evaluated by two radiologists with 20 and 3 years experience of breast MRI, and the target lesions were given a 3-point score . The RECIST-based evaluation was further classified into 3 groups (PD, SD, PR/near CR/CR). Pathological evaluation: pCR was defined as no residual invasive or noninvasive cancer in breast tissue on histopathology from surgical resection (in situ lesion was allowed). Statistical analysis was performed with MedCalc. The diagnostic performance of DWI parameters in discriminating pCR was assessed using a receiver operating characteristic (ROC) analysis. For RECIST-based criteria, they were grouped as pCR or non-pCR. Results: Among 116 cases, 40.5% (47/116) achieved pCR while 59.5% (69/116) were categorised as non-pCR on pathology. The DWI score was strongly associated with residual disease Ethics committee approval: This study protocol was approved by our institutional review board. Funding: No funding was received for this work Using only stiffness as a prognostic parameter, we obtained a sensitivity of 37%, specificity of 87%, PPV 70%, NPV 63%, and for volumes respectively: 47%, 77%, 74%, and 67%. The classifier using both parameters showed a sensitivity of 73%, specificity 75%, PPV 68%, and NPV 78%. Conclusion: The use of two US parameters, volume and stiffness, are highly predictive of tumour response after the 3 rd NAC cycle and performed better than single parameters. Limitations: A small group nothing to disclose H. Piotrzkowska-Wróblewska: nothing to disclose Z. Klimoda: nothing to disclose J. Litniewski: nothing to disclose RPS Breast MRI affects overall survival but not disease-free survival in breast cancer patients: a retrospective population-based study T. van Nijnatten 1 , L. van Tiel 2 , A. C. Voogd 1 , K. Groothuis-Oudshoorn 2 To account for missing data, multiple imputation was performed. Results: Of the 31,756 patients included in the OS-cohort (70% non-MRI and 30% MRI), 27,752 (87%) were diagnosed with invasive carcinoma NST and 4,004 (13%) with ILC. Multivariable Cox regression showed that breast MRI had a tendency towards better OS (HR 0.91, 95%-CI 0.74-1.11) and was statistically significant for patients aged >50 years (overall) Author Disclosures: T. van Nijnatten: Research/Grant Support at Siemens M. B. I. Lobbes: Research/Grant Support at GE Healthcare L. van Tiel: nothing to disclose A. C. Voogd: nothing to disclose K. Groothuis-Oudshoorn: nothing to disclose S The accuracy of a MRI scan for the prediction of pathological response (pCR) in neoadjuvant chemotherapy (NACT) Which MRI morphologic criteria can better predict the response after neoadjuvant chemotherapy (NAC) in axillary lymph nodes CTNs had a mean elastic score of 2, a shear-wave velocity of 3.2±0.5 cm/s, and an elasticity index of 33.3±11.3 kPa. The above-mentioned US CTNs features such as 1) and 4) showed a sensitivity, specificity, and accuracy of 0.95, 0.93, and 0.98, and 0.92, 1.00, and 0.97, respectively, to differentiate the CTNs from cervical metastatic or reactive lymphadenopathies. Conclusion: Continuity with the cervical plexus and the hypoechogenic internal lines with a moderate soft elastic score were the key US features for CTNs. Limitations: The small population was a study limitation Methods and materials: In this prospective study, 211 patients with 268 liver observations detected during US surveillance underwent hepatic CT and MRI examinations, histopathology, and clinical and radiological follow-up. Two radiologists evaluated the observations and independently assigned an LI-RADS category to each observation based on LI-RADS v2017 and LI-RADS v2018. We evaluated the inter-reader agreement (IRA) with Kappa statistics and compared the diagnostic performance of LI-RADS v2018 and LI-RADS v2017 by Fisher's exact test using histopathology as the reference standard. Results: LI-RADS v2018 yielded a better accuracy, sensitivity, and specificity for the categorisation of hepatic observation than LI-RADS v2017 The prognostic value of LI-RADS classification in patient candidates for orthotopic liver transplantation not viable; we excluded patients in whom it was not possible to assign a category (LR-NC) and nodules with LR-TR equivocal. To predict HCC-specific survival, we used Metroticket2calculator, which considers the number of nodules, size of the largest nodule, and α-fetoprotein value. We considered LR-5+LR-TR viable+LR-TIV nodules, then we included LR-4 nodules, and finally, we added LR-3 nodules. We used the c-index to assess prognostic accuracy in survival prediction. We compared c-indexes from the 3 models based on LI-RADS with a c-index obtained by calculating the number and maximum size of nodules obtained from a multidisciplinary tumour board meeting (MDTB) ), respectively. In turn, treatment with 3-bromopyruvate resulted in a gradual increase of pHe in all cancer cell lines, where commonly used viability assays were insensitive to detect any therapyinduced effects. Conclusion: As most liver tumours are hyperglycolytic Therefore, this study introduces a translational model to image tumour-pHe as a functional molecular biomarker for the metabolic phenotype, tumour aggressiveness, and potential use for testing of drug effects in vitro. Limitations: The limited spatial resolution; no direct pH-measurement validation (31P-MRS was performed as validation) Research/Grant Support at NIH R01 CA206180, Research/Grant Support at NIH R01 EB023366 Research/Grant Support at NIH R01 CA206180, Research/Grant Support at NIH R01 EB023366 Research/Grant Support at NIH R01 CA206180, Research/Grant Support at SIO Scientific Programme postprocedural complications, patient outcome and aneurysm recurrence after endovascular (EVT) and neurosurgical treatment (NT) The radiographic features of PAA were collected from pretreatment DSA. In addition, demographic, clinical and radiographic parameters of all patients were recorded. The outcome was measured based on the modified Rankin Scale (mRS) at six months after admission (favourable mRS score, 0-2 vs unfavourable mRS score, 3-6). Results: Overall 46 patients underwent EVT and 21 patients NT. Six months after discharge of the SAH patients 24 had a favourable outcome (mRS 0-2) and 16 patients with unfavourable outcome (mRS 3-6). The mortality rate of patients with SAH was 10.1% (n=4) and was alike compared to former studies (11%) on patients with ruptured PAA. Overall aneurysm recurrence was found in six patients (9.4%) Sunday Scientific Programme Author Disclosures: X. Gao: nothing to disclose D. Ma: nothing to disclose J. Sun: nothing to disclose RPS 1815-4 Paiva 1 , H. Leao Filho 1 ; 1 São Paulo/BR, 2 Barretos/BR (dyandramoreira@hotmail.com) Scientific Programme MRA and CTA measurements were compared with regression and Bland-Altman analysis, with valve sizing by kappa statistics. Results: The median image quality of MRA was 1.5 [IQR 1.5-2.5]. In 4 patients (12%), 1 coronary ostium each (right coronary artery 3, left main artery 1) was not clearly definable on MRA. The interobserver correlation was substantial to excellent (r=0.61-0.92) with a bias of 19 mm 2 for the annulus area (lower limit of agreement -59 mm 2 , upper limit of agreement 98 mm 2 ; p=0.009) Regarding prosthetic valve sizing, there was complete consistency between MRA and CTA-based decisions (κ=1) Conclusion: Self-navigated non-contrast 3D whole-heart MRA enables reliable aortic root measurements without a significant difference to standardised CTA. Prosthesis sizing by MRA measurements would completely match a CTA-based choice. However, in some cases, coronary ostia may not be definable. Limitations: The comparison of MRA and CTA was only performed in 19 patients. Ethics committee approval: Local ethics comittee approval was obtained. Funding: No funding was received for this work nothing to disclose C. Kranewitter: nothing to disclose C. Kremser: nothing to disclose B. Henninger: nothing to disclose G. Reiter: Consultant at Siemens Healthcare Diagnostics GmbH D. Piccini: Employee at Siemens Healthcare Switzerland G. Klug: nothing to disclose A. Mayr: nothing to disclose RPS 1815-9 The role of time-resolved magnetic-resonance-angiography (TRMRA) in the characterisation of soft-tissue vascular anomalies (VA) was conducted. TWIST (Siemens) and 4D-TRAK (Philips) sequences were acquired for a 3 minute postcontrast injection. The time-of-onset (TO), time-to-peak (TTP), the pattern of contrast enhancement, presence of feeding arteries, and draining veins were noted 16/22 cases of fibro adipose VA showed dysplastic veins, non-hypertrophied feeders, and phlebectasia. The mean TO was 8.85+/-4.89s and TTP was 73.66+/-49.38s. 18 vascular tumours enhanced early There was a good correlation with DSA findings (9/12 AVMs, 12/18 tumours). Conclusion: TRMRA provides dynamic perfusion information for characterisation of VA. It provides vascular roadmaps and helps to assess the feasibility of embolisation/sclerotherapy. Limitations: A subjective evaluation Time-resolved contrast-enhanced magnetic resonance angiography in patients with congenital heart disease: image quality using two different doses of a contrast agent Signal intensity of 6 different vascular structures, superior vena cava (SVC), ascending aorta (AAo), descending aorta (DAo), pulmonary artery (PA), left pulmonary artery (LPA), and right pulmonary artery (RPA), were reported along with the standard deviation of signal intensity of the background at each timepoint of TR-MRA. The signal-to-noise ratio (SNR) was subsequently calculated for each structure and timepoint. Peak SNRs for each structure were compared between the groups. Results: A total of 40 patients were included in this study Zanardo: nothing to disclose D. Capra: nothing to disclose F. Sardanelli: Advisory Board at Bracco, Grant Recipient at Bracco, Grant Recipient at Bayer, Grant Recipient at General Electric, Speaker at Bracco, Speaker at Bayer, Speaker at General Electric F. Secchi: nothing to disclose A. Cozzi: nothing to disclose blurring, and overall quality were assessed to test the differences between those with and without DLRecon. The signal-to-noise ratio (SNR) of liver parenchyma and the spleen-to-liver signal intensity ratios were also calculated. Results: The depiction of the anatomical details in the pancreas, blurring, and overall quality were improved with DLRecon Images were reconstructed with filtered back-projection (FBP, B30f) and iterative reconstruction (IR, I30f, level 3, SAFIRE). Subsequently, FBP-images were postprocessed using PixelShine (B30f-PS, PixelShine version 1.2.104, sharpening level 2, noise level 14, processing strength A8, soft kernel settings). CT numbers (mean and standard deviation (noise)) in 6 ROIs (background, paravertebral muscle, fat, spleen, liver, and bladder) and subjective image quality were compared for the different datasets. Results: Image noise reduced significantly for B30f-PS-datasets compared to B30f-and I30f-datasets (-38 to -50% and -12 to -30% for soft tissues, respectively). CT numbers in liver, spleen, bladder, and fat were constant for all datasets, whereas significant differences were notable for background (B30f-PS vs B30f, and B30f vs I30f) and muscle (B30f-PS vs B30f, and B30f-PS vs I30f). In general, PixelShine improved image quality of B30f datasets considerably. Compared to I30f datasets, liver tissue looked more homogeneous, confirming a lower noise level. Beam-hardening artefacts were neither reduced nor enhanced. Conclusion: PixelShine reduces noise while maintaining image information with its deep learning algorithm. Especially for older CT scanners, where IR is not available, PixelShine allows to increase image quality. For new scanners, PixelShine allows to reduce patient dose while maintaining image quality. Limitations: n/a Ethics committee approval: IRB-approved retrospective study. Funding: No funding was received for this work Deep learning reconstruction in ultra-low-dose abdominal CT: comparison with hybrid-iterative reconstruction P. Rogalla 1 3 for intra-abdominal air, and 23.1/9.9/9.3/7.25 for liver tissue, respectively, all p<0.0001 except in group B and C for intra-abdominal air and liver. Conclusion: Deep learning reconstruction provides superior subjective image quality in ultra-low-dose abdominal CT compared to the current standard-of-care iterative method. Limitations: The number of readers (2) and patients (62) Determinants of ADC in the bone marrow of healthy individuals: effects of sex, age, and fat fraction L. Bombelli 1 , A. Colombo 1 , G. Signorelli 1 , A. Rossi 2 , P. Summers 1 , R. Grimm 3 , M. Bellomi 1 , G. Petralia 1 ; 1 Milan/IT, 2 Meldola/IT, 3 Erlangen/DE (bombelli.luca@virgilio.it) Purpose: In literature, there is little documentation of ADC values in normal bone marrow and of the effects of physiological factors. The aim of this study is to evaluate how sex, age, and fat (evaluated from %FF) affect the ADC values measured in the bone marrow of healthy individuals. Methods and materials: We processed the diffusion-weighted images of WB-MRI examinations in 100 asymptomatic individuals, 50 men and 50 women aged 30-79 years. The mean value of ADC in bone marrow (ADCbm) was estimated by fitting the histogram extracted from the images with a semi-automatic segmentation technique. The %FF value was measured as the average of ROIs drawn in specific positions (iliac bone, D10, L4). Gender differences were assessed with the Mann-Whitney U test, while ADCbm correlations with age and %FF were assessed with the Spearman correlation coefficient (ρ). Results: The values of ADCbm were significantly higher in women than in men (respectively 458.1±63.3 µm 2 /s and 383.0±58.4 µm 2 /s, p<0.01). A significant negative correlation of ADCbm with age (ρ=-0.46, p< 0,001) was observed in women but was absent in men (ρ=-0.04, p=0.79). Moreover, there was a negative correlation between ADCbm and %FF in both women and men (respectively: ρ=-0.52, p<0.001 and ρ=-0.28, p<0.05). Conclusion: There was a significant difference between ADCbm values of men and women. The negative correlation between ADCbm and age in women is likely related to menopause, while that between ADCbm and %FF, present in both sexes, appears to be a separate process. Limitations: Calculate ADCbm and %FF with two different approaches. Ethics committee approval: Approved by an internal review board. Funding: No funding was received for this work. Author Disclosures: L. Bombelli: nothing to disclose A. Colombo: nothing to disclose G. Signorelli: nothing to disclose A. Rossi: nothing to disclose P. Summers: nothing to disclose M. Bellomi: nothing to disclose G. Petralia: Speaker at Siemens R. Grimm: Employee at Siemens S 16-7 17:05Radiomics signature of the human papillomavirus (HPV) status in oropharyngeal squamous cell carcinoma (OPSCC) S. P. Haider 1 , A. Mahajan 1 , T. Zeevi 1 , P. Baumeister 2 , C. Reichel 2 , K. Sharaf 2 , B. Judson 1 , B. Burtness 1 , S. Payabvash 1 ; 1 New Haven/US, 2 Munich/ DE (stefan.haider.mdf@gmail.com) Purpose: HPV-positive and HPV-negative forms of OPSCC are biologically distinct entities, with different prognosis, treatment strategies, and divergent AJCC/UICC staging schemes. Using feature selection and machine-learning classification algorithms, we identified the radiomics signature of HPV on pretreatment FDG-PET and non-contrast CT scans. All data was retrieved from the Cancer Imaging Archive. Patients with OPSCC, known HPV-status, and pre-treatment PET/CT were included. The primary tumour and metastatic cervical nodes were separately segmented on PET scans; segmentations were then copied and adapted to the coregistered CT. A set of 1,040 radiomics features was extracted from each segmentation and per imaging modality. To predict HPV-status, minimum-redundancy-maximum-relevance (MRMR) feature selection and random forest (RF) machine learning classifiers were applied in 5-fold cross-validation, repeated x10. The area-under-the-curve of receiver-operating-characteristic-curves (AUC-of-ROC) averaged across validation folds is reported. Results: Of 114 included OPSCC cases, 87 had HPV-positive and 27 had HPVnegative forms. In addition to primary tumour lesions, 119 HPV-positive and 51 HPV-negative metastatic cervical nodes were analysed. The primary tumour RF models achieved an averaged AUC-of-ROC of 0.69 (PET), 0.75 (CT), and 0.73 (PET/CT), based on 20 features selected by MRMR from the respective feature sets. In comparison, the averaged AUC of lymph node RF models was 0.66 (PET) and 0.64 (CT), based on 20 MRMR-features. PET-based lymph node models were non-predictive. Conclusion: Radiomics feature extraction from FDG-PET and non-contrast CT scans, combined with machine learning classifiers, can generate imaging biomarkers for HPV-status in OPSCC primary tumours and metastatic lymph nodes, which may aid pathologists in HPV-classification if standard immunohistochemical staining is equivocal or supplement the immunohistochemical tests in subjects requiring second-line testing. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: S. P. Haider: nothing to disclose S. Payabvash: nothing to disclose P. Baumeister: nothing to disclose C. Reichel: nothing to disclose K. Sharaf: nothing to disclose B. Judson: nothing to disclose B. Burtness: nothing to disclose A. Mahajan: nothing to disclose T. Zeevi: nothing to disclose S 16-8 17:15Repeatability of quantitative WB-MRI analysis in patients with bone metastases G. Saia 1 , A. Colombo 1 , A. Rossi 2 , A. Azzena 1 , F. Zugni 1 , P. Summers 1 , R. Grimm 3 , M. Bellomi 1 , G. Petralia 1 ; 1 Milan/IT, 2 Meldola/IT, 3 Erlangen/DE (giulia.saia.gs@gmail.com) Purpose: WB-MRI is increasingly recommended for the evaluation of patients with metastatic bone diseases. A semi-automatic technique for the segmentation of diffusion-weighted images in WB-MRI examination has been developed to allow quantitative evaluation of tumour burden in bone metastases. The aim of this study was to evaluate the intra and interobserver repeatability of quantitative analysis of ADC in healthy and metastatic bone marrow in patients with bone metastases from breast cancer (BCa) and prostate cancer (PCa). Methods and materials: 4 independent observers processed WB-MRI examinations from 20 patients with bone metastases (10 women with BCa and 10 men with PCa). Segmentation of bone marrow was performed applying a threshold on high b-value diffusion-weighted images and manually removing misclassified non-bone regions. The total volume of bone marrow (Vbm) and firstorder ADC statistics were then obtained. We measured repeatability using intraclass correlation coefficients (ICC) and Bland-Altman method. Author Disclosures: A. Marcucci: Author at University of Pisa S. Vitali: Author at University of Pisa MyT3 17-4 08:42Response evaluation of choroidal melanoma after brachytherapy using diffusion-weighted magnetic resonance imaging (DW-MRI): preliminary findings F. Bitencourt 1 , A. Bitencourt 1 , J. D. O. Souza 1 , N. Neves 2 , M. Chojniak 1 , R. Chojniak 1 ; 1 São Paulo/BR, 2 Salvador/BR (almirgvb@yahoo.com.br) Purpose: To evaluate the role of diffusion-weighted magnetic resonance imaging (DW-MRI) in the evaluation of patients with choroidal melanoma at the time of diagnosis and in the evaluation of therapeutic response after brachytherapy. Methods and materials: We performed a prospective, unicentric study approved by the research ethics committee, which included patients with choroidal melanoma and indication for brachytherapy. Three DW-MRI examinations were proposed for each patient, one before and two after treatment. The apparent diffusion coefficient (ADC) value was calculated on DW-MRI and compared with local tumour control assessed by ophthalmologic followup. Results: From 07/2018 to 06/2019, 19 patients were included, of which 13 underwent follow-up examinations. Patients' ages ranged from 24 to 78 years and 52.9% were male. At the ocular ultrasound, the mean tumour thickness and diameter were 6.3 mm and 11.5 mm, respectively. At initial MRI, most tumours presented high or intermediate signal at T1 (82.3%) and a low signal at T2 (70.6%). Two patients (15.4%) showed signs of tumour progression during follow-up. There was no statistically significant difference in tumour size between MR before and after treatment, however, there was a significant reduction in mean ADC in patients with progression (p = 0.02).Conclusion: DW-MRI has shown to be useful in assessing patients with choroidal melanoma and mean ADC values can be used for response assessment, allowing early identification of patients at risk for progression after brachytherapy. Limitations: This study was limited by the small number of cases. Ethics committee approval: The IRB approval was obtained. Funding: This study was funded by the São Paulo Research Foundation -FAPESP (grant n. 2016/05967-5). Author Disclosures: A. Bitencourt: nothing to disclose F. Bitencourt: nothing to disclose R. Chojniak: nothing to disclose M. Chojniak: nothing to disclose J. D. O. Souza: nothing to disclose N. Neves: nothing to disclose MyT3 17-5 08:46The evaluation of the maculopathy using dynamic contrast-enhanced MRI in patients with proliferative diabetic retinopathy Z. Chen, M. Liu, L. Ma; Beijing/ CN (yyqf@hotmail.com) Purpose: To evaluate the macular change in the patients with diabetic retinopathy using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) technique. Methods and materials: Twenty patients with diabetic retinopathy (DR) and 20 normal controls (NC) were included. The fast spoiled gradient recalled sequence was used to perform dynamic contrast T1WI enhancement on 3.0T MR system. The macular region, optic papilla and nasal retina were performed with quantitative DCE-MRI evaluation using Omni-Kinetics software. Results: The maximal concentration, the area under the concentration-time curve (AUCconcentration-time) and maximal slope of macular region were significantly higher in DR [0.270 (0.03, 1.20) mmol/100ml, 2.71 (0.04, 9.91 ) mmol*min and 0.38 (0.06,3.18) mmol/min, respectively] than that [0.169(0.03,0.72) mmol/、1.25 (0.13,10.41) mmol*min and 0.245 (0.06,1.34) mmol/min] in NC (U value = 515.00 and P value = 0.080, U value = 433.00 and P value = 0.000, and U value = 563.00 and P value = 0.023, respectively). The receiver operating characteristic curve (ROC) analysis demonstrated that the area under AUCconcentration-time was 0.729±0.058 with the cut-off value 1.479 mmol*min (sensitivity 80.00% and specificity 62.50%) for macular region. Conclusion: The quantitative DCE-MRI technique could be used to evaluate the maculopathy associated with the diabetic retinopathy. Limitations: The limits of this study are that the region of interest (ROI) in the retinochoroidal complex, the enhanced concentration of choroid was also included in the regions, would not truly reflect retinopathy entirely based on the current MRI technique. Ethics committee approval: This study was approved by the Chinese PLA General Hospital institutional review board, and written informed consent were obtained from all subjects. Funding: No funding was received for this work. Author Disclosures: Z. Chen: nothing to disclose L. Ma: nothing to disclose M. Liu: nothing to disclose MyT3 17-6 08:50The relationship of severity of migraine and the optic nerve sheath diameter measured by ultrasonography in patients admitted to an emergency department I. Kanbur 1 , H. Topacoglu 2 ; 1 Vienna/AT, 2 Istanbul/TR (drinciferkanbur@gmail.com) MyT3 17-19 09:42Role of diffusion tensor imaging in the evaluation of patients with cervical spondylotic myelopathy: a cross-sectional study V. S. Arunachalam, S. Saxena, R. Dev, P. Sharma, U. Chauhan, S. Sharma, N. Chatterjee; Rishikesh/ IN (drvenkat07@gmail.com) Purpose: To ascertain any correlation of diffusion tensor imaging (DTI) parameters -fractional anisotropy (FA) and apparent diffusion coefficient (ADC) with standard MR findings and clinical severity. Methods and materials: Patients presenting with cervical myelopathy to neurosurgery/orthopedic outpatient department and without recent traumatic history, neuroparenchymal pathology or contraindication to MRI and agreeing to undergo MRI were examined with MRI of the cervical spine. In addition to standard MRI sequences for imaging the cervical spinal cord, diffusion tensor imaging based on single-shot EPI with 20 diffusion directions, each encoded with b-values of 0mm2/s and 1000mm2/s was performed in both sagittal and axial planes. Postprocessing of the raw DTI data was done using ready view developed by GE (Advantage4.7). Stenotic and non-stenotic levels were compared with Kang's grading and DTI parameters. Baseline clinical severity was assessed based on modified Japanese Orthopaedic Association (mJOA) score. Using SPSS(ver23), Anova's test was done to assess different Grade individuals. Pearson's correlation of FA with ADC, Kang's grade and mJOA score were calculated. Results: 43 patients were included in the study. Mean FA values recorded at stenotic (0.346) was less than non-stenotic (0.600) levels. A negative correlation of -0.661 and -0.362 was recorded between FA and ADC and stenotic grading (Kang's), respectively. Level of significance was recorded at 0.013. Conclusion: Diffusion tensor imaging can serve as a potential tool for understanding pathophysiology thereby complementing anatomical information from standard sequences and aid in prognostication of patients with cervical spondylotic myelopathy prior to surgical and non-surgical management. Limitations: Changes in DTI parameters were not recorded following treatment. Technical limitations of DTI may be overcome by using HARDI and Q space imaging. Ethics committee approval: Institutional ethics committee approved the study. Written informed consent was obtained from all patients. Funding: No funding was received for this work. Author Disclosures: V. S. Arunachalam: nothing to disclose S. Saxena: nothing to disclose R. Dev: nothing to disclose S. Sharma: nothing to disclose N. Chatterjee: nothing to disclose U. Chauhan: nothing to disclose P. Sharma: nothing to disclose Purpose: Non-contrast-enhanced MRA techniques have experienced a renaissance due to the known correlation between the use of gadolinium-based contrast agents and the development of nephrogenic systemic fibrosis and the deposition of gadolinium in some brain regions. The purpose of this study was to assess the diagnostic performance of ungated non-contrast-enhanced radial quiescent-interval slice-selective MRA of the extracranial supra-aortic arteries in comparison with conventional contrast-enhanced MRA in patients with clinical suspicion of carotid stenosis. In this prospective study, both MRA pulse sequences were performed in 31 consecutive patients (median age, 68.8 years; 19 men) . For the evaluation, the cervical arterial system was divided into 35 segments (right and left side). Three blinded reviewers separately evaluated these segments. An ordinal scoring system was used to assess the image quality of arterial segments and the stenosis grading of carotid arteries. Results: Overall venous contamination in quiescent-interval slice-selective MRA was rated as "none" by all readers in 84.9% of the cases and in 8.1% of cases in contrast-enhanced MRA (p=0.0001). The visualisation quality of arterial segments was considered good to excellent in 40.2% for the quiescent-interval slice-selective MRA and in 52.2% for the contrast-enhanced MRA (p=0.0001). The diagnostic accuracy of ungated quiescent-interval slice-selective MRA concerning the stenosis grading showed a total sensitivity and specificity of 85.7% and 90.0%, respectively. Conclusion: Ungated quiescent-interval slice-selective MRA can be used clinically as an alternative to contrast-enhanced MRA without a significantly different image quality or diagnostic accuracy for the detection of carotid stenosis at 1.5T. Limitations: The number of patients in our single-centre study was relatively small. Ethics committee approval: The ethics committee approval was obtained (No. D 508/18 Nottingham/UK (nicholas.woznitza@nhs.net) Purpose: To investigate chest radiograph (CXR) reporting by radiographers. A prospective, single-site trial. 12-month consecutive CXRs referred from primary care were independently reported by consultant radiologists (CR;n=13) and reporting radiographers (RR;n=3). Respiratory physicians, blinded to the reporter, compared reports for agreement. Discordant cases were reviewed by thoracic radiologists, blinded to the reporter, who reached the index diagnosis and graded report elements (observation, interpretation, recommendations, and usefulness). The number of CR and RR generated CT scans and lung cancers diagnosed were recorded. Results: 8,685 of 9,136 (95.1%) CXRs were included. Agreement and insignificant disagreement between CR/RR reports occurred in 5,981 (68.9%) and 1,347 (15.5%) of cases, respectively. 1,357 (15.6%) of CR/RR reports had clinically significant disagreement. Thoracic radiology review has been performed for 908 of 1,357 (66.9%) discordant reports. Both reports were correct in 292 (32.2%), CR report correct in 255 (28.1%), RR report correct in 271 (29.8%), and neither report correct in 90 (9.9%). Thoracic radiologists were no more likely to agree with a CR or RR report (p=0.49;CI=-0.03,0.07). RR reports were non-inferior to CR reports for observation ( Purpose: To extra-validate and evaluate the reproducibility of a commercial deep learning-based automatic detection (DLAD) algorithm for pulmonary nodules on chest radiographs and to compare its performance with radiologists. Methods and materials: This retrospective study enrolled 434 chest radiographs (normal to abnormal ratio, 242:192) from 378 patients that visited a tertiary hospital. DLAD performance was compared with two radiology residents and two thoracic radiologists. Abnormality assessment (using the area under the receiver operating characteristics, AUROC) and nodule detection (using jackknife alternative free-response ROC, JAFROC) were compared among three groups (DLAD only, radiologist without DLAD, and radiologist with DLAD). A subset of 56 paired cases, having two chest radiographs taken within a 7-day period, were assessed for intraobserver reproducibility using the intraclass correlation coefficient. Independent characteristics of pulmonary nodules detected by DLAD were assessed by multiple logistic regression analysis. Results: The AUROC for abnormality detection for the three groups were 0.872, 0.929, 0.963, respectively (p<0.05), whereas the JAFROC analysis of nodule detection was 0.912, 0.834, and 0.948 (p<0.05). Reproducibility was 0.80, 0.67, and 0.80, which shows an increase in radiologists using DLAD (p<0.05). Nodules detected by DLAD were more solid, round-shaped, and well-marginated without masking (p<0.05). Conclusion: Extra validation results of DLAD showed high ROC results and there was a significant improvement in the performance when radiologists used DLAD. Reproducibility by DLAD alone showed good agreement and there was an improvement from moderate to good agreement for radiologists using DLAD. Limitations: This is a retrospective study, increasing the possibility of selection bias and limiting the ability to demonstrate a cause-and-effect relationship. Ethics committee approval: The Institutional Review Board approved this retrospective study with a waiver of informed consent. Funding: No funding was received for this work. Author Disclosures: K. E. Shin: nothing to disclose J. S. Park: nothing to disclose J. W. Lee: nothing to disclose Y. Koo: nothing to disclose S. Byun: nothing to disclose RPS 1704-4 08:48 Quantification of regional and temporal lung ventilation in xenonenhanced dual-energy CT imaging N. Buls 1 , S. Bayat 2 , G. van Gompel 1 , S. Verbanck 3 , E. Invers 3 , T. van Cauteren 1 , J. de Mey 1 ; 1 Brussels/BE, 2 Grenoble/FR, 3 Jette/BE (Gert.vangompel@uzbrussel.be) Purpose: To assess the dynamics of lung ventilation using xenon-enhanced dual-energy CT (Xe-DECT) of an animal in a normal and bronchoconstricted state. Methods and materials: Six anesthetised New Zealand white rabbits who received Xe-DECT scans using xenon gas as a contrast agent with a concentration of 70% Xe-30% oxygen, administered by mechanical ventilation via tracheostomy. Dynamic DECT scans (Revolution CT, GE Healthcare) were obtained over multiple breathing cycles during a Xe gas wash-in (WI) and washout (WO) sequence by performing 15 repetitive acquisitions with a 1,01s interscan delay. In addition, bronchoconstriction was induced by inhaled methacholine (40 mg/ml) and all scan sequences were repeated. Material decomposition was applied to isolate Xe from the image data and parametric ventilation maps were created to quantify Xe concentration in the segmented lungs. We compared the evolution of the regional Xe concentration (mg Xe/mL) in a normal and bronchoconstricted state. Results: In a normal state, the maximum Xe concentration in the lungs was 1.60±0.45 mg/mL, which was reached after 15.6±6.9 s. During WO, a 50% decrease in Xe concentration was observed after 12.1±1.3s. The parametric maps clearly showed the increase and decrease of Xe concentration during the WI and WO sequences. After induction of bronchoconstriction, local ventilation defects could be visually detected. Animals showed a reduced Xe concentration in affected areas (1.27±0.68 mg/mL) compared to a normal state (1.91±0.65 mg/mL), and also higher concentration in non-affected areas (2.26±0.71 mg/mL) Erlangen/DE (metz_c1@ukw.de) Purpose: To evaluate the feasibility of non-contrast-enhanced 3D-UTE MRI for pulmonary imaging in immunocompromised patients during haematopoietic stem cell transplantation (HSCT). Methods and materials: 28 patients with an indication for allogeneic HSCT were included in this prospective single-centre study and underwent thoracic MRI before HSCT initiation, in case of suspected pneumonia, and before discharge following completion of HSCT. MRI was acquired using a prototypical stack-of-spirals 3D-UTE sequence within a single breath-hold. Clinically indicated MDCT and a clinically used T2 weighted sequence were used as diagnostic reference. 3D-UTE MR and T2 weighted MR image sets were separately reviewed by two radiologists regarding presence of pleural effusions (PE), ground-glass opacities (GGO), and consolidations. Each MDCT used as reference for MRI results was assessed regarding the same items in consensus reading. Results: A total of 62 MRI scans were acquired without periprocedural complications. 3D-UTE MRI allowed for sufficient imaging of pulmonary consolidations (sensitivity 78%, specificity 82%, PPV 84%, and NPV 76%). Regarding PE (sensitivity 54%, specificity 99%, PPV 92%, and NPV 89%) and GGO (sensitivity 49%, specificity 87%, PPV 45%, and NPV 88%), the diagnostic performance was lower. Overall consistency rate was high (79-90%). Inter-rater agreement was moderate to substantial (PE=0.62, GGO=0.57, consolidations=0.70). 3D-UTE was equivalent (54%) or superior (43%) when compared to T2 weighted MRI in detecting ground-glass opacity areas and consolidations. Conclusion: With an acquisition time similar to MDCT, radiation-free and contrast-free 3D-UTE MRI can be expected to play a major role in future pulmonary imaging, for example, in diagnostics and follow-up of immunocompromised patients with pneumonia. Limitations: n/a Ethics committee approval: n/a Funding: The project was funded by Deutsche Forschungsgemeinschaft. The Department receives a research grant from Siemens Healthcare GmbH, which is not specifically directed towards authors. Purpose: Automated lung segmentation in chest x-rays can bring many benefits for the optimisation of image analysis pipelines, such as lung textures quantification and pathologies classification. Our goal was to improve lung segmentation by using convolutional neural networks (CNN) and a database augmented by means of a generative adversarial neural network (GAN). Methods and materials: Three public datasets of chest radiographs with manual segmentation masks of the lungs were gathered for the purpose of training CNNs. To increase the variability of the dataset, the radiographs from healthy subjects were augmented by means of a GAN. Additionally, 100 manually segmented cases from the ChestXray14 database were used for testing purposes. To evaluate different deep learning approaches, a deeply-supervised UNET and a Mask R-CNN were trained with the augmented samples. The performance of both models was measured by means of the Dice score coefficient between the predictions and the manual segmentations of the ChestXray14 database. Results: The model based on the deeply-supervised UNET architecture obtained an average Dice score of 0.94 for the right lung and 0.95 for the left lung. The model based on the Mask R-CNN architecture obtained an average Dice score of 0.96 for the right lung and 0.96 for the left lung. Conclusion: Optimising the segmentation of the lungs in chest radiographs is feasible by means of convolutional neural networks trained with augmented data. The use of these techniques allows the learning of robust segmentation models and refining them with the generation of realistic augmented data. Purpose: Using a clinical prototype for x-ray dark-field chest radiography, attenuation, differential-phase, and dark-field radiographs of the human chest are obtained to investigate the diagnostic value for chronic obstructive pulmonary disease and emphysema. Therefore, different visualisation concepts of combined dark-field and attenuation images are evaluated. Methods and materials: In the hue-saturation-visibility colour space, the two images are fused together by entering different channels of the colour space: the dark-field signal determines the hue and saturation channel while the transmission value determines the value channel. In the red-green-blue-alpha colour space, the combination is achieved by a superposition of the two images using the transparent alpha-channel. For this purpose, the dark-field signal strength is encoded by a colour map while also defining a transparency map. The colour map is subsequently superimposed on the attenuation image using the transparency map. Results: For each patient, an individually adjustable overlay is available to the radiologists as an additional tool to identify and evaluate possible diagnostic findings as well as to assess the strength of the signal of different lung regions. The combined visualisation of both the attenuation and dark-field chest radiographs allows the correlation of the strength of the dark-field signal, i.e. the amount of intact alveolar structures with the accurate spatial position in the thorax accessible in the attenuation image, and thus provides a valuable tool to further investigate the clinical benefit of dark-field imaging. Limitations: Perfectly registered images as obtained by the prototype for x-ray dark-field chest radiography are required. Ethics committee approval: The study was approved by the ethics committee and by the national radiation protection agency. Participants gave written informed consent. Purpose: To demonstrate an optimal noise reduction technique and optimal level of noise reduction in x-ray dark-field radiography of the human chest. Methods and materials: X-ray dark-field radiographs of the chest were taken of lung-healthy human subjects and patients at different stages of chronic obstructive pulmonary disease (COPD) (58 patients; 64.6 ± 12.2 years; 31% female) with a grating-based prototype dark-field scanner with a field of view similar to conventional chest radiographs (37 x 37 cm²), and the acceleration voltage set to 70 kVp. The level of noise allowed in the reconstruction was defined as σ. During the reconstruction algorithm, different levels of noise reduction were applied (σmin = 10 -5 ; σmax = 7.0) and the raw image, as well as the reconstructed images, were assessed by two experienced radiologists in terms of image definition and noise independently. Results: A noise level of σ = 1 was unanimously perceived to lead to the optimal image quality in in-vivo dark-field chest x-ray radiographs. Conclusion: Moderate noise reduction in in-vivo dark-field chest x-ray radiographs may lead to an improvement in image quality in contrast to the raw image, while a high amount of noise reduction may lead to a loss of anatomical information. Thereby, noise reduction may increase usability in a clinical setting as well as in future investigations with regard to the evaluation of diseases, e. g. improved diagnosis and staging of COPD. Limitations: n/a Ethics committee approval: This ongoing prospective study was institutional review board (IRB) approved and the dark-field contrast x-ray imaging device has been approved for use in humans by the Federal Office for Radiation Protection. Purpose: To develop a robust approach for the image reconstruction within the scopeof a first clinical patient experiment on x-ray dark-field chest radiography investigating its potential benefits for the diagnosis of pulmonary disorders. Methods and materials: Images were acquired using an x-ray grating interferometer tailored to clinical dark-field imaging of the human chest, operating at 70kVp. Three gratings were placed in the x-ray beam path, the first and second generating an intensity pattern, which was then analysed, utilising the third one. Conventionally, this is done by acquiring images at different relative positions of the gratings. Since constraints given by the fabrication process and alignment difficulties prevent the use of full-field gratings, the interferometer was moved continuously to obtain the necessary field of view of 37×37cm 2 . Consequently, moving the gratings with respect to each other needed to be avoided to keep acquisition times sufficiently short (7s). Instead, by a slight misalignment of the gratings, a so-called moiré fringe pattern was formed. The contrast of these fringes was reduced by small-angle scattering within the tissue, which constitutes the dark-field signal and could be extracted by combining several of these displaced fringe patterns. Results: An attenuation image, comparable to a conventional radiograph, and a dark-field image were computed from the overlapping raw detector images obtained from the scanning system. Conclusion: Dark-field images can be obtained from a scanning setup meeting clinical requirements concerning coverage and acquisition speed. Limitations: Due to the scanning acquisition scheme, there are some welllocalised residual fringe artifacts caused by cardiac motion. Ethics committee approval: The study was approved by the national radiation protection agency and the institutional review board. Garching/DE, 3 Hamburg/DE (tina.ml@web.de) Purpose: To evaluate the first dark-field chest x-ray images on living humans and to assess image quality and visual impression, comparing inspiration and expiration. Methods and materials: We have recently designed, constructed, and commissioned a worldwide first experimental prototype for grating-based darkfield chest radiography. 54 patients with either an overall inconspicuous chest CT or signs of pulmonary emphysema were selected and underwent conventional chest x-ray in inspiration as well as dark-field imaging in expiration and inspiration, posterior-anterior respectively. All images were evaluated by two blinded radiologists analysing the images in terms of quality and signal strength, comparing those in expiration and inspiration. Clinical information (history of lung emphysema yes/no) was later taken into account. Results: We found that both images in inspiration and expiration using x-ray dark-field radiography show good quality and appear to be feasible for clinical use. Looking at one of the dark-field images only was enough to identify pulmonary diseases such as pulmonary emphysema, although having both dark-field radiographs in inspiration and expiration helped to confirm the diagnosis and gave additional information to understand dynamic changes during breathing. Conclusion: The first dark-field x-ray images on living humans show promising results with high potential clinical value in inspiration as well as expiration. To detect pulmonary diseases such as pulmonary emphysema, it is enough to have either one inspiratory or expiratory image, however, having both gave additional information about dynamic pulmonalry changes during breathing. Limitations: In this rather early state, we only have a limited number of examinations to analyse. We expect more concrete and significant results during the course of our study. Ethics committee approval: An approval by the institutional review board (IRB) was obtained. Purpose: To compare living human emphysema quantification using the first xray dark-field chest radiograph examinations versus CT. Methods and materials: In total, 500 patients will be included in this study. So far, 79 patients were recruited after receiving a clinical CT-scan with contrast showing no pulmonary pathologies besides emphysema. For each patient, four dark-field images of the thorax (posterior-anterior and lateral, in inspiration, and in expiration) were performed using a clinical prototype. All images were evaluated by three radiologists who were blinded to the GOLD-stadium regarding signal strength. Based on the rating for signal strength, grading of emphysema was performed. The CT-examination of each patient was evaluated with dedicated software (IntelliSpace Portal, Philips GmbH) regarding the percentage of emphysematous lung tissue. Consequently, the correlation between visual dark-field signal and resulting grading of emphysema and CT based emphysema quantification was evaluated. Results: The results from this first clinical dark-field chest radiography demonstrate a high correlation of CT and dark-field images regarding the grading of pulmonary emphysema. Dark-field images obtained in inspiration showed a better correlation with CT compared to dark-field images obtained in expiration. Posterior-anterior images showed a better correlation than lateral images, but the latter are useful for anatomical assignment. Conclusion: Detection and grading of COPD is possible with dark-field x-ray radiography with a high correlation to CT, resulting in decreased radiation exposure (0.04 mSv) as well as fewer time expanses. Limitations: Only a relatively small number of patients was included. For further evaluation, more patients have to be examined. Ethics committee approval: The study was institutional review board (IRB) approved. All patients signed informed consent. Purpose: To investigate dark-field chest x-ray for the detection and staging of chronic obstructive pulmonary disease (COPD) by correlation of dark-field signal strength to lung function tests in a first patient study. Methods and materials: In this ongoing prospective clinical study, 79 of 500 adult study participants (250 patients with different stages of COPD and 250 patients without lung diseases) were enrolled. Dark-field chest x-rays were acquired using a grating-based prototype imaging setup operating at 70 kVp in scanning mode. As a reference, conventional chest x-ray, whole-body plethysmography, and COPD assessment test were performed. Clinically indicated computed tomography scans were also included in the study. Visual assessment of dark-field signal strength using a 5-point ordinal scale was performed by three independent readers. In addition, dark-field signal strength was quantitatively evaluated. Results were correlated with findings of lung function tests. Results: Dark-field signal strength correlated with the Tiffeneau-Pinelli index, which is given as the ratio of forced expiratory volume in 1 second (FEV1) to forced vital capacity (FVC), and FEV1 as a percent of predicted measured in whole-body plethysmography. Conclusion: Dark-field signal strength may serve as a diagnostic marker of pulmonary function in COPD by visualising structural changes of the lung parenchyma. As a novel imaging modality, dark-field chest x-ray may have the potential to facilitate early diagnosis of COPD and improve the monitoring of the clinical course of disease while requiring less cooperation than spirometry. Limitations: n/a Ethics committee approval: This study was approved by the institutional review board and the federal office for radiation protection. Written consent was obtained from all patients. DE (theresa.urban@tum.de) Purpose: To evaluate the first clinical dark-field chest x-ray images with respect to the breathing state of the patient. Methods and materials: We constructed a clinical prototype for dark-field chest radiography suitable for human chest imaging, with an overall acquisition time of 7 seconds and a tube voltage of 70kVp. So far, more than 80 patients have been examined in posterior-anterior orientation in inspiration and expiration, with an average effective dose of 0.04mSv per radiograph. Results: The dark-field signal in the lung can vary in strength due to both a difference in lung thickness as well as microstructural variations. In inspiration, the dark-field signal is quite homogeneous over the lung, which can be attributed to a homogeneous distribution of alveolar size and density. In expiration, the dark-field signal is higher, particularly in the lower parts of the lung. This is likely caused by an increase of air-tissue interfaces in the beam path due to the compression of lung parenchyma. The difference in dark-field signal between the two breathing states allows spatially identification of the regions of the lung that change micro-morphologically and thus participate in the ventilation of the lung. Conclusion: This technique provides spatial information about microstructural changes in the lung and thus the physiological activity of different lung regions during breathing. The change in signal strength illustrates its sensitivity to the microstructure of the lung, promising a distinct diagnostic value of dark-field images for the assessment of lung diseases. Limitations: At this rather early state of the ongoing study, there is a limited number of examinations. Ethics committee approval: The study was approved by the ethics committee and by the national radiation protection agency. Participants gave written informed consent. Purpose: To evaluate the ability of shoulder three-dimensional (3D) magnetic resonance arthrography (MRA) and additional ABER (abduction-external rotation) scans to assess bipolar bone loss and detect on-track/off-track lesions in traumatic shoulder instability, by using MPR CT images as reference. Methods and materials: 25 consecutive patients (14 male, 11 female, range 20-44) with clinical evidence of anterior shoulder instability. All patients underwent 3D-CT of the shoulder and direct shoulder MRA using threedimensional (3D) isotropic PD sequences in standard and ABER position. Two blinded independent observers randomly evaluated the images twice to assess the glenoid track, the Hill-Sachs interval, and to predict engagement by using the "on-track/off-track" method. The intra and interobserver agreement was calculated.Results: Out of 25 defect combinations, 13 were codified as non-engaging and 12 as engaging by using the "on-track/off-track" method. The intraobserver reliability was 0.921 for 3D-CT and 0.786 for MRA. The interobserver agreement ranged from "substantial" to "almost perfect" for both glenoid track and Hill-Sachs interval measurements (p<0.005). ABER MRA predicted engagement accurately in 23 cases (91.3%). Conclusion: MRA using 3D isotropic PD sequences is a reliable method in assessing defects of bone in patients affected by anterior shoulder instability and bipolar bone loss by using the "on-track/off-track" method. The same sequence in the ABER position could predict the presence of engaging lesions.Limitations: The small sample size. (felix.wuennemann@med.uni-heidelberg.de) Purpose: Using arthroscopy as a gold standard, we prospectively evaluated T2 mapping of the shoulder regarding matrix changes of the labrum in patients with and without superior-labral-anterior-to-posterior (SLAP) lesions. Methods and materials: T2 mapping was performed at 3T in 18 consecutive patients (12 men) with shoulder pain followed by shoulder arthroscopy. Two independent raters performed a region-of-interest-analysis in the superior labrum using the colour-coded T2 map. Inter-rater correlation coefficients (ICCs) were calculated. Cut-off values and their sensitivities/specificities including 95% confidence intervals (CI) for detection of SLAP lesions were assessed using arthroscopy as reference. P<0.05 was considered significant. Results: The mean age did not differ significantly between n=12 patients with SLAP lesions and n=6 patients without SLAP lesions (55.5 vs 46.2 years, p=0.22). The mean T2 value in labra with proven SLAP lesions was significantly higher compared to those with no SLAP lesions (mean, 37.7±10.6 ms vs 20.8±2.4 ms, p=0.0001). A complete data separation was present, as the maximum T2 value (21.2) in labra with no SLAP lesion was lower than the minimum value in labra with lesions ( Purpose: To investigate whether the repaired tendon thickness and sub-deltoid fluid collection after rotator cuff repair, which was seen on follow-up ultrasound (US), are correlated with clinical outcome and to assess whether these factors lead to different clinical outcomes depending on the surgical method. Methods and materials: This retrospective study included 54 patients who underwent supraspinatus tendon repair with a suture-bridge or single-row technique and follow up US. Two radiologists independently measured the thickness of the repaired supraspinatus tendons and fluid collection in subdeltoid space. We assessed the relationship between the sonographic parameters, including repaired supraspinatus tendon thickness and sub-deltoid fluid collection, with the clinical outcome represented by the Korean shoulder scoring system (KSS) score using correlation coefficients (R). Results: There was a significant relationship between sub-deltoid fluid collection and the pain of the patients (p≤0.05), although every category showed an inverse relationship with the fluid thickness. The ICC values for sonographic imaging parameters between interpreters ranged from 0.910 to 0.946, showing excellent reproducibility. Conclusion: Sub-deltoid fluid collection of repaired supraspinatus tendons measured on postoperative US can be a useful method in predicting the patient's subjective clinical outcome, especially pain, regardless of age or surgical technique. Limitations: This was a retrospective study. The measurement of sonographic parameters was based on captured images that had been taken prior to assessment and therefore it was not a real-time measurement. Ethics committee approval: n/a Funding: No funding was received for this work. Muscle mass loss after neoadjuvant chemotherapy in breast cancer: estimation on breast magnetic resonance imaging using pectoralis muscle area F. Rossi, L. Torri, S. de Giorgis, M. Calabrese, A. Tagliafico; Genoa/IT Purpose: Sarcopenia is widely considered a predictor of poor survival and toxicity in breast cancer patients. The aim of this study was to evaluate if there is pectoralis muscle area (PMA) variation, reflecting a loss of skeletal muscle mass, on consecutive MRI examinations during neoadjuvant chemotherapy. Methods and materials: A total of n=110 consecutive patients (mean age 56±11years) who were treated with NAC for histologically proven primary breast cancer between January 2017 and January 2019, and in whom tumour response was checked with standard breast MRI, were included. Two radiologists calculated the pectoralis muscle cross-sectional area before and after NAC. Results: The time between the two MRI examinations, before starting NAC, and after completion of three cycles of NAC was 154,8±34 days. PMA calculated pre-NAC (8,14 cm 2 ) was larger than PMA calculated post-NAC (7,03 cm 2 ), (p<0.001). According to RECIST criteria, there were no significant differences between responders (complete or partial response) and no responders (p=0.362). The multivariate regression analysis did not show any significant relationships between ∆PMA and age, the time between MRI exams, estrogen and progesterone receptor status, human epidermal growth factor receptor status (HER-2), Ki-67 expression, lymph nodes status, RECIST criteria, or histological type and grade. Inter-reader (k=0.72) and intra-reader agreement (0.69 and 0.71) in PMA assessment were good. Conclusion: Pectoralis muscle mass variation, reflecting loss of skeletal muscle mass, in breast cancer patients undergoing NAC can be estimated directly on standard breast MRI. Limitations: Known limitations are the retrospective nature of the study and the lack of a control-group. Ethics committee approval: The study was approved by our institutional review board. Written informed consent was obtained. Funding: No funding was received for this work. Author Disclosures: F. Rossi: nothing to disclose M. Calabrese: nothing to disclose L. Torri: nothing to disclose S. de Giorgis: nothing to disclose A. Tagliafico: nothing to disclose RPS 1802-6 11:00The value of different tumour regression of MRI in evaluating the efficacy of neoadjuvant chemotherapy for breast cancer X. Mei, J. Ma, X. Lin, C. Yi; 2 Shenzhen/CN (2635063963@qq.com) Purpose: To analyse the different shrinkage pattern of breast tumors in magnetic resonance imaging (MRI) after neoadjuvant chemotherapy (NAC) and to evaluate the efficacy of different shrinkage pattern in chemotherapy. Methods and materials: Data from 50 patients who were diagnosed with unilateral breast cancer underwent NAC before surgery after the eighth cycle of NAC was analysed retrospectively. All patients underwent MRI scan with contrast enhancement. The shrinkage pattern was classified into 3 categories: concentric shrinkage, fragment-like shrinkage, and mixed shrinkage. All the patients were divided into pathological complete response (pCR) group and N-pCR group according to the histopathologic tumour response. The χ 2 test were used to compare the baseline characteristics of pCR and N-pCR groups, as well as the tumour enhancement pattern and shrinkage pattern of the molecular types. Associations were evaluated using binary logistic regression models. Results: 16 patients achieved pCR and 34 patients had a residual disease (N-pCR). Age, menopausal status, tumour size, lymph node, T stage, molecular typing, ER, PR, ki-67 status, and tumour enhancement mode on MRI showed no significance between pCR and N-pCR groups (all p>0.05), while only HER2 status and shrinkage pattern a significant difference (p<0.05 in both). Interobserver agreement regarding shrinkage pattern was moderate after NAC (Kappa value 0.719). There were no statistically significant differences in tumour shrinkage pattern and MRI tumour enhancement among the molecular types (all p >0.05). There were statistically significant differences in MRI tumour enhancement in the tumour shrinkage pattern (p<0.05). Logistic model showed that tumour shrinkage pattern after NAC and HER2 status were independent factors for predicting pCR. Conclusion: Shrinkage pattern and HER2 status can be independent factors for predicting pCR. Purpose: To combine the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and ADC value to serve as early response markers in breast cancer after NAC. Methods and materials: Between July 2014 and July 2016, 50 patients who received NAC were considered to be eligible. All patients underwent a pretreatment breast routine ultrasonography, MRI, DCE-MRI, and DWI. 50 patients underwent these examinations four times prior to treatment. 1st, 2nd, and following the final cycle of NAC. Responses were measured by routine MRI and ultrasonography using the response evaluation criteria in solid tumours (RECIST). The areas under the ROC curve were 0.627 for the ultrasonography in early assessment of the response to NAC. The areas under the ROC curve were 0.735 for routine MRI in early assessment of the response to NAC. There was a significant difference between responders and nonresponders in K trans , Kep, and TTP after the first cycle of therapy(P=0.022＜0.05). Among various perfusion parameters, the largest areas under the ROC curve were 0.84 for Kep, then that of K trans (0.75). In a combined examination, the largest areas under the ROC curve was 0.88 and that of combined ADC and K trans was the second (0.84). Conclusion: Kep and K trans are superior to routine MRI and ADC value in the early assessment of the response to NAC. Combined ADC with Kep and ADC with K trans are the most recommended parameters that may help determine individual treatments. Limitations: The number of cases was small. All cas-es were invasive ductal carcinoma. This might affect the evalua-tion results. Ethics committee approval: Information about review or animal board approval and written informed consent. Funding: No funding was received for this work. Purpose: Determining disease extent after neoadjuvant chemotherapy (NAC) is essential for accurate surgical planning. Achieving pathologic complete response (pCR) has shown to have favourable prognosis in certain tumour subtypes, however, the definition of pCR varies between authors and can affect the interpretation of results at systematic reviews. We aimed to investigate the diagnostic accuracy of post-NAC MRI with reference to different pCR definitions. Methods and materials: Records of patients with locally advanced breast cancer who underwent NAC followed by mastectomy between January 2016 and January 2019 were retrospectively reviewed. Post-treatment MRI was correlated with residual disease on final histopathology. Two definitions for pCR were used; the first was an absence of both invasive and in situ components, and the second was an absence of invasive component only. Diagnostic accuracy of post treatment MRI was tested against both definitions separately. Results: 52 women (mean age, 47.4 years; range 28-74) with 56 breast masses were eligible for the study. Complete MRI response was noted in 22 (39%) masses. pCR was achieved in 14 (25%) and 25 (44.6%) masses using the first and second pCR definitions respectively. Negative predictive value (NPV) and overall accuracy of MRI for detecting residual disease was 50% and 75%, respectively, using the first pCR definition. With the second pCR definition, NPV and accuracy were 77.3% and 76.8%, respectively. Conclusion: MRI NPV for residual disease was higher with the second pCR definition, however, overall accuracy was not different. MRI accuracy in detecting residual disease after NAC is not adequate to replace pathological assessment. Limitations: Retrospective design, small sample size, and at single institution. Ethics committee approval: This retrospective study was approved by the institutional review board with waiving of the informed consent. We analysed 37 patients with breast cancer that underwent 18 F-FDG PET/CT prior to NST. Maximum standardised uptake values of the primary tumour (SUV-T) and the most FDG-avid lymph node (SUV-LN) and their ratio (NT ratio) were calculated. Patients with and without suspicious ALNM based on prior imaging were compared regarding their SUV values. ALNM response following NST was classified according to the Pinder classification. Patients were defined as responders (Pinder 1-3) and nonresponders (Pinder 4). The association of SUVmax with ALNM response was assessed using ROC analyses. Results: Of 37 patients, 25 (68%) had suspicion for ALNM. Patients with suspected ALNM showed significantly higher SUV-LN (1.5 vs. 6.1) and NT ratio (0.3 vs. 1.0), but no significant difference in SUV-T (6.0 vs. 8.3) was found. Nonsignificant differences between responders and non-responders of SUV-T (7.5 vs. 8.2), SUV-LN (5.6 vs. 5.1), and NT ratio (0.9 vs. 0.7) were found. The AUCs for SUV-T, SUV-LN, and NT ratio were 0.544, 0.630, and 6.11, respectively. An SUV-LN cutoff value of 3.5 had a sensitivity of 75% and a specificity of 49% in identifying patients who responded to NST. Conclusion: 18 F-FDG PET/CT showed promising results in predicting ALNM and their response to NST in breast cancer patients. Limitations: These results are preliminary, hence the low sample size. Ethics committee approval: The need for informed consent was waived by a local ethics committee. Purpose: To assess which clinicopathologic and radiologic factors influence magnetic resonance imaging (MRI) performance in the demonstration of radiologic response (false positive/negative results) in breast cancer after neoadjuvant chemotherapy (NAC). Methods and materials: A total of 80 patients with breast cancer who underwent breast MRI before and after NAC was included in this study. All patients were diagnosed with core-needle biopsy before treatment for histological tumour type and immunohistochemical markers. Pathological examination was referenced as a gold standard after complete surgical excision. Complete radiological response (rCR) to NAC was accepted when the main lesion site showed no early enhancement on dynamic MRI. Complete pathologic response (pCR) was described as the complete absence of invasive cancer in the surgical material. The clinicopathologic and radiologic factors which influence the radiologicpathologic correlation were analysed. Results: The mean age of cohort was 43.5 years (range:24-64). MRI analysis after NAC showed rCR in 44 cases (55%), while pathological analysis of surgical specimens after NAC detected pCR in 43 cases (53.8%). Based on MRI findings, the radiologic-pathologic discordance was in 23 of 80 cases. Multivariate analysis of radiologic findings to identify predictors: dense breast parenchyme ( BIRADS type C-D), multifocal multicentric lesions, high parenchyme enhancement, and non-mass enhancement in post-treatment MR imaging. Also, there were significantly more frequent cases of tumour nuclear histologic grade luminal B subtype. Conclusion: Patients with dense breast parenchyme (BIRADS type C-D), high parenchyme enhancement, multifocal multicentric lesions, luminal B, invasive ductal type cancer, and non-mass enhancement in post-treatment MR imaging are significantly associated with radiologic-pathologic discordance on MR imaging after NAC. These patient populations should be evaluated in depth. Limitations: A retrospective study, limitation in the number of patients. Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: S. Mevlütoğlu: nothing to disclose Ö. Aslan: nothing to disclose I. G. Bilgen: nothing to disclose A. O. Oktay Alfatli: nothing to disclose L. Yeniay: nothing to disclose O. Zekioğlu: nothing to disclose To determine whether a classifier based on two US parameters (volume and stiffness) is predictive of treatment responses from the first courses of neoadjuvant chemotherapy (NAC). Methods and materials: Prospective US analysis was performed in patients with breast cancer qualified for NAC, before and 7 days after the first four courses of NAC. US examinations with strain sonoelastography were performed using linear transducer L14-5. Tumour volume and tumour stiffness (using Tsukuba scale) were evaluated. The resistant malignant cells [RMC] (ranging from 0-100%) were assessed by a pathologist postoperatively. Histopathological examination classifies tumours as responding (RT) and nonresponding (N-RT). RT included those with a reduction in cellularity ≥30% while Purpose: To investigate the accuracy of MRI in evaluating response to neoadjuvant chemotherapy (NAC) in primary breast cancer and to determine if the presence of DCIS/LCIS can influence the accuracy. MRI has a high specificity for the correct detection of residual tumour after NAC, but low sensitivity (correct identification of complete pathological response (pCR) PPV 48% has been reported). Methods and materials: A cohort of 461 consecutive patients who received NAC in a single cancer centre (2007-2019) with post-NAC pre-surgery MRI and surgery pathology data were analysed. 'Gold Standard' for the response was pCR in the surgical specimen. Radiological complete response (rCR) was defined as 0 mm and non-rCR defined as ≥1 mm, using maximum tumour diameter on MRI. Results: In the whole cohort, MRI sensitivity (probability of identifying a pCR from rCR)=71.0% (88/124), and specificity (probability of correctly identifying non-pCR from non-rCR)=81.0% (273/337). The presence of DCIS/LCIS in the surgical sample reduces the accuracy of MRI assessment, sensitivity by 24.8% (with=53.1% vs without=77.9%), and specificity by 2.4% (with=78.6% vs without=81.0%), increasing overestimation of disease size. False-negative rate (probability of rCR incorrectly identifying pCR) is 46.9% with DCIS/LCIS versus 22.1% without DCIS/LCIS. Conclusion: This analysis showed that there is a need to improve the accuracy of MRI assessment of pCR. There is an urgent need to identify variables which can improve this invaluable tool to optimise treatment planning. Introducing a volumetric and texture AI-driven analysis may increase the accuracy and confidence of MRI with positive clinical benefit leading to better patient treatment and less invasive approach to surgery. Limitations: A retrospective study. Further researches needed. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Women with breast cancer receiving neoadjuvant chemotherapy undergo breast MRI as part of treatment response assessment. Incidental sternal bone marrow enhancement may be demonstrated and be confused with pathologic findings. Our goal was to evaluate bone marrow enhancement on MRI prior to and upon completion of neoadjuvant treatment. Methods and materials: A retrospective study between the years 2012-2017 included 109 consecutive women with breast cancer and treated with neoadjuvant chemotherapy. MRI was performed before and after treatment to evaluate response. The neoadjuvant regimen included AC-T in adjunct with G-CSF (bone marrow stimulating agent) and steroids. Bone marrow signal intensity was measured at the sternum, before and after contrast injection. Chest wall muscle signal intensity was used as a reference. The index was calculated as bone marrow signal intensity divided by muscle signal intensity prior, upon completion, and 1 year after completion of neoadjuvant treatment. MRI was performed on average 80 days from the end of treatment. Statistical correlation between bone marrow index, age, disease stage, and time from the end of treatment. Results: The mean patient age was 50 and 98% were diagnosed with invasive ductal carcinoma. Bone marrow index prior to neoadjuvant treatment was 1.9, upon completion of treatment 3.5, and 1.4 at 1 year follow up. Age was found to be inversely correlated to bone marrow index, the bone marrow index was found to be significantly higher in women younger than 48y. No statistical correlation was found between the time of MRI from the end of treatment and disease stage. Conclusion: MRI bone marrow signal intensity may be high in women with breast cancer upon completion of NAT. It may be secondary to treatment with GCSF. This should not be confused with pathological enhancement suspicious of metastatic disease. Limitations: n Ethics committee approval: n Funding: No funding was received for this work. Author Disclosures: I. Bokov: nothing to disclose P. M. Sklair Levy: nothing to disclose D. M. Ben-David: nothing to disclose D. O. Haisraely: nothing to disclose Purpose: To investigate which MRI morphologic criteria can predict residual disease in ALNs in patients who have undergone NAC. Methods and materials: From 2017 to 2019, 60 consecutive patients with advanced breast cancer who underwent 1,5 T MRI before and after NAC were retrospectively collected. Pre-and post-NAC MRI were retrospectively analysed by two dedicated radiologists in consensus, blinded to histological results, evaluating both quantitative (number and diameter) and qualitative criteria (irregular margins, absence of fatty hilum, cortical thickness>3mm, perifocal oedema, rim enhancement, and asymmetry comparing with contralateral side) related to ALNs. ALNs status was evaluated by sentinel ALN biopsy or axillary dissection and nodal pathological response classified according to Pinder's criteria [complete response (pCR) versus no-complete response (no-pCR)]. Statistical analysis (Chi-square or Fisher's exact tests for categorical variables, non-parametric Mann-Whitney test for continuous variables) was performed to compare pCR versus no-pCR cases. Results: At baseline MRI, no-pCR cases ALNs were characterised by loss of fatty hilum (93.3% versus 44.4% in pCR, p=0.02), presence of rim enhancement (86.7% versus 11.1% in pCR, p=0.001), asymmetry (93.3% versus 44,4% in pCR, p=0.02), average cortical thickness significantly larger (9.5 versus 3.7 mm, p=0.004) than pCR cases. At post-NAC MRI, average cortical thickness was significantly (p=0.01) larger in no-pCR patients compared to pCR (4.5 versus 1.8 mm). Considering as suspicious ALN with at least one of the above characteristics, MRI sensitivity, specificity, positive and negative predictive values, and accuracy for assessing ALNs response were 88.2%, 77.8%%, 88.2%, 77.8%, and 84.6%, respectively. Conclusion: Based on our results, supporting the use of MRI in assessing nodal response after NAC, abnormal ALNs cortical thickness seems to be the most reliable parameter associated to no-pCR. Purpose: To investigate the diagnostic value of myocardial deformation analysis based on the 17-segment heart model using non-contrast enhanced (CE) 2D tissue feature tracking (2D-FT) technique. Methods and materials: 70 patients with suspected myocarditis underwent a cardiovascular magnetic resonance (CMR) examination at 1.5 Tesla. A contrastagent-free part of this CMR protocol was additionally performed in 40 healthy volunteers (HV). Besides standard CMR data sets, 2D-FT derived segmental and global longitudinal, radial, and circumferential deformation parameters were analysed. The 2D-FT results were compared to the combined findings from CMR imaging and endomyocardial biopsy (EMB). Results: Patients were assigned to 3 groups depending on their ejection fraction (EF) (<40%, 40-55%, and ≥55%). Compared to HV, impaired EF (<55%) was significantly correlated to reduced segmental and global strain and strain rate values. The circumferential deformation analysis was more sensitive to myocardial changes than longitudinal and radial analysis. The segmental strain/strain rate had an accuracy of 84.3%/70.0% for the diagnosis of acute myocarditis, stated by EMB and CMR in 42 of 70 patients. In patients with preserved EF, acute myocarditis could be ruled out using only segmental strain analysis with a negative predictive value of 87.5%. Conclusion: In patients with suspected myocarditis, the deformation analysis based on the 17-segment heart model provides valuable information about functional myocardial inhomogeneity. This quantitative approach could be used in addition to the clinical standard CMR protocol and represents a promising tool in the framework of a prospective automatised multiparametric CMR imaging analysis. Purpose: To investigate the predictive value of head and neck CT angiography (CTA) in intraoperative haemorrhage of carotid body tumours. Methods and materials: Head and neck CT images of 36 patients with carotid body tumours confirmed by pathology were retrospectively analyzed. The intraoperative bleeding volume ＜500 ml group and intraoperative bleeding volume ≥500 ml group were compared and analysed. The patient's age, sex, Shamblin classification, size of the lesion, number of blood supply arteries, course of the disease, plain scan, and enhanced CT value were analysed. The ROC curve was established to determine the number of blood supply arteries, the maximum diameter of upper and lower, the longest diameter of the axis, and the sensitivity and specificity of the Shamblin type in judging the blood volume during the operation. The amount of bleeding during the operation of carotid body tumours is related to the age of the patients, the maximum diameter of tumours, the longest diameter of axes, Shamblin classification, and the number of blood supply arteries. The AUC of the number of blood supply artery is larger than the other parameters, which can better predict intraoperative blood loss in patients with CBT. Conclusion: The number of blood supply artery detected in CTA is the best parameter to predict intraoperative blood loss. Purpose: To determine the effect of iterative metal artefact reduction (iMAR) technique in reducing dental metallic artefacts in head and neck CT examinations. Methods and materials: 30 patients with dental prosthesis implants were prospectively collected and the original data was reconstructed respectively in an iMAR group (Group A) and a conventional iterative reconstruction group (Group B). The CT and SD values of the two groups of images were measured respectively. The difference between the two groups was compared by a paired t-test.Results: There was no significant difference in CT values on the opposite side of the same layer without artefacts (P>0.05). The CT values in Group A were significantly reduced in the high-density artefact area (p<0.05), while significantly increased in the low-density artefact area (p<0.05), and were closer to the CT values in the corresponding anatomical areas at the opposite side of the same layer. In terms of image noise, SD values in each measurement area of group A's images were lower than those of group B and significantly lower in highdensity and low-density areas (p<0.05). In terms of subjective score, the results of the two groups of image evaluation are excellently consistent (Kappa=0.945, p<0.05). The subjective score of group A was significantly higher than that of the control group ( The numbers of LR-3, LR-4, LR-5, and LR-M categories were 16 (5.4%), 28 (9.5%), 183 (62.0%), and 68 (23.1%), and the positive predictive value (PPV) of LR-3, LR-4, and LR-5 was 43.8%, 60.7%, 98.4% for HCC, respectively. The sensitivity, specificity, and positive predictive value of the LR-5 category for HCC were 73.5%, 94.0%, and 98.4%. Conclusion: CEUS LI-RADS v2017 classification standard has reliable risk prediction value for patients with high-risk factors of HCC, which the LR-5 category has a higher positive predictive value for patients at risk of HCC. However, the differential diagnosis between HCC and other non-HCC malignancies still remains to be further studied for LR-M observations. Limitations: Due to the nature of the retrospective study, LR-1 and LR-2 categories nodules were not pathologically diagnosed and were not included in this study. Due to pathological reasons, the proportion of LR-3 category nodules in this study is relatively small, which may lead to selection bias. Ethics committee approval: n/ Purpose: The liver is the most common metastatic site for neuroendocrine neoplasms. Ultrasonography is widely used to detect focal liver lesions, but it is difficult to characterise hepatic neuroendocrine neoplasms (hNEN) from hepatocellular carcinomas (HCC). We investigated the value of contrastenhanced ultrasound (CEUS) to distinguish hNEN from HCC in non-cirrhotic liver parenchyma. Methods and materials: From January 2014-June 2019, 92 patients who underwent liver CEUS examinations followed by a pathological diagnosis of neuroendocrine neoplasms in our institution were retrospectively evaluated in the study. At the same time, 92 patients with histologically-proven HCC without liver cirrhosis were randomly selected from the database as a control group. For each lesion, conventional ultrasound features and CEUS parameters, including enhancement pattern and washout time, were evaluated. In addition, the diagnostic accuracy of ultrasonic parameters for the differentiation of hNENs and HCCs was assessed using a receiver operating characteristics (ROC) curve analysis. Results: hNEN and HCC showed significant differences in the number of nodules (P<0.001), enhancement level in portal phase (P<0.001), and the start of contrast washout time (P<0.001). Compared with HCCs, hNENs more often showed multiple lesions, less likely demonstrated iso-enhancement in the portal phase, and a quicker start of the contrast washout. When using the number of nodules and washout time on CEUS as the differential diagnosis of hNEN and HCC, the area under the ROC curve was 0. Purpose: To develop and validate a radiomics nomogram for the pretreatment prediction of transcatheter arterial chemoembolisation (TACE) refractoriness with pre-therapeutic dynamic CT in patients with hepatocellular carcinoma without metastasis. This study included 80 patients with HCC treated with TACE, except for any vascular involvement, from July 2016-November 2018, after exclusion. The datasets were split into a training set (80%) and a test set (20%) for feature selection and we did 10-fold cross-validation. 40 radiomic features were extracted from arterial-phase CT using LIFEx. A lasso regression model was used for radiomics signature selection and the selected signatures were validated using a Mann-Whitney U test. The radiomics nomogram was based on the multivariate logistic regression model incorporated the rad score (radiomics score), CT image factors (enhancement, homogeneous arterial, heterogeneous portal, capsule, Bilobar, and multiple), and clinical factors (age, sex, and logAFP). For optimising the radiomics nomogram, multiple backward stepwise elimination methods were investigated. We also evaluated the radiomics nomogram using ROC and a confusion matrix. All statistical analyses were performed using RStudio software. The rad score, which consisted of GLZLM (grey-level zone length matrix), LZLGE (long-zone low grey-level emphasis), and GLZLM-GLNU (greylevel nonuniformity), was significantly associated with TACE refractioness (p<0.05) and a higher odds ratio (≥1.0). Predictors contained in the individualised prediction nomogram included rad score and Biolbar. The average age was 52±21 and 43±23 years in the first and second twin group, respectively. Concordance rates were higher in dizygotic than in monozygotic twins regarding anterior (0.47 and 0.22, respectively) and posterior (0.85 and 0.69, respectively) CoW variants. Cardiovascular risk factors did not differ significantly within monozygotic twin pairs discordant for CoW morphology (p>0.05 for body-mass index, hypertension, hypercholesterolemia and smoking, respectively). Raw heritability of CoW haemodynamics was low to moderate (0.00-0.56). Conclusion: Our results support the dominance of environmental factors on both morphology and haemodynamics of the CoW. However, no cardiovascular risk factor was identified as a predictor of CoW variants. Consequently, our results stimulate further research aiming at specifying the environmental factors affecting CoW morphology. Limitations: Major limitations include small sample size, and low spatial resolution and flow-dependence of TOF-MRA. Ethics committee approval: The regional ethical committees approved the study (approval number: 1046-260/2014; 189/2014, 2014/10/21 Mean ADC values in the cystic part showed no significant difference between the primary tumours and secondary tumours (P-value = 0.07), while mean ADC value measured in the centre of abscess shows a significant difference from tumours (P= 0.002). No statistically significant difference between the ADC values measured in the enhancing wall of the primary tumours, secondary tumour and abscess; (P= 0.09).While mean ADC values measured in peritumuoral oedema could differentiate tumour infiltration from vasogenic oedema show a slightly significant difference with (P-value = 0.04). In MR Spectroscopy, the mean Cho/Cr ratio of tumours (4.2 ± 1.23) was significantly higher than abscess (1.65 ± 0.45) (P=0.006). And it's also noted that the mean value of the intralesional Cho/Cr ratio of tumour recurrence groups were significantly higher than radiation necrosis group (P-value =0.02). ROC curve analysis demonstrated a cut off value of 1.9. As regard DTI, the mean FA measured in peritumoural oedema surrounding the primary tumours demonstrates higher value than the FA measured in the peritumoural oedema of metastases, 0.235 and 0.147 respectively. ROC curve demonstrated cut-off value of 0.165. Conclusion: DWI, MRS and DTI provided a greater confidence in differentiation of (primary from metastatic tumours), (tumours from abscesses) and (radiation necrosis from tumour recurrence). Limitations: The studies measured ADC value from a single slice of ADC maps, which may cause observer bias. The study was performed on two MRI machines which may show slight differences and variability in results. Ethics committee approval: Approved by Institutional Research Board, Mansoura University. The eligibility criteria: HIV+, performed brain MRI, an adult. Examination data of 410 patients included a history of HIV infection, opportunistic diseases, current CD4 count, plasma HIV RNA, pathogen identification in CSF. Statistic analysis included chi-square (with Yates correction), Fisher's, the Mann-Whitney, Wald criteria. Results: Regular ART was associated with a lower incidence of: large asymmetric lesions, caudate nuclei involvement, perilesional oedema, ring contrast enhancement. Duration of ART for more than five years reduces the incidence of brain involvement in HIV, contrast enhancement, white matter lesions, the involvement of the frontal lobes and basal nuclei. Lower current CD4 cells were associated with large asymmetric lesions, the involvement of the basal ganglia and posterior fossa, perilesional oedema and mass effect, blooming artefact on T2*, ring and nodular contrast enhancement. Plasma HIV RNA above 50 copies/ml was associated with diffuse white matter lesions, large asymmetric lesions, the involvement of the cerebellum, perilesional oedema. In major cases of CNS-IRIS we observed acute inflammatory demyelination, which was characterised by the appearance of new focal lesions, as well as the increase of old ones, atypical patterns of contrast enhancement: perivascular, peripheral, nodular, leptomeningeal. Conclusion: ART and dependent on its current СD4 cells, plasma HIV RNA had an influence on the severity of brain structural damage. With the disease progression, the most vulnerable regions were basal ganglia and cerebellum despite ART. The atypical PML were significantly more often visualised in CNS-IRIS. Limitations: This study was limited by its retrospective design, qualitative neuroradiological criteria, and several data were incomplete. Ethics committee approval: The study approved by the ethics committee in accordance with the criteria ICH GCP. All patients have signed informed consent. Funding: No funding was received for this work. Author Disclosures: E. Bakulina: nothing to disclose T. Trofimova: nothing to disclose MyT3 18-8 10:54Comparison of 3D DIR, 3D FLAIR and 2D FLAIR pulse sequences for imaging in demyelinating disorder (in multiple sclerosis) at 3 Tesla K. Nekar, P. P. Wali, R. Ananthasivan, U. Acharya; Bangalore/ IN (drkrishna3thi23@gmail.com) Purpose: Comparing 3D DIR sequence with that of 3D FLAIR and 2D FLAIR sequence in order to find out the h sensitivity in the detection of multiple sclerosis (MS) lesions and in addition we would like to further evaluate which among the 3D sequences would have higher sensitivity and specificity in picking up demyelinating lesions in the brain. Methods and materials: Study population (minimum of 30 cases) will include patients who are being evaluated for demyelination disorders with MRI. Imaging was performed on a 3T Philips MR system using 3D DIR, 3D FLAIR sequences with the same parameters, including FOV, matrix, slice thickness, voxel size, and a number of signals averaging (NSA) in addition to routine T1, T2, diffusionweighted and 2D FLAIR. Results: Descriptive and inferential statistical analysis has been carried out in the present study. Results on continuous measurements are presented on mean ± SD (min-max) and results on categorical measurements are presented in number (%). Significance is assessed at 5 % level of significance. The following assumptions on data are made. Wilcoxon Signed rank test has been used to find the significance of the median score in two groups. Conclusion: DIR brain imaging had the highest sensitivity in the detection of supratentorial lesions compared to3D FLAIR and 2D flair. In addition, the lesions obtained with 3D DIR images were more easily visualised. Limitations: Patients with cardiac pacemakers and metallic implants will not be subjected to MRI. Motion disorder and claustrophobia, if severe, may make the examination difficult. Ethics committee approval: The study is approved by a diplomate of the national board. Purpose: We aimed to utilise a non-invasive in-vivo MR-based imaging modality and analysis framework to detect and characterise the laminar components of the human cerebral cortex in stroke and epilepsy patients compared with healthy controls (HC). Methods and materials: Using 3T-MRI we calculated T1-component probability maps and showed that the T1-signal of the cortex can be divided into 5-6 distinct Gaussian distributions centred at different values, which represent different cortical laminar classes. For the stroke patients, we explored in-vivo if in patients (n=20) with chronic lacunar infarct, involving the corticospinal-tracts (CST), the cortical layers of the connected primary motor cortex (M1) are differently affected. We performed tractography using the infarcts as the seed areas to reconstruct the CST and identified the connected M1. For epilepsy patients (focal cortical dysplasia (FCD) and periventricular nodular heterotopia (PNH); n=9 each) we determined the laminar composition of 78 cortical regions of interest of the automated anatomical labelling (AAL) atlas, which was divided into 1000 sub-areas with equal volume. Results: For stroke patients we showed focal cortical thinning in the connected M1 and specifically only in its deepest laminar portion as compared to the nonaffected contralateral cortex (P<0.001), supporting the concept of secondary neurodegeneration. For epilepsy patients we found widespread cortical dyslamination compared to the HC, supporting evidences of epilepsy as a network disease with disrupted brain connectivity. Conclusion: Our findings provide a method to calculate and characterise the laminar architecture of the cortex, which was implemented to stroke and epilepsy patient groups. Our method may be used as a sensitive imaging biomarker, and could potentially elucidate pathophysiologic mechanisms and facilitate patient management towards developing individually tailored treatment. Limitations: Not applicable. Ethics committee approval: Approved by institutional IRB. Funding: Not applicable. Author Disclosures: E. Lotan: nothing to disclose D. Tanne: nothing to disclose Y. Assaf: nothing to disclose MyT3 18-11 11:06 Application of low radiation dose combined spectrum and ASIR-V iterative reconstruction in CT scanning of ischaemic stroke: a feasibility study Y. You; Chengdu/ CN (1390178629@qq.com) Purpose: To explore the application of low-dose combined spectrum in CT scan of ischaemic stroke. Methods and materials: Group A (n=30) used the low-dose scan parameters: 80 kV, 50 mA, and the images were reconstructed by 50% ASIR-V. Head and neck (CTA) Scanning with Spectrum CT, 80-140 kV, 200 mA，the contrast agent volume was 20ml and the flow rate was 4ml/s. The images were reconstructed with 40 keV and 80% ASIR-V. Group B (n=30) used routine scan parameter: CTP:80 kV, 150 mA, CTA :100 kV, 350 mA, the contrast agent volume was 50ml and the flow rate was 5ml/s. The images were reconstructed by FBP. The contrast agent volume was 40ml and the flow rate was 5ml/s in CTP. The quantitative evaluation included: CTP parameters of the hypothalamic, frontal and temporal lobes without perfusion defects at the basal ganglia level, the CT and SD values of the head-neck CTA segments, and the CNR values of M1. The CTP and CTA image quality were also subjectively evaluated by two senior radiologists using a four-point grading scale. Results: All CTP and CTA images were diagnostic. There were no significant differences in the subjective image, CTP parameters, CTA internal carotid artery and middle cerebral artery and CT value of common carotid artery. There were significant differences in SD and CNR in the common carotid artery. The effective dose and the contrast agent in group A were 58.5% and 23% lower than group B. Conclusion: The use of low radiation dose combined spectrum and ASIR-V iterative reconstruction in CT scanning of ischaemic stroke. Limitations: The simple size is relatively small. Ethics committee approval: Approved by the ethics committee of our hospital. Purpose: The existence of the blood-brain barrier has led to almost all chemotherapeutic drugs and gene drugs failing to effectively achieve and play a role in glioblastoma. So far, the survival rate of glioblastoma patients has hardly improved. In this study, we are aimed to construct an integrated nanoprobe based on albumin nanoparticles for targeted diagnosis and treatment of glioblastoma. Methods and materials: We constructed bovine serum albumin nanoparticles (ANG-BSA/Car/ICG MNPs) containing SPIO, carmustine and ICG coupled ANG. The in vitro and in vivo MRI/FL dual-mode imaging of glioblastomas in vitro and in vivo were systematically evaluated in order to evaluate the penetration ability of blood-brain barrier and tumour-specific targeting effect. At the same time, through experiments with the control group, we evaluated the ability of ANG-BSA/Car/ICG MNPs to inhibit tumour. Results: In vitro experiments, we verified that the targeting ability of nanoprobes to glioblastoma cells was significantly better than the control group. In addition, in vivo experiments, nanoprobes significantly increased the accumulation of brain tumours compared with the control group. This targeting imaging and drug delivery system provides an efficient strategy for targeted therapy and intraoperative localisation of glioblastoma. Conclusion: ANG-BSA/Car/ICG MNPs present a promising potential in multifunctional therapeutic of glioblastoma, and we anticipate that this nanoplatform will be a perfect candidate in glioblastoma theranostic. Limitations: This study was limited by the lack of comparison with other nanoplatforms. Ethics committee approval: All animal experiments were approved and conducted according to the principles of the Institutional Animal Care Committee from the Shenzhen Second People's Hospital. Imaging as the new yardstick for diagnosing peripheral mononeuropathies: a comparison between high-resolution ultrasound and MR neurography with an approach to diagnosis A. Agarwal Chandra, U. Jaipal, M. Bagarhatta, M. Agarwal, A. Chandra; Jaipur/ IN (a.agarwal.1992@gmail .com) Results: Overall, the diagnostic accuracy was highest for the proton density fat saturated (PDFS) MR sequence (93.89%) followed closely by HRUS (80%). The sensitivity was highest for PDFS sequence while the T1 MR sequence had the highest specificity. Combined diagnostic accuracy of both modalities was calculated to be 93.33% with a negative predictive value of 80%. HRUS and MRI equally detected the cases with nerve discontinuity, while neuromas were better identified on MRI. Conclusion: With the advent of higher frequency probes and improved MR field strength, imaging of peripheral nerves is possible with better accuracy. Imaging assessment of nerves allows anatomical delineation with the identification of the exact site of involvement. This comparative study demonstrates the role of imaging in diagnosing peripheral neuropathies with the accuracy of MRI as high as 93.89% which may serve as an imaging gold standard. HRUS, being quicker, cost-effective and a comparable accuracy of 80% can serve as a reliable screening tool. Limitations: Resource-limited setting and lack of newer sequences like DTI and DWI. Ethics committee approval: The IRB approved. Funding: No funding was received for this work. Author Disclosures: A. Agarwal Chandra: nothing to disclose U. Jaipal: nothing to disclose M. Bagarhatta: nothing to disclose M. Agarwal: nothing to disclose A. Chandra: nothing to disclose Evaluation of parameter changes in lateral lumbosacral radiography of patients with and qithout lumbar spinal stenosis in magnetic resonance imaging (MRI) N. Merd, D. Gündüz; Isparta/TR Purpose: To investigate the effect of lumbar spinal stenosis (LSS) on objective parameters reflected on radiographs and to adapt existing radiographic indexes to the diagnosis of LSS. Methods and materials: LSS was determined by measuring the midline anteroposterior (AP) spinal canal diameter on axial T2-weighted images in lumbosacral MRI. Fifty patients with LSS and 50 patients without LSS were classified as. Lateral lumbosacral radiographs of the patients were evaluated retrospectively and on radiographs, pelvic incidence (PI), pelvic tilt (PT), sacral slope (SS), lumbar lordosis (LL), L5 incidence (L5I), L5 slope (L5S), sacral table angle (STA), sagittal vertebral corpus height (SBH), sagittal vertebral corpus width(SBW), peduncular width(PW), foraminal width (FW), posterior pedicle margin (PPM) were measured, SBW/PW and SBW/PPM ratios were calculated. Results: In the LSS group, PI was significantly higher (p = 0.029). FW (p <0.001) and PW (p <0.005) values were significantly lower in the LSS group for each level. In the LSS group, SBH levels were lower (p=0.041) at L4 level, SBW levels were higher at L2-5 levels (p <0.005) and PPM values were lower at S1 level only (p <0.005). In the LSS group, the SBW / PW ratio for each level, the SBW / PPM ratio at L4 and L5 levels were significantly higher (p <0.005). In the ROC analysis of parameters with significant differences between LSS and control groups, the cut-off values of significant parameters and levels were calculated. The results of our study are useful for classifying patients with and without spinal narrow canals, but further studies are required. Limitations: Inclusion of patients with a single level of a narrow canal in the patient group, not classifying spinal stenosis as developmental or congenital. Ethics committee approval: n/a Funding: No funding was recieved for this work. Purpose: To introduce AQuaPi, a fully-automated pipeline for PET/MRI to noninvasively determine the cerebral metabolic rate of glucose (CMRGlc) images. The pipeline needs the following inputs: (1) PET listmode data, (2) attenuation correction (AC) map, (3) MR angiography (MRA) images, (4) T1-w MRI and (5) MR navigators. The processing pipeline includes two main components: (i) an image-derived input function (IDIF) component and (ii) a quantification component (QC). The IC component calculates an IDIF by first determining a volume-of-interest through automated segmentation of the MRA, followed by MR navigator-based motion correction (MC) and an MR-based partial volume correction (PVC). The QC then uses the IDIF to generate CMRGlc images. To validate the pipeline, 10 healthy volunteers underwent [18F]FDG test-retest PET/MRI examinations in an integrated PET/MR. Arterial blood samples (AIF) were collected as a reference standard. Pseudo-CT images derived from T1-w MR were used for AC. Absolute percentage difference (APD) in regional MRGlc values (IDIF vs AIF) were determined in the following 6 brain regions: corpus callosum (CC), brainstem (BS), cerebellum (CB), thalamus (TH), anterior cingulate cortex (ACC) and the superior frontal cortex (SFC). Results: The APD between CMRGlc values obtained from AIF and IDIF were: (5.9 ±3.2%) for CC, (5.9 ±3.3%) for BS, (5.8 ±3.4%) for CB, (5.5 ±3.1%) for ACC, (5 Value of routine T2WI histogram in differential diagnosis of glioblastoma Z. Ma, X. Zhao; Zhengzhou/ CN (11430884@qq.com) Purpose: To study the diagnostic value of T2 global greyscale histogram in the differential diagnosis of glioblastoma, central nervous system lymphoma and solitary metastasis of three tumour types. Methods and materials: A retrospective analysis in our hospital had brain MRI examination and 95 cases of pathologically confirmed patients, including 29 cases of glioblastoma (male 13 cases, female 17 cases, mean age 51.2 + 14.5, 10 ~ 69 years old), central nervous system lymphoma 38 cases (male 19 cases, female 21 cases. The average age of 57.1 + 12.5, range 33 to 81 years), 28 cases of solitary metastases (male 15 cases, female 13 cases, average age 59.8, range 36 to 76 years). In three groups of MR T2 axial images with Mazda software will each layer level tumour regions of interest and analyse the global grey histogram, and of the three group parameter features were analysed. The statistical significance between the various parameters. Results: Through histogram analysis of nine parameters, these nine parameters were statistically significant (all P <0.05), including mean, variance, kurtosis, skewness, Perc.01%，Perc.10%, Perc.50% , Perc.90% and Perc.99%. Conclusion: T2 global gray histogram analysis has certain value in differential diagnosis of glioblastoma, central nervous system lymphoma and solitary metastasis, and it can provide new ideas and methods for differential diagnosis. Purpose: To compare the characteristics of different types of spontaneous intracranial artery dissection on high-resolution MRI (HR-MRI) vessel wall imaging. Methods and materials: Fifty-six patients with approved spontaneous intracranial artery dissection were scanned on a 3T Siemens Skyra scanner with pre-and post-contrast 3D T1-weighted SPACE. Patients were divided into four groups based on the morphological characteristics of pre-contrast 3D T1weighted SPACE imaging. Type 1 corresponded to classic dissecting aneurysms, type 2 aneurysms were segmental ectasias, type 3 aneurysms were dolichoectatic dissecting aneurysms and type 4 aneurysms were saccular aneurysms unrelated to the branching zones. Consistency of clinical staging (acute, subacute, chronic stage) and imaging types was evaluated. The characteristics of spontaneous intracranial artery dissection were evaluated including intraluminal contrast enhancement, and artery wall enhancement (grade 0, similar to that of normal vessel walls; grade 1, less than or similar to that of muscle; and grade 2, greater than that of muscle). The characteristics of spontaneous intracranial artery dissection were compared among different types. Results: Of the 56 aneurysms, 48 were located in the vertebrobasilar system, and the other eight aneurysms were located in the anterior circulation. There were 22 patients for type 1, 10 for type 2, 14 for type 3, and 10 for type 4. Patients were classified into acute (n=18), subacute (n=18), and chronic (n=20) stages according to the time intervals from symptom onset. The clinical staging was in good consistency with imaging types (p=0.000, r=0.732). There was significantly different of intraluminal contrast enhancement and vessel wall enhancement among the four types groups (p=0.030). Conclusion: The 3T HR-MRI reveals the vessel wall characteristics and provides distinguishing findings between different types of spontaneous intracranial artery dissection. Limitations: This study was limited by a small number of study groups. Ethics committee approval: n/a Funding: No funding was received for this work. Author Disclosures: B. Tian: nothing to disclose X. Tian: nothing to disclose Z. Shi: nothing to disclose J. Lu: nothing to disclose Purpose: Subependymal giant cell astrocytoma (SEGA) is a brain tumour developed in 10-26% of individuals with tuberous sclerosis complex (TSC). Since they are located near the foramen of Monro, SEGAs can cause obstructive hydrocephalus, leading to increase morbidity and mortality. It is recommended to follow TSC patients with contrast-enhanced (CE) MRIs, but the repetitive use of gadolinium-based contrast agents (GBCAs) may cause gadolinium deposits. Our aim was to compare the performance of CE and non-CE MRI to diagnose SEGA. Methods and materials: Thirty-five TSC patients were enrolled in this retrospective single-centre study from September 2007 to January 2019 (15/35 (43%) males and 20/35 (57%) females). The inclusion criteria were: diagnosis of TSC on clinical and genetic arguments (according to international TSC consensus conference held in 2012) and at least three-brain MRIs performed for follow-up with a contrast agent injection. In total, 70 MRIs were selected (25 MRIs in the SEGA group and 65 MRIs in the non-SEGA group), two consecutive MRIs per patient were anonymised. Three radiologists performed a doubleblinded review of the non-CE and CE MRI sequences. The diagnostic performances were calculated (sensitivity, specificity, positive/negative predictive values and inter and intra-observer agreements). Results: The performances for the detection of SEGA were good and similar between the CE and non-CE MRIs. Inter-observer agreement was excellent for the two groups (range of Kappa coefficient: 0.81-0.93, p<0.001). The size of nodules near the Monro foramen, intense and heterogeneous contrast enhancement and modifications between two consecutive MRI were significantly associated with the presence of SEGA. Conclusion: The performances of CE and non-CE MRIs are similar and excellent for detecting SEGA, raising the question of reducing unnecessary gadolinium-chelates injection for TSC patients. Limitations: n/a Ethics committee approval: n/a Funding: No funding was received for this work. Role of transcranial ultrasound with Doppler and strain elastography in neonatal hypoxic-ischaemic encephalopathy with magnetic resonance imaging as the gold standard A. Singh, U. Jaipal, D. A. Bhandari; Jaipur/IN (DrAsthaSingh@outlook.com) Purpose: To compare and correlate the diagnostic role of neurosonography with spectral Doppler and strain elastography in neonatal hypoxic-ischaemic encephalopathy with MRI as the gold standard. To assess the role of diffusionweighted Imaging in early identification and extent of ischemic injury. Methods and materials: A prospective study was conducted on 50 patients under one month age with a clinical picture of neonatal encephalopathy, a history of perinatal asphyxia and low Apgar score. All patients were graded by Sarnat and Sarnat grading. Neurosonography with spectral Doppler of the anterior cerebral artery and strain elastography was done with 3-7 MHz probe on anterior fontanelle. Elastographic scores were assigned on a five-point colour scale.3 Tesla MRI was done with DWI (diffusion-weighted images). Ultrasound findings were correlated with MRI in various areas of the brain parenchyma and the role of DWI was evaluated by SPSS v.20.0. Results: The diagnostic accuracy, sensitivity and specificity of neurosonography compared to MRI were 82%, 82.6% and 75% respectively. DWI detected mild ischaemic changes when other MR sequences were negative. Neurosonography detected all cases of germinal matrix haemorrhage and showed better sensitivity for thalamic, basal ganglia and periventricular white matter lesions. Altered RI values of ACA were seen in 59% of moderate and severe cases. Diagnostic accuracy of Elastographic score was 76.67% with a specificity of 86.67% (95%CI;59.54%-98.34%) and sensitivity of 66.67% (95%CI;33.38%-88.18%). Conclusion: Neurosonography is an effective screening tool for detecting ischaemic encephalopathy. Raised elasticity of periventricular white matter in neonatal encephalopathy points towards its promising role in early diagnosis. DWI detects ischaemic changes earlier than conventional MRI which is important in initiating early management. Limitations: This study was limited by its small sample size. Ethics committee approval: Approval was taken from the ethical committee of SMS Medical College. A total of 33 patients with a diagnosis of cerebral hemiatrophy at our institution between January 2012 to July 2019 were retrospectively review. Patients of this study must fulfil the following criteria: (i) clinical presentation of seizure; (ii) radiological manifestations with ipsilateral sulcal enlargement, ipsilateral lateral ventricular dilatation, and ipsilateral calvarial thickening. The patients were distributed into two groups based on the present or absent of childhood head injury. Clinical and imaging characteristics were compared and analysed. Raw data are presented as frequencies and percentages for categories and mean for the age at presentation. Fisher's exact test was used to assess the association between two categorical variables. Results: There are 14 patients (eight males, six females) aged 1 to 46 years (mean age±SD=20.7±13.8) were included. There are eight patients with childhood head injury and six patients without history of childhood head injury. Clinical features recognised in these two groups are hemiplegia (87.5% vs 66.7% respectively, p=0.539), mental retardation (87.5% vs 66.7%,p=0.539) and psychiatric disorders (37.5% vs 16.7%, p=0.580). Of these patients, radiological appearances of wallerian degeneration and hyperpneumatisation of paranasal sinus (75% vs 50%, p=0.580 ); elevation of orbital roof (75% vs 33.3%,p= 0.277); hperpneumatisation of the mastoid and elevation of petrous ridge (62.5% vs 16.7%, p=0.138) were found in these two groups. The formation of porencephaly Conclusion: The clinical-radiological manifestations of DDMS vary widely. The patients with childhood head injury were more likely to have the formation of porencephaly and midline structural shifts. Limitations: This study was limited by a small number of patients and by the lack of detailed history and clinical information due to the retrospective nature of this study. Ethics committee approval: n/a Funding: No funding was received for this work. For objective image quality, the contrast-to-noise ratio (CNR) and signal-to-noise (SNR) of two MRA images were compared. The subjective image quality was assessed using a 5-point score scale by two experienced observers and the interobserver agreement was determined. With DSA as the reference, the neck, height, and width of aneurysms in two MRAs were measured and compared on maximum intensity projection (MIP) image datasets, respectively. Purpose: We proposed a new index of haemodynamic severity of atherosclerotic stenosis of carotid arteries: gradient of narrowing of arterial lumen GNL={(Snorm-Sstenosis )/Dnorm-stenosis}, mm 2 /mm, i.e. as ratio of difference between the areas of cross-sections of the artery at the levels of most prominent stenosis and of normal non.stenosed artery closest to it, to the distance between them. Methods and materials: 25 patients were studied, all with a stenosis of the internal carotid artery (ICA) >50%, (ECST), mono (n=22), or bilateral (n=3). 11 people without ICA stenosis served as the control group. In all participants, MRangiography was carried out, with the calculation of the GNL on the ICA stenoses. Brain MRI was also carried out on all participants in Т1-, Т2-, PD-, and flair-weighted protocols. Results: Patients were assigned to group 1 (n=12, without brain damages), and group 2 (n=13, evidence of earlier ischaemic damages Purpose: To explore intracerebral perfusion and territory shifts conducted by each main artery in Moyamoya disease pre-and post-surgery. Methods and materials: 27 MMD patients underwent non-enhanced 3D t-ASL, 3D-TOF MRA, and 3D pCASL within one week before and after surgery. T-ASL maps and perfusion shifts of unilateral ICA and bilateral VA were obtained. Perfusion shifts were further subdivided and compared according to MRA. T-ASL images were coregistered to each ASL-CBF maps and created into binary masks. The CBF value was extracted by ASL-CBF maps and all these masks, respectively. A paired t-test was used to compare the surgical side, non-surgical side ICA-CBF, and bilateral VA-CBF, respectively, pre-and post-surgery. Independent two samples t-test was used to compare territorial CBF and its changes before surgery between the haemorrhagic group and the ischaemic group. Results: 27 (7 haemorrhagic, 13 ischaemic) were enrolled. 4 distinct categories of perfusion territory shifts were observed on t-ASL. Perfusion territory shifts were divided into 2 categories. 7 perfusion territory shifts vanished and 6 appeared after surgery, including both 2 categories. There were significant CBF improvement in the surgical side ICA territory ( Purpose: In Marfan patients, accurate aortic diameter measurements are crucial for life-long monitoring. We compared the influence of overall image quality, amount of image artefacts, contrast-to-noise-ratio (CNR), and aortic root surgery on the reliability of aortic measurements of non-contrast MRA in Marfan patients at 1.5T and 3T. Methods and materials: 40 Marfan patients were prospectively evaluated by non-contrast 2D-bSSFP MRA of the aorta at 1.5T and after 12 months at 3T before (n=24) and after aortic surgery (n=16). Two readers independently rated the image quality and the presence/amount of artefacts on Likert scales at 9 (8) aortic levels before (after) aortic surgery. The CNR of the delineated aortic wall was separately calculated for each aortic level. Results: Aortic imaging at 3T provided significantly superior CNR and more reproducible aortic diameters compared to 1.5T. Interobserver variances of aortic measurements were significantly smaller at the sinuses of the valsalva, sinotubular junction, mid-ascending aorta, and proximal descending aorta in presurgical patients (p<0.05). After aortic surgery, the majority of the advantages of 3T were mitigated by a significantly impaired image appearance due to the higher number of susceptibility/flow artefacts at 3T compared with 1.5T. This resulted in significantly higher aortic diameter variances at the mid-graft and distal anastomosis. Purpose: To automate the scan range delimitation in topograms of chest CT exams using deep learning techniques. Automation of this process would free the radiographer from this routine task and could potentially reduce the radiation exposure for the patient due to a stronger restriction of the scan range. Methods and materials: A total of 1,150 chest topograms of patients (mean age=64.6 years) with pleural effusion (N=360), with atelectasis (N=82), with both diseases (N=21), or without both diseases (N=411) were extracted retrospectively from our PACS, and the scan range was delimitated in each topogram by two radiologists in consensus. A conditional generative adversarial neural network (CGAN) for image-to-image translation was trained on 1,000 topograms to generate virtual scan range delimitations. Quantitative evaluation was performed on 150 randomly selected topograms by computing the absolute difference of the predicted scan delimitations and those of the two expert radiologists. Results: Automated scan range delimitation showed an excellent agreement with the scan range delimitation of the radiologists. The mean absolute difference between expert radiologists and the prediction of the neural network was 3.77 +/-0.74 mm and 5.87 +/-1.29 mm at the upper and lower limit, respectively. Conclusion: An automated scan range delimitation of chest CT scans using deep learning techniques was trained and evaluated by comparing the scan range with the annotations of two senior radiologists. The results indicate that the conditional generative adversarial network can generate scan range delimitations that are comparable to those of expert radiologists. Limitations: More training data could potentially increase overall performance. Purpose: In the quest for increasing image quality, MR throughput can suffer and manoeuvre, which creates faster scans trade-off quality. There is a need to enhance images without prolonging scan time. Recently, deep learning-based reconstruction methods have shown promise to enhance image value. We evaluated the impact of a new deep learning image reconstruction (DLR) method for both noise reduction and improved image sharpness in clinical MR exams of the brain and spine. The investigational DLR leverages a deep convolutional residual encoder network trained on a >10K image database to create images with enhanced SNR and spatial resolution. 28 patients were scanned using clinical 2D brain (3T-7 and 1.5T-4) or spine (3T-12 and 1.5T-5) protocols. K-space data was reconstructed with both conventional and DLR (tuned to 75% noise reduction). Two neuroradiologists independently rated 93 pairs of conventional and DLR images side-by-side. Ratings were based on overall IQ, lesion conspicuity, perceived SNR and resolution, CNR, image texture, and artefact using a 5-point Likert scale (5=excellent, 1=non-diagnostic). A Wilcoxon signed-rank test was used to compare the ratings and inter-rater reliability between readers was assessed using the Bennett S score. Purpose: To investigate the effect of a deep learning image reconstruction algorithm on image quality in chest CT scans. Methods and materials: With institutional ethics approval, 100 consecutive patients underwent chest CT at standard radiation doses on a 320-detector row CT scanner with the following scan parameters: Helical scan, 0.5 mm x 80 detector rows, 120 or 100kV, and automatic exposure control, 0.275s rotation speed and standard pitch. Each scan was reconstructed at 0.5 mm volume and 3 mm axial, coronal, and sagittal slices with both lung and soft tissue kernels using the clinical standard hybrid IR (AIDR3D) and deep learning reconstruction (AiCE) techniques. Images were evaluated for overall image quality, noise, presence of artefacts, contrast, visibility of small structures, and diagnostic confidence using a 4-point Likert scale. SNR and CNR were calculated for each reconstruction. Data was analysed by a paired t-test. weighted fast spin-echo (Discovery MR750 3.0T; GE Healthcare) with the following parameters: standard T2WI (std-T2WI, matrix, 320x192; thickness, 5 mm; NEX, 2), high-resolution T2WI (HR-T2WI, NEX, 1; matrix, 452x192), and super high-resolution T2WI (sHR-T2WI, NEX, 1; matrix, 452x192; thickness, 2.5 mm). The acquired data was reconstructed with and without DLRecon. DLRecon is a new deep learning-based MR reconstruction which comprises of a deep convolutional residual encoder network trained using a database of over 10,000 images to achieve images with high SNR and high spatial resolution. The depiction of anatomical details of the pancreas and the liver, motion artefact, Objective and qualitative IQ analyses of deep learning image reconstruction in multiphasic CT imaging of the liver: a patient and phantom study F. Legou 1 , P. Roux 1 , V. Purpose: To evaluate the clinical benefits on image quality (IQ) of TrueFidelity, a deep learning image reconstruction (DLIR), in multiphasic liver CT compared to adaptive statistical iterative reconstruction V (ASIRV) in patients and on a phantom. Methods and materials: 66 patients underwent a multiphasic liver CT during a 1-month-period. The protocol was tailored according to patient morphology: 80 kV for a body mass index (BMI) <25; 100 kV for 2530. IQ of patient images reconstructed with DLIR and ASIRV50 was assessed on portal phase by measuring liver parenchyma contrast and signalto-noise ratios (CNR, SNR), and qualitatively by two radiologists using a 5-point Likert-scale. Phantom images acquired with similar protocols were evaluated by computing the noise power spectrum and the task-based modulation transfer function (MTFtask). Results: Compared to ASIRV50, CNR and SNR were significantly improved with DLIR by 71% and 56% at 80, 83% and 77% at 100, and 61% and 60% at 120 kV, respectively (p<0.01 or less). Qualitative IQ was also improved with DLIR for each patient morphology (p<0.0001). On the phantom, compared to ASIRV50, DLIR reduced the noise magnitude by 31% at 80, 100, and 120kV while maintaining a coarser texture (noise mean frequency was 0. Evaluation of automated quality control of multicentre clinical trial CT data using spine localisation based on a machine learning method S. Lee 1 , C. Page 2 , P. Galette 3 , P. Murphy 2 , B. Glocker 1 ; 1 London/UK, Purpose: Consistent image quality across multiple centres in clinical trials is crucial. The quality control is typically done by centralising data followed by manual checks prior to radiological review. For evaluating the feasibility of automating the QC process, we developed a software tool, Automatic Visual Quality Control (AutoVQC), that automatically locates the spine, calculates the anatomical field of view (FOV), and detects imaging artefacts using machine learning. AutoVQC was applied to in-house CT data from a multicentre clinical trial. The quality of CT scans (in total, 459 series from 62 subjects acquired at 7 sites) was evaluated using AutoVQC, which recognises the anatomical FOV of the images using random forests and detects artefacts such as missing slices. AutoVQC returns two values between 0 (foot) and 1 (top of the head) indicating the FOV in the canonical body axis and the outcome: pass or artefacts. The QC outcome was reviewed and classified as true-positive, true-negative, false-positive, and false-negative, where positive and negative indicate QC pass and artefacts, respectively. Results: AutoVQC returned the spine localisation result within one minute per series. The result was reviewed for sufficient data quality for subsequent analyses and accurate corresponding canonical body axis range. From all 459 series evaluated, accuracy 0.93, precision 0.89, recall 0.98, and specificity 0.87 were achieved. The causes of artefacts included missing slices, non-axial image orientation, poor image quality, and the presence of multiple series in one single folder. The initial evaluation of AutoVQC on multicentre study data demonstrated the potential of incorporating it into clinical trial image data management/analysis workflow for quality control of images based on automated spine localisation. Purpose: To develop artificial intelligence for evaluating chest radiograph image-quality. We considered 3 different features of the image quality: inclusion, rotation, and inspiration. The inclusion was further divided into 4 edges: sin, dex, top, and bottom. The data comprised of 2,019 posteroanterior chest radiographs in an upright position. We annotated the images based on the European Commission's guidelines on quality criteria for diagnostic radiographic images. The inclusion criteria were divided into three classes: too tight, correct, and too wide. The rotation and inspiratory were divided into two classes: ok and not ok. We increased the amount of the image data for the inclusion by cropping the correct images to meet the too-tight criteria and for the inclusion and the rotation by flipping the images horizontally. The image histograms were equalized and the images were resized to a resolution of 512x512 pixels. Approximately 100 and 200 images were extracted for validation and test data, respectively. We trained ResNet50 and DenseNet121 networks with the remaining images. Results: Resnet50 and Densenet121 both performed accurately for the inclusion detection in sin-and dex-edges. Densenet121 performed better for the inclusion detection in top-and bottom-edges as well as in the patient rotation and inspiration detection. The AUC was > 0.92 for the inclusion detection in all four edges in three classes. The AUC was > 0.71 and > 0.89 for the rotation and the inspiration, respectively. Conclusion: Artificial intelligence can be used in a clinical setting as instant feedback of the chest radiograph image-quality. Additionally, the trained networks provide a tool for long-term quality control of a radiography unit. Limitations: Data from only one centre was used. Ethics committee approval: n/a Funding: No funding was received for this work. Purpose: Preprocessing CT material density images is a common step for noise reduction, which improves diagnostic image quality. Our study aimed to determine the utility of adaptive artificial intelligence (AI)-based preprocessing for adjusting parameters across each CT slice compared to a similar "fixed" AIbased version that applies the same parameters across all slices. We hypothesised that the adaptive AI-based approach would decrease noise levels across different imaging exams more than the fixed approach. Methods and materials: This retrospective single-site observational study used head (n=569) and lung-nodule screening (n=276) CT exams acquired over the month prior to analysis on 64-MDCT scanners at standard-dose parameters.Original data was reconstructed as thin-section CT: 1) without preprocessing, 2) with fixed AI-based preprocessing, and 3) with adaptive AI-based preprocessing.The noise was computed as the dependent variable in each case using a manually coded software script and between-group analyses was conducted.Results: There was a significant (p<0.05) between-group reduction in noise level variation, with decreases in noise for the AI-processed compared to the non-processed CT data and greater for the adaptive versus fixed AI-based algorithm. In particular, the standard deviation (SD) for the mean noise was markedly lower for the adaptive versus fixed AI-based algorithm, indicating decreases in image variability across scans. The adaptive AI-based preprocessing algorithm outperformed the fixed method. In particular, SD differences between methods were more substantial, suggesting less variability in noise levels across cases for the adaptive method. Future investigation is necessary to determine if this algorithm produces image quality that improves sensitivity/specificity for the detection and diagnosis of clinically-relevant pathology. Limitations include relitively small sample-size and retrospective single-site design. Ethics committee approval: IRB-approved. Purpose: CT images often have artefacts due to subject motion. These cause mimics leading to false-positives (FPs) for AI models. In this study, we train an artefact-detector model and use this to make a subdural haemorrhage (SDH) detection model robust to artefacts. In FP's of the SDH model deployed at an urban outpatient centre, we found that 54% had artefacts. Visually inspecting these scans, we found that the models recognised motion-artefacts as SDH because both were hyperdense areas close to the cranium. We used a manually annotated dataset of 452 scans to train a convolutional neural network (CNN) to detect if a CT slice has artefacts or not. The original SDH model was also a CNN that predicted if a slice contained SDH. This model predicted scan as SDH positive if any slice was predicted positive. We modified this model by discounting the slices predicted as containing artefacts. A validation dataset collected from the above centre from 06/18 to 01/19 contained 712 scans, of which 42 had SDH and 1,447 (total 22,131) slices had artefacts. Sensitivity and precision were used as metrics for SDH models, while slice-wise AUC for an artefact model.