key: cord-0812858-xmjmwxfn
authors: nan
title: 55. Jahrestagung der Deutschen Gesellschaft für Neuroradiologie e.V.
date: 2020-09-23
journal: Clin Neuroradiol
DOI: 10.1007/s00062-020-00951-w
sha: 036ec7e98b4adb50337d2eae5a31d78012767028
doc_id: 812858
cord_uid: xmjmwxfn

nan

Background & Purpose: Directional Deep Brain Stimulation (dDBS) electrodes allow to steer the electrical field in a specific direction. When implanted with torque they may rotate for a certain time after implantation. Aim of this study was to evaluate whether and to which degree leads rotate in the first 24 h after implantation using a sheep brain model.

Rotational deviation of implanted electrodes after applied torsion 180° applied torsion depicted in GREEN, 360° in RED and controls (0°) in BLUE. a Initial rotational deviation (timepoint 0 h) and in follow-up. b Absolute deviation in relation to the rotation after applied torsion penic vertebral fractures and whether PDFF values can differentiate between low-energy osteoporotic/osteopenic and high-energy traumatic vertebral fractures. Methods: Of 52 study participants, 25 presented with one or multiple acute low-energy osteoporotic/osteopenic vertebral fractures of the lumbar spine (BMD<120 g/cm 3 ) and 7 patients with acute high-energy traumatic vertebral fractures (BMD>120 g/cm3). These patients were frequency-matched for age and sex to subjects without vertebral fractures (N = 20) . Bone mineral density (BMD) values were derived from quantitative CT. Chemical shift encoding-based water-fat MR imaging of the lumbar spine was performed considering a single T2* decay and PDFF maps were calculated. Segmentations of the vertebral bodies were performed excluding fractured vertebrae. Associations between category/fracture status and PDFF were assessed using multivariable linear regression models. Results: A significant correlation between mean PDFF and BMD (r = -0.664, P < 0.001) was found. In the osteoporotic/osteopenic group, those patients with osteoporotic/osteopenic fractures had a significantly higher PDFF than those without osteoporotic fractures after adjusting for age, sex and BMD (adjusted mean difference [95 % confidence interval]: 19.87 % [13.23 %, 26.50 %]; P < 0.001). When evaluating all patients with acute vertebral fractures, those with high-energy traumatic fractures had a significantly lower mean PDFF than those with low-energy osteoporotic/osteopenic vertebral fractures (P < 0.001). Conclusion: MR-based PDFF enables the differentiation between patients with and without osteoporotic/osteopenic vertebral fractures as well as between low-energy osteoporotic/osteopenic and high-energy traumatic vertebral fractures, suggesting its potential as a biomarker for bone fragility. Background & Purpose: On MRI, hippocampal sclerosis (HS) is often associated with gray/white matter blurring (GMB) of the temporal pole. In order to evaluate the microstructural substrate of GMB we acquired a MRI protocol including diffusion mesoscopic imaging and MP2RAGE sequences. Methods: 13 patients with temporal lobe epilepsy and histologically proven hippocampal sclerosis were studied. Standardized temporal pole white-matter regions were segmented and diffusion measures and T1 relaxation times were compared in temporal poles ipsilateral to HS to values obtained from contralateral temporal poles and normal controls. Results: Temporal poles on the HS side showed higher T1 values, lower white-matter intraaxonal volumes, lower fractional anisotropy and higher extraaxonal diffusivities compared to the contralateral temporal poles and normal controls. Changes were less pronounced in patients without visually obvious GMB but clearly different from the contralateral side and normal controls (Fig. 1) . Conclusion: Hippocampal sclerosis is associated with axonal loss of the ipsilateral temporal pole, even in patients in which these changes are not visible on FLAIR or T2-weighted sequences. Background and Purpose: Preclinical evidence points towards a metabolic reprogramming in isocitrate dehydrogenase mutated tumor cells with downregulation of the expression of genes that encode for glycolytic metabolism. We non-invasively investigated lactate and Cr axial; d, e coronal) and subtraction images (c axial; f coronal) in a patient with MS. One new T2-lesion in the left frontal deep white matter was initially missed on routine imaging without subtraction maps (c, f arrow*). Four other new/growing lesions are also easily depictable on subtraction images (c, f arrows). The examination interval is easily readable in the output images (#) concentrations, as well as intracellular pH using 1 H/ 31 P MRS in a glioma patient cohort. Materials and Methods: 30 prospectively enrolled, mostly untreated glioma patients met the spectral quality criteria (WHO °II n = 7, °III n = 16, °IV n = 7; IDHmut n = 23, IDHwt n = 7; 1p/19q codeletion n = 9). MRI protocol included 3D 31 P CSI and 1 H single voxel spectroscopy (PRESS at TE 30 ms and TE 97 ms with optimized echo spacing for detection of 2-hydroxyglutarate) from the tumor area. Values for absolute metabolite concentrations were calculated (phantom replacement method). Intracellular pH was determined from 31 P CSI. Results: At TE 97 ms, lactate peaks can be fitted with little impact of lipid/macromolecule contamination. We found a significant difference in lactate concentrations, lactate/Cr ratios, and intracellular pH comparing tumor voxels of IDHmut to IDHwt patients, with reduced lactate levels and near normal intracellular pH in IDHmut patients. We additionally found evidence for codependent effects of 1p/19q codeletion and IDH mutations with regard to lactate concentrations for tumor grades WHO °II and °III, with lower lactate levels in patients exhibiting the codeletion. There was no statistical significance comparing lactate concentrations between IDHmut WHO °II and °III gliomas. Conclusion: We found indirect evidence for metabolic reprogramming in IDHmut tumors with significantly lower lactate concentrations compared to IDHwt tumors and a near normal intracellular pH.

Background & Purpose: So far it remains unclear whether smoking-induced microangiopathic changes contribute to the damage of peripheral nerves in diabetic polyneuropathy (DPN) . The aim of this study was to investigate the correlation between cigarette smoking and peripheral nerve damage in diabetic polyneuropathy. Methods: We performed 3 T MR-neurography of the right leg in 77 patients (34 never smokers, 32 ex-smokers, 11 active smokers; female: 32, male: 35). Semi-automated analysis of lesion load and nerve average diameter was performed. In addition, clinical, serological and electrophysiologic parameters were assessed. Results: A significantly higher load of T2w-hyperintense sciatic nerve lesions were found in smokers compared to ex-smokers (20.8 % ±7.2 vs. 9.%±1.1 respectively 7.4 % ± 0.9, respectively p < 0.001). Also, we identified a significantly larger diameter of sciatic nerve in smokers (162 mm 2 ± 18) in comparison to ex-smokers (130.2 mm 2 ± 5) and to never smokers (127 mm 2 ± 5; p = 0.024). Lesion load and diameter of sciatic nerve also showed a negative correlation between peroneal nerve conduction velocity (r = -0.29; p = 0.008 respectively r = -0.28; p = 0.012) and Neuropathy Disability Score (r = 0.32; und r = 0.33; respectively p < 0.001). We did not find a correlation for HbA1c-values. Conclusion: Our results reveal a significant association of smoking on peripheral nerve damage in diabetic neuropathy. Both peripheral nerve diameter and increased lesion load of sciatic nerve are further associated with a lower nerval conduction velocity.

Schmitz-Koep B, MD 1,2 ; Bäuml J G, PhD 1,2 ; Menegaux A, PhD 1,2 ; Nuttall R, MSc 1,2 ; Zimmermann J, MSc 1,2 ; Schneider S C, BSc 1,2 ; Daamen M, PhD 3, 4 ; Boecker H, MD 3 ; Zimmer C, MD 2 ; Wolke D, PhD 5,6 ; Bartmann P, MD 4 ; Sorg C, MD 1, 2, 7 ; Hedderich D M, MD, MHBA 1, 2 and prospectively collected cohort of 101 very premature-born adults (< 32 weeks of gestation and/or birth weight below 1500 g) and 108 full-term controls at 26 years of age.

We found significantly lower whole amygdala volumes in premature-born adults. Also, amygdala composition differed in premature-born adults, which showed significantly lower volumes of the accessory basal nucleus (pertaining to the basolateral amygdala), adjusted for whole amygdala volume. Moreover, association with variables of premature birth (gestational age, birth weight and duration of ventilation) was moderate for basolateral and superficial amygdala nuclei and absent for left centromedial amygdala nuclei.

Our data suggests differentially affected development of the basolateral amygdala after premature birth, possibly through disturbance of distinct developmental pathways. 1. Swanson, L. W., & Petrovich, G. D. (1998) . What is the amygdala? Trends in Neurosciences. 2. Peterson, B. S. et al. (2000) . Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. Journal of the American Medical Association. 3. Cismaru, A. L. et al. (2016) After controlling for age, Wahlund score, vascular risk factors and log transformation of mean K trans values, participants with cognitive impairment as defined by the CDR in combination with neuropsychological deficits showed a disruption of the hippocampal BBB (left p = 0.04, right p = 0.035). An example of a bilateral BBB permeability disruption is given in Fig. 1 . Conclusion: Using DCE-MRI early BBB disruption can be detected in the AD continuum independently of vascular risk factors and white matter lesions.

Group comparison of accessory basal nucleus volumes in premature-born individuals compared to controls. These Boxplots show mean bilateral volumes of the accessory basal nucleus in individuals born very preterm and/or with very low birth weight (VP/ VLBW) ans in full-term controls. Marginal means as estimated by the general linear model with left or right whole amygdala volume, respectively, sex and scanner as covariates were significantly smaller in premature-born individuals compared to controls (left 239.5 mm 3 vs. 244.5 mm 3 , p = 0.003; right: 241.2 mm 3 vs. 246.7 mm 3 , p = 0.003) Results: Interrater-agreement was only moderate for PBV-ASPECTS (w-Kappa = 0.53), while it was substantial for CBV-ASPECTS (w-Kappa = 0.63) and best for NECT-ASPECTS (w-Kappa = 0.74). Accuracy, as assessed by spearman correlation between acute and follow-up AS-PECTS in patients with successful recanalization (mTICI 2b or better), was best for NECT-ASPECTS (rho = 0.86 (0.65-0.97), p < 0.001), while it was comparable for PBV-ASPECTS (rho = 0.55 (0.24-0.81), p = 0.01) and CBV-ASPECTS (rho = 0.56 (0.18-0.85), p < 0.001). Noteworthy, cases of relevant infarct overestimation occurred in both acute PBV-and CBV-ASPECTS evaluation. Conclusion: NECT-ASPECTS prior to mechanical thrombectomy outperformed both PBV-ASPECTS and CBV-ASPECTS in accuracy and reliability, while the latter two were found to be comparable in accuracy but not in reliability.

[42] Accelerating Multiple Sclerosis Imaging with Attention to Detail [1] and Deep Learning-based image synthesis [2] synergistically accelerate high-resolution MR imaging of Multiple Sclerosis (MS) patients. Here, we present preliminary results on a novel, lesion-attention GAN (Generative Adversarial Network) for full-resolution image synthesis with a special attention on (small) lesion translation, allowing a full examination (T1w, T2w, FLAIR, DIR, all 3D 1 mm isotropic) in less than seven minutes. Methods: To refine our DiamondGAN architecture [3] with special emphasis on the reliable reconstruction of (small) MS lesions, we added an additional L1 loss based on automatically created lesion segmentation maps for the cycle-consistency loss. This network was trained on 30 subjects to generate FLAIR and DIR sequences from input T1w and T2w (both acquired with compressed sensing in 1 mm isotropic resolution in less than seven minutes). For comparison, we also trained a standard CycleGAN without lesion attention. Results: Median Contrast-Noise-Ratio (CNR) for MS lesions was comparable between acquired and synthetic (AttentionGAN, Cycle-GAN) DIR images (27.3 vs. 24.2 vs. 20 .5), while CNR was lower for synthetic FLAIR images (12.9 vs. 10.3 vs. 9.7). However, in particular small lesions were visually better translated in AttentionGAN (Fig. 1) . Conclusion: Combining compressed sensing and image synthesis potentially allows to acquire a full MS MR examination with 3D 1 mm isotropic sequences in less than seven minutes by synthesizing DIR and FLAIR contrasts from T1w and T2w sequences. Future work will focus on optimizing the loss term for balancing lesion translation and halluzination. Background and Purpose: Focal enhancement on MR vessel wall imaging is frequently encountered in unruptured intracranial aneurysms, but its implication for risk stratification and patient management remains unclear. This study investigates the association of focal wall enhancement with established clinical, hemodynamic and morphological risk factors and histologic markers of wall degradation. Methods: Patients with an unruptured middle cerebral artery aneurysm who underwent 3 T MR vessel wall imaging and 3D rotational angiography (RA) were included. Segmentations of enhanced MR areas and the 3D aneurysm model based on RA were carried out and co-registered.

Hemodynamic parameters were calculated based on flow simulations and compared between enhanced regions and the entire aneurysm surface. Morphological parameters were semi-automatically extracted and quantitatively associated with wall enhancement. Histological analysis included detection of vasa vasorum, CD34 and myeloperoxidase staining in a subset of patients. Results: Twenty-two aneurysms were analyzed. Enhanced regions were significantly associated with lower cycle-averaged wall shear stress, lower maximum oscillatory shear, and increased low shear area. Higher PHASES score and histological signs of wall inflammation and degeneration were significantly associated with focal enhancement. Higher ellipticity index was an independent predictor of wall enhancement. Conclusion: Focal wall enhancement is co-localized with hemodynamic factors that have been related to a higher rupture risk. It is correlated with morphological factors linked to a higher rupture risk, higher PHASES scores and histologic markers of wall destabilization. The results support the hypothesis that focal enhancement could serve as a surrogate marker for a higher risk of rupture.

Background: Data on the frequency and outcome of repeated mechanical thrombectomy (MT) in patients with short-term re-occlusion of intracranial vessels is limited. Addressing this subject, we report our multicenter experience with a systematic review of the literature. Methods: A retrospective analysis was conducted of consecutive acute stroke patients treated with MT repeatedly within 30 days at 10 tertiary care centers between January 2007 and January 2020. Baseline demographics, etiology of stroke, angiographic outcome and clinical outcome evaluated by the modified Rankin Scale (mRS) at 90 days were noted. Additionally, a systematic review of reports with repeated MT due to large vessel occlusion (LVO) recurrence was performed. Results: We identified 30 out of 7844 (0.4 %) patients who received two thrombectomy procedures within 30 days due to recurrent LVO. Through systematic review, three publications of 28 participants met the criteria for inclusion. Combined, a total of 58 participants were analyzed: cardioembolic events were the most common etiology for the first (65.5 %) and second LVO (60.3 %), respectively. Median baseline NIHSS (National Institutes of Health Stroke Scale) was 13 (IQR 8-16) before the first before the second MT (p = 0.031). Successful reperfusion was achieved in 91.4 % after the first MT and in 86.2 % patients after the second MT (p = 0.377). The rate of functional independence (mRS 0-2) was 46 % at 90 days after the second procedure. Conclusion: Repeated MT in short-term recurrent LVO is a rarity but appears to be safe and effective. The second thrombectomy should be pursued with the same extensive effort as the first procedure as these patients may achieve similar good outcomes. Background & Purpose: Expansion of MRI T2-weighted and/or T1-weighted lesion volume after radiotherapy (RT) may indicate pseudoprogression (PsPD). The differentiation between true progression and PsPD is a diagnostic and therapeutic challenge and it is underinvestigated particularly in pediatric low-grade glioma (LGG). The aim of this study was to apply radiological and clinical criteria for PsPD in pediatric LGG following three RT-modalities to estimate the incidence of PsPD. Methods: 133 pediatric LGG patients (68 [51.1 %] male, median age at therapy start 11.36 years [range 0.78-25.92]) of the SIOP-LGG 2004 study and registry with primary RT (iodine-seed-brachytherapy [IS; n = 51], photon-beam [PT; n = 60] or proton-beam RT [PBRT; n = 22]) were included. Initial and regular follow-up brain MRI over 2 years were evaluated. Increasing 1) total tumor-associated T2-lesion, 2) focal tumor-associated T2-lesion and 3) T1-weighted contrast-enhancing tumor were the evaluated items according to which PsPD suspicion was raised. Their imaging behavior over a follow-up of 2 years finally determined true PsPD. Results: True PsPD was radiologically determined in 54/119 (45.4 %) and was not dependent on the RT-modality applied (IS 43.8 %; PT 47.2 %; PBRT 44.4 %; p = 0.939). True PsPD occurred at a median of 6.1 months after RT initiation and persisted for a median of 7 months (IS 7.8 months; PT 6.25 months; PBRT 8.1 months). Finally, intratumoral necrosis within the focal tumor-associated T2-lesion predicted true progression (p < 0.001). Conclusion: PsPD in irradiated pediatric LGG is substantially frequent and seems not to depend on the modality of primary RT applied (IS vs. PT vs. PBRT).

Harlan M 1 , Pfaff G 2 , Bendszus M 1 , Pfaff JAR 1 1 Department of Neuroradiology, Heidelberg, Germany 2 Visiting Lecturer in Epidemiology and Psychiatric Epidemiology, Protestant University of Applied Sciences Ludwigsburg

Background & Purpose: Social distancing and stay-at-home advisories aimed at reducing COVID-19 spread may inadvertently affect emergency medical care. We analyzed the time course of neuroradiological emergency consultations (NECs) in a teleradiological network before and after implementation of COVID-19 pandemic lockdown measures.

We performed an ambispective observational study of NECs in a teleradiological network connecting a tertiary care university hospital and thirteen hospitals in Southwest Germany. The study period covered prepandemic calendar weeks (CW) 01/2019-11/2020, and COVID-19 pandemic weeks 12-28/2020. Descriptive data on NEC computed tomography imaging from the prepandemic period were compared with prospective observations for the pandemic period. Results: During the prepandemic study period, the number of NECs per week remained stable around a median of 103 (interquartile range [IQR]: 97-115). After Germany entered COVID-19 lockdown in CW 12/2020, teleconsultations declined sharply, followed by a slow rebound (median: 80, IQR: 67-88; p-value < 0.001). Following gradual loosening of the lockdown starting in CW 20/2020, the weekly number of NECs rose to exceed the comparison figures for 2019 (median: 111, IQR: 100-129; p-value: 0.13, see Fig. 1 ). Conclusion: After the implementation of COVID-19 pandemic lockdown measures in Germany, we observed a temporary massive decline in neuroradiological emergency consultations. This depression may correspond to a period of delayed or missed opportunities for early diagnostic workup and treatment of neurological emergency situations. We recommend educating the public about both COVID-19 precautions and the overriding importance to seek immediate medical care in health emergencies such as stroke. Methods: Consecutive patients with CRAO between 01/2010 and 12/2019 and stroke-MRI performed within 2 weeks of clinical onset were included. Patient data, including visual acuities (VA), fundoscopic features, medical history, laboratory findings, and intravenous thrombolysis were recorded. DWI was evaluated for RDR by a neuroradiologist blinded for CRAO side and clinical data. Results: 127 patients (mean age 69.6 ± 13.9; 59 female) were included. RDR were present in 67.2 % of patients overall and in 78.6 % within the first 24 h (arrows Fig.) . Only in one case (0.8 %) was RDR falsely attributed to the wrong side. There was a trend for RDR to be more frequent in patients with blindness compared to less severe visual impairment (p = 0.07). Absence of RDR was more frequent in patients with complete VA restitution (75 %) vs. without remission (28.4 %; p = 0.006). Patients without retinal opacity or cherry red spot on ophthalmoscopy were more likely to show no RDR compared to those with these findings (60 % vs. 27.1 %; p = 0.004). Overall detection rates of RDR did not differ significantly within the first week, but dropped significantly in the second week (>7-14 d 10.0 %; p = 0.0006). Conclusions: Retinal diffusion restrictions are present in a majority of CRAO patients and are time dependent. They tend to match clinical and ophthalmoscopic severity. Further work should explore the potential of stroke DWI in the hyperacute phase for patient selection for recanalizing therapies. Background and Purpose: Visualization of the endolymphatic hydrops (EH) by MRI is used additionally for the diagnosis of Morbus Menière. However an EH can also be found in other inner ear diseases with different otologic symptoms. Changes of the hydrops over time are rarely examined so far since in most cases imaging is only performed once. Method: MRI of the temporal bone to detect EH was performed in 200 patients with otologic symptoms. We correlated their clinical diagnosis, the time since onset of symptoms and the number of experienced sudden hearing losses with the grade of the EH. We used four degrees to describe the cochlear hydrops: 1 = apical hydrops visible, 2 = apical and middle turn, 3 = 2 and part of the basal turn involved, 4 = complete cochlea involved. Results: 38 patients were excluded for tumor, insufficient image quality (movement) and enhancement, leaving 162 patients (324 ears) to be finally evaluated. The extent of the hydrops was increasing with the time since onset of symptoms (p< = 0.05) as well as with the number of the events with sudden hearing loss (p< = 0.05). The correlation of hydrops extent and the diagnosis of Morbus Menière see Fig. 1 . Conclusion: With time since onset of symptoms the extent of the EH increases. A positive correlation of hydrops extent and the certainty of the clinical diagnosis is given.

Background & Purpose: In patients with cranial cerebrospinal fluid leaks, precise identification of leakage site is crucial for surgical approach. High-resolution CT cisternography (CTC) is limited in the ability to demonstrate the site of a CSF leak, particularly in patients with multiple or small osseous defects or inactive leaks during imaging. We aimed to test the feasibility of a novel high-resolution gadolinium-enhanced compressed-sensing SPACE technique for MR cisternography (MRC) and to compare findings to CTC and intraoperative results. Methods: Between November 2019 and March 2020, seven patients with CSF rhinorrhea were studied with CTC and MRC. For MRC, a highly accelerated CS T1 SPACE sequence was applied on a 3 T whole-body MR scanner using a 64-channel head/neck coil. Syngo. via software is used to overlay 3D CS T1 SPACE and CTC images to delineate areas of leaks. Findings of CS SPACE MRC were compared to standard CTC images and intra-operative results. Results: All CSF leaks were precisely depicted on CS T1 SPACE images ( Fig. 1) . In five CTC studies, leaks were missed (n = 1), falsely located (n = 1), or only suspected (n = 3). All CSF leaks detected on MRC correlated with findings at surgical repair. Conclusion: High-resolution gadolinium-enhanced CS T1 SPACE MRC is a promising method for detection of CSF leaks in patients with CSF rhinorrhea. In our pilot experience, this technique appears superior to standard CTC. Purpose: The quality of cerebral microperfusion (CM) is strongly related to vessel occlusion location and the robustness of arterial intracranial collaterals (IC) in patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO). Robust CM allows for transit of blood through the ischemic brain tissue into the veins. The venous microcirculation profile (VMP) may more accurately reflect tissue perfusion compared to arterial IC, but it is unclear to what extent the venous CM profile is affected by arterial clot localization during AIS-LVO. We determined, if the arterial vessel occlusion localizations predict VMP profile in AIS-LVO patients.

We performed a multicenter retrospective cohort study of consecutive patients who underwent thrombectomy for AIS-LVO treatment. Patient details were obtained from prospectively maintained stroke databases and the electronic medical record. Baseline CT angiography was used to localize vessel occlusion, which was dichotomized into proximal (internal carotid artery and proximal M1) and distal (distal M1 and M2) occlusions. The primary outcome measure was VMP, which was determined on baseline CTA by the cortical vein opacification score (COVES). COVES venous opacification was scored for the vein of Labbé, sphenoparietal sinus, and superficial middle cerebral vein were scored as: 0, not visible; 1, moderate opacification; and 2, full opacification. Results: 374 patients met inclusion criteria. Median age was 76 (IQR: 65-82) and 49 % were female. 196 patients (52 %) had a proximal occlusion and 178 patients (48 %) had a distal occlusion. Median COVES was 1 (range 0-5) for proximal occlusion and 3 (range 0-6) for distal occlusion patients. Mann-Whitney-U tests indicated a significant difference between proximal and distal occlusions (p < 0.001). Ordinal logistic regression showed that patients with more distal vs proximal occlusions had increased odds of having higher COVES (OR = 12.62, Fig. 1 A 55-year-old female who was re-admitted for recurrent spontaneous left-sided CSF rhinorrhea and persistent slight headaches after previous endoscopic sinus surgery. CT cisternography is unremarkable (a, b arrows). CS SPACE MR cisternography demonstrates a subtle CSF leak originating from the anterior rim of the left cribriform plate near foramen caecum (c, d arrows) with a thin CSF collection extending to the anterior nasal cavity (arrowheads in c, e)

[95 % CI 8.02-20.22]; p < 0.001), independent of age or presentation NIHSS.

The distinct arterial clot localization in AIS-LVO patients affects the cortical venous microperfusion profile. Venous microperfusion was found to be impaired in patients with proximal versus distal vessel occlusions.

Purpose: Robust pial arterial collaterals (PAC) preserve blood flow to critically hypoperfused brain tissue in patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO). CT angiography (CTA) based methods of pial collateral assessment do not provide tissue level perfusion information, and prior studies have shown that PAC assessment on CT perfusion imaging strongly predicts outcome in AIS-LVO patients treated by thrombectomy. Patients with favorable pial collaterals and brain tissue perfusion also likely have robust cortical venous drainage relative to patients with more impaired cerebral perfusion. We determined the venous microperfusion profile (VMP) in AIS-LVO patients. We hypothesized that robust PAC on CT perfusion predict robust cortical venous contrast opacification on pre-treatment CTA and that a favorable VMP is associated with good clinical outcomes in AIS-LVO patients.

We performed a multicenter retrospective cohort study of consecutive AIS-LVO patients who underwent thrombectomy. Included patients had interpretable pre-thrombectomy CT angiography (CTA) and CT perfusion (CTP) studies and clinical outcome data. Patient details were obtained from prospectively maintained stroke databases and the electronic medical record. Pre-thrombectomy CTA and CTP studies were reviewed and scored for tissue-level collaterals using the Hypoperfusion Intensity Ratio (HIR). HIR was defined as the volume ratio of brain tissue with [Tmax>10 sec/Tmax>6 sec] such that a lower HIR correlates with favorable collaterals. HIR was automatically calculated by RAPID (iSchemaView). VMP was determined by opacification of the vein of Labbé, sphenoparietal sinus, and superficial middle cerebral vein on CTA as: 0, not visible; 1, moderate opacification; and 2, full. Primary outcome measure was VMP. Secondary outcome measure was ordinal modified Rankin Scale (mRS). Ordinal linear regression models were performed to predict the effect of HIR on VMP, as well as the effect of VMP on mRS. Results: 186 patients met inclusion criteria. HIR was dichotomized into lower (≤ 0.4, good collaterals) and higher (≥ 0.5, poor collaterals) ratios. Mann-Whitney-U test indicated that subjects with higher HIR (median COVES = 1) had lower VMP than patients with lower HIR (median COVES = 3) (p < 0.001). In an ordinal logistic regression model, we tested the effects of VMP on mRS at 90 days after discharge while controlling for HIR (non-dichotomized), age, and TICI score. High (favorable) VMP predicted lower (favorable) mRS (OR = 0.544, [95 % CI 0.4-0.7]; p = 0.032), which indicates that patients with robust VMP had better neurological outcomes 90 days after discharge. Conclusion: A robust cerebral venous microperfusion profile reflects greater tissue microperfusion, good arterial collateralization status and is associated with improved clinical outcome in patients with AIS. Background & Purpose: In order to estimate the brain and head size the head circumference is typically used as a surrogate parameter. As MRI images can be freely zoomed, visual analysis often relies on "impressions" such as the craniofacial ratio or a simplified gyral pattern. Aim of this study was to validate a MRI-based method to measure the head circumference. Methods: Head circumferences of 85 children (41 micro-, 22 macrocephalies, 22 normal controls) were retrospectively calculated using a 3D-T1w (MPRAGE) data set. Three readers independently placed an ovoid ROI in an axial plane starting from the supraorbital bulge and covering the largest supra-auricular head circumference. Clinical measurements of the head circumference served as ground truth. Results: Mean deviation from clinical measurements was 6 mm or 1.3 %, respectively. Inter-observer deviations were between 5 and 9 mm. 41 of 41 microcephalies and 19 of 22 macrocephalies were reliably detected. Two patients having head circumferences at the 93rd and 95th percentiles were falsely classified as macrocephalies. Conclusion: The head circumference can be reliably determined with a simple measurement on 3D sequences using multiplanar reformations. This approach may help to diagnose a micro-or macroecephaly, especially when the head circumference is not reported by the referring clinician. 3 T MRI with TOF and contrast-enhanced MPRAGE were performed at baseline to evaluate the extent of thrombosis and vessel segments affected. Baseline and follow-up 4D combo MRV data were assessed for signs of a DAVF. Inter-rater reliability of DAVF detection and the extent of recanalization were analyzed with kappa statistics. Results: DAVF were detected in 4/24 CVT patients (16.7 %). 2/24 (8.3 %) had coincidental DAVF with CVT on admission. At follow-up, de novo formation of DAVF following CVT were seen in 2/24 patients (8.3 %). Both de-novo DAVF were low grade and benign fistulae (Cognard type 1, 2a), which had developed at previously thrombosed segments. Endovascular treatment was required in two DAVF detected at baseline for high degree lesions (Cognard 2a,b) and in one de-novo DAVF (Cognard 1) due to debilitating headache and tinnitus. Thrombus load, vessel recanalization, and frequency of cerebral lesions (hemorrhage, ischemia) were not associated with the occurrence of DAVF. Conclusions: De-novo DAVF formation occurred more frequently than previously described. Although de-novo DAVF were benign, 75 % of all detected DAVF required endovascular treatment. Therefore, screening for DAVF seems worthwhile in patients with CVT and can be performed using dynamic MRV, such as 4D-combo-MRV. Background & Purpose: Due to its high sensitivity MRI is an often used tool for cerebral staging in tumor patients. Contrary to this the relatively long examination times and the limited availability of MRI slots might lead to delayed examinations. Aim of this study was to compare an ultra-short MRI protocol to the routinely used standard protocol. Methods: Two radiologists retrospectively evaluated two sequences of a cerebral MRI (Flair images and contrast enhanced T1 MPR images) of 147 patients with malignant melanoma. The results were compared to the report of the full MRI examination and a statistical testing for non-inferiority was performed. Results: 12.93 % of the patients had cerebral metastases. Overall 79 Metastases were detected, 65 were located supratentorial and 14 were located infratentorial. Concerning the detection of cerebral metastases, the ultra-short MRI examination was not inferior to the full MRI protocol in general (p = 0.001) and separated by location for supratentorial (p < 0.001) and infratentorial (p = 0.014) metastases. Conclusion: No general recommendations for a MRI screening protocol of neurologic asymptomatic patients with cancer for cerebral metastases exist. Our study shows that an ultra-short MRI protocol for staging purpose is not inferior in detecting cerebral melanoma metastases to the routine MRI protocol. Even though a cerebral staging in neurologic asymptomatic patients is not necessary in early tumor stages, a rising demand for cerebral imaging can be expected due to a rising incidence of neoplastic diseases (1) . Shorter and faster MRI examinations could be the key for a more efficient use of the often restricted MRI capacities. "tumor" = contrast enhancing part; "edema" = edema with-out contrast enhancement and "normal appearing white matter" (NAWM) = next to the edema. Results: 8 MRI of histological proven GBM with IDH-wildtype were evaluated. In all cases lactate was detected in the defined regions, even though the measured lactate concentrations were on different levels for each patient. Taken the "tumor" area as 100 %, relative concentrations of 47.78 % (SD 9.52 %) in the "edema" and of 6.11 % (SD 1.94 %) in the "NAWM" were comparable. Conclusion: Lactate accumulates in cancer and can reflect the tumor infiltration. To get a better lactate signal, a TE of 288 ms was chosen (1) . Interestingly, different levels of lactate concentrations in between patients were detectable, whereas the relative changes of the lactate concentrations among the defined regions are comparable.

The infiltrative growth of GBM is reflected by a lactate gradient as monitored by MVS. Further studies must clarify in more detail whether these finding have a prognostic impact and can be used for therapy monitoring. Background & Purpose: While it is well known that the brain functions as a network, little is known concerning metabolic connectivity within the human brain. We aimed to uncover possible brain metabolic connectivity by determination of correlations between regional metabolite concentrations. Methods: Based on the data acquired with short echo-time wholebrain MR spectroscopic imaging from 55 healthy subjects at 3 T, N-acetylaspartate (NAA), total choline (tCho), total creatine (tCr), glutamine and glutamate (Glx), and myo-Inositol (mI) concentrations were measured in 12 regions of interest (ROIs). Pearson's correlation test was performed to assess spatial correlations between regional metabolite concentrations.

Results: Significant spatial correlations were found in several brain regions for metabolites NAA, tCho, Glx and mI. The most frequent metabolic correlations were occurred in the ROI of posterior limb of the internal capsule, i. e. to putamen for NAA and Glx, to centrum semioval for tCho and mI, to subcortical motor areal and parietal white matter for tCho, and to frontal white matter for mI; and the most spatial metabolic correlations were found for metabolite mI. Conclusion: Our preliminary results indicated possible metabolic connectivity networks in human brain. Future studies are needed to validate present results.

Patrick Nösel MRI studies and neuropathological findings in patients with 22q11.2 deletion syndrome (22q11.2DS), which is the most common microdeletion syndrome [1] , suggest anomalous early brain development [2, 3] . We aimed to evaluate morphological abnormalities of the brain in patients with 22q11.2DS and to correlate these with the most common neuropsychiatric impairments. Methods: Morphological abnormalities were assessed based on 3D T1-weighted images in 75 patients with 22q11.2DS and in 53 demographically matched controls. Three raters, blinded for disease status and for expected MR findings, identified all individuals with gray matter heterotopias and other morphological brain abnormalities. Moreover, we examined the association between the most frequent morphological findings, general cognitive performance, and co-morbid neuropsychiatric conditions. Results: Nodular heterotopia (periventricular or in the white matter) were the most frequent findings in patients (n = 33; controls n = 4, p < 0.001), followed by cavum septi pellucidi et vergae (n = 20; controls n = 0, p < 0.001), dysmorphic small neurocranium (n = 17; controls n = 1, p = 0.002), and periventricular cysts (n = 11; controls n = 0, p < 0.005). Three patients had unilateral polymicrogyria (Fig. 1) . There was no difference in psychiatric or cognitive behavior between patients with and without these morphological brain abnormalities. Conclusion: MR morphological manifestations in patients with 22q11.2DS are frequent and not related to psychiatric or cognitive manifestations. MR morphological signs of impaired brain development seem to be indicative for the pathogenesis of 22q11.2DS, but they are not a surrogate for its clinical severity. require the estimation of remaining cerebrovascular reactivity (CVR), for example by breath-hold(bh)-triggered-fMRI [1] . Recent findings suggest the use of resting-state (rs)-fMRI [2] . The aim of this study was to compare rs-fMRI to bh-fMRI. Methods: rs-and bh-fMRI data sets of 7 MMD patients were realigned, normalized, segmented into 6 standardized ROIs [3] and spatially smoothed. The bh-images were additionally slice-time corrected. The rs-data was temporally band-pass filtered (0.02-0.04 Hz). bh-CVR-maps (Fig. 1b) were calculated by voxel-wise integrating the signal time-course and rs-CVR-maps (Fig. 1a) were calculated by linear regression analysis in which the cerebellar time-course was the regressor. We compared the mean CVR of the 6 ROIs ( Fig. 2a) of all patients.

The CVR-maps of both modalities showed high correlation (correlation coefficient = 0.80, p < 0.001, Fig. 2b ). Conclusion: rs-fMRI seems to be a promising method for hemodynamic evaluation. It requires minimum patient compliance and no complex equipment. The musculature of patients suffering from neuromuscular diseases (NMD) is mainly affected by atrophy/hypertrophy, fatty infiltration, and/or edematous changes [1] . Therefore, MRI is an important tool for diagnosis and monitoring. Concerning fatty infiltration, standard T1-weighted or T2-weighted DIXON TSE sequences enable a qualitative assessment of muscle involvement [2] . To achieve higher comparability semi-quantitative grading scales, such as the 4-point Mercuri scale [3] , can be applied. However, the evaluation remains dependent on the reader's judgment. Therefore, effort is being invested to develop quantitative MRI techniques, such as proton density fat fraction (PDFF) mapping. The present work aims to assess the diagnostic value of PDFF mapping in correlation to Mercuri grading in patients with DM2, LGMD2A, and Pompe disease. Methods: T2-weighted DIXON TSE and PDFF mapping were performed in 13 patients (DM2: n = 5; LGMD2A: n = 5; Pompe disease: n = 3). Nine different thigh muscles were rated in all patients according to the Mercuri grading and segmented to extract PDFF values.

Mean PDFF values ranged from 7 to 37 % in Pompe and DM2 patients and up to 79 % in LGMD2A patients (Fig. 1) . In all three groups a high correlation of the Mercuri grading and PDFF values was observed (Table 1) .

In the investigated patient groups PDFF mapping offers the same diagnostic value as the clinically established Mercuri grading. With its greater dynamic range (enabling the assessment of more subtle changes) and the increased objectivity, PDFF should be considered a potential biomarker and alternative to Mercuri grading in the assessment of fatty infiltration of muscle tissue. Background & Purpose: Combination therapy for melanoma brain metastases (MM) using stereotactic radiosurgery (SRS) and immune checkpoint inhibition (ICI) or targeted therapy (TT) is currently of high interest 1,2 . In this collective, time evolution and incidence of imaging findings indicative of pseudoprogression is sparsely researched. We therefore investigated time-course of MRI characteristics in these patients. Background & Purpose: The application of Deep Learning for medical diagnosis is often hampered by two problems. First, the amount of training data may be scarce, limited by the number of patients diagnosed with the condition. Second, the training data may be corrupted by various types of noise. Here, we study the problem of brain tumor diagnosis with magnetic resonance spectroscopy (MRS) data, where both problems are prominent. To overcome these challenges, we propose a new method for training a deep neural network that distills par- 3388 of them localized in normal appearing brain tissue and 4054 in brain tumor. These spectra were labeled and used for analysis. Data distillation and data augmentation are used to improve labeling of the spectra and to increase the number of training samples, respectively (Fig. 1) . The deep convolutional residual neural network (ResCNN) 1 was used to classify the spectra in the tumor and non-tumor group on the validation set. Human experts also assessed the same validation set.

The proposed algorithm with data augmentation achieved an AUC of 0.77 (solid blue), which encompasses most of the neuroradiologists in the ROC plot (Fig. 2) . Background & Purpose: APT-CEST imaging allows for assessing tumor infiltration, providing information on protein concentrations and intracellular pH changes 1 . The method acquires a Z-spectrum. This can be time-consuming, especially for a large number of frequency offsets and full tumor coverage. Here, we present a fast multi-slice CEST-EPI sequence, extending the work by Sun et al. 2 via further optimizing the saturation scheme. Methods: The proposed sequence employs a pre-saturation CEST pulse train, driving the magnetization into a steady-state, followed by a secondary module comprising one CEST pulse and an EPI acquisition, being embedded in a slice and a frequency offset loop. This maintains the steady-state throughout the whole measurement, being faster than the standard scheme which employs pre-saturation for each single frequency offset. Up to 16 slices per volume can be acquired in 8 s (80 mm coverage, in-plane resolution: 3×3 mm 2 ). The sequence was tested with a maximum increment of 0.5 ppm. Both schemes were compared in-vitro on a phantom with different T1 times and in-vivo. Result: In-vitro, no significant difference was found via MTRasym(3.5 ppm) (Fig. 1) . The comparison in-vivo over 16 slices shows a similar contrast for the white matter as well as for the tumor. Only CSF has a higher signal compared to the proposed sequence (Fig. 2) . Conclusion: The proposed sequence yields a speedup by a factor of two via restriction to a single pre-saturation module. Comparison with the results of the standard scheme revealed no significant differences. [1] . These algorithms clearly carry great scientific and clinical potential. Here, we evaluate BraTS Toolkit [2] , designed to facilitate the use of state-of-the-art brain tumor segmentation.

We collected a total of 68 preoperative glioma MR exams (WHO Grades II-IV) from the TUM and REMBRANDT repositories. These were manually segmented and cross-validated by two neuroradiology residents into necrosis, contrast-enhancing tumor and edema. Cases were processed with BraTS Toolkit in a fully automated fashion and the resulting segmentations evaluated. Results: All cases were successfully segmented. A majority voting fusion of the top 5 algorithms (Maj5; by BraTS performance) outperformed all single candidate segmentations for whole tumor segmentation both in Dice score and Hausdorff's distance (95th percentile) metrics (Fig. 1) . This was corroborated when looking at structure segmentation (in particular edema and contrast-enhancing tumor). Also, false-positive detections were relevantly reduced in fusions. Conclusion: Candidate segmentation fusions of BraTS algorithms relevantly improve single algorithm segmentation performance. In combination with BraTS Toolkit, objective glioma assessment by fully-automated segmentation is readily available for everybody. Future research will focus on improving fusions through Deep Learning or taking into account patch-wise semantic information. 

The device performance contributes significantly to radiation exposure in CT-supported lumbar punctures of SMA patients. The radiation exposure of the most modern scanner is lower than the nation diagnostic reference values in all cases and results in the lowest amount of radiation exposure for needle-guidance.

Background & Purpose: Brain metastases are common in patients with lung cancer. A non-invasive imaging biomarker with the ability to distinguish small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) could accelerate the initiation of adequate therapy. Hypothetically, more cell membranes are found in solid tumor parts of Fig. 1 an SCLC than in NSCLC. Thus, the diffusion of water molecules might be more restricted in SCLS than in NSCLC. This study was undertaken with the aim of comparing ADC maps in SCLC and NSCLC. Methods: We retrospectively analyzed 1500 cranial MR scans of 410 patients (298 NSCLC, 112 SCLC) from 2008 to 2019. The ADC-ratio was calculated by dividing the ADC value of the solid part of the metastasis by a reference value, which was measured in the ADC map using a symmetric position in the healthy contralateral hemisphere, as shown in Fig. 1 .

In 188 patients (54 with SCLC) brain metastases were detected. We included 65 pre-therapeutic patients (25 with SCLC) with histologically confirmed brain metastasis of lung cancer and an available pre-therapeutic MRI. In 100 % of the pre-therapeutic patients with SCLC the ADC value was significantly lower than the reference value, ADC-ratio(SCLC) was 0.69 ± 0.13 for the solid part of the metastasis. ADC-ratio(NSCLC) was 1.17 ± 0.36. AR<0.9 (AR<0.8) shows a sensitivity of 100 % (76 %) and a specificity of 75 % (90 %) in detecting SCLC vs. NSCLC. Conclusion: In pre-therapeutic patients with lung cancer and brain metastases with solid tumor parts, ADC-ratio enables a good differentiation of SCLC and NSCLC.

Pilatus U, Wachter L, Ludin N, Matura S, Silaidos C, Pantel J, Eckert GP, Hattingen E Institut für Neuroradiologie, Klinik für Allgemeinmedizin und Brain-Imaging-Center, Goethe-Universität Frankfurt Institut für Ernährungswissenschaften, Justus-Liebig-Universität Gießen Background & Purpose: Cerebral phospholipid membranes are involved in signaling functions and cell aging processes. Estrogens can influence the fluidity and function of membrane phospholipid layers. In vivo 1H/31P MR-spectroscopic imaging (MRSI) of the brain allows quantifying different compounds of the membrane lipid metabolism to investigate sex-and/or age-related differences in healthy subjects. Methods: 1 H-MRSI and 31 P-MRSI was acquired in 130 healthy volunteers (33 young females, 35 old females, 32 young males, 30 old males; mean age 26.58 ± 0.5 (young) and 70.92 ± 1.0 (old)). Data were analyzed from the target regions (Fig. 1 ). According to [1] the sum of 31P-MRS detectable Cho-containing compounds (PCho+GPC) will amount to 1.6 mMol/l while tCho obtained from 1H-MRS will be 1.9 mMol/ l. Based on this assumption, bars scaled to the averaged signal intensity for each modality are shown in Fig. 2 . Results: MRSI revealed lower tCho in young women compared to men and to older women (p < 0.01 for GM, WM). Phosphorylated PCho and GPC were not significantly different between the groups (p = 0.45 for WM, p = 0.06 for GM). In general, the sum of PCho and GPC was less than the tCho concentration hinting to a residual choline fraction (rCho) invisible with 31P-MRSI. This fraction was lower in young females (p < 0.01 for GM; WM) and it may account for the lower tCho concentration. Conclusion: Increased rCho may be attributed to increased lipid mobility [1] . Thus, lower rCho levels may indicate higher integrity of cerebral membrane phospholipids of young females. A possible reason is the higher estrogen level in young females. Background & Purpose: To test and compare repeatability and diagnostic accuracy of brain volumetry using mdbrain's DeepVol and FreeSurfer Methods: Brain volumetry was carried out with FreeSurfer and compared to DeepVol. Both algorithms were tested on the MIRIAD* data set (45 patients with confirmed Alzheimer's disease (age 69.4 y+/-7.1 y) and 23 healthy controls (age 69.7 y+/-7.2 y) scanned over a course of 2 y including two back-to-back scans (n = 178)). Images were acquired on a 1.5 T MR-scanner using 3D-T1w images. Volumetry was performed for: whole brain, grey&white matter, frontal, parietal, occipital, temporal lobe, hippocampus and ventricles. Both algorithms were compared in terms of (a) repeatability, (b) performance and (c) sensitivity/specificity. For (a) only the back-to-back scans were used.

(b) Performance was tested with respect to the algorithm's ability to correctly identify healthy patients and those with confirmed Alzheimer's. For that, ROC was used to calculate the corresponding AUC, and (c) sensitivity/specificity on the best performing regions. Results: For (a), DeepVol showed a significantly higher stability (mean deviations of 0.25 % vs. 1.03 %). Performance analysis yielded higher AUC value of up to 0.96 for the Hippocampus compared to 0.94. Additionally, slightly increased sensitivity/specificity of 0.96/0.98 for DeepVol as compared to 0.93/0.96 were calculated when both hippocampus and temporal lobe were taken into account. Conclusion: DeepVol shows better results for (a)-(c) independent of the evaluated regions. This is reflected by the improved mean values and a decreased error. Taking into account the shorter evaluation time of ~3 min vs. ~10 h, DeepVol appears to be a valuable tool for a daily application in clinical practice.

Background & Purpose: Diffusion-weighted imaging in stimulated echo detection mode (STEAM-DWI) is an interesting alternative compared to the most commonly used diffusion-weighted echo-planar imaging (EPI-DWI). A novel STEAM-DWI, described by Merrem et al. 2017 , was routinely performed together with the "gold standard" EPI-DWI (See Fig. 1 for an example). Methods: EPI-and STEAM-DWIs with 3 mm layer thicknesses were performed between 01 July 2019 and 30 June 2020 by means of 3-T MRI in patients with suspected subacute stroke. Three neuroradiologists independently and separately assessed both the EPI-and the STEAM-DWI, stating (i) whether there was a stroke, (ii) which vessel it was associated with, (iii) the presence of artifacts and (iv) whether it was infra-or supratentorial. In case of an embolic shower, the count of impacts should be detected. The sensitivity and specificity of the STEAM-DWI compared to the EPI-DWI for detecting a stroke was determined. Results: In 53 (23 right, 21 left, 9 both hemispheres) of 85 patients a subacute stroke was confirmed using the EPI-DWI. The following territories were mainly affected: ACA 8 %, MCA 45 %, PCA 27 %, Brainstem 10 %, Cerebellum 10 %. In 51/53 cases the STEAM-DWI detected a stroke (96.2 %), in 35 of 37 patients microembolic events were noticed (94.6 %). Results show a sensitivity and specificity of 100 % (74/74) for major infarcts (> 4 mm 2 in-plan) and a sensititvity of 90.5 % (124/137) and specificity of 100 % for detecting subacute microembolic lesions. Less artifacts were noticed in the STEAM DWI. Conclusion: STEAM DWI can be used for diagnosis of subacute strokes and could provide additional information, especially in cases with a high level of susceptibility artifacts. Background & Purpose: Mechanical recanalization of acute intracranial occlusions of the anterior and posterior cerebral artery and middle cerebral artery segments distal to M2 remains a matter of debate. Embolisms to new territories (ENT) occur in up to 4-9 % according to Fig. 1 a Reproducibility and b performance tests of brain volumetry (mean +/-CI95) carried out on the MIRIAD* MR datasets using mdbrain-DeepVol (blue) and FreeSurfer (grey). The error bars show the variability using the backto back scans for a whereas for b error bars are based on recalculating the AUC multiple times by sampling with replacement over the used dataset. mdbrain-DeepVol showed significantly higher repeatability over all regions (p < 0.05). A higher performance was also calculated, although not statistically significant randomized prospective trials where distal embolisms in the territory of the primarily occluded vessel are not even included as they usually are not detected until the target vessel is recanalized. Navigation and retrieval maneuvers in distal segments may lead to higher rates of intracranial hemorrhage. Due to the clinical relevance of distal embolism to eloquent brain areas, an atraumatic and efficient system is required. Purpose: To report the experience with the Tiger 13 as an adjustable clot retriever with a low crossing profile that allows distal navigation through a 0.013" microcatheter. Methods: Distal thrombectomy with the Tiger 13 was performed in either primarily or secondarily occluded vessels that were ineligible for the use of regular stentretrievers. Results: From 2019 to 2020 43 patients from two neurovascular centers were included. 50 occlusions in M2, M3, A2, A3 and distal PCA segments were treated with Tiger 13. In 37 occlusions Tiger13 was used for recanalization of the initial occlusion or for ENT (intention-to-treat group). In 9 occlusions Tiger13 was used as a bail-out device. Successful recanalization (TICI 2b-3) was achieved in 94 %. Intracranial hemorrhage (SAH or ICH) occurred in 6 patients where none of those were classified as symptomatic. Conclusion: Tiger 13 together with a 0.013" microcatheter is safe and effective and allows for atraumatic navigation in distal occlusions of the anterior and posterior circulation.

Dominik Sepp 1 , Claus Zimmer 1 , Maria Berndt 1 , Sebastian Mönch 1 , Silke Wunderlich 2 , Benjamin Friedrich 1 , Tobias Boeckh-Behrens 1 , Christian Maegerlein 1 1 Department of Diagnostic and Interventional Neuroradiology, Klinikum rechts der Isar, Technical University of Munich, Germany 2 Department of Neurology, Klinikum rechts der Isar, Technical University of Munich, Germany

Background & Purpose: Endovascular therapy of acute ischemic stroke has proven highly effective in selected patients. But the patients' selection criteria are still under discussion. Collateral flow is known to be an important factor, but its evaluation is often subjective and time-consuming. CBV index is a presumed indicator of collateral capacity and can be provided fast and easily by automated quantitative analysis.

We evaluated in this study the relationship of the CBV index from the affected region with the clinical outcome in acute ischemic stroke patients after endovascular therapy. Methods: We included consecutive patients admitted to our hospital with acute ischemic stroke of the anterior circulation treated with end-ovascular therapy. CBV index was automatically analyzed by RAP-ID software by dividing the average of CBV from the affected region (Tmax<6 s) by normal CBV. Results: 155 patients were included in this study. The rate of successful recanalization (TICI≥2b) was 89.1 %. 66 patients (42.58 %) had a good clinical outcome (90-day mRS ≤2) and a significant improvement of the NIHSS. Median CBV index was 0.7 ± 0.16. Higher CBV index was associated with good clinical outcome (p < 0.001) and with significant NIHSS improvement (p = 0.001) also after adjustment for NIHSS at baseline, age and Aspect-score (p = 0.005; p = 0.009). Conclusion: Higher CBV index at baseline is associated with good clinical outcome in patients with acute ischemic stroke after endovascular therapy. CBV index is a presumed indicator for collateral capacity that can be obtained fast and automatically using CT-Perfusion imaging. This could improve current selection criteria for endovascular treatment especially in complex cases.

Purpose: Endovascular treatment (ET) in occlusions of the M1-and proximal M2-segment of the middle cerebral artery (MCA) is an established procedure. In contrast, ET in distal M2-occlusions has not yet been sufficiently evaluated (1) (2) . The purpose of this study was to assess relevant parameters for clinical outcome, efficacy and safety of patients undergoing ET in M1-, proximal M2-and distal M2-occlusions. Methods: One-hundred-seventy-four patients undergoing ET in acute ischemic stroke with an occlusion of the M1-or M2-segment of the MCA were prospectively enrolled (a-priori power-analysis showed power = 75 % for a moderate effect-size). Group comparisons in threemonths modified Rankin Scale (mRS), TICI scale and NIHSS were performed. Binary-logistic-regression-models were calculated for each occlusion site concluding age, NIHSS at admission, Maas Score, onset-to-recanalization-time and complication rate as independent variables and dichotomized three-months mRS as dependent variable. Results: There were no significant group differences in three-months mRS, TICI scale or complication rates between M1-and M2-occlusions nor between proximal and distal M2-occlusions. Binary-logistic-regression in patients with M1-occlusions showed a substantiate explanation of variance (NR 2 = 0.35) of mRS and significantly contributing factors were NIHSS (p = 0.009) and Maas Score (p = 0.01). Binary-logistic-regression in M2-occlusions showed a high explanation of variance (NR 2 = 0.50) of mRS but no significant results. Conclusions: Clinical outcome and procedural safety of patients with M2-occlusions undergoing ET are comparable to those of patients with M1-occlusions. Clinical outcome of patients with M1-occlusions undergoing ET is primarily influenced by the initial neurological deficit and the collateralization of the occlusions. By contrast, clinical outcome in patients with M2-occlusions is more multifactorial. tacts connected to distinct cortical areas like the supplementary motor area, fibers of the internal capsule, and structures of the basal ganglia-thalamo-cortical circuitry. Connectivity-patterns differed for the different adverse effects. Conclusion: Certain side-effects of subthalamic deep brain stimulation seem to be associated with electrode-contacts maintaining specific connectivity-patterns. We conclude that considering a symptom-specific and connectivity-based approach may improve the outcome of deep brain stimulation for Parkinson's disease by helping to achieve more individual targeting in deep brain stimulation surgery. Background & Purpose: The aim of this study was to investigate the diagnostic value of dual-layer spectral detector computed tomography (SDCT) in the detection of posttraumatic prevertebral hematoma of the cervical spine by using optimized imaging reconstructions. Methods: 38 patients with posttraumatic imaging of the cervical spine were included and underwent both SDCT and MRI. MRI was set as reference and combined conventional/electron density (C + ED) images were compared to conventional CT (CCT) images. Images were evaluated by two blinded readers. Results: 18 prevertebral hematomas were identified. The mean age of the patients was 63 ± 22.9 years. Reader 1 identified 14 of 18 and reader 2 15 of 18 prevertebral hematomas by using C + ED reconstructions. 6 of 18 and 9 of 18 hematomas were seen on CCT by reader 1 and 2, respectively. CCT showed a sensitivity of 33-50 % and a specificity of 75-80 %, while C + ED reconstructed images had a sensitivity of 77-83 % and a specificity of 85-90 %. Accuracy increased from 55-66 % to 84 % by using C + ED images. The minimum thickness for the detection of hematoma on C + ED images was 3 mm. Readers showed an excellent inter-rater reliability (kappa = 0.82) for C + ED images and a moderate inter-rater reliability (kappa = 0.44) for CCT. Conclusion: SDCT allows an increased accuracy for the detection of posttraumatic prevertebral by using combined conventional and electron density reconstructions, compared to conventional images. Background & Purpose: Treatment of cerebral vasospasm after subarachnoid hemorrhage with compliant and non-compliant balloon catheters remains controversial since it bears in principle the risk for devastating acute complications with both balloon types having different mechanical properties, which can lead to vessel injury. As a late complication, high grade stenoses have been reported sporadically [1, 2] . We analyzed the radiological follow-up of vasospastic cerebral arteries treated with different balloon types to obtain data about the incidence and relevance of long-term vessel changes [3] . Methods: We retrospectively analyzed 30 arterial vessel segments treated with compliant (n = 23) and non-compliant (n = 7) balloons for cerebral vasospasm after subarachnoid hemorrhage concerning radio- Fig. 1 Bone mineral density in patients with and without follow-up fracture. The BMD was calculated using the initial baseline CT scans logical follow-up as well as patients' clinical characteristics and functional outcome after 3 months. Results: Only mild delayed vessel narrowing was detected in 13 % of balloon treated vessels with no evident differences between both balloon types. Moderate or high grade late occurring stenoses were not observed after treatment with compliant or non-compliant balloons. Conclusion: Our data support transluminal balloon angioplasty as a safe treatment option for cerebral vasospasm concerning long-term complications with no differences between compliant and non-compliant balloons.

Background & Purpose: Patient selection for endovascular thrombectomy (EVT) is still a challenging task for neuroradiologists. Identifying patients at the earliest stage of presentation that might benefit the most from by EVT or vica versa is an imperative 1 . Here, we investigated whether machine learning (ML) workflows can support interventionalists in patient selection based on early-phase clinico-radiological and laboratory data by predicting poor outcome 2 . Methods: A single-center retrospective cohort of 172 (90 M; 52.3 %) consecutive patients undergoing EVT in 2017-2018 was retrieved from local RIS/PACS. Admission ASPECTS was extracted from reports using NLP 3 and re-evaluated by two blinded readers on imaging. Explanatory variables included age, sex, comorbidities and blood rheology parameters as well as neuro-interventional procedural data on time, retrieval count and final Thrombolysis in Cerebral Infarction following angiography. The primary outcome was the modified Rankin Scale (mRS) score at hospital discharge. Poor outcome was defined as mRS 5-6 (98; 56.9 %). Previously described multistage 5-fold cross-validated ML-workflows using random forests (RF) were applied to subsets of the features available at pre-and post-EVT 2 . Results: All pre-and post-EVT features were available for 140 cases. Eighty-five cases (60.7 %) had poor outcome. The pre-EVT-RF model showed an accuracy of 65 % while the post-EVT-RF model achieved slightly higher performance of 67.9 %. Conclusion: ML-supported patient selection for optimized EVT outcome is feasible, however, this is a hard task at the earliest stage of diagnosis even when considering several clinico-radiological and laboratory parameters.

Objective: The aim of this study was to evaluate the safety and efficacy of a manually expandable stenttriever (Tigertriever, Rapid Medical, Yoqneam, Israel) in the treatment of acute ischemic stroke caused by intracranial large vessel occlusions (LVO). Methods: We performed a single center retrospective analysis of all patients treated by mechanical thrombectomy due to LVO using the Tigertriever solely or in combination with other thrombectomy devices. The angiographic and clinical success was evaluated by the modified thrombolysis in cerebral infarction score (mTICI) and the modified Rankin score (mRS). Results: 68 acute intracranial arterial occlusions in 61 patients (42 female, median age 77, range 43-92 years) were treated by mechanical thrombectomy using the Tigertriever. The successful reperfusion rate (mTICI 2b-3) was 85.3 % (58/68 procedures) with a first past effect (mTICI 3) of 23.5 % (16/68 occlusions). In the 46 cases carried out with the Tigertriever alone (absence of other thrombectomy devices) the success rate was 91.3 % (42/46 occlusions) with a first pass effect (mTICI 3) of 34.8 % (16/46 occlusions). In seven patients a mild subarachnoid hemorrhage occurred (11.5 %). None of these patients experienced a clinical sequel. At discharge, 39.3 % of the patients (24/61) had a favourable clinical outcome (mRS 0-2). Conclusion: The Tigertriever offers a safe and effective treatment option for intracranial LVOs with promising reperfusion and low complication rates comparable to other stenttriever devices. Further comparative trials will help to prove the value of the Tigertriever among the existing technologies for mechanical thrombectomy.

[54] WEB for Atypical Aneurysm Locations Background: The safety and efficacy of the Woven Endobridge Device (WEB) has been shown in multiple good clinical practice (GCP) trials, whereas aneurysm locations in these trials were restricted to bifurcation aneurysms located at the circle of Willis (MCA bif, ICA bif, A com A, BA tip). Objective: To evaluate angiographic and clinical results with the WEB 17 in aneurysm locations that were excluded from the GCP trials, assuming that the angiographic and clinical results are similar to those of the GCP trials for aneurysms in traditional locations. Methods: Retrospective analysis of immediate and follow-up results of aneurysms in locations outside the GCP trials were the WEB 17 was applied on an intention-to-treat approach. 

The WEB 17 is safe and effective in aneurysm locations different from the traditional bifurcation aneurysms included in the GCP trials. Further studies will help to define the entire spectrum of aneurysm morphologies and locations suitable for the WEB 17.

Maus V 1 , Weber W 1 , Berlis A 2 , Maurer C 2 , Fischer S 1 1 Department of Diagnostic and Interventional Neuroradiology and Nuclear Medicine, Ruhr University, Knappschaftskrankenhaus Bochum, Bochum, Germany 2 Department of Diagnostic and Interventional Neuroradiology, University Hospital Augsburg, Augsburg, Germany Background and Purpose: The principle of flow diversion has revolutionized the therapy of brain aneurysms. In this study, we report our experience of the new Surpass Evolve (SE) flow diverter in the treatment of intracranial aneurysms. Materials and Methods: Inclusion criteria were patients suffering from wide-necked, blister-like, or fusiform/dissecting aneurysms in the anterior and posterior circulation who were treated with the SE as first-line therapy between May 2019 and June 2020 at two experienced institutions. Primary endpoint was technical success defined as favorable navigation to the target vessel and successful deployment of the SE. Secondary endpoints were favorable aneurysm occlusion defined as O'Kelly Marotta (OKM) scale C1-3+D on follow-up, procedure-related complications and retreatment. Results: Forty-six aneurysms in 42 patients were treated with 57 SE flow diverters. Median aneurysm size was 6.6 mm (IQR 4.0-12.2 mm) with a median neck width of 4 mm (IQR 2.2-5.4 mm). Forty-one aneurysms (89 %) were located in the anterior circulation and six (13 %) were ruptured. The primary endpoint was reached in 96 %. Median follow-up was 116 days (IQR 92-134 days) and available for 36/46 (78 %) aneurysms. Favorable aneurysm occlusion was seen in 31/36 (86 %) aneurysms and 27/36 (75 %) were occluded completely. An acute in-stent thrombosis occurred in one (2 %) patient. Two aneurysms (6 %) required additional treatment due to insufficient closure. Conclusion: The new SE flow diverter is safe and seems to be effective with promising occlusion rates at short-term follow-up.

Background and Purpose: Intracranial dural arteriovenous fistulas (DAVFs) are abnormal shunts between dural arteries and dural venous sinus or cortical veins. The authors report their experience with endovascular therapy of primary complex DAVFs using modern embolic agents. Methods: This is a retrospective analysis of patients with DAVFs treated between 2015 and 2019. Patient demographics and technical aspects including the use of embolic agent, access to the fistula, number of treatments, occlusion rates, and complications were addressed. Angiographic treatment success was defined as complete occlusion (CO) of the DAVF. Results: Fifty patients were treated endovascularly. Median age was 61 years and 66 % were men. The most common symptom was pulsatile tinnitus in 17 patients (34 %). The most frequent location of the DAVF was transverse-sigmoid sinus (40 %). Thirty-six fistulas (72 %) had cortical venous reflux. Non-adhesive and adhesive liquid agents were used in 92 % as single material or in combination. Complete occlusion was achieved in 48 patients (96 %). In 28 individuals (56 %) only one procedure was necessary. Non-adhesive liquid agents were exclusively used in 14 patients (28 %) with CO attained in every case. For CO of tentorial DAVFs, multiple sessions were more often required than for the other locations (55 % vs. 14 %, p = 0.0051). Among 93 procedures, the overall complication rate was 3 %. The procedure-related mortality rate was 0 %. Conclusion: Endovascular treatment of intracranial DAVFs is feasible, safe and effective with high rates of CO. In more than half of the patients the DAVF was completely occluded after a single procedure. However, in tentorial DAVFs, multiple sessions were more often required.

Background & Purpose: Visual radiological assessment of nigrosome-1 (N-1) on 3 T magnet resonance (MR) susceptibility-weighted imaging (SWI) shows a high diagnostic accuracy for Parkinson's disease (PD) versus control group 1 . Differentiation of PD from other diseases with similar clinical profile is difficult. We aimed to evaluate the accuracy of N-1 assessment as compared to 18F-DOPA positron emission tomography (PET) to diagnose PD in a highly pre-selected group of patients presenting with neurodegenerative parkinsonism in a tertiary referral center. Methods: We enrolled consecutive patients who received 18F-DOPA PET-MR imaging with SWI between 01/2014 and 06/2019. PD versus non-PD parkinsonism was determined retrospectively according to the Movement Disorder Society (MDS) Clinical Diagnostic Criteria 2 . We performed a blinded visual assessment of N-1 for each hemi-mesencephalon as described previously and scans were divided accordingly in normal, abnormal and non-diagnostic 1 . We quantified 18F-DOPA uptake and classified as normal and abnormal using a set of spherical volumes of interest manually placed in bilateral basal ganglia and occipital reference. Sensitivity, specificity and predictive values of N-1 assessment, 18F-DOPA PET and the combined approach to PD detection were calculated. Results: 50 patients (median age 65 (IQR 57.8, 74) years, 29 male, median duration of parkinsonism 2 (1, 4.5) years) met the inclusion criteria. Prevalence of PD according to MDS criteria was 38 % (n = 19/50 clinically established and probable PD). Table 1 shows the analysis of diagnostic accuracy. Background & Purpose: Recently, we found an association of direct contact aspiration thrombectomy for middle cerebral artery occlusion and higher rates of successful first pass reperfusion in patients with a regular thrombus phenotype 1 . This study aimed to assess whether thrombus surface morphology has an impact on reperfusion results in thrombectomy of acute basilar artery occlusion (BAO). Methods: We enrolled consecutive stroke patients treated by thrombectomy for acute BAO between 01/2016 and 12/2019. We assessed patients' characteristics, procedural data and thrombectomy results. We retrospectively categorized thrombus surface into regular versus irregular phenotype and analyzed first pass and final reperfusion and local recanalization results (modified treatment in cerebral ischemia (mTICI) and arterial occlusive lesion (AOL)) by blinded 3-reader-consensus as described before 1,2 . Data analysis was stratified according to thrombus phenotype. Results: 100 patients (median age 74 (IQR 65, 80) years, 63 males, NI-HSS 15 (5, 32), time from symptom onset to groin puncture 285 (190, 360) minutes) met the inclusion criteria. After consensus, 27 patients had a regular and 18 patients an irregular thrombus phenotype. Thrombus surface was not evaluable in 55 patients due to (1) poor delineation of thrombus surface secondary to bilateral inflow from vertebral arteries or early outflow via cerebellar collaterals, (2) stenosis, (3) spontaneous recanalization, (4) missing pretreatment images and (6) low-flow application of contrast medium to avoid vessel-rupture. Thrombectomy results according to thrombus phenotype are summarized in Table 1 . Conclusion: In BAO, thrombus surface phenotyping is poorly practicable. Nevertheless, regular phenotype was associated with the highest rates of successful reperfusion and local recanalization.

CT perfusion is an important tool in stroke imaging, as it enables thrombectomy in cases with an unknown or extended time window by providing dynamic 4D-data to assess both penumbra and infarct core. We present a case with accidental intra-arterial injection of contrast to demonstrate the possible effects on brain CT perfusion imaging. Methods: We present imaging findings of a stroke patient from our emergency department who was presenting with reduced vigilance and with a mild unilateral paresis that had resolved on arrival. Neurological examination was limited due to vigilance, patient history revealed a known paraplegia. Because of difficult conditions a peripheral venous catheter was placed cubital guided by ultrasound from an experienced emergency physician. Multimodal imaging including non-enhanced CT, CT angiography and CT perfusion was performed. Results: CT perfusion showed an early enhancement of the posterior circulation vessels similar to the pattern of bilateral ACI occlusion, but with quick wash out followed by a normal attenuation curve of all intracranial vessels. This was interpreted as rapid contrasting via brachial and vertebral artery with retrograde flow because of high injection rate followed by "normal" contrasting through venous circulation. Post-processing was possible because of quick wash out and no overlap between both phases of attenuation. The peripheral "venous" catheter was checked again and was found to be intra-arterial. Conclusion: Accidental intra-arterial injection should be recognized immediately to prevent delay in the time sensitive setting of stroke imaging and to circumvent repetition without resolution of the underlying problem. 

The cerebral collateral circulation has an important impact on lesion progression and clinical outcome in ischemic stroke and may even modify the effect of endovascular treatment. The purpose of this study was to quantify the effect of vessel recanalization on lesion pathophysiology and clinical outcome in patients with a poor collateral profile. Methods: 129 acute ischemic stroke patients with anterior circulation artery occlusion and a collateral score (CS) of 0-2 were included. Collateral profile was defined using an established 5-point scoring system in CT-angiography. Lesion progression was determined using quantitative lesion water uptake measurements in the admission and follow-up CT (FCT), and clinical outcome was assessed using modified Rankin Scales (mRS) scores after 90 days. Results: In patients with persistent vessel occlusion, the mean mRS after 90 days was 5.2 (95 %CI: 4.6-5.7), which was significantly higher than in patients following successful vessel recanalization (mean mRS 4.0, 95 %CI: 3.7-4.4; p < 0.001). Edema formation in FCT was significantly lower in patients with vessel recanalization versus persistent vessel occlusion (mean 19.5 %, 95 %CI: 17-22 % versus mean 27 %, 95 %CI: 25-29 %; p < 0.0001). Conclusion: Although poor collaterals are known to be associated with poor outcome, endovascular recanalization was still associated with significant edema reduction and comparably better outcome in this patient group. Patients with poor collaterals should not generally be excluded from thrombectomy. . Two peaks of contrast with the first around 3 to 6 seconds post injections and the second around 20 secons. Because there is no overlap manual post-processing was possible using only dara from later than the wash out of first peak Background and Purpose: We report on our experience with ruptured intracranial aneurysms treated with flow diverters with hydrophilic coating (p48 and p64 MW HPC, phenox, Bochum, Germany) under single-antiplatelet therapy. Methods: Patients were either treated with flow-diverter placement alone or flow-diverter and additional coiling. Due to the severity of the hemorrhage, the potential for peri-procedural re-hemorrhage and the potential for additional surgical interventions, a single-antiplatelet regimen was used in all of the patients.

The majority of the patients received an ASA single antiplatelet protocol, one patient was treated with prasugrel only, one patient was treated with tirofiban first and then switched to the ASA single-antiplatelet protocol. One device related complication occurred in form of a thrombosis of an over-stented branch. All stents however remained open at DSA, CTA or MRA follow-up. Conclusion: Single-antiplatelet therapy seems to be an option in carefully selected cases of SAH due to aneurysm rupture when the aneurysm cannot be treated otherwise. 

Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction

A new scale for age-related white matter changes applicable to MRI and CT

Ischemic stroke: experimental models and reality

Thrombo-inflammation in acute ischaemic stroke-implications for treatment

Local Leukocyte Invasion during Hyperacute Human Ischemic Stroke

Determining Lactate Gradients in Glioblastomas with MRI

PHANTOM-S: the prehospital acute neurological therapy and optimization of medical care in stroke patients-study

Prähospitale Versorgung von Patienten mit Schlaganfall

Effect of the use of ambulance-based thrombolysis on time to thrombolysis in acute ischemic stroke: a randomized clinical trial

Effects of golden hour thrombolysis: a Prehospital Acute Neurological Treatment and Opti

Combined 1H and 31P spectroscopy provides new insights into the pathobiochemistry of brain damage in multiple sclerosis

Comparison of the Diffusion-Weighted STEAM-DWI and the EPI-DWI in Ischemic Stroke

Deep Learning Segmentation for 3D Intracranial Aneurysm References

Restenosis after balloon angioplasty for cerebral vasospasm

Delayed progressive bilateral supraclinoid internal carotid artery stenosis in a patient with a ruptured basilar artery aneurysm

Non-Compliant and Compliant Balloons for Endovascular Rescue Therapy of Cerebral Vasospasm after Spontaneous Subarachnoid Hemorrhage: Experiences of a Single Center Institution with Radiological Follow-up of the Treated Vessel Segments

Machine Learning to Support Interventionalist in Patient Selection for Endovascular Thrombectomy: Early Identification of Cases

Maros 1,2 , Tabea Gerdes 1 , Chang Gyu Cho 1 , Victor Saase 1 , Benedikt Kämpgen 6 , Fabian A. Flottmann 5

Predicting clinical outcomes of large vessel occlusion before mechanical thrombectomy using machine learning

Machine learning workflows to estimate class probabilities for precision cancer diagnostics on DNA methylation microarray data

Comparative analysis of machine learning algorithms for computer-assisted reporting based on fully automated cross-lingual RadLex® mappings

Mechanical Thombectomy in Acute Ischemic Stroke Using a Manually Expandable Stenttriever (Tigertriever): Preliminary Single-Center Experience

The 'Swallow Tail' Appearance of the Healthy Nigrosome-A New Accurate Test of Parkinson's Disease: A Case-Control and Retrospective Cross-Sectional MRI Study at 3T

MDS Clinical Diagnostic Criteria for Parkinson's Disease, Movement Disorders

Impact of thrombus surface on first pass reperfusion in contact aspiration and stent retriever thrombectomy

Efficacy and safety of direct aspiration first pass technique versus stent-retriever thrombectomy in acute basilar artery occlusion-a retrospective single center experience

Pitfalls in Stroke Imaging: Accidental Intra-Arterial Injection of Contrast in Brain CT Perfusion Sondermann, Stefan

Andre 65 Kender

Jürgen 17 Klisch

Volker 64 R Raithel

Conclusion: Diagnostic accuracy of N-1 assessment and 18F-DOPA PET were comparable. Interestingly, the combination of both methods slightly enhanced the accuracy.

Little is known about catheter-based endovascular treatment of vertebrobasilar artery branch occlusion (VE-BABO) in acute ischemic stroke (AIS). Nonetheless, the experience of mechanical thrombectomy (MT) in distal small sized arteries of the anterior circulation seems promising in AIS. In this multi-center study, we report feasibility, efficacy and safety of MT in VEBABO. Methods: Retrospective analysis of consecutive AIS patients treated with MT due to VEBABO including posterior and anterior inferior cerebellar artery (PICA, AICA) and superior cerebellar artery (SCA) occlusions at seven tertiary care-centers between 01/2013-05/2020. Baseline demographics and angiographic outcomes including recanalization success of the affected cerebellar arteries and procedural complications were recorded. Clinical outcomes were evaluated by the modified Rankin Scale (mRS) at discharge and 90 days. Results: Out of 668 endovascularly treated posterior circulation strokes we identified 16 (0.02 %) cases with VEBABO MT. Most frequently, MT of the SCA was done (13/16; 81 %). Most VEBABOs occurred after MT of initial basilar/posterior cerebral artery occlusion (9/16; 56 %). In 10/16 (63 %) procedures, the affected VEBABO was recanalized successfully. Three out of four patients (75 %) with isolated VEBABO had benefited from endovascular therapy. Subarachnoid hemorrhage was observed in 3/16 (19 %) procedures. The rate of favorable outcome (mRS 2) was 40 % at discharge and 47 % at 90 days follow-up. Mortality was 13 % (2/15). Conclusion: Mechanical thrombectomy for VEBABO seems to be feasible and effective. However, the comparatively high rate of procedure-related hemorrhage highlights that the indication for MT in these occlusion sites should be weighed carefully. Background and Purpose: Evaluating the extent of cerebral ischemic infarction is essential for treatment decision and assessment of possible complications in patients with acute ischemic stroke. Patients are often triaged according to image-based early signs of infarction, de-fined by ASPECTS. Our aim was to evaluate interrater reliability in a large group of readers. Methods: We retrospectively analyzed 100 investigators who independently evaluated 20 NCCT scans as part of their qualification program for the TENSION study. Test cases were chosen by four neuroradiologists who had previously scored CT scans with ASPECTS between 0 and 8 and high interrater agreement. Percent and interrater agreements were calculated for total ASPECTS, as well as for each ASPECTS region. Results: Percent agreements for ASPECTS ratings was 28 %, with interrater agreement of 0.13 (CI95%: 0.09-0.16), at zero-tolerance allowance and 66 %, with interrater agreement of 0.32 (CI95%: 0.21-0.44), at tolerance allowance set by TENSION inclusion criteria. ASPECTS region with highest level of agreement was the insular cortex [percent agreement = 96 %, interrater agreement = 0.96 (CI95%: 0.94-0.97)] and with lowest level of agreement the M3 region [percent agreement = 68 %, interrater agreement = 0.39 (CI95%: 0.17-0.61)]. Conclusion: Despite relatively low exact interrater agreement for total ASPECTS, consensus for the decision for or against study enrollment was acceptable. Individual region analysis suggests some are particularly difficult to evaluate, with varying levels of reliability. Potential impairment of the supraganglionic region must be examined carefully, particularly in regard to the decision whether or not to perform mechanical thrombectomy.