RPS 1811 - Non-ischemic vascular pathology of the brain

Author Block: F. J. A. Meijer, O. Van Hulst, S. Pegge, H. D. Boogaarts; Nijmegen/NL
Purpose: The aim of this study was to evaluate the indications for four-dimensional computed tomography angiography (4D-CTA), and to assess its diagnostic accuracy for detecting and localizing cerebrovascular arteriovenous shunting, in a real-world clinical setting.
Methods or Background: A single-center retrospective cohort study was performed, where all patients who underwent 4D-CTA in the period January 2010 and January 2025 were retrieved. Demographic and clinical parameters including age, sex, and presenting symptoms were obtained from the medical records. The indications and results of angiographic studies (shunt presence, type, location, and classification) were retrieved from the study reports. Diagnostic accuracy of 4D-CTA for the detection of AV-shunting was evaluated in a subgroup of patients who underwent digital subtraction angiography (DSA) within 6 months of 4D-CTA, with unchanged clinical status.
Results or Findings: Among 366 patients, the most common indications for 4D-CTA were pulsatile tinnitus (37%) and the diagnostic work-up of cerebral hemorrhage (19%). In the subgroup of patient who also underwent DSA (n=71), 4D-CTA demonstrated a sensitivity of 96.2%, positive predictive value of 98.0%, specificity of 94.7%, and negative predictive value of 90.0% for the detection of AV-shunting. 4D-CTA demonstrated high accuracy (98%) in distinguishing between arteriovenous malformations (AVM) and dural arteriovenous fistulas (AVF).
Conclusion: 4D-CTA exhibits high sensitivity and specificity for the detection of cerebrovascular arteriovenous shunting, and high accuracy to differentiate between AVM or AVF, supporting its potential role as a first-line imaging modality prior to invasive DSA.
Limitations: The lack of confirmatory DSA following negative 4D-CTA examinations may have introduced a selection bias, as a majority of patients with positive findings on 4D-CTA underwent subsequent DSA.

The generalizability of our findings may be constrained to tertiary neurovascular centers, where referred populations often differ from broader clinical cohorts.
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: F. YAO, D. Zhu, Y. Yang; Wenzhou/CN
Purpose: Current research on intracerebral aneurysms (IAs) rupture risk assessment mainly focuses on deep learning and radiomics models, both of which rely on accurate segmentation of aneurysm regions. This study aims to create a robust, fully automated segmentation network for multi-center use and compare deep learning and radiomics models for rupture prediction.
Methods or Background: The study utilized data from the MIRACLE cohort, applying the DGIS method for automatic aneurysm segmentation. Five machine learning (ML) and deep learning models were developed for rupture prediction. The models were trained on 593 aneurysms and validated on 75 aneurysms from four hospitals. Model performance was evaluated using AUC, accuracy, sensitivity, specificity, and F1 score, along with SHAP analysis for feature importance and Grad-CAM for interpretability.
Results or Findings: The DGIS segmentation method achieved high accuracy, with Dice coefficients of 0.98 in the source domain and 0.75 in the target domain. The DRE deep learning model outperformed other models, achieving an AUC of 0.995 in training and 0.851 in external testing, surpassing radiomics AdaBoost (AUC, 0.983 and 0.634) and SVM (AUC, 0.712 and 0.656). SHAP analysis revealed key rupture risk features, and Grad-CAM highlighted the model’s focus on the aneurysm region.
Conclusion: This study presents an automated, generalizable approach for IA rupture risk analysis, offering an efficient and clinically applicable method for aneurysm detection and rupture prediction across multi-center datasets.
Limitations: First, the study primarily focused on a single type of aneurysm (middle cerebral artery aneurysms), which may limit the generalizability of the findings to other types of aneurysms. Even though the attention regions of the model can be visualized, fully understanding the internal workings of the model is still challenging.
Funding for this study: This study has received funding by the Key Laboratory of Novel Nuclide Technologies on Precision Diagnosis and Treatment & clinical Transformation of Wenzhou (Grant No. 2023HZSY0012), the Discipline Cluster of Oncology, Wenzhou Medical University, China (Grant No. z1-2023008), the Summit Advancement Disciplines of Zhejiang Province (Wenzhou Medical University – Pharmaceutics).
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: G. Jung, J. Y. Lee, H. Kim, S-M. Gho, Y. Yim, J-B. Lee, T-K. Nam, C-Y. Park, K-Y. Park; Seoul/KR
Purpose: Reliable detection of intracranial aneurysms on MRA requires substantial expertise, yet diagnostic variability among physicians remains a challenge. As clinical use of MRA expands, there is a growing demand for consistent and accurate aneurysm detection. This study aimed to assess interobserver variability in detection performance across physicians with different levels of experience and to explore the potential role of artificial intelligence as a decision-support tool.
Methods or Background: TOF-MRA studies with and without intracranial aneurysms were retrospectively collected. An inexperienced physician and a neurology resident with one year of training independently reviewed the images and marked suspected aneurysms. Their findings were compared with reference annotations provided by two board-certified neuroradiologists. After a washout period, the same readers reinterpreted the MRAs with the assistance of deep learning–based software. Diagnostic accuracy and interpretation time were compared between readings with and without AI support.
Results or Findings: A total of 642 patients were included, among whom 321 patients had 398 aneurysms (up to five per patient). Without AI assistance, the inexperienced physician required an average of 85 seconds per case, achieving a sensitivity of 63.6% (253/398) and an accuracy of 74.5%. The neurology resident required 33 seconds per case, with a sensitivity of 60.0% and an accuracy of 66.8%. With AI support, the inexperienced physician improved to 42 seconds per case, with a sensitivity of 83.4% and an accuracy of 80.2%. The resident improved to 29 seconds per case, with a sensitivity of 85.9% and an accuracy of 86.1%.
Conclusion: Deep learning–based AI assistance improved sensitivity and accuracy of intracranial aneurysm detection on MRA while reducing interpretation time.
Limitations: As the study is ongoing, the findings represent a preliminary analysis of the available data.
Funding for this study: This work was supported by 'Supporting Project to Clinical evaluation Domestic Medical Devices in Hospitals' funded by 'Ministry of Health and Welfare (MOHW)' and 'Korea Health Industry Development Institute (KHIDI)'
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approved by the Chung-Ang University Hospital Institutional Review Board. (IRB No. 2504-007-19570)

Author Block: D. Wang, S. Wang, C. Xia; Beijing/CN
Purpose: Mounting evidence indicates that early treatment of the aneurysm and aggressive prevention and management of complications can both improve patients’ clinical outcomes. This study intends to verity the effectiveness of a CT-angiography-based auxiliary triage and evaluation system for intracranial aneurysms via a clinical trial.
Methods or Background: A dual-center, retrospective, single-arm target-value design was used to evaluate the triage performance of a deep learning-based intelligent system (InferRead CTA Stroke. A total of 485 subjects were enrolled from two centers. Of these, 481 subjects completed the trial: 321 from Center 1 and 160 from Center 2. Reference standard was established by three experienced cerebrovascular imaging experts who were independent of the AI system. Primary endpoints are aneurysm triage sensitivity and specificity both over 80%. Secondary endpoint is saved time by the AI system compared to the hospital’s routine imaging-report turnaround time.
Results or Findings: The system achieved a triage sensitivity of 88.14 % (95 % CI: 82.44 %–92.50 %) and specificity of 88.16 % (95 % CI: 83.98 %–91.57 %). Because the lower bound of the confidence interval exceeded the pre-specified targets of 80% (sensitivity, p = 0.0068 < 0.01; specificity, p = 0.0004 < 0.001), the system can be considered effective in identifying positive cases and ruling out negative cases. Median software reading time was 843s, which was 3 545s faster than the hospital’s routine reporting time (p < 0.0001). In terms of mean time consumption, the system increased diagnostic efficiency by 94%. Thus, the software is expected to significantly improve reading efficiency.
Conclusion: The validated AI auxiliary diagnostic system is effective in identifying aneurysm cases in a more efficient way, exhibiting great potential to improve the clinical workflow for early detection of intracranial aneurysm.
Limitations: N/A
Funding for this study: N/A
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics Committee of Zhongnan Hospital of Wuhan University and Affiliated Zhongshan Hospital of Dalian University

Author Block: A. Brancasi, D. Parisi, A. Perna, G. Marziali, R. Di Bonaventura, M. Galeazzi, R. Russo, A. Albanese, S. Gaudino; Rome/IT
Purpose: This study investigates the diagnostic value of high-resolution preoperative MRI for unruptured intracranial aneurysms and the association between wall bleb enhancement and aneurysm instability.
Methods or Background: Intracranial aneurysm rupture is life-threatening, yet non-invasive risk assessment remains difficult. Blebs—focal protrusions of the aneurysmal sac—are considered markers of instability, though their predictive value is inconsistent. High-resolution vessel wall imaging (VWI) with post-contrast black-blood sequences suppresses intraluminal flow and depicts wall thickening and enhancement as indicators of inflammation and remodeling. Although enhancement has been linked to unstable phenotypes, its association with blebs remains uncertain. We conducted a prospective single-center study of 30 patients scheduled for surgery for unruptured aneurysms. All patients underwent preoperative 3T MRI (GE, 48-channel coil) with Gadovist, including 3D TOF and pre- and post-contrast black-blood sequences. Imaging findings were compared with intraoperative appearance, and wall tissue was collected for histopathology in nine cases.
Results or Findings: Of 30 aneurysms, 22 (73.3%) had at least one bleb confirmed by imaging and surgery. Contrast enhancement was significantly associated with blebs (73% vs. 25%; P = 0.0492). Blebs were classified as red, yellow, or white. Enhancement was most frequent in white blebs (77%), followed by red (63%) and yellow (60%), with no link between morphology and inflammation. In all histological samples, wall enhancement matched inflammatory infiltration.
Conclusion: High-resolution VWI with black-blood sequences is a valuable tool for preoperative aneurysm assessment. Bleb wall enhancement correlates with histological inflammation, supporting its role as a non-invasive biomarker of wall instability and a tool for rupture risk stratification.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: W. L. KAO¹, Y-S. Hu², C-J. Lin¹, J-K. Loo², H-C. Yang¹, C-C. Lee¹, H-J. Chiou¹, F-C. Chang¹, K-D. Liu¹; ¹Taipei/TW, ²New Taipei/TW
Purpose: Radiation-induced changes (RICs) may cause neurological deficits in patients with brain arteriovenous malformations (BAVMs) after radiosurgery. The present study investigated quantitative angiographic parameters contributing to symptomatic RICs.
Methods or Background: A total of 131 patients with supratentorial BAVMs who had not received prior treatment and underwent radiosurgery between 2011 and 2020 were included. Patients completed ≥ 24 months of MRI and clinical follow-up. MRIs and angiograms taken before radiosurgery were analyzed for morphological characteristics and quantitative angiographic parameters. Symptomatic RICs were defined as neurological symptoms attributed to RICs. The vein-artery (VA) ratio was defined as the sum of all draining vein diameters divided by the sum of all supplying artery diameters. The modified cerebral circulation time (mCCT) was defined as the interval between the bolus arrival time of the ipsilateral cavernous internal carotid artery and the parietal vein. Logistic regression models were used to evaluate associations between these parameters and symptomatic RICs.
Results or Findings: Symptomatic RICs developed in 27 (20.6%) of 131 patients. Nine patients with symptomatic RICs were hospitalized. Multivariable analysis revealed that a lower VA ratio and shorter mCCT were independently associated with symptomatic RICs. Furthermore, the quantitative angiographic model exhibited a higher performance in association with symptomatic RICs than the angioarchitectural model did.
Conclusion: A lower VA ratio and shorter mCCT were quantitative angiographic parameters of venous outflow impairment and high blood flow of BAVMs, respectively. These parameters may quantify the hemodynamic effect that contributes to symptomatic RICs development in patients with BAVMs after radiosurgery.
Limitations: Our single-center, retrospective study may have overestimated symptomatic RIC rates due to follow-up bias and excluded certain BAVM cases, limiting generalizability to other patient groups and treatments.
Funding for this study: Funding was provided by Taiwan’s Ministry of Science and Technology (grant number: MOST 112-2314-B-A49-064- to C.J.L. and MOST 113-2314-B-A49-051- to C.J.L.) and Melissa Lee Cancer Foundation.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the Institutional Review Board (3) of Taipei Veterans General Hospital and supervised by the Institutional Review Board (3) of Taipei Veterans General Hospital. After the review by Human Research Protection Center, the implementation of the study was approved.

Author Block: T. Nemirovskaya¹, A. Nemirovskiy²; ¹Kazan, Moscow/RU, ²Kazan/RU
Purpose: Extracranial-intracranial bypass (ECIC) surgery may be useful for internal carotid artery (ICA) chronic occlusion. Most cases are secondary to atherosclerosis, however other mechanisms exist, e.g. post-radiotherapy damage, vasculopathy, developmental anomalies. We use CT-perfusion (CTP) as key modality for revascularization candidates selection. Purpose of study was assessment of perfusion patterns in unilateral, bilateral atherosclerotic and moya-moya disease/syndrome ICA occlusion.
Methods or Background: Retrospective assessment of pre-operation CTP-maps of 271 patients, later underwent ECIC in our Neurosurgery department, was performed. All underwent CTP according to standard institutional protocol. Perfusion maps: cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT). Six regions of interest were manually outlined bilaterally in main supply areas. Mean hemispheric values and mean ICA-area values were calcualted with interhemispheric ratios assessment.
Results or Findings: A total 271 patients with chronic unilateral (234) and bilateral (27) ICA occlusion and moya-moya disease (10) were selected for ECIC surgery. Candidates criteria were ipsilateral MTT elevation, CBF reduction, increased CBV. Unilateral occlusion group CBF was 34.785 ml/100g/min, interhemispheric asymmetry 76.59% , ICA areas asymmetry 65.159% . No significant difference of interhemispheric CBF in bilateral occlusion group, 37.612 and 35.933ml/100 g/min and symmetry 104.67%, were observed, however anterior-posterior asymmetry was detected. Moya-moya group demonstrated highest CBF, 49.251 and 57.913ml/100 g/min, with asymmetry 81,806%. P-value <0.0001. The highest MTT values were observed in unilateral occlusion group (P = 0.05). CBV values were high and not significantly differed. We established perfusion patterns for each group with highest interhemispheric asymmetry in unilateral occlusion, anterior-posterior pattern in bilateral occlusion and asymmetric increased values in moya-moya group.
Conclusion: Specific CTP patterns depending on different aetiology of chronic ICA occlusion should be considered surgical treatment candidates evaluation.
Limitations: Retrospective assessment of CT perfusion maps
Funding for this study: No
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: L. Fetscher¹, E. Siebert¹, C. Csapo-Schmidt¹, N. Mahmoudi¹, O. Reihani¹, D. Kupka², M. P. Wattjes¹, J. Nawabi¹; ¹Berlin/DE, ²Falun/SE
Purpose: Endovascular treatment of chronic subdural hematoma (cSDH) by middle meningeal artery embolization (MMAE) is rapidly expanding, yet treatment success remains difficult to predict. Existing grading tools on angiographic imaging neglect branch anatomy, dominance, and collateral supply. We therefore developed the BE-EMMA Score, a branch enhanced, angiographic grading system, intended to improve reliability and enable prognostic correlation.
Methods or Background: We retrospectively reviewed 50 MMAEs in 38 patients with cSDH (Charité, 2023-2024). Initial inter-rater testing of a preliminary version of the BE-EMMA Score (capturing branch dominance, occlusion, and collaterals) in 19 cases was followed by refinement in consensus with four interventional (neuro-)radiologists. The refined scoring system integrates pre-, intra- and post-embolization parameters: (i) branch-specific occlusion grade (0-3; grade 4 = proximal stasis without distal penetration, liquid agents only), (ii) branch dominance (F/f = frontal, P/p = posterior), (iii) aplasia or surgical interruption, (iv) main trunk diameter (D/d for ≥/<1.5 mm), and (v) collateral status (C+ = relevant supply, C− = hazardous collaterals prohibiting embolization).
Evaluation of inter-rater reliability (four inter-raters) and correlation with treatment outcome is ongoing in a final dataset of 50 MMAE-cases.
Results or Findings: Using the preliminary version, near-complete occlusion (grade ≥2) was achieved in 84% of divisions; full bilateral occlusion in 53%; relevant collaterals (C+) in ~20%. Inter-rater agreement was strong (κ = 0.84 overall; occlusion κ = 0.88; collaterals κ = 0.85).
Conclusion: The BE-EMMA Score provides an accurate, imaging based, reproductible, and anatomically detailed evaluation of MMAE. By integrating multiple anatomical and procedural dimensions, it extends beyond descriptive reporting toward a standardized framework that may facilitate future correlation with treatment success and clinical outcome.
Limitations: Broader validation is needed to confirm prognostic value and clarify potential use in clinical decision-making.
Funding for this study: Future prospective research including validation of the BE-EMMA Score is funded by the ESMINT Pilot Research Grant. Development of the score itself was not funded.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics approval for the retrospective study of MMAE cases (2022–2024) was granted by the institutional review board of Charité – Universitätsmedizin Berlin (EA4/181/24).

Author Block: D. Turilli, L. Piscopo, A. Dessì, C. Pinna, L. M. Fattacciu, E. Solinas, I. Conti, S. Masala, M. Scaglione; Sassari/IT
Purpose: Neonatal birth-related intracranial subdural hemorrhages (SDHs) represent a form of bleeding inside the skull that occurs in newborns. This condition includes the extravasation of blood both in the encephalic parenchyma and in the extra-axial spaces. Recent studies have shown that SDH and particularly post-traumatic birth-related hemorrhages represent a frequent occurrence, but they are often asymptomatic. The gold standard for the diagnosis and follow-up of patients with SDH is multiparametric Magnetic Resonance Imaging. The aim of this study is to describe our experience by reporting several cases of SDH with different distribution and Central Nervous System involvement by the MRI of this pathology in infants up to 30 days of age.
Methods or Background: We analyzed the age and sex of the patients included in this study, the localization of SDH in different CNS areas, and their frequency using distribution plots and pie charts.
Results or Findings: About the analysis of the SDH locations in the 32 patients, the most common location was the cerebellum (31/32, 96.9%), followed by parietal and occipital lobes (19/32, 59.4%; 18/32, 56.2%, respectively), falx cerebri (11/32, 34.4%), tentorium cerebelli (10/32, 31.2%), temporal lobes (6/32, 18.7%), and finally cervical and dorsal spine in the same patients (4/32, 12.5%). According to SDH locations, the patients were divided into supratentorial, infratentorial, both, and Spinal Canal.
Conclusion: Our study confirmed the literature data regarding the neonatal birth-related SDH high frequency, but also allowed us to focus our attention on the rarest spinal SDH localizations with the same benign evolution.
Limitations: The small sample object of the study.
Funding for this study: This research received no external funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The local Institutional review board approved the study. The Ethical
Committee’s approval was obtained (2022030) (28 December 2023).

Author Block: V. L'ALLINEC; Angers/FR
Purpose: Reversible Cerebral Vasoconstriction Syndrome (RCVS) is characterized by thunderclap headaches and reversible segmental cerebral arterial narrowing. In the absence of guidelines, radiological management remains heterogeneous. This study aimed to evaluate diagnostic and follow-up imaging practices across French hospitals.
Methods or Background: A multicenter cross-sectional survey was conducted between February and May 2024 in 25 French neuroradiology centers. A self-administered online questionnaire explored initial imaging strategies, CT and MRI protocols, confirmatory and long-term follow-up imaging, and the role of invasive techniques. Responses were analyzed descriptively.
Results or Findings: Twenty centers (80%) responded, mostly university hospitals. CT was the preferred first-line modality (85%), mainly to exclude aneurysmal subarachnoid hemorrhage. MRI was used in all centers at some stage, with a common protocol including DWI, FLAIR, SWI, and TOF-MRA. Practices diverged beyond this: CTA coverage varied (head and neck vs intracranial only), advanced MRI sequences such as vessel wall imaging (40%) and ASL (15%) were inconsistently used. After positive initial imaging, 40% of centers did not perform confirmatory exams, while 75% repeated imaging after negative but clinically suspicious cases. Long-term follow-up was performed by 90% of centers, most often at 3 months, to confirm reversibility. No center reported invasive endovascular treatments.
Conclusion: This first national survey highlights consensus on initial non-invasive imaging but substantial heterogeneity in subsequent diagnostic and follow-up pathways. Such variability reflects the low incidence of RCVS and the absence of formal guidelines. These findings emphasize the need for national recommendations and collaborative prospective studies.
Limitations: Declared practices may differ from real-world management. The predominance of university hospitals may limit generalizability, and descriptive design precludes outcome comparisons. Rapidly evolving imaging technologies could also impact practices
Funding for this study: Investigator-initiated study with no external funding.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: B. Li¹, Y. Liu¹, X. Du¹, Z. Hong-Tao¹, R. Xie¹, M. Lu², X. Zhao¹, J. Cai¹; ¹Beijing/CN, ²Tianjin/CN
Purpose: This study aims to examine the determinants associated with persistent dizziness (RD) in adult patients diagnosed with Moyamoya disease (MMD) following successful Encephalo-Duro-Arterial-Synangiosis (EDAS) surgery.
Methods or Background: In patients with dizziness as the initial symptom of MMD, post-surgery dizziness improvement is noted, but mechanisms remain unclear. This study included 78 adults with MMD who initially experienced dizziness. We recorded patient characteristics, comorbidities, and DHI scores. CVR was assessed pre- and post-EDAS surgery with breath-modulated BOLD-fMRI. DSC-MRI measured TTP, MTT, CBV, and CBF in middle cerebral artery (MCA) regions. White matter was evaluated with LA-score MRI. Patients were categorized into RD and non-RD groups based on dizziness within 3 months post-surgery. We then assessed the impact of CVR, perfusion, and white matter on RD in MMD patients.
Results or Findings: In the RD group, pre- and post-surgical CVR was markedly lower than in the non-RD group, accompanied by elevated DHI and LA scores, with statistically significant differences (P < 0.05). Decreased CVR in MCA territory, along with increased DHI and LA scores, were identified as risk factors for early RD post-indirect surgery in adults with MMD. Multivariate logistic regression confirmed CVR, DHI, and LA scores as independent RD risk factors. ROC analysis revealed AUCs of 0.705 for CVR, 0.720 for LA score, and 0.782 for DHI score. Combined, these indicators achieved an AUC of 0.821, indicating high diagnostic accuracy (sensitivity 68.93%, specificity 89.54%).
Conclusion: Reduced CVR in MCA region, along with increased DHI and LA scores, predict short-term residual dizziness in adult MMD patients post-EDAS surgery. The combined assessment of CVR, DHI, and LA scores provides superior predictive value for RD.
Limitations: The study was conducted at a single medical center, which may limit the generalizability.
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: the fifth medical center of Chinese PLA general hospital; Institutional Review Board (IRB number: KY-2022-4-29-1)