Magnetic resonance imaging (MRI) is an important imaging modality in stroke. Computer based automated medical image processing is increasingly finding its way into clinical routine. The Ischemic Stroke Lesion Segmentation (ISLES) challenge is a continuous effort to develop and identify benchmark methods for acute and sub-acute ischemic stroke lesion segmentation. Here we introduce an expert-annotated, multicenter MRI dataset for segmentation of acute to subacute stroke lesions (https://doi.org/10.5281/zenodo.7153326). This dataset comprises 400 multi-vendor MRI cases with high variability in stroke lesion size, quantity and location. It is split into a training dataset of n = 250 and a test dataset of n = 150. All training data is publicly available. The test dataset will be used for model validation only and will not be released to the public. This dataset serves as the foundation of the ISLES 2022 challenge (https://www.isles-challenge.org/) with the goal of finding algorithmic methods to enable the development and benchmarking of automatic, robust and accurate segmentation methods for ischemic stroke.

Background: The optimal endovascular strategy for reperfusing distal medium-vessel occlusions (DMVO) remains unknown. This study evaluates angiographic and clinical outcomes of thrombectomy strategies in DMVO stroke of the posterior circulation. Methods: TOPMOST (Treatment for Primary Medium Vessel Occlusion Stroke) is an international, retrospective, multicenter, observational registry of patients treated for DMVO between January 2014 and June 2020. This study analyzed endovascularly treated isolated primary DMVO of the posterior cerebral artery in the P2 and P3 segment. Technical feasibility was evaluated with the first-pass effect defined as a modified Thrombolysis in Cerebral Infarction Scale score of 3. Rates of early neurological improvement and functional modified Rankin Scale scores at 90 days were compared. Safety was assessed by the occurrence of symptomatic intracranial hemorrhage and intervention-related serious adverse events. Results: A total of 141 patients met the inclusion criteria and were treated endovascularly for primary isolated DMVO in the P2 (84.4%, 119) or P3 segment (15.6%, 22) of the posterior cerebral artery. The median age was 75 (IQR, 62–81), and 45.4% (64) were female. The initial reperfusion strategy was aspiration only in 29% (41) and stent retriever in 71% (100), both achieving similar first-pass effect rates of 53.7% (22) and 44% (44; P =0.297), respectively. There were no significant differences in early neurological improvement (aspiration: 64.7% versus stent retriever: 52.2%; P =0.933) and modified Rankin Scale rates (modified Rankin Scale score 0–1, aspiration: 60.5% versus stent retriever 68.6%; P =0.4). In multivariable logistic regression analysis, the time from groin puncture to recanalization was associated with the first-pass effect (adjusted odds ratio, 0.97 [95% CI, 0.95–0.99]; P <0.001) that in turn was associated with early neurological improvement (aOR, 3.27 [95% CI, 1.16–9.21]; P <0.025). Symptomatic intracranial hemorrhage occurred in 2.8% (4) of all cases. Conclusions: Both first-pass aspiration and stent retriever thrombectomy for primary isolated posterior circulation DMVO seem to be safe and technically feasible leading to similar favorable rates of angiographic and clinical outcome.

Background Evidence supporting a potential benefit of thrombectomy for distal medium vessel occlusions (DMVOs) of the anterior cerebral artery (ACA) is, to the knowledge of the authors, unknown. Purpose To compare the clinical and safety outcomes between mechanical thrombectomy (MT) and best medical treatment (BMT) with or without intravenous thrombolysis for primary isolated ACA DMVOs. Materials and Methods Treatment for Primary Medium Vessel Occlusion Stroke, or TOPMOST, is an international, retrospective, multicenter, observational registry of patients treated for DMVO in daily practice. Patients treated with thrombectomy or BMT alone for primary ACA DMVO distal to the A1 segment between January 2013 and October 2021 were analyzed and compared by one-to-one propensity score matching (PSM). Early outcome was measured by the median improvement of National Institutes of Health Stroke Scale (NIHSS) scores at 24 hours. Favorable functional outcome was defined as modified Rankin scale scores of 0-2 at 90 days. Safety was assessed by the occurrence of symptomatic intracerebral hemorrhage and mortality. Results Of 154 patients (median age, 77 years; quartile 1 [Q1] to quartile 3 [Q3], 66-84 years; 80 men; 94 patients with MT; 60 patients with BMT) who met the inclusion criteria, 110 patients (median age, 76 years; Q1-Q3, 67-83 years; 50 men; 55 patients with MT; 55 patients with BMT) were matched. DMVOs were in A2 (82 patients; 53%), A3 (69 patients; 45%), and A3 (three patients; 2%). After PSM, the median 24-hour NIHSS point decrease was -2 (Q1-Q3, -4 to 0) in the thrombectomy and -1 (Q1-Q3, -4 to 1.25) in the BMT cohort (P = .52). Favorable functional outcome (MT vs BMT, 18 of 37 [49%] vs 19 of 39 [49%], respectively; P = .99) and mortality (MT vs BMT, eight of 37 [22%] vs 12 of 39 [31%], respectively; P = .36) were similar in both groups. Symptomatic intracranial hemorrhage occurred in three (2%) of 154 patients. Conclusion Thrombectomy appears to be a safe and technically feasible treatment option for primary isolated anterior cerebral artery occlusions in the A2 and A3 segment with clinical outcomes similar to best medical treatment with and without intravenous thrombolysis. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Zhu and Wang in this issue.

Diffusion-weighted MRI is critical for diagnosing and managing ischemic stroke, but variability in images and disease presentation limits the generalizability of AI algorithms. We present DeepISLES, a robust ensemble algorithm developed from top submissions to the 2022 Ischemic Stroke Lesion Segmentation challenge we organized. By combining the strengths of best-performing methods from leading research groups, DeepISLES achieves superior accuracy in detecting and segmenting ischemic lesions, generalizing well across diverse axes. Validation on a large external dataset (N = 1685) confirms its robustness, outperforming previous state-of-the-art models by 7.4% in Dice score and 12.6% in F1 score. It also excels at extracting clinical biomarkers and correlates strongly with clinical stroke scores, closely matching expert performance. Neuroradiologists prefer DeepISLES’ segmentations over manual annotations in a Turing-like test. Our work demonstrates DeepISLES’ clinical relevance and highlights the value of biomedical challenges in developing real-world, generalizable AI tools. DeepISLES is freely available at https://github.com/ezequieldlrosa/DeepIsles.

Background Numerous questions regarding procedural details of distal stroke thrombectomy remain unanswered. This study assesses the effect of anesthetic strategies on procedural, clinical and safety outcomes following thrombectomy for distal medium vessel occlusions (DMVOs). Methods Patients with isolated DMVO stroke from the TOPMOST registry were analyzed with regard to anesthetic strategies (ie, conscious sedation (CS), local (LA) or general anesthesia (GA)). Occlusions were in the P2/P3 or A2–A4 segments of the posterior and anterior cerebral arteries (PCA and ACA), respectively. The primary endpoint was the rate of complete reperfusion (modified Thrombolysis in Cerebral Infarction score 3) and the secondary endpoint was the rate of modified Rankin Scale score 0–1. Safety endpoints were the occurrence of symptomatic intracranial hemorrhage and mortality. Results Overall, 233 patients were included. The median age was 75 years (range 64–82), 50.6% (n=118) were female, and the baseline National Institutes of Health Stroke Scale score was 8 (IQR 4–12). DMVOs were in the PCA in 59.7% (n=139) and in the ACA in 40.3% (n=94). Thrombectomy was performed under LA±CS (51.1%, n=119) and GA (48.9%, n=114). Complete reperfusion was reached in 73.9% (n=88) and 71.9% (n=82) in the LA±CS and GA groups, respectively (P=0.729). In subgroup analysis, thrombectomy for ACA DMVO favored GA over LA±CS (aOR 3.07, 95% CI 1.24 to 7.57, P=0.015). Rates of secondary and safety outcomes were similar in the LA±CS and GA groups. Conclusion LA±CS compared with GA resulted in similar reperfusion rates after thrombectomy for DMVO stroke of the ACA and PCA. GA may facilitate achieving complete reperfusion in DMVO stroke of the ACA. Safety and functional long-term outcomes were comparable in both groups.

For patients with symptomatic acute occlusions of the internal carotid artery below the circle of Willis, endovascular and best medical treatment were both effective and safe therapeutic options.