Duarte Armindo, Rui

BACKGROUND: Intracranial atherosclerosis (ICAS) is associated with an increased risk of early recurrent ischemic stroke. We evaluated whether biomarkers of impaired distal perfusion—specifically, anterior circulation borderzone infarct, and hypoperfusion mismatch volume—were associated with recurrent ischemic stroke within 90 days. METHODS: The BIORISK ICAS (Biomarkers and Recurrence Risk in Symptomatic Intracranial Atherosclerosis) is a multicenter retrospective international study (35 sites) that included hospitalized patients with symptomatic ICAS (50%–99% luminal stenosis of the intracranial vertebral, basilar, distal internal carotid, or proximal middle cerebral artery) from January 2019 to June 2024. The primary outcome was recurrent ischemic stroke in the territory of the symptomatic artery within 90 days. In the primary analysis, the exposure was an acute anterior circulation borderzone infarct. In secondary analysis of the subset with presentation within 72 hours of last known normal and perfusion imaging completed, the exposure of interest was prespecified as hypoperfusion mismatch volume at T max (time to maximum) threshold of 6 seconds, dichotomized at the Youden index. We performed multivariable Cox regression to test associations between exposure variables and the outcome, adjusting for clinically relevant variables and those associated with the outcome ( P <0.1). RESULTS: Of 2050 patients with symptomatic ICAS, 1737 (84.7%) presented within 72 hours of symptom onset, among whom 509 (29.3%) underwent perfusion imaging. The primary analysis included 1891 patients; 174 (9.2%) patients had recurrent ischemic stroke in the symptomatic arterial territory. In adjusted Cox regression models, there was an association between anterior circulation borderzone infarct and recurrent ischemic stroke at 90 days (adjusted hazard ratio, 1.40 [95% CI, 1.02–1.93]). In the perfusion imaging analysis, hypoperfusion mismatch of ≥10 mL was associated with recurrent ischemic stroke (adjusted hazard ratio, 1.83 [95% CI, 1.03–3.28]). CONCLUSIONS: Biomarkers of impaired distal perfusion, anterior circulation borderzone infarct, and hypoperfusion mismatch were associated with increased risk of recurrent ischemic stroke. These findings support the use of perfusion imaging in ICAS as well as future trials investigating early reperfusion in high-risk patients with ICAS.

The translation of AI-generated brain metastases (BM) segmentation into clinical practice relies heavily on diverse, high-quality annotated medical imaging datasets. The BraTS-METS 2023 challenge has gained momentum for testing and benchmarking algorithms using rigorously annotated internationally compiled real-world datasets. This study presents the results of the segmentation challenge and characterizes the challenging cases that impacted the performance of the winning algorithms. Untreated brain metastases on standard anatomic MRI sequences (T1, T2, FLAIR, T1PG) from eight contributed international datasets were annotated in stepwise method: published UNET algorithms, student, neuroradiologist, final approver neuroradiologist. Segmentations were ranked based on lesion-wise Dice and Hausdorff distance (HD95) scores. False positives (FP) and false negatives (FN) were rigorously penalized, receiving a score of 0 for Dice and a fixed penalty of 374 for HD95. The mean scores for the teams were calculated. Eight datasets comprising 1303 studies were annotated, with 402 studies (3076 lesions) released on Synapse as publicly available datasets to challenge competitors. Additionally, 31 studies (139 lesions) were held out for validation, and 59 studies (218 lesions) were used for testing. Segmentation accuracy was measured as rank across subjects, with the winning team achieving a LesionWise mean score of 7.9. The Dice score for the winning team was 0.65 ± 0.25. Common errors among the leading teams included false negatives for small lesions and misregistration of masks in space. The Dice scores and lesion detection rates of all algorithms diminished with decreasing tumor size, particularly for tumors smaller than 100 mm3. In conclusion, algorithms for BM segmentation require further refinement to balance high sensitivity in lesion detection with the minimization of false positives and negatives. The BraTS-METS 2023 challenge successfully curated well- annotated, diverse datasets and identified common errors, facilitating the translation of BM segmentation across varied environments and providing the tools for future development of personalized volumetric reports to patients undergoing BM treatment.