AI-Collateral CTA Tool Rapidly and Accurately Detects Anterior Circulation Large Vessel Occlusions

An AI algorithm for CTA detected anterior circulation large vessel occlusions with 95.8% sensitivity and 88.6% specificity. In a multicenter study of 379 patients, mean time-to-notification was just 3.30 minutes, supporting its use as a rapid triage tool.

• The multicenter study included 379 patients (142 LVO, 237 non-LVO) from four institutions, with ground truth established by expert neuroradiologists.
• Performance was robust across different institutions and CT vendors, with sensitivity by occlusion subtype of 98.5% for M1, 89.7% for M2, and 97.1% for ICA occlusions.
• The authors noted that error analysis suggested several apparent false positives and negatives were due to ground-truth ambiguity or severe stenosis rather than pure algorithmic error.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.