Abstract TP129: Automated Vessel Density Maps on CT Angiography Identify M2 Occlusions

Abstract

Introduction: Patients with occlusions of the M2 branch of the middle cerebral artery are frequently considered for endovascular thrombectomy. Most AI software programs focus on identification of large vessel occlusions and frequently miss M2 inclusions. An automated software program with good sensitivity for detecting more distal occlusions could help rapidly identify patients for reperfusion therapies. Methods: Consecutive patients with a final diagnosis of occlusion of an M2 branch of the middle cerebral artery were identified from the Johns Hopkins Stroke database. All occlusions were verified by a neuroradiologist (VY). The baseline CTA from these patients was evaluated with the Rapid CTA vessel density software. This software identifies a reduction in vessel density within the middle cerebral artery territory and classifies the vessel density reduction as as: 20-25%, 26-40%,41-55% or >55%. A reduction of greater than 20% or greater was consider positive as a potential indicator of M2 occlusion. Adding a sample of about 50 pts with no evidence of vessel occlusion would be desirable to be able to report specificity. Results: Forty-four of 62 patients (71%) with M2 occlusions had a reduction in vessel density of at least 20% on automated CTA. Among the positive cases, 20 percent had a 20-25% reduction, 20% had 26-40%, 30% had 41-55% and 30% had a >55% reduction. Conclusion: The Rapid CTA hyperdensity program identified 71% of M2 occlusions with most patients having more than a 25% reduction in vessel density. Automated software has the potential to quickly identify patients with M2 branch inclusions.