To determine if utilization of Viz.ai artificial intelligence software improves detection of large vessel occlusions (LVO) in a single primary stroke center (PSC) for patients requiring transfer for mechanical thrombectomy. 

Viz.ai artificial intelligence (AI) software uses AI powered detection of large vessel occlusions (LVO) using CT angiogram imaging to alert on-call stroke teams. Our previous study showed significantly improved treatment time at our comprehensive stroke center (CSC). We performed an analysis to determine if Viz.ai software improves detection of LVO patients in single primary stroke center (PSC) requiring transfer for mechanical thrombectomy.

We compared acute ischemic stroke (AIS) patients with LVO from a single PSC transferred to our CSC before and after the implementation of Viz.ai software. Pre Viz.ai study phase included patients from November 2015 to October 2018 and post Viz.ai study phase from November 2018 to August 2021. Using our CSC stroke database we analyzed demographic information, NIH stroke scale on admission and discharge, and rate of hemorrhagic transformation of ischemic stroke.

We analyzed 20 patients in the pre Viz.ai study phase (average age 69.5 yrs +/- 12.68, 35% females, 90% Hispanic) and 39 patients in the post Viz.ai study phase (average age 70.1 +/- 14.08 yrs, 48.7% females, 97.4% Hispanic). Baseline demographics were similar between groups with no statistically significant difference in age, NIH stroke scales on admission and discharge. Mean rate of hemorrhagic transformation was decreased in the post Viz.ai study phase patients (pre 0.25 vs post 0.06, p=0.0247). After the utilization of Viz.ai, there was a 95% increase in LVO patients transferred for mechanical thrombectomy for treatment of ischemic stroke.

Our study supports the use of Viz.ai software for improved detection of LVO. We found an increase in transfers of AIS patients with LVO from a primary stroke center in a largely underserved area.