Artificial Intelligence–Parallel Stroke Workflow Tool Improves Reperfusion Rates and Door‐In to Puncture Interval

Abstract

Background The Viz.ai artificial intelligence (AI) software module Viz large‐vessel occlusion (LVO) utilizes AI‐powered LVO detection and triage technology to automatically identify suspected LVOs through computed tomographic angiogram imaging and alert on‐call stroke teams. We performed this analysis to determine whether the Viz LVO software can reduce the door‐in to puncture time interval within a comprehensive stroke center (CSC) for patients requiring endovascular treatment. Methods We conducted a retrospective chart review that compared the time interval between patient arrival in the emergency department (door‐in) to puncture for patients who presented consecutively with a stroke code to our CSC between November 2016 and May 2020. The implementation of the AI software (Viz LVO) at the CSC was in November 2018. Using a prospectively collected database at the CSC, demographics, and outcomes were examined. This study was based on evaluating real‐world practice. Results We analyzed 86 patients from the pre‐AI study phase (average age, 68.53±13.13 years, 40.7% female) and 102 patients from the post‐AI study phase (average age, 69.87±15.75 years, 43.1% women). Following the implementation of the software, the mean door‐in to puncture time interval within the CSC significantly improved by 86.7 minutes (206.6 versus 119.9 minutes; P <0.001); significant improvements were also noted in the rate of reperfusion (modified Thrombolysis in Cerebral Infraction 2B‐3) for patients in the post‐AI population ( P =0.036). Conclusion The incorporation of the software was associated with a significant improvement in treatment time within the CSC, as well as significantly higher rates of adequate reperfusion. Prospective, multicenter, controlled studies with a larger cohort are warranted to expand on the ability of Viz LVO to improve treatment times and outcomes for patients with an LVO stroke.

Ähnliche Arbeiten