The Implementation of an Artificial Intelligence-Based Software for Large Vessel Occlusion Stroke Improves Door-To-Puncture Time at a Comprehensive Stroke Center

Division

East Florida

Hospital

Kendall Regional Medical Center

Document Type

Manuscript

Publication Date

11-20-2025

Keywords

Artificial intelligence, Large vessel occlusion, Stroke, Thrombectomy

Disciplines

Cardiovascular Diseases | Emergency Medicine | Nervous System Diseases | Quality Improvement

Abstract

Endovascular thrombectomy (EVT) is the standard treatment for large vessel occlusion (LVO) acute ischemic stroke (AIS), with a target door-to-device time of under 90 min. While prior reports have indicated that artificial intelligence (AI) software can reduce door-to-procedure times, we set out to study its impact in our own patient population. A retrospective cohort study was conducted at a comprehensive stroke center located in a densely populated, urban region consisting of a majority Hispanic population. Data were analyzed from 222 patients who underwent EVT between July 2018 and July 2022. The study compared workflow and patient outcomes before (n = 84) and after (n = 138) the implementation of an AI software (Viz LVO, Viz.ai, Inc.) in July 2020. The primary outcome was the fraction of EVT patients who received thrombectomy within 90 min. The proportion of patients undergoing thrombectomy within 90 min increased significantly from 26.5% (95% CI 17.9-36.8%) to 68.1% (95% CI 60.0-75.5%, p <  0.001). The median door-to-groin (DTG) time decreased from 109 min pre-Viz to 75 min post-Viz (p <  0.001). An interrupted time-series analysis confirmed a significant reduction in DTG time following the AI software implementation (p = 0.001). The percentage of AIS patients who received EVT went up between pre- and post-AI periods. No differences were found in TICI scores, length of stay, or mortality/hospice rates. Implementation of an AI-powered platform was associated with a significant increase in the fraction of patients who received EVT within 90 min and decrease in DTG time. An AI-based system helped optimize stroke care workflow in an urban center with a diverse patient population.

Publisher or Conference

Internal and Emergency Medicine

Link to Full Text

Share

COinS