Validation of Artificial Intelligence-Assisted Imaging for Stroke 

Data from several new studies presented at the American Heart Association International Stroke Conference validate the sensitivity and specificity of an artificial intelligence-assisted imaging system (Viz LVO and Viz LCH; Viz.ai, San Francisco, CA) for evaluation of stroke in real-world settings. 

Studies presented by researchers from 3 different institutions found positive impacts on workflow and patient care related to use of multiple Viz.ai platform modules, including Viz LVO (for large vessel occlusion) and Viz ICH (for intracranial hemorrhage).
 
Real-world data from University of California at San Diego showed reductions in door-to-groin (DTG) time of 41 minutes for individuals with LVO who arrived directly to a comprehensive stroke center. Data from Ohio State University demonstrated the accuracy of Viz ICH in identifying and alerting stroke teams in 82 of 83 ICH cases detected by radiologists.
 
A recent publication from Mount Sinai Hospital presented data from a study of 1,822 CT angiography (CTA) scans demonstrating detection rates of 100% and 93% for ICA terminus and M1 occlusions, respectively. An additional abstract from the same center studied 682 patients analyzed with Viz ICH and found an overall accuracy rate of 99%. 

According to the publication, “In this work, the diagnostic accuracy of Viz LVO was tested in real life and real time across one of the largest prospective consecutive cohorts to date, in a multiple-tiered healthcare system. Viz LVO is a promising AI-driven software that can reliably detect ICA-T and M1 LVOs with impressive negative predictive value (NPV), sensitivity, and overall accuracy. It is a useful adjunct in triaging patients with a LVO stroke at varying levels of stroke centers.”