Skip to main content
SIR 2025 Event Main banner
BSCI - Consideration Cases Ad 1
Merit/DPSSME Siege
Back

SIR 2025

Venous Interventions

Traditional Poster

Session: Traditional Poster Presentations

118 - Artful Research: Using Artificial Intelligence to Screen Potential Research Patients

Monday, March 31, 2025
12:10 PM – 1:10 PM CT
Location: Music City Center, Expo Poster Area
Purpose: This study uses Viz.AI, an artificial intelligence application to screen potential subjects for inclusion criteria in the PE-TRACT trial.

Materials and Methods: In this study we use the VizRecruit version 1.93.10 application provided by Viz.AI San Francisco, CA. This AI screening application filters computed tomography chest exams performed at Weill Cornell Medicine (WCM) for inclusion criteria including presence of main and lobar pulmonary embolism, and measured RV/LV ratio >1.0. This research is ongoing, the data for this submission covers subjects screened from study inception at WCM from November 2023 to September 2024.

Results: To date, this study site has screened 285 subjects and enrolled 23. Of those screened, 217 were identified by the Viz.AI application alone, and 68 were identified through traditional pulmonary embolism response team (PERT) notifications through Epic. 39 were identified through both Viz.AI and PERT. Of the 23 subjects enrolled 11 were identified through PERT alone; 2 were identified only by Viz.AI; and 10 were identified by Viz.AI and PERT simultaneously. In those cases where Viz.AI and PERT identify the same subject, the Viz.AI notification reached researchers approximately 20 minutes before the PERT notification was triggered. The Viz.AI application also provides a real-time measurement for RV/LV ratio for inclusion purposes.

Conclusion: The Viz.AI application for screening enhances efficiency of screening large numbers of potential subjects, cuts down on time to notification of researchers, and expedites screening measures for inclusion into the study. It is not as effective as PERT notifications in identifying subjects who ultimately enroll but does provide a larger population of subjects for screening.