Multi-valve detection — mitral, tricuspid, and aortic — built on the same research foundation that produced our cleared products.
The research foundation behind our cleared products extends directly into multi-valve detection. Each valvular pathology is being developed on the same infrastructure, the same training discipline, and the same evidence standards that produced Precision LVEF and Precision Cardiac Amyloid.
Mitral regurgitation detection was published in Circulation (2024). A fully automated deep learning pipeline identifies apical-4-chamber view videos with color Doppler across the mitral valve and assesses regurgitation severity — developed on 58,614 studies from Cedars-Sinai Medical Center and externally validated at Stanford Healthcare. The model correctly identifies clinically significant (moderate or severe) MR with strong agreement to expert cardiologist readers.
Tricuspid regurgitation detection was published in JAMA Cardiology (2025). Led by Amey Vrudhula and colleagues, the workflow identifies A4C color Doppler tricuspid videos and characterizes TR severity. It was developed on 47,312 studies from Cedars-Sinai and validated across temporally distinct (CSMC, 2022) and geographically distinct (Stanford Healthcare) test cohorts, detecting moderate-or-severe TR with AUC 0.928 and severe TR with AUC 0.956.
Aortic valve work is in active development, building on the multi-view architecture established across the rest of our platform. Clearance pathways will follow the same disciplined route: peer-reviewed publication, prospective validation, and regulatory submission.