Montreal Heart Institute (HeartWise.AI) (Nolin-Lapalme, Avram et al.)
Open-weight video-based deep neural network that predicts reduced right-ventricular systolic function (RVSF) directly from routine left and right coronary angiogram videos, enabling real-time RV-dysfunction screening in the catheterization lab when echocardiography is unavailable. Built on an X3D-M spatiotemporal video architecture (Kinetics-400 pretrained) that aggregates per-video probabilities into a study-level normal-vs-reduced RVSF classification, with Grad-CAM/Guided-Backpropagation explainability confirming attention to RV-specific coronary motion rather than left-ventricular signal. Trained on 8,053 angiographic studies from 6,923 Montreal Heart Institute patients (2017-2023), externally validated at UCSF, and prospectively deployed at MHI via the PACS-AI platform, where AI assistance improved reader accuracy from 72.1% to 77.6% for cardiologists and 43.5% to 64.0% for medical students.
Architecture
CNN (3D)
X3D-M video spatiotemporal CNN (Kinetics-400 pretrained), aggregating per-video probabilities across multi-view coronary angiograms into a study-level normal-vs-reduced RVSF classification; an MViT-2B-S transformer architecture was also explored
Framework
PyTorch
Added to catalog
2026-08-10
Montreal Heart Institute Angiography-RVSF Cohort
8,053 coronary angiography studies from 6,923 patients at the Montreal Heart Institute, labeled for right-ventricular systolic function (RVSF, normal vs. reduced) from paired transthoracic echocardiography per 2025 ASE guidelines; used to train DeepRV.
UCSF External Angiography-RVSF Validation Cohort
2,247 coronary angiography studies from UCSF (27.7% reduced-RVSF prevalence), used as an external validation cohort for DeepRV.
Reduced right ventricular systolic function (RVSF) classification from coronary angiography video (normal vs. reduced)