Using Deep-Learning Algorithms to Simultaneously Identify Right and Left Ventricular Dysfunction From the Electrocardiogram
Problem
Rapidly assessing left and right ventricular (heart pump) function from a routine ECG was not possible; no deep learning tools existed for right ventricular function and left ventricular tools covered only very low function.
Why it matters
Undetected ventricular dysfunction delays diagnosis and treatment; ECG-based screening could speed clinical workflow.
What the AI did
Deep learning models were developed to classify LVEF into categories from the universal definition of heart failure, estimate LVEF by regression, and predict RV systolic dysfunction or dilation, using multicenter data with external validation at a fifth hospital.
Results
External validation performed at 1 of 5 New York City hospitals; concrete performance metrics not stated in source
Impact
The models could assist diagnostic workflow by estimating ventricular function from ECGs.
Record
| Venue | JACC. Cardiovascular imaging |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Akhil Vaid, Kipp W. Johnson, Marcus A. Badgeley, Sulaiman S. Somani, Mesude Bicak, Isotta Landi, Adam J. Russak, Shan Zhao et al. |
| Citations | 124 (OpenAlex, at import time) |
| References | 76 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1016/j.jcmg.2021.08.004 |
| Open access | hybrid |
| Keywords | deep learning, ECG, ventricular dysfunction, LVEF, echocardiography, diagnosis |
Importance rating
significant — thousands to millions affected
ECG-based screening for ventricular dysfunction could speed diagnosis for a common, serious condition; multicenter study across 5 hospitals with ~148k patients and strong AUCs (0.94 for LVEF≤40%).