External validation of a deep learning electrocardiogram algorithm to detect ventricular dysfunction
Problem
Detecting left ventricular systolic dysfunction (LVSD), which raises morbidity and mortality even when asymptomatic, using ECG rather than echocardiography.
Why it matters
Undetected LVSD increases risk of death and complications; cheap ECG-based screening could catch it earlier.
What the AI did
A deep learning AI-ECG algorithm, derived at Mayo Clinic, screened ECGs to detect LVSD; it was externally validated on 4277 Russian adults who also had echocardiography.
Results
Robust performance in 4277-subject external validation (0.6% LVSD prevalence); exact metrics not stated in source (abstract truncated)
Impact
Independent external validation supports AI-ECG as a low-cost LVSD screening tool in general populations.
Record
| Venue | International Journal of Cardiology |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Itzhak Zachi Attia, Andrew S. Tseng, Ernest Diez Benavente, José R. Medina‐Inojosa, Taane Gregory Clark, Sofia K. Malyutina, Suraj Kapa, Henrik Schirmer et al. |
| Citations | 80 (OpenAlex, at import time) |
| References | 19 |
| Model or method | AI-ECG (deep learning algorithm) |
| Organization | Mayo Clinic (derivation); Know Your Heart Study (validation population) |
| Primary source | https://doi.org/10.1016/j.ijcard.2020.12.065 |
| Open access | hybrid |
| Keywords | deep learning, ECG, ventricular dysfunction, external validation, screening, cardiology |
Importance rating
significant — thousands to millions affected
Abstract states LVSD, even asymptomatic, confers increased morbidity and mortality; AI-ECG offers low-cost screening validated externally on 4277 subjects in a general population.