Automatic Detection of Left Ventricular Dilatation and Hypertrophy from Electrocardiograms Using Deep Learning
Problem
Detecting left ventricular dilatation (LVD) and left ventricular hypertrophy (LVH) from 12-lead ECGs to improve heart failure screening.
Why it matters
LVD and LVH are risk factors for heart failure, and earlier detection enables earlier intervention.
What the AI did
Developed a deep learning model using paired ECG and echocardiography data from 18,954 patients, and compared it against logistic regression, random forest, and conventional ECG criteria.
Results
AUROC 0.810 (95% CI 0.801-0.819) for LVD detection, significantly higher than logistic regression and random forest (P < 0.001)
Impact
Deep learning outperformed conventional ML methods for ECG-based LVD detection in a large patient cohort.
Record
| Venue | International Heart Journal |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Takahiro Kokubo, Satoshi Kodera, Shinnosuke Sawano, Susumu Katsushika, Mitsuhiko Nakamoto, Hirotoshi Takeuchi, Nisei Kimura, Hiroki Shinohara et al. |
| Citations | 27 (OpenAlex, at import time) |
| References | 30 |
| Model or method | deep learning model (architecture not stated in source) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1536/ihj.22-132 |
| Open access | diamond |
| Keywords | deep learning, electrocardiogram, left ventricular hypertrophy, dilatation, heart failure screening |
Importance rating
significant — thousands to millions affected
Abstract states LVD and LVH are heart failure risk factors and their detection improves screening; ECG-based deep learning could scale low-cost screening, though no population scale is stated.