Deep learning analysis of resting electrocardiograms for the detection of myocardial dysfunction, hypertrophy, and ischaemia: a systematic review
Problem
Detecting structural cardiac pathologies (LV systolic dysfunction, hypertrophy, ischaemia) from resting electrocardiograms.
Why it matters
ECG-based detection could enable earlier, lower-cost diagnosis of serious heart disease.
What the AI did
A systematic review assessed 12 studies of end-to-end deep learning analysis of resting ECG signals for detecting structural cardiac pathologies.
Results
LV systolic dysfunction: AUC 0.89-0.93, accuracy 98% (3 studies); LV hypertrophy: AUC 0.87, accuracy 87% (1 study); 6 further studies on acute myocardial pathology (details truncated).
Impact
Aggregated peer-reviewed evidence shows DL detects these pathologies with high AUC, but this is secondary review evidence.
Record
| Venue | European Heart Journal - Digital Health |
|---|---|
| Published | 2021 |
| Work type | review |
| Authors | Ghalib Al Hinai, Samer Jammoul, Zara Vajihi, Jonathan Afilalo |
| Citations | 52 (OpenAlex, at import time) |
| References | 53 |
| Model or method | end-to-end deep learning (specific architectures not stated in source) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1093/ehjdh/ztab048 |
| Open access | gold |
| Keywords | deep learning, ECG, cardiac pathology, systematic review, AUC |
Importance rating
significant — thousands to millions affected
ECG-based detection of structural cardiac pathologies could enable cheaper, earlier screening; abstract claims clinical applicability for more effective screening but states no scale of affected population.