Deep learning for electrocardiogram interpretation: Bench to bedside
Problem
Automating ECG analysis to detect and predict rhythm, conduction, ischemic and structural heart diseases, where clinical integration of algorithms remains limited.
Why it matters
The ECG is a core diagnostic tool, and automation could improve cardiovascular disease recognition at scale.
What the AI did
Deep learning algorithms analyze 12-lead ECGs to detect and predict cardiac abnormalities, and the review assesses randomized controlled trials of their clinical effect.
Results
DL algorithms show performance comparable to physicians, but deployment into routine clinical practice remains limited.
Impact
The review documents DL applications and RCT evidence, but concludes bedside integration is still limited.
Record
| Venue | European Journal of Clinical Investigation |
|---|---|
| Published | 2025 |
| Work type | article |
| Authors | Bas B.S. Schots, Camila S. Pizarro, Bauke K. O. Arends, Marish I.F.J. Oerlemans, Dino Ahmetagić, Pim van der Harst, René van Es |
| Citations | 26 (OpenAlex, at import time) |
| References | 92 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1111/eci.70002 |
| Open access | hybrid |
| Keywords | deep learning, ECG, cardiology, clinical deployment, diagnosis |
Importance rating
significant — thousands to millions affected
ECG is described as a valuable tool for diagnosing cardiovascular conditions; automating its analysis could improve recognition at scale, though clinical integration remains limited.