Arrhythmias & electrophysiologyDiagnostic or prognostic study
Reclassifying Single Lead Sinus Rhythm ECGs for Improved Atrial Fibrillation (AF) Recurrence Detection With Deep Learning
- development and validation of a novel deep learning model
- single-lead ECG strips from handheld devices
Authors' conclusion
Our dynamic deep learning model reclassifies 30-second single-lead sinus rhythm ECGs based on AF events within ±3 days, transforming a brief ECG snapshot into a summarized indicator of short-term AF risk.
Study context
- Population
- Patients after catheter ablation for atrial fibrillation (AF), with sinus rhythm strips from handheld single-lead ECG devicesno documented recurrence (Benign Sinus strips) 308 · documented recurrence (Severe Sinus strips) 250
- Participants
- 558 participants
- Comparison
- Both clinical feature-based models and a deep learning model were used to classify these ECGs according to their proximity to AF recurrence events
- Benign Sinus stripsn = 308
- Severe Sinus stripsn = 250
Further findings
| Outcome | strip-level pretrained deep learning model | participant-level clinical feature-based Random Forest benchmark |
|---|---|---|
| Classification performance (Severe vs Benign Sinus strips) | ||
| area under the receiver operating characteristic curve | 0.86 | 0.71 |
| sensitivity | 0.79 | 0.58 |
| specificity | 0.76 | 0.69 |
Feature-based analyses
- Associations involving heart-rate variability, mean R-R interval, female sex, left atrial volume, and mean P-wave duration
Original publication
Reclassifying Single Lead Sinus Rhythm ECGs for Improved AF Recurrence Detection With Deep Learning. Circ Arrhythm Electrophysiol. Published 7 October 2026. doi:10.1161/circep.125.014659
Summarised from the published abstract; every number is checked against it