Arrhythmias & electrophysiology
Enhanced detection of atrial fibrillation in single-lead electrocardiograms using a Cloud-based artificial intelligence platform
Authors' conclusion
The result of this external validation indicates that the AI platform can match cardiologist-level accuracy in AF detection from 1-lead ECGs. Such tools are promising for AF screening and have the potential to improve accuracy in noncardiology expert health care professional interpretation and trigger further tests for effective patient management.
Study context
- Population
- Patients with 30-second electrocardiogram (ECG) records from a single-lead handheld ECG devicepatients analyzed
- Participants
- 8,528 patients
- Index test or model
- Willem Artificial Intelligence (AI) platform (and a rules-based algorithm) against ground truth established through a benchmark algorithm and expert manual labeling
- Target condition
- atrial fibrillation (AF)
Methodological detailsCloud-based deep-learning platform, not trained on these ECGs
- Cloud-based deep-learning platform
- Not trained on these ECGs
Performance
As reported in the abstract
| Test or model | Sensitivity | Specificity | Positive predictive value | Negative predictive value | Accuracy |
|---|---|---|---|---|---|
| AI platform (initial labels) | 83.3% | 97.3% | 75.2% | 98.4% | 96.1% |
| AI platform (expert review) | 84.2% | 97.6% | 78.0% | 98.4% | 96.4% |
Sensitivity
AI platform (initial labels)83.3%
AI platform (expert review)84.2%
Specificity
AI platform (initial labels)97.3%
AI platform (expert review)97.6%
Positive predictive value
AI platform (initial labels)75.2%
AI platform (expert review)78.0%
Negative predictive value
AI platform (initial labels)98.4%
AI platform (expert review)98.4%
Accuracy
AI platform (initial labels)96.1%
AI platform (expert review)96.4%
Further findings
AI platform vs rules-based algorithm
- Performance of the AI platform largely exceeded the performance of the rules-based algorithm for all metrics
Other arrhythmias
- The AI also detected other arrhythmias, such as premature ventricular complexes, premature atrial complexes along with 1-degree atrioventricular blocks
Original publication
Enhanced detection of atrial fibrillation in single-lead electrocardiograms using a Cloud-based artificial intelligence platform. Heart Rhythm. Published 10 January 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →