Image based artificial intelligence-enhanced ECG prediction of incident atrial fibrillation
File(s)
Published version
Author(s)
Type
Journal Article
Abstract
Background
Early prediction of atrial fibrillation (AF) is crucial for reducing adverse outcomes. While artificial intelligence-enhanced ECG (AI-ECG) analysis shows promise in predicting AF, most approaches require digital ECG signals, limiting their application in settings where ECGs are stored as images.
Objective
We aimed to develop and validate an image-based AI-ECG approach for predicting incident AF across multiple datasets.
Methods
We used 1,163,401 ECGs from 189,539 patients in the Beth Israel Deaconess Medical Center (BIDMC) dataset and 70,655 ECGs from 65,610 participants in the UK Biobank. The AI-ECG model was trained on ECG images processed to 310x868 pixels.
Results
The model achieved C-statistics of 0.754 (95% CI: 0.747-0.761) in the BIDMC dataset and 0.723 (95% CI: 0.704-0.741) in the UK Biobank for predicting incident AF. Performance was maintained across key subgroups including outpatients, females, and non-White individuals. Compared to the CHARGE-AF risk score, the AI-ECG model showed superior performance (c-statistic 0.696 vs 0.667, p<0.05) and provided significant additive value when combined (c-statistic 0.711, p<0.0001). The model also performed well on smartphone-photographed ECGs (c-statistic 0.736). Saliency mapping indicated the model primarily focused on P-wave morphology and PR interval regions.
Conclusion
This image-based approach enables AI-ECG prediction of AF in settings without digital ECG infrastructure and provides additive value to known clinical risk scores
Early prediction of atrial fibrillation (AF) is crucial for reducing adverse outcomes. While artificial intelligence-enhanced ECG (AI-ECG) analysis shows promise in predicting AF, most approaches require digital ECG signals, limiting their application in settings where ECGs are stored as images.
Objective
We aimed to develop and validate an image-based AI-ECG approach for predicting incident AF across multiple datasets.
Methods
We used 1,163,401 ECGs from 189,539 patients in the Beth Israel Deaconess Medical Center (BIDMC) dataset and 70,655 ECGs from 65,610 participants in the UK Biobank. The AI-ECG model was trained on ECG images processed to 310x868 pixels.
Results
The model achieved C-statistics of 0.754 (95% CI: 0.747-0.761) in the BIDMC dataset and 0.723 (95% CI: 0.704-0.741) in the UK Biobank for predicting incident AF. Performance was maintained across key subgroups including outpatients, females, and non-White individuals. Compared to the CHARGE-AF risk score, the AI-ECG model showed superior performance (c-statistic 0.696 vs 0.667, p<0.05) and provided significant additive value when combined (c-statistic 0.711, p<0.0001). The model also performed well on smartphone-photographed ECGs (c-statistic 0.736). Saliency mapping indicated the model primarily focused on P-wave morphology and PR interval regions.
Conclusion
This image-based approach enables AI-ECG prediction of AF in settings without digital ECG infrastructure and provides additive value to known clinical risk scores
Date Issued
2026-03-01
Date Acceptance
2025-10-09
Citation
Heart Rhythm, 2026, 23 (3), pp.515-524
ISSN
1547-5271
Publisher
Elsevier BV
Start Page
515
End Page
524
Journal / Book Title
Heart Rhythm
Volume
23
Issue
3
Copyright Statement
© 2025 Published by Elsevier Inc. on behalf of Heart Rhythm Society.
License URL
Identifier
10.1016/j.hrthm.2025.10.024
Subjects
Artificial intelligence
Electrocardiography
Atrial fibrillation
Prediction
Image analysis
Deep learning
Cardiovascular disease
Machine learning
Publication Status
Published
Date Publish Online
2025-10-14