Arrhythmias & electrophysiology
Prediction of incident atrial fibrillation: A comprehensive evaluation of conventional and artificial intelligence-enhanced approaches
Authors' conclusion
We present the first comprehensive evaluation of methodologies for predicting incident AF. Risk prediction with a model including AIRE-AF and CHARGE-AF resulted in similar performance to more complex models.
Study context
- Population
- Patients from the Beth Israel Deaconess Medical Center (development) and UK Biobank volunteers (external validation)Training dataset of 1,163,401 electrocardiograms from 189,539 patients, external validation n = 38,892 · Beth Israel Deaconess Medical Center patients (training) 189,539 · UK Biobank (external validation) 38,892
- Predictors
- AIRE-AF (artificial intelligence-enhanced electrocardiogram risk estimator-AF), compared with CHARGE-AF, left atrial size, N-terminal pro-B-type natriuretic peptide and a polygenic risk score
- Predicted outcome
- incident AF
Methodological detailsconvolutional neural network with a discrete-time survival loss function, externally validated in the UK Biobank
- Convolutional neural network with a discrete-time survival loss function
- Externally validated in the UK Biobank
Model performance
As reported in the abstract
C-index
AIRE-AF
0.750(0.743-0.758)
CHARGE-AF (best single predictor)UK Biobank
0.750(0.734-0.769)
AIRE-AF and CHARGE-AF (best combination of 2 predictors)UK Biobank
0.768(0.743-0.792)
AIRE-AF, CHARGE-AF and polygenic risk scoreUK Biobank
0.791(0.766-0.816)
Further findings
Incident atrial fibrillation (AF): AIRE-AF vs other predictors (Beth Israel Deaconess Medical Center cohort)
- Addition of CHARGE-AF and left atrial size: C-index improvement
- 0.017
- AIRE-AF superior to CHARGE-AF, left atrial size, and N-terminal pro-B-type natriuretic peptide
- No additive value of N-terminal pro-B-type natriuretic peptide in combination with AIRE-AF
Incident AF (UK Biobank)
- Addition of a polygenic risk score provided a further significant improvement in performance
Original publication
Prediction of incident atrial fibrillation: A comprehensive evaluation of conventional and artificial intelligence-enhanced approaches. Heart Rhythm. Published 22 August 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →