What question did this study set out to answer?

The study aims to evaluate whether atrial fibrillation ablation outcomes can be predicted using an ECG-based deep learning algorithm.

February 14, 2026Open Access

Can Atrial Fibrillation Ablation Outcomes be properly predicted with Electrocardiography and Artificial Intelligence?

Key Result

An ECG-based deep learning model predicted repeat AF ablation with limited accuracy (AUC 0.61), similar to clinical variables alone (AUC 0.59).

Key Points

The study aims to evaluate whether atrial fibrillation ablation outcomes can be predicted using an ECG-based deep learning algorithm.
Included 865 patients undergoing AF ablation with follow-up data.
Trained a deep neural network on raw data from standard 12-lead ECGs.
Employed stratified 9-fold nested cross-validation for model evaluation.
Compared model performance with a random forest model based on clinical variables.
The deep learning model achieved an AUC of 0.61, indicating limited prediction capability.
The random forest model based only on clinical variables presented a similar AUC of 0.59.
The most relevant ECG features identified included P-wave, QRS-complex, and T-wave characteristics.

Structured PICO

Does an ECG-based deep learning algorithm accurately predict the need for repeat ablation in patients undergoing atrial fibrillation ablation?

Population

865 patients undergoing atrial fibrillation (AF) ablation

Intervention

Deep neural network trained on raw data of the standard 12-lead ECG obtained within 3 months prior to the index ablation

Comparator

Random forest model using only clinical variables (age, sex, body mass index, AF pattern)

Outcome

Need for a repeat ablation procedure during follow-uphard clinical

A deep learning model based on standard pre-procedural 12-lead ECGs demonstrated limited ability (AUC 0.61) to predict the need for repeat atrial fibrillation ablation, offering no significant advantage over basic clinical variables.

Main Result

Absolute Event Rate: 0% vs 0%

Abstract

Abstract Aims The success of ablation for atrial fibrillation (AF) varies, often leading to repeat ablation. Reliable prediction of repeat ablation remains challenging. This study aimed to investigate if AF ablation outcomes can be predicted with an electrocardiogram (ECG)-based deep learning algorithm. Methods and Results We included 865 patients undergoing AF ablation, of whom 163 (18.8%) required a repeat procedure during a minimum follow-up of 572 days. A deep neural network was trained on the raw data of the standard 12-lead ECG obtained within 3 months prior to the index ablation, using stratified 9-fold nested cross-validation. Unfortunately, the model achieved a nested-cross-validation area under the receiver operating characteristic curve (AUC) of only 0.61 (95% CI: 0.57-0.64). For comparison, the same analytic approach achieved significantly higher accuracy for sex classification (AUC = 0.87, 95% CI: 0.86-0.89). A random forest model only using clinical variables (age, sex, body mass index, AF pattern) yielded a similar performance for a repeat ablation (cross-validated AUC = 0.59, 95% CI: 0.55-0.63), suggesting limited added value of ECG-based prediction. SHapley Additive exPlanations was used to pinpoint the most relevant ECG segments and highlighted contributions from P-wave, QRS-complex, and T-wave features. Conclusion The deep learning model demonstrated limited ability to predict repeat AF ablation based on the standard 10-second 12-lead ECG. Ablation outcomes may be influenced more by non-ECG parameters or require larger datasets or long-term ECG monitor data, and multi-modality inputs to be accurately predicted.