Key result
Machine learning predicts acute AF termination better than long-term ablation success, reflecting distinct underlying phenotypes.
Why the study?
Identifying AF patients most likely to respond to ablation remains difficult, and acute versus long-term outcomes may reflect distinct physiology.
Can machine learning models using multimodal data predict acute and long-term response to atrial fibrillation ablation?
Population
561 consecutive AF registry patients plus an independent external cohort of n=77
Comparison
6 machine learning models predicting acute vs long-term ablation outcomes
Design
Registry-based cohort study with external validation
Follow-up
1-year
Authors
Loading...
ML models should not yet guide AF ablation decisions; hypothesis-generating for distinct acute versus long-term response phenotypes.
Cohort (n=561)
Yes
Can machine learning models using multimodal data predict acute and long-term response to atrial fibrillation ablation?
Effect estimate: AUC 0.86 (acute) vs 0.67 (long-term)
p-value: p=<0.001
Machine learning reveals that long-term and acute responses to AF ablation reflect distinct clinical and electrical physiology, respectively, with acute termination being more predictable than long-term success.
Ganesan et al. (2025) conducted a cohort in Atrial fibrillation (n=561). Machine learning models was evaluated on Prediction of acute termination and 1-year success after atrial fibrillation ablation (AUC 0.86 (acute) vs 0.67 (long-term), p=<0.001). Machine learning predicted acute termination of atrial fibrillation better than long-term success after ablation (AUC 0.86 vs 0.67; P<0.001), reflecting distinct electrical and clinical phenotypes.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: