Key result
Multimodal ensemble AI achieves 0.84 AUC for AF subtype classification and predicts 3-year recurrence.
Why the study?
A one-time binary classification may not appropriately capture atrial remodeling characteristics as AF progresses, limiting AF subtype classification and post-ablation recurrence risk stratification.
Does a multimodal ensemble AI model improve atrial fibrillation subtype classification and prediction of post-ablation recurrence compared to standard clinical classification in patients with atrial fibrillation?
Cohort (n=4,067)
Does a multimodal ensemble AI model improve atrial fibrillation subtype classification and prediction of post-ablation recurrence compared to standard clinical classification in patients with atrial fibrillation?
A multimodal AI model combining structural, electrical, and clinical data improves atrial fibrillation subtype classification and better predicts post-ablation recurrence risk than standard clinical classification.
AI-predicted PeAF probability may refine recurrence risk stratification after ablation; leaves open prospective validation before clinical use.
Background Clinical classification of atrial fibrillation (AF) subtypes plays an important role in treatment decisions and catheter ablation strategies, but a one-time binary classification may not appropriately capture the characteristics of atrial remodeling as AF progresses. We aimed to determine whether a multimodal ensemble artificial intelligence (AI) model combining structural, electrical, and functional features could predict AF subtype and provide prognostic utility for AF recurrence after catheter ablation. Methods A multimodal ensemble AI model was developed to classify AF subtypes using the Yonsei AF cohort (n = 4,067). The model incorporated left atrial wall thickness (LAWT), full-length and median 12-lead electrocardiograms (ECGs), ECG-derived features, transthoracic echocardiographic parameters, and clinical variables. The dataset was randomly divided into training (75%) and internal holdout test sets (25%, n = 1,017); all performance metrics were computed on the internal holdout test set. We also assessed the association between AI-clinical subtype discordance and markers of atrial remodeling. Results The ensemble model achieved the highest performance (AUC, 0.84; F1 score, 0.68) compared with single-modality models (AUC range, 0.73–0.78). SHapley Additive exPlanations (SHAP) analysis indicated that median ECG and LAWT contributed most significantly. AI-predicted persistent AF (PeAF) probability showed slightly higher discrimination for 3-year recurrence than clinical PeAF classification (AUC, 0.61 vs. 0.59; C-index, 0.59 vs. 0.57, respectively). In survival analysis, AI-predicted PeAF probability (per 1 standard deviation increase) remained independently associated with recurrence even after adjustment for clinical factors and AF subtype (P = 0.008). Consistently, patients with clinical paroxysmal AF predicted using AI as PeAF showed a significantly higher AF recurrence risk (log-rank, P = 0.005). Conclusion Our findings suggest that AI-predicted AF subtypes may better reflect atrial remodeling and post-ablation recurrence risk. Continuous AF subtype monitoring using AI may improve risk stratification and support personalized post-ablation monitoring strategies.
No takes yet. Share an insight, caveat, or question.
Kwon et al. (2026) conducted a cohort in Atrial fibrillation (n=4,067). Multimodal ensemble artificial intelligence (AI) model vs. Single-modality models and clinical classification was evaluated on AF subtype classification and 3-year recurrence (p=0.008). A multimodal ensemble AI model achieved an AUC of 0.84 for AF subtype classification, and its predicted persistent AF probability was independently associated with 3-year recurrence (P=0.008).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: