Machine learning based prediction of 1-year arrhythmia recurrence after ventricular tachycardia ablation in patients with structural heart disease

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Project no. NVKP_16-1–2016-0017 (’National Heart Program’) - National Research, Development and Innovation Fund of Hungary Thematic Excellence Programme (2020-4.1.1.-TKP2020) of the Ministry for Innovation and Technology in Hungary Background Ventricular tachycardia (VT) is a life-threatening condition. Catheter ablation treatment is often successful, however post procedural VT recurrence still remains an issue. Therefore, there is a demand for an accurate risk stratification system to assess the probability of arrhythmia recurrence after the procedure. Objective We aimed to implement a machine learning (ML) pipeline to predict 1-year VT recurrence in patients with structural heart disease (SHD) undergoing VT ablation. Methods For 297 patients who underwent VT ablation, we collected medical history, laboratory, echocardiography, and procedural data. Following manual and ML-based feature selection, we trained several supervised machine learning models to predict 1-month and 1-year recurrence. The area under the receiver operating characteristic curve (AUC) was calculated to quantify the models’ performance. Results 1-year VT recurrence was observed in 107 (36%) cases. The best predictions of VT recurrence were demonstrated by random forest models utilizing 7 input features [1-month AUC: 0.74; 1-year AUC: 0.76]. These models significantly outperformed a previously published risk score, the I-VT score [AUC: 0.63, p=0.024] on our data. The most important predictors of recurrence were the number of VT morphologies during the procedure, electrical storm, left ventricular ejection fraction, left ventricular end systolic diameter, and the severity of mitral regurgitation. Conclusion Our machine learning model can efficiently predict VT recurrence in SHD patients undergoing VT ablation. Thus, it could facilitate the prompt identification of high-risk patients and the personalization of treatment and follow-up strategies, ultimately leading to improved outcomes.