Machine learning model including left ventricular strain analysis for sudden cardiac death prediction in hypertrophic cardiomyopathy

Abstract Funding Acknowledgements Type of funding sources: None. Background The excess mortality in hypertrophic cardiomyopathy (HCM) patients is mainly attributed to the occurrence sudden cardiac death (SCD). prediction ventricular arrhythmias remains challenging and could be improved. Purpose This study evaluated added predictive value a machine learning-based model combining clinical conventional imaging parameters with information from left strain analysis predict SCD HCM. Methods A total 434 HCM (65% men, mean age 56 years) were retrospectively included two referral centers different countries followed longitudinally (mean duration 6 years). Strain automatically extracted ventricle longitudinal segmental curves each patient Ridge Regression alongside data. composite endpoint sustained tachycardia, appropriate implantable cardioverter defibrillator therapy, aborted arrest, or death. Results 34 (7.8%) met an incidence 0.9%/years. Among 18 most discriminating parameters, 7 derived (figure 1). After n=200 rounds cross-validation, final showed superior performance compared models area under curve (AUC) 0.83 ± 0.8 AUC 0.56 0.61 for 2014 ESC risk score 2020 AHA/ACC model, respectively. Conclusion learning including strain-derived was existing models. improve stratification patients.