Arrhythmic sudden death survival prediction using deep learning analysis of scarring in the heart

Abstract

Sudden cardiac death from arrhythmia is a major cause of mortality worldwide. In this study, we developed a novel deep learning (DL) approach that blends neural networks and survival analysis to predict patient-specific survival curves from contrast-enhanced cardiac magnetic resonance images and clinical covariates for patients with ischemic heart disease. The DL-predicted survival curves offer accurate predictions at times up to 10 years and allow for estimation of uncertainty in predictions. The performance of this learning architecture was evaluated on multi-center internal validation data and tested on an independent test set, achieving concordance indexes of 0.83 and 0.74 and 10-year integrated Brier scores of 0.12 and 0.14. We demonstrate that our DL approach, with only raw cardiac images as input, outperforms standard survival models constructed using clinical covariates. This technology has the potential to transform clinical decision-making by offering accurate and generalizable predictions of patient-specific survival probabilities of arrhythmic death over time.

Interpretable machine learning of action potential duration restitution kinetics in single-cell models of atrial cardiomyocytes

“…In addition, we developed the ML models for predicting Conventional parameter sensitivity analysis could determine the linear relationships between ion channel profiles and 'static' arrhythmia markers, such as APD, RMP, Vmax, and alternans threshold (26). Recently, several ML approaches have been utilized to evaluate the arrhythmogenicity from electrical/structural remodeling status (27)(28)(29)(30). We previously demonstrated that statistical physics-derived methods, such as a maximum entropy model, can intuitively describe functional coherence properties in the spatiotemporal wave dynamics of cardiac fibrillation (31); however, those methods typically require huge computational costs for systemic population-scale data analyses.…”

Interpretable machine learning of action potential duration restitution kinetics in single-cell models of atrial cardiomyocytes

“…In addition, we developed the ML models for predicting Conventional parameter sensitivity analysis could determine the linear relationships between ion channel profiles and 'static' arrhythmia markers, such as APD, RMP, Vmax, and alternans threshold (26). Recently, several ML approaches have been utilized to evaluate the arrhythmogenicity from electrical/structural remodeling status (27)(28)(29)(30). We previously demonstrated that statistical physics-derived methods, such as a maximum entropy model, can intuitively describe functional coherence properties in the spatiotemporal wave dynamics of cardiac fibrillation (31); however, those methods typically require huge computational costs for systemic population-scale data analyses.…”

The future of resuscitation

“…Recent studies investigating risk factors for SCD suggest that healthy lifestyle modifications are associated with low odds of SCD occurrence. [22][23][24] Machine learning-based analysis of large population-based clinical data and omics technology using minimal blood samples have been attracted attention in terms of SCD risk factor analysis. 25,26 These new technologies can predict the risk of SCD, not only in the individual but also in the population, which can be applied to establishing regional/national healthcare policies to reduce SCD.…”