Improving Risk Stratification in Sudden Cardiac Death Using Interpretable Machine Learning: A Clinical Perspective
Hana Ivandic1, Branimir Pervan1, Vedran Velagic2,3
1Faculty of Electrical Engineering and Computing, University of Zagreb, Unska 3, 10000 Zagreb, Croatia.
Insights
Machine learning improves sudden cardiac death (SCD) risk prediction for implantable cardioverter-defibrillator (ICD) selection. This approach refines patient stratification beyond traditional metrics, enhancing protection for those at risk.
Area of Science:
- Cardiology
- Medical Informatics
- Machine Learning
Background:
- Sudden cardiac death (SCD) is a significant cause of cardiovascular mortality.
- Current implantable cardioverter-defibrillator (ICD) selection lacks precision, leading to suboptimal patient outcomes.
- Interpretable machine learning (ML) offers a potential solution for improved risk stratification.
Purpose of the Study:
- To develop and validate an interpretable ML model for predicting appropriate ICD activation.
- To refine patient selection for ICD implantation using diverse clinical data.
- To enhance the transparency of clinical reasoning in SCD risk assessment.
Main Methods:
- Retrospective analysis of 607 patients undergoing ICD or CRT-D implantation.
- Collection of comprehensive baseline data: demographic, clinical, echocardiographic, laboratory, and device-related variables.
- Development of a logistic regression (LR) model to predict appropriate ICD activation, using patient follow-up data.
Main Results:
- The LR model achieved strong predictive performance (AUC-ROC 0.74, sensitivity 86.50%).
- Key predictors identified include ventricular tachycardia (VT) burden, sustained VT, longer follow-up, and secondary prevention.
- Combination therapy with furosemide and spironolactone was associated with lower predicted SCD risk.
Conclusions:
- ML applied to routine clinical data can significantly enhance SCD risk stratification.
- This approach complements existing guideline criteria by confirming known predictors and identifying novel associations.
- Interpretable ML models can support clinical decision-making for ICD implantation, improving patient selection and outcomes.
Abstract:
Background: Sudden cardiac death (SCD) remains a major cause of cardiovascular mortality. Implantable cardioverter-defibrillators (ICDs) reduce arrhythmic mortality, but current selection based largely on left ventricular ejection fraction (LVEF) lacks precision. Many patients undergo device implantation without ever receiving therapy, while others at risk remain unprotected. Interpretable machine learning (ML) can integrate diverse clinical variables and refine patient selection while maintaining transparency in clinical reasoning. Methods: We retrospectively analyzed 607 patients who underwent ICD or CRT-D implantation at a Croatian tertiary care center. Baseline demographic, clinical, echocardiographic, laboratory, and device-related variables were collected. Patients were followed through routine device interrogations, with appropriate ICD activation serving as a surrogate for SCD prevention. A logistic regression (LR) model was trained to predict appropriate device activation. Results: LR model demonstrated strong predictive ability (AUC-ROC 0.74, sensitivity 86.50%). Significant predictors included ventricular tachycardia (VT) burden, sustained VT, longer follow-up, and secondary prevention. The combination of furosemide and spironolactone therapy was linked to lower predicted SCD risk. Conclusions: ML applied to routinely collected data can support risk stratification in SCD and complement existing guideline criteria by reinforcing known predictors and uncovering novel associations.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:30Heart Failure IV: Classification and Diagnostic Evaluation
01:25Cardiomyopathy I: Introduction and Classification