Machine learning approach to predict 1-year mortality after heart transplantation: a single-centre study.

Area of Science:

• Cardiology
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Heart transplantation is a vital treatment for end-stage heart failure.
• Predicting patient survival post-transplant remains a significant clinical challenge.
• Developing accurate mortality prediction models is crucial for improving patient outcomes.

Purpose of the Study:

• To evaluate the efficacy of machine learning (ML) models in predicting 1-year mortality in heart transplant recipients.
• To compare the performance of Support Vector Machine (SVM) and Logistic Regression (LR) models for this prediction task.
• To identify critical factors influencing post-transplant mortality in Saudi Arabia.

Main Methods:

• A retrospective observational study analyzed data from 419 heart transplant cases (2007-2022).
• Support Vector Machine (SVM) and Logistic Regression (LR) models were developed and validated.
• Model performance was assessed using accuracy, precision, recall, F1 score, and AUC.

Main Results:

• Logistic Regression (LR) achieved a testing accuracy of 96.43%, with weight and Body Mass Index (BMI) as significant predictors.
• Support Vector Machine (SVM) demonstrated a testing accuracy of 95.24%.
• Ischemia time, mechanical support devices (LVAD, ECMO), and BMI were identified as key mortality predictors.

Conclusions:

• Machine learning models, specifically LR and SVM, are highly effective for predicting 1-year mortality post-heart transplantation.
• These models can identify significant mortality predictors, enabling tailored healthcare strategies.
• This research underscores the value of advanced computational methods in optimizing cardiac transplant care within the Saudi population.