Improving Prognostic Risk Assessment of Cardiovascular Events with Machine Learning: An Evaluation Using PET Myocardial Perfusion Imaging

Fares Alahdab, Radwa El Shawi, Ahmed Ibrahim Ahmed, Mahmoud Al Rifai, Mouaz Al-Mallah,
Improving Prognostic Risk Assessment of Cardiovascular Events with Machine Learning: An Evaluation Using PET Myocardial Perfusion Imaging,
Journal of Nuclear Cardiology,
2025,
102539,
ISSN 1071-3581,
https://doi.org/10.1016/j.nuclcard.2025.102539.
(https://www.sciencedirect.com/science/article/pii/S1071358125004131)
Abstract: Background
Machine learning (ML) holds potential for improving risk assessment in patients with suspected or confirmed coronary artery disease (CAD). However, certain approaches offer greater benefit than others for this task, particularly to capture non-linearity between variables as well as case-by-case explainability.
Methods
We included consecutive patients who underwent clinically indicated positron emission tomography (PET) imaging. Using automated ML (AutoML) and unseen data for performance testing, clinical and PET variables were used to train the predictive models. A logistic regression (LR) and a deep feed-forward neural network (DNN) were trained on the same data for comparison. Major adverse cardiovascular events (MACE) included death, myocardial infarction, or coronary revascularization >90 days after imaging.
Results
We included 8,357 patients (80% for development and 20% held out for testing), 46.3% females, with a mean (SD) age of 67.2 (11.7) years. The median (IQR) myocardial flow reserve (MFR) was 2.1 (1.6 to 2.6). After an average follow-up of 589 days, a total of 852 patients (10.2%) experienced MACE. The AutoML achieved an AUROC of 0.82 (95% CI 0.79 to 0.85) vs. 0.79 (0.76 to 0.82) and 0.76 (0.73 to 0.80) for the LR and the DNN models, respectively. Model explainability showed that MFR topped the list of most impactful features, followed by total perfusion defects, serum creatinine, and diastolic blood pressure.
Conclusion
An AutoML model integrating clinical and PET data discriminated MACE risk in CAD more accurately than LR or DNN and provides interpretable patient-level explanations that can inform personalized care.
Keywords: coronary artery disease; positron emission tomography; nuclear cardiology; machine learning; explainable AI; automated machine learning