Machine Learning in Cardiovascular Risk Diagnosis

Accurate quantification of ASCVD risk is essential for early and effective cardiovascular risk management. Conventional models rely solely on traditional risk factors (TRFs). These often fail to incorporate newer, non-traditional risk variables, leading to potential underestimation or overestimation of risk, especially across diverse ethnic populations. This book introduces a novel machine learning (ML)-based framework that integrates TRFs with non-traditional ultrasound-based markers like carotid intima-media thickness (cIMT) and carotid plaque (cP) features, to enhance the predictive accuracy. It covers the development of a diagnostic architecture that uses hybrid intelligent models optimized using different Meta-heuristic algorithms. The chosen framework has the advantage due to the ability to include additional newer risk variables without methodological reconstruction and thereby contribute to the development of reliable, efficient, and customizable solutions for ASCVD risk prediction in public healthcare settings.