Advancements in Personalized Machine Learning Models Enhance Coronary Artery Disease Risk Prediction

Scientists from the Scripps Research Translational Institute have developed a groundbreaking machine learning model that significantly improves the accuracy of predicting the risk for coronary artery disease (CAD). Unlike traditional methods that primarily consider age, this new approach incorporates a multitude of factors such as genetics, lifestyle, medical history, and other biomarkers. The model was trained using data from the UK Biobank, analyzing around 2,000 potential risk factors, which were refined down to 53 key indicators, including physical measurements, blood biomarkers, family medical history, sleep patterns, and genetic variants.

Published in Nature Medicine on April 16, 2025, the study demonstrates that the new model outperforms the conventional clinical risk assessment tools, doubling the ability to predict CAD events. Over a 10-year follow-up, 62.9% of individuals categorized as high risk by the model developed CAD, whereas only 0.3% of those in the low-risk group did.

This personalized approach enables clinicians to identify at-risk individuals more precisely, including those who might be underestimated by traditional tools—such as younger persons and women. Notably, genetic predisposition emerged as the strongest predictor among all factors, highlighting the importance of genetics in cardiovascular risk. The model also showed robust performance across diverse populations, including European, African, and Hispanic groups, when validated with the NIH All of Us dataset.

The researchers are now planning long-term clinical trials to assess whether informing patients of their individual risk can facilitate early interventions and prevent the onset of CAD. The ultimate goal is to motivate patients to adopt healthier lifestyles and receive tailored treatments based on their unique genetic and clinical profiles.

This innovation represents a significant step toward personalized medicine in cardiology, promising to improve outcomes through early detection and targeted prevention strategies.

Source: https://medicalxpress.com/news/2025-04-personalized-coronary-artery-disease.html