Back

A Multi-Dimensional Machine Learning Approach For Cardiovascular Disease Prediction In The UK Biobank Study

Hernandez-Morante, J. J.; Murcia-Garcia, E.; Martinez-Cortes, C.; Piernas, C.; Perez-Sanchez, H.

2025-11-17 cardiovascular medicine
10.1101/2025.11.14.25340257 medRxiv
Show abstract

Cardiovascular diseases (CVD) are complex disorders involving the impaired function of blood vessels or the heart. Several risk factors contribute to CVD, including diet, body composition and physical activity. We developed an interpretable machine learning model integrating nongenetic factors from 502,134 UK Biobank participants, of whom 53,378 experienced a CVD event over eighteen years of follow-up. Age, cystatin C, and self-perceived health status emerged as the most relevant predictors of CVD. By combining these variables into a multidimensional model, we achieved excellent predictive performance, with an area under the ROC curve of 0.99 and both sensitivity and specificity above 0.97 in the validation cohort. Although cardiovascular events arise from multiple factors, our results show that data integration and machine learning can accurately predict individual risk using simple measurements. This approach could also support the prediction of other cardiovascular outcomes and aid personalized risk assessment and preventive care.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.