Predicting early-stage coronary artery disease using machine learning and routine clinical biomarkers improved by augmented virtual data

Koloi A, Loukas VS, Hourican C, Sakellarios AI, Quax R, Mishra PP, Lehtimäki T, Raitakari OT et al.

Machine learning models using routine clinical biomarkers, age, sex, and smoking status were trained on 3,316 LURIC patients to predict angiographic coronary artery disease. Augmentation with synthetic virtual patient data improved random forest accuracy from 0.75 to 0.79 and specificity from 0.55 to 0.70, while gradient boosting reached accuracy 0.80 and specificity 0.74. Validation in the Young Finns Study confirmed generalizability and the potential to limit invasive diagnostic procedures.

European Heart Journal. Digital Health, 9. August 2024 | PMID 39318697