Quantitative coronary calcification analysis for prediction of myocardial ischemia using non-contrast CT calcium scoring

In three linesML study on 1,375 patients: predicting myocardial ischemia from non-contrast CT calcium scans. XGBoost+SHAP model combining Agatston score, 8 calcium-omics features, and age. Results: 98.9% precision, 79.2% sensitivity, 87.7% F1. Calcium-omics features significantly improve performance vs clinical variables alone (p<0.05).

Summary generated by Claude — human-verified