Coronary Model Shows Limits in Cath-Referred Patients
An interpretable machine-learning model classified angiographic coronary artery disease in patients referred for coronary angiography, but high disease prevalence and unclear inflammatory signals limited clinical interpretation.
By
Andrea Surnit
June 22, 2026
Clinical Scorecard: Coronary Model Shows Limits in Cath-Referred Patients
At a Glance
Category Detail
Condition Coronary Artery Disease (CAD) Key Mechanisms Machine-learning model incorporating cardiovascular risk factors and inflammatory biomarkers. Target Population Patients referred for coronary angiography without acute myocardial infarction. Care Setting Retrospective single-center study.
Key Highlights
Study included 3,482 patients with suspected CAD. Final model achieved 81% accuracy and 82% sensitivity. Hypertension, hyperlipidemia, sex, diabetes, and triglyceride levels were key predictors. Model not validated in independent external cohort. Study did not assess medication use directly. Guideline-Based Recommendations
Diagnosis
CAD defined as at least 50% stenosis in at least one major coronary artery. Management
Model should be interpreted as a classifier of angiographic stenosis. Monitoring & Follow-up
Further validation needed to determine clinical utility. Risks
Model performance not established in lower-prevalence outpatient or primary care populations. Patient & Prescribing Data
Patients undergoing coronary angiography for suspected CAD.
Prior treatment, particularly statin therapy, may affect inflammatory biomarker levels.
Clinical Best Practices
Consider class imbalance when interpreting model performance metrics. Use multiple modalities for comprehensive assessment of CAD. Related Resources & Content
