Key result
AI models improve obstructive CAD and adverse event prediction in stable angina, achieving AUCs >0.95.
Why the study?
Traditional risk scores rely on linear, population-oriented assumptions, while the utility of AI approaches for predicting cardiovascular events in stable angina needed evaluation.
Do artificial intelligence models improve the prediction of cardiovascular events in adults with stable angina compared to traditional risk scores?
Systematic Review (n=39,655)
Do artificial intelligence models improve the prediction of cardiovascular events in adults with stable angina compared to traditional risk scores?
Artificial intelligence models show promise in predicting cardiovascular events in patients with stable angina, but current evidence is limited by a small number of studies, high heterogeneity, and a lack of external validation.
Answers are grounded in this paper and cite their sources.
AI may enhance risk stratification in stable angina via ECG and clinical data; supports broader AI integration but requires prospective validation.
Stable angina pectoris is a prevalent condition with concerning symptoms and an increased risk of myocardial infarction (MI), stroke, heart failure, and mortality. Risk stratification is important in preventive care because the severity of the condition significantly impacts prognosis. With the availability of digital data such as electrocardiogram (ECG) and clinical variables, artificial intelligence (AI) approaches, including machine learning and deep learning, can be beneficial for risk prediction of chronic diseases, such as coronary artery disease (CAD), in patients with stable angina. At the same time, traditional approaches such as Diamond–Forrester, Framingham Risk Score, and PROCAM lag because they rely on linear and population-oriented assumptions. This systematic review, guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020, evaluates AI models that have been developed to predict cardiovascular events in adults suffering from a stable angina condition. Out of 1,250 articles, only five research studies have been included because they predict the risk of cardiovascular diseases in patients suffering from stable angina. Data modalities in these studies include longitudinal electronic health record (EHR) variables, patient-reported outcomes (Seattle Angina Questionnaire), invasive angiography indices (e.g., Gensini score), and 12-lead ECG. With area under the curve (AUC) varying from moderate (about 0.78–0.83) to excellent (> 0.95) based on outcome and data richness, AI enhances discrimination for obstructive CAD and adverse outcomes across studies. This review highlights that there are a limited number of studies on this topic and a lack of clinical validity. Also, the final set of features in model development varies highly. Hence, heterogeneity, a lack of external validation, and implementation limitations exist.
No takes yet. Share an insight, caveat, or question.
Kumar et al. (2026) conducted a systematic review in Stable angina (n=39,655). Artificial intelligence models (machine learning and deep learning) vs. Traditional risk scores was evaluated on Prediction of cardiovascular events and obstructive coronary artery disease (AUC 0.78 to >0.95). Artificial intelligence models enhanced discrimination for obstructive coronary artery disease and adverse cardiovascular outcomes in patients with stable angina, achieving AUCs ranging from 0.78 to >0.95.
Synapse has enriched 10 related papers. Consider them for comparative context: