Artificial Intelligence in Preventive Cardiology: Risk Prediction, Exposomics, and Personalized Interventions.
文献摘要
Cardiovascular disease is the leading cause of morbidity and mortality worldwide. Prevention has shifted from a narrow focus on atherosclerosis to the cardiovascular-kidney-metabolic (CKM) framework, which requires integrating data across organ systems, specialties, and care settings. Artificial intelligence (AI) can work with the high-dimensional, weakly structured data this integration produces, and its use in prevention is now endorsed by major cardiovascular societies. This review covers two applications. The first encompasses risk prediction, spanning electrocardiogram-based models that reframe routine tracings as general biomarkers of cardiovascular health, AI-enhanced cardiac imaging that extracts biological signals invisible to standard reads, opportunistic analysis of scans acquired for unrelated indications, and emerging efforts to quantify environmental and neighborhood-level exposomic risk. The second is the management of modifiable risk factors across early CKM stages: lifestyle behavior, obesity phenotyping, hypertension monitoring, dyslipidemia detection, glycemic management, and chronic kidney disease risk stratification. In both domains, retrospective discriminative performance is frequently excellent, while prospective evidence that AI-guided intervention improves patient-relevant outcomes remains limited.