Applications of artificial intelligence in cardiovascular risk detection and prediction among adults with obesity: a scoping review

Background Obesity is a chronic and multifactorial disease that substantially increases cardiovascular risk, the leading cause of mortality worldwide. Conventional cardiovascular risk prediction tools are largely derived from general populations and often fail to capture the metabolic heterogeneity and complex pathophysiology associated with obesity. Artificial intelligence (AI) has been proposed as an approach to improve cardiovascular risk detection and prediction through the integration of large and heterogeneous clinical datasets. Objective To map and characterise the available evidence on the application of artificial intelligence for cardiovascular risk detection, prediction, and stratification in adults with obesity. Methods scoping review was conducted following the Joanna Briggs Institute methodology and reported according to the PRISMA ScR guidelines. PubMed, Scopus, Web of Science, LILACS, and IEEE Xplore were searched for studies published between January 2015 and January 2026. Eligible studies included observational designs and investigations describing the development or validation of AI based models applied to cardiovascular risk assessment in adults with obesity. Data were synthesised using narrative and tabulated approaches. Results Thirty studies were included, most of which were retrospective and characterised by heterogeneous populations. Tree based ensemble methods, particularly Random Forest and gradient boosting algorithms, were most frequently used, followed by support vector machines and artificial neural networks. Outcomes mainly focused on cardiovascular risk stratification and disease detection, whereas prediction of incident cardiovascular events and mortality was less common. External validation was infrequently reported, and model performance was generally moderate when longitudinal outcomes were assessed. Conclusions Artificial intelligence shows potential as a complementary tool for cardiovascular risk assessment in adults with obesity. However, methodological heterogeneity, limited external validation, and inconsistent outcome definitions currently limit clinical implementation. Future research should prioritise prospective designs, robust external validation, and standardised outcomes to define the clinical value of AI based cardiovascular risk models in obesity.