This Arkangel AI whitepaper is a narrative review of artificial intelligence in cardiology across electrocardiography, cardiovascular imaging, risk stratification and remote monitoring, with a focus on what it means for Latin America. The authors searched PubMed/MEDLINE, Crossref and official WHO and PAHO sources, mostly from 2021 to 2026, and built a corpus of 25 sources: 20 articles with a DOI and 5 official or policy documents. Each application was placed on a map of five evidence-maturity stages, from retrospective development to routine implementation, and the version includes external review by a cardiologist.
Development and technical validation dominate. AI-ECG models reported C-indexes of 0.638 to 0.787 for mortality and cardiovascular outcomes, and an echocardiography model reached an AUC of 0.96 in external validation, but clinical utility, outcome impact and routine use were rarely shown. The clearest prospective example is a pragmatic randomized trial in Nigeria, where AI-assisted digital stethoscope screening detected left ventricular systolic dysfunction in 24 of 587 participants versus 12 of 608 with usual care (AUC 0.976 for LVEF below 50%). Direct Latin American evidence was limited, so regional recommendations are labeled as implementation inferences. It is a whitepaper, not peer reviewed, and the authors are mostly affiliated with Arkangel AI.
