J Cardiovasc Imaging. 2026 Aug 20;34(1):24. doi: 10.1186/s44348-026-00064-x.
ABSTRACT
Artificial intelligence (AI) is emerging as a transformative tool in cardiovascular imaging, particularly in coronary angiography. With the growing interest in precision medicine, AI offers the potential to enhance clinical decision-making by improving diagnostic accuracy, predicting outcomes, and guiding interventions. This scoping review aims to examine the current landscape of AI applications in predicting clinical outcomes related to angiographic procedures, including diagnosis, risk stratification, and postprocedural monitoring. A total of 59 relevant studies published between 2015 and 2025 were included in this review. The selection was based on predefined inclusion criteria focused on AI-based models applied in coronary artery disease diagnosis, outcome prediction and restenosis detection. The studies were reviewed for the AI techniques used, clinical endpoints targeted, model performance, validation strategies, and limitations. Four primary themes were identified: (1) prediction of adverse cardiovascular events such as myocardial infarction and mortality; (2) detection of hemodynamically significant stenosis; (3) prediction of in-stent restenosis; and (4) identification of coronary arteries. Machine learning algorithms, especially random forests, support vector machines, and convolutional neural networks, have been commonly used. Many models have demonstrated superior performance compared to traditional statistical methods. However, limitations such as small sample sizes, lack of external validation, retrospective designs, and concerns about data privacy were frequently observed in these studies. AI demonstrates promising capabilities in outcome prediction within coronary angiography. Despite current limitations, its integration into clinical practice is feasible with prospective, multicentric studies and standardized validation frameworks. Ethical considerations and regulatory oversight will be crucial in ensuring safe and effective implementation.
PMID:42625241 | DOI:10.1186/s44348-026-00064-x