Using Artificial Intelligence to Detect Cardiovascular Disease through Retinal Imaging: A Review

Abstract

Cardiovascular diseases cause a large number of deaths and disabilities worldwide. To reduce the related losses and burdens, it is especially important to promote early non-invasive screening methods. In recent years, with the development of AI, retinal images can become a more important tool in cardiovascular risk assessment. This article reviews the application of artificial intelligence in the screening of cardiovascular diseases using retinal images and finds that AI can accurately predict cardiovascular disease risks from retinal images and can achieve better performance by integrating multimodal data and using large-scale databases. However, the heterogeneity of data, cross-population applicability and ethical issues still limit the discoveries in this field. In the future, larger-scale model training, multi-task learning and different-level data fusion can be used to improve the predictive ability and accuracy of the model. At the same time, this method has potential in remote medical care, personalized medical care, and public health fields.