AI in Ophthalmology (Book Chapter), October 2021
The eye is the sole organ in the body that allows for the direct observation and imaging of the neurological and vascular systems. In recent years, researchers have harnessed the noninvasive nature of color fundus photographs (CFPs) to examine changes in the retina as a possible marker of systemic disease risk. Building on large-scale epidemiological studies that have reported relationships of retinal features such as retinal vascular caliber with systemic diseases, the application of artificial intelligence (AI) technology, specifically in deep learning (DL), on CFPs is advancing new research that focuses on retina-systemic disease relationships. In this relatively new field, current studies fall into two basic groups: 1) cross-sectional studies that use AI-DL technology on CFP to detect or estimate systemic risk factors (e.g., age, blood pressure, smoking) or other biomarkers (e.g., coronary artery calcium); 2) longitudinal studies that use AI-DL technology on CFP to predict the incidence or risk of systemic disease (e.g., cardiovascular event or mortality). The range of systemic factors studied from CFP via AI-DL approaches is reviewed based on these cross-sectional and longitudinal studies, and areas of future research are discussed while acknowledging the limitations that AI-DL on CFP presents.