Retinal Microvascular Signatures as Early Predictors of Cardiovascular Risk: Integrating Pathophysiology, Molecular Insights, and Artificial Intelligence.

Cardiovascular diseases (CVDs) continue to be a significant public health burden and public health emergency in the world, underscoring the importance of early and easily accessible cardiovascular risk prediction strategies. While systemic clinical and biochemical markers have been the mainstay of traditional risk assessment models, there is growing evidence that microvascular dysfunction is an early marker of vascular injury that is present prior to the onset of overt cardiovascular disease. The retinal microvasculature offers a unique non-invasive window into systemic vascular health, given its structural and physiological similarities with the coronary and cerebral circulation. Microvascular retinal changes such as changes in vascular caliber, tortuosity, branching and patterns of perfusion, are associated with hypertension, diabetes mellitus, stroke, coronary artery disease and heart failure. Advanced retinal imaging techniques like optical coherence tomography angiography (OCT-A) also allow to analyse the architecture of retinal vessels and microcirculatory function in detail. In addition, emerging molecular evidence shows shared pathways of endothelial dysfunction, inflammation, oxidative stress, and activation of the renin-angiotensin system, that link retinal vascular remodeling to cardiovascular pathology. Over the past few years, the use of artificial intelligence (AI) and deep learning (DL) has dramatically increased the capability of retinal imaging for translation, leading to automated acquisition of high dimensional vascular features and prediction of cardiovascular risk from retinal photographs. Retinal imaging biomarkers could complement clinical and molecular parameters to enhance risk stratification for cardiovascular disease in an individual basis. Overall, retinal microvascular phenotyping is a promising tool for early cardiovascular risk assessment, scalable, and non-invasive, and could play a role in future precision cardiology and preventive cardiovascular care strategies.