Book Title: Multidisciplinary Research Nexus: Exploring Intersections of Knowledge
Editor: Prof. Amos R
Chapter: 11
DOI: https://doi.org/10.59646/559/11
Authors: Druthi S, and Amos R
Abstract
Cardiovascular disease (CVD) remains the leading global cause of mortality and morbidity. Retinal imaging—particularly fundus photography and optical coherence tomography (OCT)—provides a noninvasive window into the microvasculature and neural tissue that reflect systemic vascular health. Recent machine-learning (ML) and deep-learning (DL) approaches have demonstrated promising ability to extract cardiovascular risk markers and predict outcomes directly from retinal images. This paper reviews the current state of the field, synthesizes representative model architectures and reported performance, discusses methodological and clinical challenges, and outlines directions to advance clinical translation. Representative works show that DL models can predict patient age, blood pressure, smoking status and major adverse cardiovascular events from fundus images with moderate accuracy, and emerging OCT-based DL systems show promise for improved event prediction. However, variability in datasets, limited external validation, and unclear clinical utility impede immediate adoption. Future work should prioritize large multi-ethnic cohorts, multimodal retinal imaging, explainability, and prospective outcome studies.