Aims: The overall goal of this project is to develop and validate a screening test for predicting ocular diseases, systematic metabolic diseases and mental health problems through eye imaging. Scientific rationale: The modern lifestyle has invariably contributed to a significant increase in the prevalence of systemic metabolic disorders and mental health issues. Predicting an individual’s disease risk with precision is crucial for administering preventative treatments to asymptomatic individuals who are at a heightened risk. Presently, the clinical prediction instruments at our disposal are hampered by a lack of accuracy, the necessity for invasive procedures, substantial costs, and various other constraints. Often referred to as ‘the windows to the soul,’ the eyes offer a unique glimpse into the overall state of one’s body. In recent decades, research in this domain has advanced swiftly, with the discovery of innovative ocular biomarkers. However, much of the existing research has concentrated on a limited number of features, identified through standardized assessment systems by ophthalmic specialists. Such processes are not only labor-intensive but also fall short in yielding ‘personalized’ image features and patterns that cater to individual variations. The ability of learning the “personalized” image features of the deep learning technology is impressive and investigate these image features’ correlation and prediction value of various diseases is meaningful. Project duration: 36 months. Public health impact: Our project will benefit hundreds of millions of people worldwide who are suffering from visual impairment and other metabolic or mental problems. By leveraging the rich information contained within fundus images, our exploration into the relationships among metabolomics, genomics, radiomics, and disease has the potential to reveal new therapeutic targets and establishing follow-up indicators.
