AI Meets the Eye: A Multimodal Artificial Intelligence Approach for the Detection of Ocular and Systemic Diseases

Background: Artificial intelligence (AI) is increasingly playing an important role in the detection and diagnosis of various ocular and systemic diseases. Multimodal retinal imaging together with AI has enabled early identification of disease with potential for earlier management and treatment. The retina represents a non-invasive window through which the vascular, neurological, and metabolic condition of the eyes and the body as a whole can be examined, making it an ideal medium for AI diagnostic platforms. Objective: This paper aims to investigate the performance of artificial intelligence for the detection of various ocular and systemic diseases using multimodal retinal imaging, and to compare AI-based screening against clinician-based screening through a systematic review and meta-analysis. Methods: The databases searched were PubMed, Medline, Web of Science, ScienceDirect and CENTRAL. Included studies involved AI and deep learning screening of diabetic retinopathy applied to fundus images and OCT scans. Meta-analysis of sensitivity and specificity was conducted using Meta-DiSc software with a bivariate random-effects model, and quality was assessed using QUADAS-2. Results: A total of eighteen studies from 2,065 initial articles were included. The overall sensitivity and specificity of AI-based screening for retinopathy were 0.877 and 0.906 respectively, while for clinicians they were 0.75 and 0.941 respectively, indicating greater sensitivity for AI with comparable specificity. A combined model integrating fundus photos, OCT, and patient data achieved 95.1% diagnostic accuracy. Retinal AI screening also showed utility in assessing systemic disease biomarkers for cardiovascular disease, stroke risk, chronic kidney disease, and several neurological disorders. Conclusion: Artificial intelligence demonstrates high diagnostic accuracy for the identification of ocular diseases as well as screening for systemic diseases. Combining AI and multimodal imaging offers an accurate, affordable, and scalable means of screening with the ability to significantly reduce the global impact of avoidable blindness.