AI Analyzes 62,876 Retinal Photos to Flag Alzheimer's Risk Years Before Diagnosis
Updated
Updated · ScienceAlert · Jun 19
AI Analyzes 62,876 Retinal Photos to Flag Alzheimer's Risk Years Before Diagnosis
2 articles · Updated · ScienceAlert · Jun 19
Summary
62,876 retinal photographs from more than 40,000 UK Biobank participants let researchers train AI models to flag people at higher risk of Alzheimer's years before diagnosis.
The system did not diagnose future Alzheimer's directly; it predicted 12 linked risk factors, including age, smoking, sleep, alcohol use, depression, body mass and blood pressure.
Retinal patterns tied to later Alzheimer's included vascular stiffening, lower blood-vessel density, optic-nerve thinning and constricted small arteries and arterioles.
University of Florida-led researchers said routine retinal imaging used for diabetes, glaucoma and cataracts could become a low-cost early warning tool, though the findings remain associative and need further study.
Researchers have made a major breakthrough by using artificial intelligence to detect Alzheimer's disease risk factors from routine retinal scans. This advancement, led by Ruogu Fang and her team, applies deep learning to analyze detailed images of the retina, allowing the AI system to spot subtle, biologically relevant changes that may signal Alzheimer's long before symptoms appear. Announced in June 2026, this innovation marks a pivotal step toward non-invasive, early detection of Alzheimer's, offering hope for earlier intervention and improved patient outcomes by leveraging technology already used in eye exams.