AI retinal scan predicts osteoporosis risk, linking eye images to bone health

Researchers in Singapore and the United Kingdom have created an AI system, called RetiAGE, that analyses retinal photographs to estimate a person’s risk of osteoporosis. The tool was trained on more than 100,000 eye images, learning to recognise microscopic patterns in blood vessels and tissue that reflect biological ageing.

In the initial study, 1,965 older participants in Singapore provided retinal images and underwent standard DEXA bone‑density scans. The AI predictions were compared with the DEXA results to establish a correlation between retinal signs and weakened bones. The model was then validated on 43,938 individuals from the UK Biobank, followed for about twelve years to see who later developed the disease.

The researchers wrote, "Retina imaging could therefore serve as a simple, non‑invasive and readily accessible method for opportunistic screening of osteoporosis risk." If further confirmed, this approach could offer a low‑cost alternative to expensive DEXA scans and enable earlier detection before serious fractures, which carry a 20‑24 % mortality rate within the first year.