The retina may offer a surprisingly detailed glimpse into how the body is ageing.

Researchers at the University of Lausanne have developed an artificial intelligence model that can estimate a person’s retinal age from a photograph of the back of the eye — and the gap between that estimate and chronological age appears to carry broader health information.

Published in Nature Communications, the study suggests that retinal ageing is associated with future disease risk, inflammation, metabolism and lifestyle. It also revealed notable differences between men and women, including a shift in women around the age of menopause.

How The AI Estimated Age

The researchers adapted RETFound, a foundation model designed for analysing retinal images, and trained it using photographs from more than 70,000 UK Biobank participants aged 40 to 79.

The model learned to predict chronological age from a retinal photograph with an average error of less than three years. Researchers then calculated the “retinal age gap” — the difference between the age predicted by AI and a person’s actual age.

A positive gap meant the retina appeared older than expected, while a lower retinal age suggested comparatively slower biological ageing.

The retina is particularly useful for this type of research because its blood vessels can be observed directly. It is also part of the central nervous system and can reflect processes associated with inflammation, diabetes and vascular health.

An Older Retina Was Linked To Risk

A larger retinal age gap was associated with a higher risk of several major health outcomes.

People whose retinas appeared older had an increased risk of cardiovascular and respiratory diseases, cancer, dementia and death during the following 15 years. Retinal age was also associated with measures of metabolism and inflammation, cognitive performance and lifestyle factors.

These findings do not mean that an older-looking retina causes disease. Instead, retinal age may capture biological changes occurring elsewhere in the body at the same time.

Men And Women Showed Different Patterns

One of the study’s most striking findings was that retinal ageing did not appear to reflect the same biological processes in men and women.

In men, an older retinal age was more strongly associated with characteristics of metabolic syndrome, including high blood pressure, diabetes, high cholesterol and excess body weight.

In women, the associations were more closely connected with vascular health. A higher retinal age was linked with an increased risk of thrombosis, an association that researchers did not observe in men.

Age also appeared to change the pattern. Before menopause, women tended to have retinas that appeared younger than those of men of a similar age. After menopause, this pattern reversed, with women showing higher retinal ages and a less favourable health profile.

The researchers stressed that the study cannot prove menopause itself caused this change. Hormonal and vascular changes occurring during this stage of life may contribute, but further research is needed.

What Is The Model Detecting?

AI does not necessarily examine retinal images in the same way as a clinician.

The model appeared to focus on particular retinal regions and detect subtle differences in brightness, colour and blood-vessel density that may be difficult for the human eye to recognise. The connection with vascular patterns seemed especially important in women.

Some of the model’s internal signals remain difficult to interpret, which means researchers still need to understand precisely why particular image features correspond with ageing and disease risk.

A Marker, Not A Diagnosis

The researchers emphasised that retinal age should not be interpreted as a diagnosis. Having a retina that appears older does not mean someone currently has, or will inevitably develop, a particular disease.

Instead, the retinal age gap is being investigated as a potential prognostic biomarker — an additional indicator that could help estimate future health risk alongside established clinical measurements.

Retinal photography is relatively simple, inexpensive and minimally invasive, which makes the approach attractive for future research. However, clinical use remains some way off. The findings still need to be tested in more diverse populations and researchers must establish whether retinal age can provide information useful enough to change medical decision-making.

For now, the study adds to growing evidence that the eye can reveal far more than vision alone. A single retinal image may eventually help researchers understand how ageing is progressing across the body — but its greatest value will depend on how reliably those signals translate into meaningful information about health.