Can Retinal Age Help Predict Future Disease?
A retinal photograph captures much more than the appearance of the eye.
Using artificial intelligence, researchers can detect age-related patterns in retinal blood vessels and other structures. Increasingly, scientists are asking whether these patterns might also provide information about future health.
Recent evidence suggests that they may.
What does retinal age measure?
Retinal age is an estimate generated by artificial intelligence from a photograph of the retina.
Researchers compare this estimate with a person’s chronological age. The difference is known as the retinal age gap.
If a 60-year-old person’s retina is estimated to be 65, the retinal age gap is +5 years.
This does not mean that the person’s entire body is biologically five years older. Instead, the retinal features detected by the model appear older than expected for that person’s chronological age.
Researchers can then investigate whether people with an older-appearing retina experience different health outcomes over time.
What happened during long-term follow-up?
In a large 2026 study, researchers analysed retinal images from more than 71,000 UK Biobank participants and examined health outcomes during long-term follow-up.
A higher retinal age gap was associated with a greater likelihood of several future outcomes, including:
cardiovascular disease,
stroke,
chronic obstructive pulmonary disease,
dementia,
thrombosis,
ischemic heart disease,
cancer,
and all-cause mortality.
When researchers studied the development of a particular disease, people who already had that condition were excluded from the analysis.
This allowed them to examine whether retinal age was associated with new disease developing later, rather than simply reflecting a disease that was already known.
Does retinal age add information beyond traditional risk factors?
Age, smoking, blood pressure, cholesterol, body weight, diabetes and previous disease already provide substantial information about future health risk.
The researchers therefore took established clinical risk factors into account.
Retinal age gap remained associated with several future outcomes even after these factors were considered.
This suggests that retinal images may contain additional information about long-term health risk that is not completely captured by conventional risk factors.
But there is an important distinction.
A marker can be associated with future disease without substantially improving our ability to predict which individual person will become ill.
When retinal age gap was added to conventional risk factors, the overall improvement in prediction was generally modest. For some outcomes it was statistically significant; for others it was not.
So retinal age may add information, but the evidence does not show that it should replace established clinical risk assessment.
Can retinal age predict what will happen to you?
No single risk marker can tell us with certainty what will happen to an individual person.
A person with an older retinal age may remain healthy for many years. Someone with a younger retinal age may still develop cardiovascular disease, cancer or another condition.
The study also cannot tell us whether changes in the retina contribute to disease or whether they reflect biological processes already occurring elsewhere in the body.
This is why retinal age should be understood as a risk-related signal, not as a forecast of an individual’s future.
Where could retinal age become useful?
The attraction of retinal imaging is clear.
A retinal photograph is rapid and non-invasive and provides information about vascular and neural tissue. Artificial intelligence may allow additional health information to be extracted from this relatively simple examination.
In the future, retinal age could potentially become one additional layer of health assessment, interpreted alongside medical history, blood pressure, laboratory results, imaging, physical fitness and other established measures.
Research now suggests that the retina may contain information relevant to future health. The next question is whether using that information can meaningfully improve clinical decisions and outcomes.
