Researchers at Tohoku University have developed an artificial intelligence model that estimates “retinal age” from fundus images. This advance, noted in a study published in Communications Treatmentcould revolutionize the early detection of serious diseases such as diabetes.
The model was trained with more than 50,000 images, achieving higher accuracy than other previous methods. It focuses on the identification of specific retinal patterns, incorporating the glucose marker HbA1c to refine its estimates. Photos of the retina are easy to obtain during medical checkups.
Initial studies show that the difference between estimated retinal age and chronological age is greater in people with health conditions such as diabetes and heart disease. This suggests a potential for the use of retinal screening as a frequent health indicator.
Paths to the future
Despite the promising results, the researchers point out that more longitudinal studies are necessary to validate the relationship between retinal age and disease development.
A study is planned in which a cohort of 10,000 individuals will be followed for three years to evaluate these connections.
Retinal age to predict diseases
“Retinal age” (the biological age that an algorithm calculates from an image of the retina) can serve as a biomarker for the body’s frequent aging, and it is not only related to diabetes. When the retina appears older than chronological age, an increased risk of other diseases and premature death has been observed, even in people without diagnosed diabetes.
Risk of cardiovascular and cerebrovascular diseases. The vascular characteristics of the retina (caliber, curvature, branching of vessels) reflect the state of blood vessels throughout the body. Studies have shown that certain retinal patterns may be associated with increased risk of acute myocardial infarction and stroke, suggesting that “retinal age” may capture systemic vascular damage.
Neurological and neurodegenerative diseases. The retina is an extension of the central nervous system, so its structural changes (loss of neurons, alterations in the nerve fiber layer) can be related to neurodegenerative processes such as dementia or similar diseases. Although accurate prediction of specific diseases is still under investigation, the fact that the retina “ages faster” is interpreted as an indication of broader neuronal damage.
Frequent mortality and other systemic diseases. Various studies have found that a retina “older” than the current age is associated with a greater risk of death from non-cardiovascular and non-oncological causes, such as respiratory, liver or kidney diseases. For example, for each year that the retina appeared older, the risk of dying in subsequent years increased statistically significantly, suggesting that retinal age integrates the wear and tear of multiple organs.
How is it used in practice? Currently, retinal age is calculated using artificial intelligence algorithms that analyze fundus photographs (retinographs) and compare the vascular and tissue pattern with databases of people of different ages. In the future it could be used as a rapid screening test to identify people with accelerated biological aging, who in turn could receive closer controls for blood pressure, cholesterol, kidney function, etc., beyond the defense of diabetes.
You may also be interested in:
· Although widely consulted, most popular AI chatbots often give poor health advice
· New guidelines suggest a more intense, personalized approach to LDL cholesterol management
· These Popular Medications Are Affecting Your Kidneys and You Might Not Know It
