When researchers from the University of Edinburgh analysed the health data and retinal images of more than 9,300 UK Biobank participants, they uncovered a powerful new way to predict people’s long-term risk of serious retinal damage. 

The team employed fundus refraction offset, which uses artificial intelligence (AI) to compare how a person’s retina looks structurally to how it should normally look according to their glasses prescription. It could help clinicians better identify patients at risk of retinal disease and lead to more tailored management to prevent vision loss, the researchers said (Yii et al, 2025).

Such is the beauty, and significance, of biomarkers. Defined as an objective and quantifiable measure of a physiological process, pathological process or response to a treatment (European Medicines Agency, 2025), biomarkers “may represent a variety of things”, says Michael Cook, Executive Director of Science at Our Future Health, including “environmental exposures, endogenous biological processes, interactions of exposures and biology, risk of disease, presence of disease and propensity for disease to progress”. The types of biomarkers clinicians may use range from height and weight to blood chemistry, MRI features and, in optometry, retinal optical coherence tomography (OCT).