Age estimates differed between two types of eye photographs
Average age-prediction errors were 7.09 years for photos centered on the retina’s middle, versus 7.77 years for photos centered where the optic nerve exits.
Limit: Differences between views varied across prediction ranges and patient subgroups, limiting interchangeability
Edited by Boris Djordjevic · AgingResearch.News
A retrospective cohort study examined 5,895 patients aged 40 or older at Moorfields Eye Hospital in London between 2008 and 2018. Researchers analyzed 12,230 pairs of same-day photographs of the retina, the eye’s light-sensing tissue. Each pair included one image centered on its central region and another centered where the optic nerve exits the eye.
Researchers used a previously validated artificial intelligence model to predict age from each photograph. They calculated a retinal age gap: predicted age minus actual age. Predictions from the two views agreed closely overall but differed systematically. A larger age gap was associated with age-related macular degeneration, a disease affecting central vision. The researchers found no statistically significant evidence that photographic framing modified the age gap’s associations with the diseases examined.
Why it matters
Research on retinal age estimates asks whether eye images can help identify age-related disease. The authors supported using either photographic view when researchers need imaging flexibility.
Caveats
This retrospective hospital-based study did not establish that the views are interchangeable: differences varied across prediction ranges and patient subgroups. The authors called for studies following people from the general population before clinical use.
The paper
University College London; Moorfields Eye Hospital NHS Foundation Trust
Investigative Ophthalmology & Visual Science, Jun 1, 2026, Open access
