Author profile picture

Researchers from the University College London (UCL) and Western Eye Hospital (London, UK) have developed a test using artificial intelligence (AI) to predict so-called wet age-related macular degeneration (wet AMD) three years before the first symptoms appear.

AMD is a common eye disease affecting approximately 2.3% of adults over 65 years of age. It is characterized by the death of cones in the yellow spot (the macula lutea) of the retina, which reduces visual acuity.

There is a dry form as well as a wet form, which is worse. In addition to dying cones, the wet form also involves the growth of small blood vessels under the retina, out of which fluid leaks into the retina. This leads to impaired vision within a short period of time.


Over the last 20 years, techniques have been developed to counteract this, including laser treatment, but often the treatment comes too late to be able to minimize any damage.

That is why UCL sought a solution and eventually found it in a DARC test (Detection of Apoptosing Retinal Cells).

With the DARC test, a fluoridating substance is injected into the bloodstream that binds to retinal cells. The weak cells light up during an eye scan (see photo).


This image has an empty alt attribute; its file name is amd_pr_image_rgb_8x5-678x424.jpg

But you are still not there yet. The problem with scans of this kind is that specialists often reach different conclusions. Artificial intelligence has provided the solution by comparing large quantities of photographs with each other. This has taught the computer to recognize patients with early-onset wet AMD.

The results are “promising” according to Professor Francesca Cordeiro of the UCL Institute of Ophthalmology. “Our new test is capable of predicting wet AMD up to 36 months before symptoms appear.”
Clinical trials are now required to improve the results, she states.

Check out our dossier on artificial intelligence here.