International, News, Optometry, Research

Artificial intelligence used to diagnose uveitis

Artificial intelligence is being used to extract hidden information from retinal images of patients with birdshot uveitis, in a hope it could lead to more timely diagnosis for the rare condition.

The team of researchers from Birmingham and London are examining the capability of artificial intelligence to extract information from images that is invisible to human examiners. It is hoped the technology could eventually lead to more accurate and earlier diagnosis of the disease, leading to better treatments and outcomes.

Professor Alastair Denniston, project lead from the University of Birmingham, said with current systems patients are often diagnosed late and frequently miss out on the treatment necessary to prevent permanent vision loss.

“It is such a difficult condition to assess reliably so patients may be under-treated and risk losing sight, or over-treated and experience side effects. Our project aims to combine the use of modern imaging techniques with cutting-edge data analytics to develop a set of robust new measures that can help doctors diagnose the condition earlier and personalise treatment decisions.”

The project will also involve the development of a new birdshot image-bank linked to the UK Birdshot Registry and Biobank, which is expected to be useful for both current and future research.

Dr Neil Ebenezer, director of research, policy and innovation at Fight for Sight, said the organisation is pleased to fund this research.

“If doctors can diagnose birdshot uveitis at an early stage, this will mean that a patient will receive treatments sooner when they are more likely to prevent sight loss. And if they have better tools that enable them to more accurately monitor changes in the eye linked with the condition, this will help tailor treatment for each patient – protecting them from sight loss while reducing the risk of side effects.”

The project is one of many around the world that is currently investigating applications for artificial intelligence in diagnosing eye conditions.

The project is being funded by the Birdshot Uveitis Society and Fight for Sight UK.