Ophthamologist wearing goggles

AI ophthalmology triage service could be in NHS hospitals in three years

Image credit: Dreamstime

A neural network trained with data collected from NHS patients has proved itself capable of identifying the markers of eye disease at least as well as world-leading clinicians, the Press Association (PA) has reported.

The researchers behind the system hope that it could help reduce sight loss, which is often preventable by early detection and intervention. Neural networks have previously been built which can identify early signs of diabetic eye disease, skin cancer and other health conditions.

“The number of eye scans we’re performing is growing at a pace much faster than human experts are able to interpret them,” Dr Pearse Keane, a Moorfields clinician, told PA. “There is a risk that this may cause delays in the diagnosis and treatment of sight-threatening diseases, which can be devastating for patients.”

“The [artificial intelligence] technology we’re developing is designed to prioritise patients who need to be seen and treated urgently by a doctor or eye care professional. If we can diagnose and treat eye conditions early, it gives us the best chance of saving people’s sight,” he added.

Researchers from Moorfields Eye Hospital, London, University College London and Google’s DeepMind research centre worked together on the system.

The artificial neural network – a type of machine learning system loosely inspired by the structure and processes of biological brains – used optical coherence tomography (OCT) retinal scan data from nearly 15,000 NHS patients. This large dataset enabled the network to identify abnormalities in these scans which could indicate eye disease.

The network was able to identify patients with more than 50 different eye conditions in more than 94 per cent of cases from their retinal scans, and recommend that they are sent on to a specialist or put under observation. This rate matches the accuracy of world-leading ophthalmologists.

The researchers are making plans for clinical trials for the system. They believe that the system could be made available for clinical use in 30 UK hospitals within three years.

Ophthalmologists are highly unlikely to find themselves out of work due to the network, which is unable to make its own diagnoses. It is likely that such a system will assist with triaging patients, who may subsequently be diagnosed by a human clinician. The clinician will then decide on treatment, potentially preventing irreversible sight loss.

“We set up DeepMind Health because we believe artificial intelligence can help solve some of society’s biggest health challenges, like avoidable sight loss, which affects millions of people across the globe,” said Mustafa Suleyman, head of Applied AI at DeepMind Health.

“These incredibly exciting results take us one step closer to that goal and could, in time, transform the diagnosis, treatment and management of patients with sight-threatening eye conditions, not just at Moorfields, but around the world.”

This week, it was reported that an “unorthodox” approach to reinforcement learning (a category of machine learning) had resulted in a neural network capable of determining optimal cancer treatment plans for patients which minimised side effects.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them