Powered by Google’s AI company DeepMind, the system was able to identify more than 50 types of eye disease in seconds, with 94-per-cent accuracy. The hospital also said that in around 1,000 tests performed by the system, not one case that would lead to an urgent referral was missed. The system is due to begin clinical trials next year, and could possibly be within hospitals in three years, according to Moorfields.
Nearly 15,000 histrionic and anonymised 3D, high definition optical coherence tomography retinal scans were examined by the AI system in order to ‘teach’ it to diagnose conditions. It then scanned a further 1,000 patients currently at Moorfields and its results were compared to the findings of the hospital’s consultant ophthalmologists and specialist optometrists.
Pearse Keane, a consultant ophthalmologist at Moorfields and scientist at the UCL Institute of Ophthalmology, told The Evening Standard: “They are pretty stunning results and could, I think, transform the whole speciality of ophthalmology in the next few years.”