Detecting A Range Of Eye Diseases Through DeepMind’s AI Tool

A prototype medical device that scans a patient’s retina for eye diseases in real time has been developed by Moorfields Eye Hospital in London and AI algorithms from Google’s DeepMind. The DeepMind algorithm is trained to spot indicators of eye disease and recommends how patients should be referred for treatment and care. After the retina is scanned, the device analyses the images and after approximately 30 seconds returns a diagnosis and an “urgency score”. The diseases it can diagnose include diabetic retinopathy, glaucoma, and age-related macular degeneration. The accuracy of the diagnosis itself is said to be on par with top eye specialists according to DeepMind; early results for the system show a correct referral rate of 94 per cent for over 50 types of eye disorders. Developed over the last three years, it is yet to be submitted for regulatory approval. However, the core benefit of such a device is that it can address the shortage of specialists who can interpret high volumes of eye scans performed every year.