DeepMind, the world leader in AI research, is training an algorithm to diagnose eye diseases by analysing medical images.
The AI has been applied to data collected from thousands of anonymised 3D retinal scans, supplied by Moorfield’s Eye Hospital, a world leader in Ophthalmology based in London.
Early results showed promising signs the technology can quickly and efficiently detect three of the most common and serious eye conditions: diabetic retinopathy, glaucoma, and age-related macular degeneration.
Representatives of DeepMind, which is based in the UK and was acquired by Google in 2014, have said their AI is ‘general’, meaning it could be applied to other types of medical images in the future.
Medical experts say image recognition technology is one of the most promising applications of AI in medicine.
At the AIMed conference in December 2017, Jeremy Howard, Founder and Deep Learning Researcher at fast.ai, said, “You should assume every kind of an image a human looks at, a computer can look at more accurately and faster.
“Medical imaging through AI should not be looked at, at this point, as something weird or controversial. It has been two years since it was definitively proven that an AI system is better at looking at a CT scan than a human is.”
The big question for Howard was: how can image recognition be more quickly and effectively implemented into the clinic?
DeepMind’s technology could enter clinical trials in a few years if results pass a peer review by academics, and the company is planning to enter into further partnerships with University College London and Imperial College London.
Legal constraints could stymie progress, as in 2017 when a UK government watchdog ruled that a National Health Service (NHS) trust had broken the law by giving DeepMind access to 1.6 million patient medical records.