Could Artificial Intelligence be the answer to the early diagnosis of eye diseases such as macular degeneration, glaucoma and diabetic retinopathy? One consultant at Moorfields is so convinced that the answer lies in AI that he contacted Google’s AI research subsidiary DeepMind directly to talk about the idea. And the result is a collaboration which could lead to a simple digital scan to recognise sight-threatening conditions.
Pearse Keane, a consultant ophthalmologist at Moorfields, was familiar with the strides that DeepMind had made in the area of “deep learning” - a subfield of machine learning concerned with algorithms inspired by the structure and function of the brain called artificial neural networks – and he recognised that there was a potential application in the field of eye disease diagnosis.
The new research involves producing an algorithm, based on a million digital eye scans, which would allow a neural network to assess eye scans. This kind of deep learning would effectively make individual expert experience available on a mass scale. As Professor Peng Tee Khaw, the head of Moorfield’s ophthalmology research centre, says; “It takes me my whole life experience to follow one patient’s history. And yet patients rely on my experience to predict their future. If we could use machine assisted deep learning, we could be so much better at doing this, because then I could have the experience of 10,000 lifetimes.”
This is not the first time a collaboration between DeepMind and the NHS has been attempted. Two years ago a project involving The Royal Free Hospital ran aground amidst controversy over the use of confidential healthcare data. This time, potential confidentiality issues have been circumnavigated by the use of anonymised data.
AI could spot eye disease more accurately than doctors, study suggests.