How are AI models in ophthalmology advancing healthcare?
In a recent presentation to the Clinical AI Interest Group at The Alan Turing Institute, INSIGHT director and chief data officer Pearse Keane explained how ophthalmology is spearheading the advancement of AI-enabled research in healthcare.
As evidence, the first autonomous AI system as a medical device was in ophthalmology, approved by the FDA in 2018 for detecting diabetic retinopathy – the commonest cause of blindness in working-age people in developed countries, and increasingly in low- and middle-income countries.
Since then, a growing body of research is demonstrating that by training AI models to process hundreds of thousands of eye scans it is possible to create diagnostic tools for detection not only of eye disease, but also systemic disease such as dementia and cardiovascular disease. This emerging field of ‘oculomics’ — a term coined in 2020 by INSIGHT strategic advisor Professor Alastair Denniston — is being pioneered by Pearse and his AI research group across Moorfields NHS Foundation Trust, University College London's Institute of Ophthalmology, University Hospital Birmingham NHS Foundation Trust and other institutions.
Despite rapid progress in the field, there are significant challenges to meet — including information governance, clinical validation, regulatory approval and business modelling — before a fledgling AI research technology makes it all the way to a care setting where it can benefit patients.
However, in assessing these hurdles, Pearse highlights exciting new developments in AI modelling within ophthalmology. Enabled by cloud infrastructure, coupled with data curation and aggregation tools, these novel models hold the potential to accelerate the translational timeline and expand the scope of healthcare research.
Watch the presentation below to find out more.
(Pearse’s presentation starts at 12’ 45”)
Shared with the kind permission of The Alan Turing Institute’s Clinical AI Interest Group