AI could help to reduce diabetic screening backlog in wake of COVID-19
Scientists highlight that machine learning could safely halve the number of images that need to be assessed by humans
The study, which was published in British Journal of Ophthalmology, used the AI technology, EyeArt, to analyse 120,000 images from 30,000 patient scans in the English Diabetic Eye Screening Programme.
The technology had 95.7% accuracy for detecting eye damage that would require specialist referral, and 100% accuracy for moderate to severe retinopathy.
The researchers from St George’s, University of London, Moorfields Eye Hospital, UCL, Homerton University Hospital, Gloucestershire Hospitals and Guy’s and St Thomas’ NHS Foundation Trusts highlight that the introduction of the technology to the diabetic screening programme could save £10 million per year in England alone.
Professor Alicja Rudnicka, from St George’s, University of London, said that using machine learning technology could safely halve the number of images that need to be assessed by humans.
“If this technology is rolled out on a national level, it could immediately reduce the backlog of cases created due to the coronavirus pandemic, potentially saving unnecessary vision loss in the diabetic population,” she emphasised.
Moorfields Eye Hospital consultant ophthalmologist Adnan Tufai, highlighted that most AI technology is tested by developers or companies, but this research was an independent study involving images from real-world patients.
“The technology is incredibly fast, does not miss a single case of severe diabetic retinopathy and could contribute to healthcare system recovery post-COVID,” he said.