Research

CSIRO first to commercialise AI screening for diabetic retinopathy

Rarkably, the five-mber team required only two years to complete and trial the technology, in the process bypassing similar efforts being undertaken by Google’s DeepMind and IBM Research. Now, following the success of the recent trials, they are in a position to put their hard work to practical use.“It moved very fast. Initially we tested on large-scale number of images and then rolled it out to clinics to do the trials, and now we’re ready to move further,” the technology’s creator and trial co-lead, Professor Yogi Kanagasingam said.”“Google has got a syst, but first to market is a huge thing. It’s very difficult to compete with Google but we’ll see how it goes.”{{quote-A:R-W:400-I:2-Q: There’s about 415 million people with diabetic issues, so you imagine it’s difficult for ophthalmologists to screen each and every one, there’s not enough of th. So this is the only option for the future, -WHO:Professor Yogi Kanagasingam}}The new technology, nicknamed ‘Dr Grader’, not only screens at risk patients for DR, but also provides a grading of the severity of the condition. According to Kanagasingam, the difference between their syst and the competition’s is in the software.“It is a combination of artificial intelligence and rule-based technology, so that is the difference between ours and Google’s – Google’s is purely artificial intelligence,” he said.“We didn’t miss anybody with severe disease, we didn’t miss anybody who should be referred. We are getting a sensitivity of almost 97%, which is very high.”The technology was first tested on thousands of images, before being put to work at a GP clinic in Western Australia. During the trial, GPs successfully screened 187 diabetic patients, taking high-resolution images of their eyes, which were then analysed by the technology for signs of diabetic retinopathy.Importantly, as a basis for comparison the images were also analysed by an ophthalmologist, and the technology was found to be as effective as the specialist in detecting signs of diabetic retinopathy and grading its severity.As a result, the syst has been given the green light to be rolled out across WA and Kanagasingam said the innovation would provide a more efficient pathway for DR patient referrals.“Patients at risk of this condition would usually be referred to a specialist for screening, waiting six weeks or more – now it can potentially be done in a single 30-minute visit to a GP,” he explained.“Around the world there’s about 415 million people with diabetic issues, so you imagine it’s difficult for ophthalmologists to screen each and every one, there’s not enough of th. So this is the only option for the future.“It could help avoid unnecessary referrals to public hospitals, potentially reduce waiting periods for patients and enable ophthalmologists to focus on patients needing treatment and surgery. It could also help reduce the financial impact of diabetes on the Australian economy, which is estimated to cost up to $14 billion a year.”US business TeledC has already acquired the licence for the technology, and will seek to install it at a further 20 GP clinics in WA over the next few months, with a view to expanding across Australia and potentially Singapore.