In a first-of-its-kind project in Mumbai, which has found mention in the Journal of the American Medical Association (JAMA)’s ophthalmology journal, a private hospital in association with the BMC used Artificial Intelligence (AI) and smartphones to screen diabetic retinopathy in dispensaries run by the civic body.
The technique, which can be used by semi-skilled healthcare workers, will not only help in early diagnosis but also find wide use in rural areas to diagnose this disorder. Diabetic retinopathy is caused by damage to blood vessels in retina. It may cause blindness. Diabetes is a major cause of the disorder.
Last year, the BMC signed an MoU with Aditya Jyot Hospital to screen patients in BMC dispensaries using a portable gun-shaped, smartphone-mounted device. The smartphone takes high-quality retinal pictures and uses AI to determine whether a patient has to be referred to an ophthalmologist for diabetic retinopathy treatment.
According to Dr Radhika Krishnan, CEO of Aditya Jyot Foundation, while the technology has been in use in developed countries, this is the first time that AI was used to detect the disorder without using WiFi or Internet connection.
“The AI functions offline. In other techniques used so far, the device would upload the pictures online and a software would detect retinopathy. With offline technology, this device can be used in rural areas,” she said.
The software has been developed by a Bengaluru-based start-up. In the last one year, 1,688 people were detected with the disorder in BMC dispensaries.
The study, published in JAMA journal last week, provides details of 231 patients. “We found 100 per cent sensitivity in detecting retinopathy through this device,” said Dr Sundaram Natarajan, medical director at Aditya Jyot.
Traditionally, a patient has to visit an ophthalmologist for diagnosing this disorder. An ophthalmologist uses infra-red device in a typical clinic set-up to diagnose the disorder.
The hand-held device can be used by non-medical staff who just need to hold the device and take a photo of eye. It will specially be helpful in rural areas.