BENGALURU: With treatable diseases like Diabetic Retinopathy, one of the leading causes of blindness in Indian patients going undetected, Bengaluru-based Sankara Eye Foundation is helping in developing Nethra.AI, an artificial intelligence enabled software to detect eye diseases. It uses computer vision and machine learning.
India has an ophthalmologist: patient ratio of 1:10,000. With even fewer retinal specialists, eye diseases are going increasingly unnoticed.
Singaporean company Leben Care reached out to Sankara six months ago, asking it to help in machine learning. Their product has so far been fed with one lakh eye images of patients and will soon be not only commercially deployed here but also in South Eastern China to screen diseases, although the exact date is not known yet.
“For sensitivity specificity we are looking at glaucoma and age-related degeneration too which bear a significant burden for India. Teaching the machine what it should see, correlate it with other things is our main job. Machine finally works on an algorithm where it does image analysis. It also requires someone to validate what it detects saying an opthalmologist would also have made the same detection,” Dr Kaushik Murali, paediatric ophthalmologist and president, Sankara Eye Foundation.
Complete confidentiality of patient data is maintained as the images are not identified with patient details. “Diabetes is going to be the main reason for blindness soon. We cannot have an ophthalmologist detect it in all patients. We need to have a machine that at least alerts the patient that he needs to see an ophthalmologist soon from a screening standpoint. It will never replace a retinal surgeon, it will largely act as a referral trigger,” Murali said.
The early stages of both diabetic retinopathy and glaucoma are asymptomatic. “There comes a tipping point where the retina or optic nerve is so damaged that it causes irreversible blindness so you can only salvage whatever is left behind. The earliest referral the patient has can save his potential vision. That is why we are looking at these two diseases,” he said.
Imran Akhtar, CTO and founder, Leben Care, said, “With retinal diseases the challenges are that one needs to dilate the eye, use a special camera called Fundus Camera to study the retina and then identify what the abnormalities are. Only 2,25,000 eye specialists are there in the world and only 15 per cent are retina experts. We want to enable disease detection with as little intervention from the specialist as possible.”