AI-Powered Smartphone App Identifies

Diabetes-Related Vision Problem

 

 

The Indian medical regulator has approved the app developed by Remidio Innovative Solutions.

The Indian medical technology startup Remidio Innovative Solutions has created a smartphone app that employs artificial intelligence (AI) to identify diabetic retinopathy (DR), an eye ailment that can result in blindness in diabetics.

After downloading the Remidio Medios DR AI app to a smartphone, the device can be linked to the company’s portable retinal camera. According to the company, the setup enables DR screening in remote locations without access to servers based on graphics processing units or even the internet. It is presently being used in Himachal Pradesh, rural West Bengal, and other areas with limited screening availability. It additionally has the advantage benefit of not requiring the patient’s pupils to be dilated.

High blood sugar levels can harm retinal blood vessels, which is the cause of diabetic retina. They may then enlarge, leak, or obstruct. The National Institutes of Health state that this can lead to the development of new, poorly functioning blood vessels in the eye that are prone to leaking or bleeding. Diabetic-related blindness can be avoided with early DR diagnosis.

Recently, India’s Central Drugs Standard Control Organization (CDSCO) certified Remidio Medios DR. Additionally, the Health Sciences Authority of Singapore and the European Commission have approved it. But as of right now, it’s not accessible in the US.

“The CDSCO’s approval helps us bring care closer to patients, but we are particularly excited to see India become a global leader in leveraging AI to eliminate preventable blindness caused by DR,” Divya Rao, the chief medical officer of Remidio,
Utilizing two convolutional neural networks based on Google’s Inception-V3, a deep learning technique is employed by the AI for computer vision tasks such as image analysis and identification.

According to a study published in the Indian Journal of Ophthalmology, the program was found to be highly successful in detecting DR. The algorithm was trained on a set of more than 50,000 photographs of retinas, some with DR and others without.

“A science-first, public-health approach has helped Remidio not only assess the performance of its AI in real world settings, but also for the AI research ecosystem to learn from clinical outcomes associated with primary care-led DR screening programs, care-gap closure pathways and the cost-benefit analysis of bringing screening closer to the patient,” said Rajiv Raman, a senior vitreo-retinal consultant with Sankara Nethralaya, a specialty eye-care hospital in India.