CHENNAI: Mass diabetic retinopathy screening with high predictive accuracy could soon become a reality in high-burden countries such as India, eliminating the need for individual appointments and multiple referrals. This shift is also being enabled by a breakthrough artificial intelligence system developed by a Bengaluru-based Indian med-tech company.
Remidio Innovative Solutions on Thursday said its regulatory-approved diabetic retinopathy (DR) artificial intelligence system has demonstrated strong real-world clinical performance in the largest independent evaluation of DR AI algorithms published to date. The study, published in The Lancet Digital Health, assessed how well AI systems perform when deployed at population scale rather than in controlled research settings.
Led by Professor Alicja Rudnicka and Professor Adnan Tufail in collaboration with the UK National Health Service, the evaluation analysed more than 200,000 diabetic eye screening encounters covering around 1.2 million retinal images. Of the 32 DR AI algorithms initially assessed, eight met the criteria for final analysis, including Remidio’s CE (Class IIa)-approved system.
What makes the findings notable is their focus on real-world screening relevance. The study found that Remidio’s AI achieved an overall sensitivity of 87 percent for referable diabetic retinopathy, which includes moderate non-proliferative DR and more advanced disease. Sensitivity rose sharply for more serious conditions, reaching about 99 percent for moderate-to-severe non-proliferative DR and approximately 97 percent for proliferative DR. This indicates that the system is highly unlikely to miss sight-threatening disease, a critical requirement for large-scale screening programmes.
Equally important from a health-system perspective, the AI demonstrated high specificity and positive predictive value, helping to reduce false positives and unnecessary referrals. With a negative predictive value of around 98 percent, the system offers strong reassurance for patients who screen negative, potentially enabling wider screening coverage and longer intervals between examinations without compromising safety.
Clinicians involved in large public health programmes said this balance between sensitivity and specificity is essential.
Dr R. Kim, senior consultant at Aravind Eye Hospital, says the findings are clinically meaningful because they show AI can reliably detect sight-threatening disease without overwhelming referral systems. He noted that such operational discipline is crucial for sustainable screening, particularly in high-volume public health settings.
The evaluation also highlighted the broader health-system impact of AI deployment. According to the study, Remidio’s AI could reduce reliance on human graders by up to 80 percent, allowing specialist ophthalmologists to focus on patients who need treatment rather than routine screening. This has implications for cost efficiency and workforce optimisation, especially as diabetes prevalence continues to rise globally.
Another significant finding was the stability of AI performance across demographic subgroups, including age, sex and ethnicity. This consistency supports the equitable deployment of AI-based screening tools across diverse populations, addressing concerns that algorithmic performance may vary between groups.
Dr Divya, Chief Medical Officer at Remidio Innovative Solutions, said the results validate the company’s focus on designing AI for real-world screening environments rather than idealised clinical trials. She said the system’s ability to identify sight-threatening disease while avoiding unnecessary referrals is critical for safe and efficient public health use.
As health systems grapple with growing diabetes-related eye disease and constrained specialist capacity, the study underscores the potential role of well-validated AI tools in strengthening screening pathways. By combining high clinical sensitivity with operational efficiency, such systems could play a central role in building scalable, equitable and sustainable models for diabetic eye care worldwide.