Researchers at Wenzhou Medical University and the University of Coimbra have made a breakthrough in detecting early-stage retinal damage in diabetes using AI-inspired texture analysis. This method can identify subtle changes in retinal tissue before any visible signs of diabetic retinopathy (DR) appear, potentially allowing for earlier intervention and reducing the risk of blindness.
The study used optical coherence tomography (OCT) images to analyze retinal texture in diabetic rats and found significant changes in texture metrics, such as autocorrelation and homogeneity, even when structural and molecular damage was minimal. These changes occurred before any major inflammation or vascular leakage was detectable.
The researchers believe this technology could lead to the development of AI-assisted diagnostic tools that can automatically screen for preclinical DR based on retinal texture signatures. This could enable ophthalmologists to identify high-risk patients before permanent vision damage occurs, allowing for earlier treatment and better outcomes.
The study’s findings have significant implications for the early detection and treatment of DR, which affects over 130 million people worldwide and is a leading cause of blindness among working-age adults. Further clinical trials are needed to validate the results in human subjects.