An AI-assisted model based on 71 different blood proteins could help doctors better predict retinal degeneration in diabetic patients before symptoms occur, according to a study published June 2nd in the open access journal PLOS Medicine by Huangdong Li from the Guangdong Provincial Clinical Research Center for Ocular Diseases in Guangzhou, China, and colleagues.
More than half a billion people around the world are now affected by diabetes. People with the disease are at risk of different neurodegenerative conditions, including the breakdown of the retina, the part of the eye that detects light, in a condition called diabetic retinal neurodegeneration (DRN). It can cause severe visual impairment and vision loss, and scientists believe that DRN is a "window" into the diabetic degeneration of other parts of the nervous system, including cognitive impairment and dementia, as well as degradation of nerves in peripheral areas like the fingers and toes.
Unfortunately, DRN is only detected after symptoms appear, when damage is already irreversible. To better predict who might suffer from DRN and when, the researchers sampled the blood plasma from 1,492 patients in the Guangzhou Diabetic Eye Study with type 2 diabetes who did not yet have DRN, and examined the eyes of 1,218 of them through scans over a six-year period. They compared their results with another 502 people with diabetes in the United Kingdom BioBank.
The researchers identified 71 different plasma proteins associated with DRN. The proteins were part of cell pathways for processes like inflammation and cellular maintenance. Using machine learning, the scientists used the protein levels in plasma to develop a predictive model called Pro-DRN which was able to improve on the best-performing model by 26 percent. The scientists have already put the model online to allow doctors to assess the risk. While Pro-DRN is based on plasma protein levels and relies on associations between protein levels and DRN and not direct causes, the authors hope that it could help doctors predict and potentially prevent neurodegeneration, using a simple blood test analyzed by AI.
The authors add, "Our study suggests that early retinal nerve damage in diabetes leaves measurable signals in the blood. By combining plasma proteomics, longitudinal retinal imaging, and explainable AI, Pro-DRN may help move diabetic eye care from detecting established damage toward earlier, molecularly informed risk stratification, so that closer monitoring and future neuroprotective interventions can be directed to the people most likely to benefit."
In your coverage, please use this URL to provide access to the freely available paper in PLOS Medicine: https://plos.io/4nczi6C
Citation: Li H, Zhu Z, Yang S, Cheng W, Tan S, Xin Z, et al. (2026) Proteomic signatures of early retinal neurodegeneration in type 2 diabetes mellitus. PLoS Med 23(6): e1004868. https://doi.org/10.1371/journal.pmed.1004868
Author countries: China, United States of America, Australia
Funding: This work was supported by the GBRCE for Major Blinding Eye Diseases Prevention and Treatment, the Hainan Province Clinical Medical Center, the Science and Technology Projects in Guangzhou (2025A04J7150, W.W.), the National Natural Science Foundation of China (82371086 [W.W.], 82301253 [W.C.], 82571271 [S.C.]), the Natural Science Foundation of Guangdong Province (2026A1515010675, S.C.), the Guangdong Basic and Applied Basic Research Foundation (2022A151511, W.C.), the Projects of Research Center for Sharp Vision at The Hong Kong Polytechnic University (P0057931, S.T.), the Health and Medical Research Fund-Research Fellowship Scheme, Health Bureau, Hong Kong (07210207, S.T.), and the Lumitin Vision to Brightness Research Funding for the Young and middle-aged Ophthalmologists (BCF-KH-YK-20230803-03, S.C.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
