New research from the University of Dundee has shown that artificial intelligence (AI) can analyze routine eye-screening photos to detect early signs of kidney disease in people with type 2 diabetes—years before symptoms appear.
This breakthrough, presented at the Diabetes UK Professional Conference 2025, could lead to earlier diagnosis and treatment, potentially preventing severe kidney damage.
The Link Between Diabetes and Kidney Disease
Type 2 diabetes occurs when the body either doesn’t produce enough insulin or doesn’t use it effectively. Over time, high blood sugar levels can damage the body’s organs, leading to serious complications, including heart disease, blindness, and kidney failure.
Kidney disease related to diabetes develops slowly and often goes unnoticed until it reaches an advanced stage.
One in five people with diabetes will eventually need treatment for kidney disease, and diabetes is a leading cause of end-stage kidney failure, requiring dialysis or a kidney transplant. The challenge has been identifying those at risk early enough to slow or prevent kidney damage.
How AI Uses Eye Photos to Detect Kidney Disease
In the UK, everyone over the age of 12 with diabetes is invited to regular eye screenings. These screenings involve taking photographs of the retina, the light-sensitive layer at the back of the eye, to check for signs of diabetes-related eye disease.
Researchers at the Universities of Dundee and Glasgow wondered whether these images could also provide insights into kidney health.
Led by Dr. Alexander Doney, the team developed an AI tool that analyzed nearly 1 million eye-screening photos from almost 100,000 people with type 2 diabetes in Scotland. The AI was trained to recognize patterns in the images linked to existing kidney disease by comparing them with medical records. It was then tested using data from another 30,000 people.
The AI tool achieved impressive results:
- It correctly identified people with existing kidney disease 86% of the time.
- In people without kidney disease, it predicted who would develop it within five years with 78% accuracy.
- It outperformed traditional kidney function tests, detecting risks even in individuals who showed no warning signs through standard testing.
What This Means for Future Treatment
The findings suggest that AI could transform the way kidney disease is detected in people with diabetes. By analyzing routine eye-screening photos, doctors may be able to identify patients at risk long before conventional tests reveal any problems. This could allow for earlier lifestyle changes, medications, or other interventions that may prevent or slow the progression of kidney disease.
Dr. Elizabeth Robertson, Director of Research at Diabetes UK, emphasized the significance of this breakthrough. “Kidney damage often progresses silently until it becomes severe, and early detection is critical.
This research has uncovered a new way to assess kidney health—through the eyes. By detecting invisible patterns in eye-screening images, AI could alert doctors to early signs of kidney disease, giving them a chance to intervene before it’s too late.”
Dr. Doney, the study’s lead researcher, explained why this method is so promising. “The retina is the only place in the body where we can directly observe blood vessels, which play a crucial role in kidney health. AI can detect subtle changes in these images that humans can’t see, offering a glimpse into a patient’s overall health before standard tests show any problems.”
The Potential for a New Diagnostic Tool
Beyond kidney disease, researchers believe this AI approach could be expanded to predict other diabetes-related complications, such as heart disease and nerve damage. By harnessing the power of AI, routine diabetic eye screenings could become a multi-purpose tool for detecting and preventing serious health issues early.
As this technology develops, it could help millions of people with diabetes avoid severe kidney disease and improve their long-term health outcomes. The next step is further research and clinical testing to integrate AI-driven diagnostics into everyday medical practice.
Copyright © 2025 Knowridge Science Report. All rights reserved.
