AI 94% accurate in detecting diabetic retinopathy

Some 45 percent of those with diabetes experience retinopathy, which can cause blindness if not detected early. Researchers from the Byers Eye Institute at Stanford University have published a study in Ophthalmology detailing how artificial intelligence (AI) and deep learning helped create an algorithm to detect diabetic retinopathy (DR).

Early detection of DR is crucial, but it often goes unnoticed in routine eye exams. In response, the researchers developed a deep leaning algorithm that was capable of detecting DR with 94 percent accuracy.

“What we showed is that an artificial intelligence-based grading algorithm can be used to identify, with high reliability, which patients should be referred to an ophthalmologist for further evaluation and treatment,” said Theodore Leng, MD, an assistant professor at Standford and the study's lead author. “If properly implemented on a worldwide basis, this algorithm has the potential to reduce the workload on doctors and increase the efficiency of limited healthcare resources. We hope that this technology will have the greatest impact in parts of the world where ophthalmologists are in short supply." 

Using 75,000 images of patients with varying severity of DR, researchers were able to teach the computer to identify DR patients and differentiate them from healthy patients. With 98 percent specificity, the program was able to identify different stages of DR progression.

“A fully data-driven artificial intelligence–based grading algorithm can be used to screen fundus photographs obtained from diabetic patients and to identify, with high reliability, which cases should be referred to an ophthalmologist for further evaluation and treatment,” wrote Leng and colleagues. “The implementation of such an algorithm on a global basis could reduce drastically the rate of vision loss attributed to DR.”