Study: AI successfully identifies eye diseases from retinal images

A team of international researchers developed a deep-learning system to screen patients for diabetic retinopathy and related eye diseases, according to a study published in JAMA Dec. 12.

The researchers trained a deep-learning system to identify diabetic retinopathy and possible glaucoma based on thousands of patients' retinal images.

The system, which was evaluated using 494,661 retinal images, demonstrated 90.5 percent sensitivity and 91.6 percent specificity for referable diabetic retinopathy. For vision-threatening diabetic retinopathy, the deep-learning system's sensitivity was 100 percent and its specificity was 91.1 percent.

The system exhibited a sensitivity of 96.4 percent and specificity of 87.2 percent with regard to possible glaucoma diagnoses.

"In this evaluation of retinal images from multiethnic cohorts of patients with diabetes, the [deep-learning system] had high sensitivity and specificity for identifying diabetic retinopathy and related eye diseases," the study authors concluded.

More articles on artificial intelligence:
Microsoft previews AI-powered chatbot to select healthcare orgs
Austrian researchers say digital scan of patient's eye may accurately depict general health
AI will force up to 800M people to find new jobs by 2030: 5 things to know

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months