Comparison of Early Diabetic Retinopathy Staging in Asymptomatic Patients Between Autonomous AI-Based Screening and Human-Graded Ultra-Widefield Colour Fundus Images

AI Generated Summary*

Diabetic retinopathy (DR) staging from an autonomous artificial intelligence system (IDx-DR) was compared with specialists' readings of ultra-widefield colour images in a pilot study of 54 asymptomatic patients with diabetes (107 eyes) and no known DR. Graders assessed a projected ETDRS 7-field area and the whole gradable retina. The AI called 16 patients no or mild, 28 moderate and 10 vision-threatening; 7-field grading gave 23 no DR, 11 mild, 19 moderate, 1 severe. Patient-level agreement reached 66.6 percent (kappa 0.4) with 7-field and 66.7 percent (kappa 0.38) with full-field, sensitivity 100 and 95 percent but specificity only 47 percent. Peripheral lesions worsened manual grades in 11.1 percent of eyes. Accuracy was deemed adequate for screening in this small tertiary-centre sample.

*This summary was generated by AI and is published unedited. Oku does not alter these summaries. It may contain errors or omissions and is provided for general informational purposes only. Oku does not guarantee its accuracy, completeness, or reliability. For authoritative information, please refer to the original, peer-reviewed article.