Artificial intelligence in early detection of diabetic retinopathy: a study conducted in Uzbekistan
Аннотация
Diabetic retinopathy (DR) is one of the leading causes of preventable blindness in Uzbekistan. Early detection is critical, yet access to ophthalmologists is limited, especially in rural regions. In this pilot study, we evaluated an AI-assisted diagnostic system for DR using 1,200 retinal images collected from patients in Tashkent and Andijan. The AI model demonstrated 94.2% sensitivity and 91.7% specificity for referable DR, showing strong concordance with ophthalmologists (kappa = 0.90). Patient feedback highlighted high acceptance, with 92% reporting comfort and 88% trusting the AI results. Cost-effectiveness analysis indicated AIassisted screening could reduce unnecessary clinic visits by 35% and cut national screening costs by 28%. These results indicate that AI-assisted DR screening is accurate, practical, and highly suitable for Uzbekistan, bridging urban-rural healthcare gaps and enabling timely intervention.