Diagnostic Accuracy of Community-Based Diabetic Retinopathy Screening With an Offline Artificial Intelligence System on a Smartphone

These pilot study results show promise in the use of an offline AI system in community screening for referable diabetic retinopathy with a smartphone-based fundus camera. The use of AI would enable screening for referable diabetic retinopathy in remote areas where services of an ophthalmologist are unavailable. This study was done on patients with diabetes who were visiting a dispensary that provides curative services to the population at the primary level. A study with a larger sample size may be needed to extend the results to general population screening, however.