Artificial intelligence, specifically deep learning (DL), has garnered substantial attention in the medical world in recent years because it has been used to accurately classify medical images, often at a human expert level, across different medical disciples. Within ophthalmology, the prospect of applying DL in screening for diabetic retinopathy (DR) is particularly exciting, because DR is the most common cause of visual acuity loss among the working-age population in the United States1 and many countries around the world, and early detection of DR is cost-effective.2 While numerous studies have demonstrated the ability of DL to detect DR in color fundus photographs with high sensitivity and specificity, most of these studies were performed using retrospectively collected data. Also, there is evidence to suggest that artificial intelligence–based diagnostic tools, when deployed in a clinical practice setting, can have greatly diminished performance because of variation in disease prevalence and image quality.3
Liu TYA. Smartphone-Based, Artificial Intelligence–Enabled Diabetic Retinopathy Screening. JAMA Ophthalmol. Published online August 08, 2019. doi:10.1001/jamaophthalmol.2019.2883
Browse and subscribe to JAMA Network podcasts!
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: