Assessing the Value of Artificial Intelligence in Retinopathy of Prematurity Screening—Be Specific With Regard to the Data

Morrison et al describe an intriguing application of artificial intelligence (AI) to screen for retinopathy of prematurity (ROP), an area of substantial lifetime comorbidity. The authors note that ROP causes blindness in up to 30000 patients diagnosed with the condition per year worldwide, including in economically low-resource communities in the US. ROP screening is stressful, even for ophthalmologists, owing to the difficulty in performing a complete retinal examination on such young, often uncooperative patients; AI could be a force multiplier, leveraging scarce talent by improving the screening process. If ROP could be screened more effectively, vision could be preserved in many premature babies annually, leading to large health benefits. This is an important topic considering the large unmet need, and the authors do a commendable job in addressing it.