Prospective clinical study validates automated AMD screening

A peer-reviewed study recruited 845 adults across six New York clinics. Using non-dilated colour fundus photographs, the system achieved author-reported subject-level AUC 0.92, sensitivity 90.27% and specificity 83.36% for more-than-early age-related macular degeneration against ophthalmologist grading.

Why it matters: Prospective, multi-site clinical validation is more decision-relevant than retrospective benchmark performance.

Impact: The result supports further evaluation for primary-care screening, but does not itself establish regulatory approval, cost effectiveness or performance outside the sampled setting. Scientific Reports