Telemedical Diagnosis of Stage 4 and Stage 5 Retinopathy of Prematurity

Abstract

Purpose : To determine the accuracy of image-based diagnosis for stage 4 or worse retinopathy of prematurity (ROP) disease.

Methods : This is a prospective cohort study. 1220 eye examinations from 230 infants were prospectively obtained from 8 major ROP centers. An ophthalmologist at each center provided a clinical diagnosis using indirect ophthalmoscopy. Wide-angle retinal images (RetCam; Clarity Medical Systems, Pleasanton, CA) were then obtained, which were independently read by two ROP experts using a web-based system for an image-based diagnosis. Sensitivity and specificity of image-based diagnosis from the ROP experts were calculated using the clinical diagnosis as the reference standard.

Results : 28/1220 (2%) eye examinations had a clinical diagnosis of stage 4 or worse disease. Among these 28 examinations, expert 1 had a 21/28 (75%) agreement with the clinical diagnosis, and for the 7 discrepancies, the expert provided a diagnosis of stage 0 in 5 eye examinations and stage 3 in 2 eye examinations. Among the 28 eye examinations with a clinical diagnosis of stage 4 or worse disease, expert 2 had agreement in 24/28 (85%) eye examinations, and for the 4 discrepancies, the expert provided a diagnosis of stage 0 in 4 eye examinations. Using the clinical diagnosis as the reference standard, for the image based detection of stage 4 or worse disease, experts 1 and 2 had a sensitivity of 75% and 86%, respectively. For the image based detection of stage 5 disease, both experts 1 and 2 had a sensitivity of 69%.

Conclusions : There are inconsistencies in the accuracy of image-based diagnosis of stage 4 and stage 5 ROP when compared to the clinical diagnosis.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.