Quantitatively comparing weekly changes in retinal vascular characteristics of eyes eventually treated versus not treated for retinopathy of prematurity

Abstract

Purpose To quantitatively compare retinal vascular characteristics over time in eyes eventually treated versus not treated for retinopathy of prematurity (ROP), using ROPtool analysis of narrow-field retinal images. Methods This longitudinal study used prospectively collected narrow-field retinal images of infants screened for ROP, prior to treatment, if needed. Images were analyzed using a methodology that combines quadrant-level measures from several images of the same eye. For the longitudinal analysis, one examination per postmenstrual age (PMA) was included per eye. We compared the following ROPtool indices and their change per week between eyes eventually treated versus not treated for ROP: tortuosity index (TI), dilation index (DI), sum of adjusted indices (SAI), and tortuosity-weighted plus (TWP). Analysis was performed on three levels: eye (mean value/eye), quadrant (highest quadrant value/eye), and blood vessel (highest blood vessel value/eye). Results Of 832 examinations (99 infants), 745 images (89.5%) had 3-4 quadrants analyzable by ROPtool. On the eye level, ROPtool indices differed between eyes eventually treated versus not treated at PMA of 33-35 and 37 weeks for TI, SAI, and TWP, and at PMA of 33-34 and 37 weeks for DI (P ≤ 0.0014), and change per week differed between eyes eventually treated versus not treated only for SAI at PMA of 32 weeks (P < 0.001). Conclusions Quantitative analysis of retinal vascular characteristics using ROPtool can help predict eventual need for treatment for ROP as early as 32 weeks PMA. ROPtool index values were more useful than change in these indices to predict eyes that would eventually need treatment for ROP.