Quantification of fibrovascular ridge thickness in retinopathy of prematurity using swept-source optical coherence tomography

Abstract

Purpose : Retinopathy of prematurity (ROP) stage is classified based on the appearance of the ridge on ophthalmoscopy or imaging. Using ultra-widefield (UWF) optical coherence tomography (OCT) we hypothesized that we could measure the axial thickness of the peripheral ridge, and that there is a positive correlation between stage category and ridge thickness.

Methods : We captured 26 unique ridge portions (showing at least 3 continuous clock hours of disease) in 11 patients with stage 1 to 3 ROP using portable, swept-source OCT with up to a 140° field of view (FOV). Area from the retinal pigment epithelium to the inner ridge surface was manually segmented (Figure 1), and axial thickness was calculated pixel-wise (1 pixel=4.4 μm) along the transverse dimension. A masked grader provided diagnosis of max stage using International Classification of ROP (ICROP) guidelines, as well as a second label representing a continuous spectrum of stage (e.g. 1.5). We compared ridge thickness (mean and standard deviation [SD]) to stage categories (analysis of variance [ANOVA]) and the stage spectrum (linear regression).

Results : The mean (+SD) ridge thickness increased with stage classification, with thickness of 252 (+49), 340 (+77), and 380 (+95) μm for stages 1, 2, and 3, respectively (p=0.04 by ANOVA). There was a significant relationship between the max ridge thickness and the continuous stage labels (Figure 2, R²=0.566; p<0.001).

Conclusions : UWF OCT measurement of retinal thickness at the peripheral ridge correlated with ICROP stage classification and with continuous approximation of stage labels. These results suggest that OCT may provide objective and quantifiable biomarkers for ROP stage in the future, and that ROP severity overall may be better represented as a spectrum, since it varies both axially and circumferentially in eyes with ROP.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Figure 1. Ridgeline thickness measurement. OCT en face (A) with manual tracing of the fibrovascular ridge (B) to obtain axial scan running the length of the ridge (C). Region between the RPE and the inner ridge surface was manually segmented (D) to obtain thickness measurements.

 

Figure 2. Continuous stage versus thickness. On linear regression, we found a significant relationship between continuous stage rating and max ridge thickness (p<0.001), with a linear model explaining over half of the variance (R²=0.566).