Dynamic Range and Measurement Floor for Optical Coherence Tomography Angiography in Glaucoma

Abstract

Purpose : To determine the dynamic range and measurement floor for vessel area density (VAD) from an optical microangiography (OMAG) based optical coherence tomography angiography (OCTA) device and compare to that of the corresponding OCT circumpapillary retinal nerve fiber layer (cpRNFL) measurements in a population of subjects with and without glaucoma. The goal of this study is to evaluate whether OCTA has the potential to improve how moderate-to-severe primary open-angle glaucoma (POAG) is clinically monitored.

Methods : 6x6mm RNFL thickness scans and 6x6mm OCTA scans (Cirrus 5000 with Angioplex, Zeiss) centered on the optic nerve and Humphrey 24-2 visual field testing were obtained on patients from an academic clinic population and from the African American Eye Disease Study (AFEDS), a population-based study based in Inglewood, CA. Informed consent was obtained for each subject according to an IRB-approved protocol. Exclusion criteria were based upon Moghimi et al. (2019) (Ophthalmology. 2019;126:980–988) as follows: visual field results with ≥33% fixation losses and false negatives and ≥15% false positives; subjects with dementia and history of stroke; subjects with comorbid narrow angles and history of other retinopathies and ocular trauma; and cataract or glaucoma procedure <3 months prior to study date. ONH images were graded according to a custom image quality grading protocol. VAD versus visual field mean deviation (MD) and cpRNFL (µm) versus MD scatterplots with LOWESS curves and splines were generated using R. Change-point analyses were performed using the R segmented package.

Results : Based on our inclusion/exclusion criteria, 120 POAG subjects were selected and age-gender-matched with 110 glaucoma suspect and healthy eyes to compare the VAD and RNFL dynamic ranges. AIC comparison between the LOWESS curves and splines indicated that a linear relationship adequately models the dynamic range. Change-point analysis indicated break-points at -22.549 (SEM = 2.067) and -12.087 (SEM = 4.445) for OCTA and SD-OCT, respectively. There are 8 and 5 steps to the floor of the dynamic range for OCTA and SD-OCT, respectively.

Conclusions : OCTA was shown to have a significantly lower measurement floor than SD-OCT with a greater number of steps to this floor. These findings suggest that OCTA may improve our ability to monitor progression in patients with moderate-to-severe POAG.

This is a 2020 ARVO Annual Meeting abstract.