The Ability of SD-OCT to Differentiate Early Glaucoma With High Myopia From Highly Myopic Controls and Nonhighly Myopic Controls.

Abstract

Optical coherence tomography (OCT) instruments do not embed a normative database from highly myopic normal (HMN) eyes. The abilities of three OCT instruments to detect early glaucoma with high myopia were compared using the two controls with or without high myopia. A total of 52 early glaucomatous eyes (mean deviation > -6.0 dB) with high myopia (spherical equivalent ≤ -6.0 diopters [HMG]), 54 HMN eyes, and 90 nonhighly myopic normal (NHMN) eyes were enrolled. Each participant was imaged using Cirrus, RTVue, and Topcon 3D OCT to evaluate the thicknesses of the circumpapillary retinal nerve fiber layer (cpRNFL), the macular retinal nerve fiber layer (mRNFL), ganglion cell layer + inner plexiform layer (GCL/IPL), and mRNFL + GCL/IPL (GCC). The covariate-adjusted areas under the receiver operating characteristic curves (AUCs) for detecting HMG were compared among the instruments and between the two normal groups (HMN or NHMN). Highly myopic normal eyes showed higher AUCs for the temporal quadrant cpRNFL thickness but lower AUCs for the superior and inferior RNFL thicknesses compared with NHMN. We found the AUCs for the GCC thickness showed no significant difference between the two control groups, but the GCL/IPL and mRNFL thicknesses had differences. The abilities of the three OCT instruments to detect early glaucomatous eyes with high myopia were different if the normal eyes were associated with high myopia or not. A normative database that includes data from patients with high myopia should be established for accurate diagnosis of glaucoma with high myopia. (www.umin.ac.jp/ctr number, UMIN000006900.).