The Ability of Macular Parameters and Circumpapillary Retinal Nerve Fiber Layer by Three SD-OCT Instruments to Diagnose Highly Myopic Glaucoma

Abstract

Purpose.: To compare the ability of circumpapillary retinal nerve fiber layer (cpRNFL) thickness and macular parameters obtained by three spectral-domain optical coherence tomography (SD-OCT) instruments to detect highly myopic glaucoma.

Methods.: In this study, 84 glaucomatous eyes, 53 normal eyes with high myopia, and 86 normal eyes (not highly myopic) were enrolled. Each participant was imaged using Cirrus, RTVue, and 3D OCT to evaluate the average and quadrant cpRNFL thicknesses. The macular retinal nerve fiber layer (mRNFL), ganglion cell layer + inner plexiform layer (GCL/IPL), and mRNFL + GCL/IPL (GCC) thicknesses were analyzed. The areas under the receiver operating characteristic curves (AUCs) were compared between the instruments. In addition, the best parameters for the AUC were compared between the cpRNFL parameters and macular parameters in each instrument.

Results.: These analyses revealed similar AUCs for the average cpRNFL and GCC thicknesses between the instruments. RTVue displayed a significantly higher AUC in the nasal cpRNFL thickness than both Cirrus (P = 0.0004) and 3D OCT (P = 0.0006). 3D OCT showed higher AUCs than Cirrus in the mRNFL parameters. There were no significant differences between the best cpRNFL parameters and macular parameters in each instrument.

Conclusions.: The average cpRNFL and GCC thicknesses that were measured using these OCT instruments exhibited similar abilities for diagnosing highly myopic glaucoma, and RTVue exhibited better diagnostic abilities than Cirrus and 3D OCT for nasal cpRNFL. 3D OCT had better ability than Cirrus in the mRNFL. Both cpRNFL and GCC in each instrument were comparable, and their measurements offered good ability for diagnosing glaucoma with high myopia. ( http://www.umin.ac.jp/ctr/index/htm9, UMIN000006900.)

Introduction

Glaucoma is a progressive, irreversible optic neuropathy causing a gradual loss of ganglion cell axons. Early detection is therefore essential to begin administering pressure-reducing treatment to stop or delay progressive loss of visual function. ¹ In clinical practice, glaucoma diagnosis is performed using ophthalmoscopic examination of the optic nerve head, examination of the retinal nerve fiber layer (RNFL), and visual field testing with standard automatic perimetry. The evaluation of the structural changes using spectral-domain optical coherence tomography (SD-OCT) plays an important role in the diagnosis and treatment of glaucomatous patients. Because glaucoma primarily affects retinal ganglion cells and their axons, up to this point, most OCT studies have used circumpapillary retinal nerve fiber layer (cpRNFL) thickness measurements to detect glaucoma. Furthermore, some studies have already shown that the macular ganglion cell complex (GCC) has a good glaucoma-discriminating power that is comparable with that of the RNFL. ^2–6 Recent advances in OCT technology have allowed for an automatic segmentation between the RNFL and the ganglion cell layer (GCL) at the macula. ⁷  

SD-OCT devices are commercially available from several different companies. The speed and resolution of image acquisition varies between instruments, despite similar working principles. Two studies revealed that the cpRNFL measurements from healthy controls using various devices were different and not comparable. ^8,9 Additionally, Leite et al. and one of our previous studies made interinstrumental comparisons between SD-OCT devices to diagnose glaucomatous optic neuropathy. However, both studies excluded highly myopic eyes. ^10,11  

Myopia is reported as a common ocular abnormality worldwide, and it is an independent risk factor for glaucoma. ¹² Although diagnosing glaucoma in myopic subjects is important in clinical practice, structural changes related to myopia—such as tilting, optic disc deformation, shallow cup, and large peripapillary crescent—disturb the precise diagnosis of glaucoma. ^13,14 The clinical diagnosis of glaucoma in such patients, especially in these patients without advanced visual field loss, is challenging because the optic disc appearance is often unusual and difficult to assess. As the normative database of SD-OCT instruments largely comprises data collected from normal eyes with no or low myopia, interpreting the RNFL thickness (RNFLT) deviation map in eyes with high myopia can be problematic. Several studies have demonstrated that myopic eyes have thinner RNFL measurements and a unique pattern of RNFL distribution, leading to inaccurate diagnosis by OCT. ^15–20 However, a few studies have demonstrated the diagnostic abilities of OCT measurements for glaucoma in highly myopic eyes. ^21–24 One group reported that the GCC thickness measured with RTVue was superior to the cpRNFL thickness for detecting glaucoma in highly myopic patients, ^22,23 while another group reported no significant difference in the detection ability between cpRNFL and GCC thickness. ^21,24 There are no reports that have compared the cpRNFL thickness and macular parameters at the same time among several SD-OCT instruments in a single population consisting of highly myopic patients. 

This study assessed the diagnostic ability of the cpRNFL thickness and the macular parameters evaluated by Cirrus, RTVue, and 3D OCT to detect highly myopic glaucomatous eyes. 

Materials and Methods

The subjects were recruited at the Kobe University Hospital (Kobe, Japan) for this observational, cross-sectional study. The institutional review board of Kobe University approved the study protocol, which adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from each subject after an explanation of the study protocol was given. 

All subjects received a full ocular examination. A visual field analyzer (Humphrey Field Analyzer [HFA] 30-2 SITA standard program; Carl Zeiss Meditec, Inc., Dublin, CA) performed the visual field (VF) test. Subjects with a best-corrected visual acuity of 20/40 or better, a cylinder correction within ± 3.0 diopters (D) and gonioscopically open angles were included. The subjects were divided by the spherical equivalents into a highly myopic group (< −6.00 D) and a nonhighly myopic group (≥ −6.0 D). Axial length was acquired using a commercial biometer (IOLMaster; Carl Zeiss Meditec, Inc.). No subject had undergone any ocular surgeries. VF tests, cpRNFL thickness, and macular parameters were measured using three SD-OCT instruments within 3 months. A total of 84 highly myopic glaucomatous (HMG) eyes from 62 subjects and 53 highly myopic normal (NMN) eyes from 40 subjects were enrolled. Additionally, we included the measured thicknesses of 86 nonhighly myopic normal (NHMN) eyes from 66 subjects. 

Glaucomatous optic neuropathy (GON) was defined as a vertical cup-disc asymmetry between fellow eyes of 0.2 or more with neuroretinal rim damage, such as excavation, rim thinning, and notches with or without parapapillary hemorrhages, or RNFL defects with a reproducible VF defect. The glaucomatous VF defect was based on the following criteria: two or more contiguous points with a pattern deviation sensitivity loss of P < 0.01, three or more contiguous points with sensitivity loss of P < 0.05 in the superior or inferior arcuate areas, or a 10-dB difference across the nasal horizontal midline at two or more adjacent locations and an abnormal result on the glaucoma hemifield test. ²⁵ For controls, healthy subjects at least 20 years of age were also invited to participate in the study. The inclusion criteria for normal eyes were as follows: intraocular pressure < 21 mm Hg, reliable HFA results (fixation loss, false positive or false negative >33%), abnormal findings in HFA suggestive of glaucoma (as mentioned above), no retinal diseases, and no GON. Two independent masked glaucomatous specialists (AA and AK) defined normal appearance of the optic disc. In the case of a disagreement, a third specialist (MN) reviewed the stereophotographs and made a determination. 

cpRNFL Measurements

The optic disc cube protocol was adopted for Cirrus HD-OCT (software version 6.1.0.96; Carl Zeiss Meditec, Inc.). This protocol is based on a 3-dimensional scan of a 6- × 6-mm² area that is centered on the optic disc. A 3.46-mm diameter circular scan was performed automatically around the optic disc, which provided measurements of parapapillary RNFL thickness. Images with signal strength <6 were excluded. RNFL thickness values at 256 measurement points on the circular scan were exported and evaluated, as described below. 

The optic nerve head map protocol was applied for RTVue-100 (software version 4.0.5.39; Optovue, Inc., Fremont, CA). This protocol generated an RNFL thickness map that was measured along a circle of 3.45 mm in diameter and centered on the optic disc. 3D disc protocol was used to register the edge of the optic nerve head. Only good quality images, as defined by a signal strength index of >30, were included. RNFL thickness parameters calculated by the built-in software were utilized. 

3D OCT 2000 (software version 8.00; Topcon, Inc., Tokyo, Japan) used a 7 × 7-mm scan disc protocol. The magnification effect in each eye was corrected according to the formula (modified Littman's method) provided by the manufacturer, which was based on the refraction, corneal radius, and axial length, to obtain more accurate circle sizes during measurements. Images with a quality factor >60 were included in analyses. RNFL thickness data with 1024 points of resolution on a 3.46-mm circle diameter were exported by the software provided by Topcon, Inc., and evaluated as described below. 

The RNFL thicknesses at 256 points from Cirrus and 1024 points from 3D OCT were converted to the following parameters. The mean 360° RNFL thickness was defined as the average of RNFL thicknesses. Starting from a point on the temporal margin, which was designated as 0°, the mean quadrant RNFL thickness between 315° and 45° was defined as the temporal RNFLT: between 45° and 135° was superior, between 135° and 225° was nasal, and between 225° and 315° was inferior. 

Macular Inner Retinal Layer Thickness Measurements

The ganglion cell analysis (GCA) was used to process the data produced by Cirrus OCT. GCA measured the thickness of the macular RNFL (mRNFL), ganglion cell layer + inner plexiform layer (GCL/IPL), and mRNFL + GCL/IPL (GCC) within a 14.13-mm² elliptical annulus area that is centered on the fovea. The special software provided from Zeiss was used to export the data. The superior (0–180°) and inferior (180–360°) segments were calculated from the corresponding sectors. 

The GCC was measured when the RTVue-100 OCT was used. The GCC protocol explored parameters within a circle with a 6-mm diameter; the center of the GCC scan was shifted approximately 1 mm temporal to the fovea to improve the sampling of temporal peripheral nerve fibers. The variables generated by the GCC analysis included the average, superior, and inferior hemiretinas. 

Raster scanning of a 7 × 7-mm area that was centered on the fovea with a scan density of 512 (vertical) × 128 (horizontal) scans was performed using 3D OCT. The built-in protocol measured a 6 × 6-mm area that was centered in the fovea using embedded 3D OCT measurement software. The data divided into 10 × 10 grids were exported by the software provided from Topcon, Inc. The average thickness and superior and inferior hemiretina thicknesses of the mRNFL, GCL/IPL, and GCC were calculated. 

Table 1 summarizes the protocol for the measurements in each instrument. 

Table 1

 

Scan parameters for Cirrus, RTVue, and 3D OCT

	Cirrus	RTVue	3D OCT-2000
Scan speed (1 scan/s)	27,000	26,000	50,000
Software version	6.1.0.96	4.0.5.39	8.00
Scan program of cpRNFL thickness	Optic disc cube 200 × 200	3D disc and ONH	3D disc
Obtained data of cpRNFL thickness	256 points	16 sectors	1024 points
Scan program of macular area	Macular cube 200 × 200	GCC	3D macular
Analyzed macular area	4 × 4.8 mm, oval	6 × 6 mm, circle	6 × 6 mm, square
Obtained macular parameters	GCC, mRNFL, GCL/IPL	GCC	GCC, mRNFL, GCL/IPL
Obtained sectors of macular area	6 sectors	Superior/inferior	10 × 10 grids

Statistical Analysis

All numerical data showed normal distributions confirmed by the Kolmogorov-Smirnov test. Bilateral eyes were included in the analyses if they matched the inclusion criteria. Because measurements from both eyes of the same subject were likely to be correlated, the standard statistical method for parameter estimation could lead to underestimation of standard errors (SEs). ²⁶ To compare the means between the groups, a generalized linear model was used to evaluate the mean difference between the groups. This method of compensating for age imbalance has been used in previous ophthalmology studies. ^2,27 The generalized linear model accounted for the intereye correlation in the study sample. 

Receiver operating characteristic (ROC) curves were constructed for the cpRNFL, GCC, GCL/IPL, and mRNFL thickness to investigate the ability of the devices to differentiate glaucomatous eyes with high myopia from highly myopic normal eyes. To account for the potential correlation between the eyes, the cluster of the data for the subject was considered as the unit of resampling when calculating SEs. This procedure has been used in the literature to adjust for the presence of multiple correlated measurements from the same unit. ²⁸ ROC curves were adjusted for differences in age and gender using covariate-adjusted ROC curves, as established by Pepe. ²⁹ A bootstrap resampling procedure was used (n = 1000 resamples). The area under the ROC curve (AUC) was calculated for each parameter. A pairwise comparison of the AUCs was performed using a method proposed by Dodd and Pepe. ³⁰ When the estimated correlation between the two instruments set at 0.7, a minimum of 62 cases was required in positive and negative groups to detect a 0.1 difference in the AUCs at values of more than 0.85. In the same conditions, a minimum of 41 cases was required to detect a 0.1 difference in the AUCs at values of more than 0.95. These findings were achieved at a statistical power of 80% and a type I error of 5%. 

The sensitivity of the detection of highly myopic glaucomatous eyes was determined using the average cpRNFL thickness and the average GCC, as measured by each instrument with a target specificity ≥95%. 

Statistical analyses were performed using computer programs (Stata, version 12.0; StataCorp, College Station, TX; and SPSS, version 20.0, Japan IBM, Tokyo, Japan). P values of less than 0.05 were considered statistically significant. 

Results

Table 2 presents the demographics and ocular characteristics of the analyzed subjects. Refraction and axial length revealed no significant differences between the normal and glaucomatous eyes with high myopia. However, there were significant differences in age and sex, so the AUCs were adjusted by these covariates. 

Table 2

 

Characteristics of the Studied Eyes (mean ± SD)

	HMN, n = 53	NHMN, n = 86	HMG, n = 84	HMN vs. HMG	NHMN vs. HMG
Age, y	38.8 ± 12.3	45.6 ± 12.6	45.4 ± 10.8	0.0013*	0.16*
Range	24–61	24–68	27–64
Sex, % (female)	77.4	66.3	50.6	0.0059†	0.054†
Refraction (spherical equivalent), D	−7.76 ± 1.87	−2.23 ± 2.1	−7.87 ± 1.34	0.612*	<0.0001
Range	−6.25 to −13.0	2.25 to 6	−6.25 to −13.0
Axial length, mm	26.8 ± 0.89	24.8 ± 1.1	26.8 ± 0.87	0.184*	<0.0001
Range	25.8–28.8	22.39–25.5	25.12–29.01
Mean deviation, dB	−0.70 ± 1.23	0.12 ± 1.49	−7.36 ± 6.52	<0.0001	<0.0001
Range	0.64 to −3.11	2.66 to −4.37	−0.11 to −27.96	<0.0001	<0.0001

*  Unpaired t-test.

†  Fisher's exact test.

Table 3 summarizes the distribution of the cpRNFL thickness in the three groups. The cpRNFL thickness in highly myopic glaucomatous eyes was significantly thinner than those in normal myopic eyes except for the nasal quadrant. In the nasal quadrant, only RTVue was found to be thinner in highly myopic glaucomatous eyes than in highly myopic normal eyes, but Cirrus and 3D OCT could not detect this. 

Table 3

 

cpRNFL Thickness Measured by SD-OCT Instruments (mean ± SD)

cpRNFL Thickness, μm	HMN, n = 53	NHMN, n = 86	HMG, n = 84	P Value
				HMN vs. HMG	NHMN vs. HMG
Cirrus
Average	90.3 ± 7.9	93.4 ± 6.8	67.6 ± 10.6	<0.0001	<0.0001
Superior	107.6 ± 15.4	115.5 ± 12.7	78.2 ± 16.5	<0.0001	<0.0001
Nasal	64.1 ± 9.6	67 ± 9.2	63.0 ± 10.5	0.601	0.012
Inferior	104.9 ± 16.2	117.3 ± 13.5	68.4 ± 15.7	<0.0001	<0.0001
Temporal	94.2 ± 19	81.8 ± 12.1	60.8 ± 15.6	<0.0001	<0.0001
RTVue
Average	96.7 ± 7.9	102.3 ± 7.7	74.5 ± 10.1	<0.0001	<0.0001
Superior	107.6 ± 15.4	124.1 ± 14	88.8 ± 15.0	<0.0001	<0.0001
Nasal	64.1 ± 9.6	70.2 ± 10.1	57.7 ± 8.3	0.0002	<0.0001
Inferior	104.9 ± 16.2	132.1 ± 13.5	84.4 ± 16.2	<0.0001	<0.0001
Temporal	94.2 ± 19	83 ± 10.2	64.6 ± 20.1	<0.0001	<0.0001
3D OCT
Average	102.6 ± 10	104.3 ± 6.3	74.9 ± 13.6	<0.0001	<0.0001
Superior	122.3 ± 17.2	128.3 ± 12.7	91.1 ± 21.4	<0.0001	<0.0001
Nasal	64.8 ± 15.8	76.7 ± 13.9	64.1 ± 13.0	0.673	<0.0001
Inferior	128.5 ± 16.6	131.6 ± 11.2	79.0 ± 22.0	<0.0001	<0.0001
Temporal	94.2 ± 19.1	81.8 ± 12.1	65.8 ± 15.7	<0.0001	<0.0001

Table 4 summarizes the mean thicknesses of the macular parameters in the three groups. All of the parameters in the highly myopic glaucomatous eyes were significantly thinner than those in highly myopic normal eyes and in nonhighly myopic normal eyes. 

Table 4

 

Inner Retinal Layer Parameters by SD-OCT Instruments (mean ± SD)

Macular Thickness, μm	HMN, n = 53	NHMN, n = 86	HMG, n = 84	P Value
				HMN vs. HMG	NHMN vs. HMG
Cirrus
GCC
Average	111.8 ± 10.8	115.5 ± 8.5	89.0 ± 14.2	<0.0001	<0.0001
Superior	112.5 ± 10.8	121.4 ± 12.6	93.8 ± 15.0	<0.0001	<0.0001
Inferior	111.4 ± 10.7	116.3 ± 7.3	84.3 ± 15.7	<0.0001	<0.0001
GCL/IPL
Average	76.5 ± 6.5	81.7 ± 5.7	62.5 ± 8.5	<0.0001	<0.0001
Superior	76.8 ± 7	81.9 ± 6.6	65.0 ± 8.3	<0.0001	<0.0001
Inferior	75.6 ± 6.8	81.4 ± 5.9	59.9 ± 8.3	<0.0001	<0.0001
mRNFL
Average	35.4 ± 4.6	33 ± 3.5	26.7 ± 6.8	<0.0001	<0.0001
Superior	35.2 ± 4.5	33.3 ± 3.7	28.8 ± 7.3	<0.0001	<0.0001
Inferior	35.8 ± 4.7	34.8 ± 3.1	24.4 ± 8.0	<0.0001	<0.0001
RTVue
GCC
Average	90 ± 4.8	93.9 ± 6.9	71.5 ± 10.0	<0.0001	<0.0001
Superior	90.8 ± 5.7	93.6 ± 6.9	75.1 ± 10.8	<0.0001	<0.0001
Inferior	89.1 ± 6.1	93 ± 7	67.6 ± 11.1	<0.0001	<0.0001
3D OCT
GCC
Average	104.3 ± 6	105.6 ± 6.9	80.2 ± 12.0	<0.0001	<0.0001
Superior	104.0 ± 6	103.2 ± 6	85.0 ± 13.2	<0.0001	<0.0001
Inferior	104.2 ± 6.8	105.1 ± 14.1	75.1 ± 13.0	<0.0001	<0.0001
GCL/IPL
Average	69.1 ± 4	69.4 ± 9.3	57.4 ± 7.3	<0.0001	<0.0001
Superior	69.8 ± 4.1	70.3 ± 9.4	59.8 ± 7.5	<0.0001	<0.0001
Inferior	67.9 ± 4.2	68.6 ± 9.4	54.9 ± 7.5	<0.0001	<0.0001
mRNFL
Average	35.2 ± 3.2	35.3 ± 3.7	22.7 ± 7.8	<0.0001	<0.0001
Superior	34.2 ± 3	33.7 ± 3.9	25.2 ± 8.4	<0.0001	<0.0001
Inferior	36.3 ± 4	36.9 ± 4.5	20.2 ± 8.5	<0.0001	<0.0001

 NA, not applicable.

Table 5 demonstrates the age- and sex-adjusted AUCs of the parameters for each instrument to distinguish highly myopic glaucomatous eyes from normal eyes with high myopia. The average cpRNFL thickness had the highest AUC among all of the parameters in Cirrus (0.969) and RTVue (0.975). 3D OCT had the highest AUC in the average mRNFL thickness (0.974). There were no significant differences among these parameters with the highest AUC in each instrument. The AUCs for the average thickness were 0.969, 0.975, and 0.957 for Cirrus, RTVue, and 3D OCT, respectively. Between the instruments, no significant differences were observed in the cpRNFL thickness of the average and four quadrants except for in the nasal quadrant. RTVue exhibited a significantly higher AUC for the nasal quadrant compared with Cirrus (P = 0.0004) and 3D OCT (P = 0.0006). 

Table 5

 

AUC Analysis Using Cirrus, RTVue, and 3D OCT for HMG Eyes (mean ± SD)

Measured Thickness Parameters	SD-OCT Instruments			P Value
	Cirrus	RTVue	3D OCT	Cirrus vs. RTVue	RTVue vs. 3D OCT	Cirrus vs. 3D OCT
CpRNFL
Average	0.969 (±0.016)	0.975 (±0.132)	0.957 (±0.018)	0.56	0.65	0.45
Quadrant
Superior	0.931 (±0.026)	0.902 (±0.030)	0.904 (±0.030)	0.17	0.414	0.94
Temporal	0.889 (±0.039)	0.902 (±0.034)	0.879 (±0.036)	0.73	0.55	0.72
Inferior	0.944 (±0.021)	0.953 (±0.022)	0.964 (±0.017)	0.62	0.53	0.15
Nasal	0.540 (±0.065)	0.737 (±0.053)	0.599 (±0.060)	0.0004*	0.006*	0.3
GCC
Average	0.945 (±0.028)	0.957 (±0.025)	0.968 (±0.014)	0.71	0.38	0.14
Superior	0.914 (±0.036)	0.936 (±0.031)	0.953 (±0.024)	0.51	0.52	0.11
Inferior	0.935 (±0.027)	0.949 (±0.026)	0.969 (±0.020)	0.66	0.38	0.088
GCL/IPL
Average	0.946 (±0.032)	NA	0.954 (±0.028)			0.07
Superior	0.925 (±0.037)	NA	0.917 (±0.038)			0.076
Inferior	0.942 (±0.040)	NA	0.954 (±0.026)			0.58
mRNFL
Average	0.901 (±0.025)	NA	0.974 (±0.017)			0.021*
Superior	0.837 (±0.034)	NA	0.926 (±0.030)			0.006*
Inferior	0.903 (±0.026)	NA	0.961 (±0.017)			0.034*

Regarding the macular parameters, the AUCs for the average GCC were 0.945, 0.957, and 0.968 for Cirrus, RTVue, and 3D OCT, respectively. There were no significant differences in the average and two-hemifield GCC thicknesses between the instruments. RTVue was not able to discriminate the GCL/IPL and mRNFL. Therefore, the segmented parameters were only compared between Cirrus and 3D OCT. No significant differences in the AUCs of the average and two-hemifield GCL/IPL thicknesses were observed between the instruments. 3D OCT exhibited significantly higher AUCs for the average and two-hemifield mRNFL measurements compared with Cirrus. Figure 1 illustrates the ROC curves for detecting highly myopic glaucomatous eyes from highly myopic normal eyes. 

Figure 1

 

ROC curves of average cpRNFL thickness (A), the average GCC thickness (B), the average GCL/IPL thickness (C), and the average mRNFL thickness (D) measured using Cirrus, RTVue, and 3D OCT for discriminating highly myopic glaucomatous eyes from highly myopic normal eyes.

We also compared the highest AUC in cpRNFL parameters with the highest AUC in macular parameters in each instrument. There were no significant differences in these parameters (Cirrus, P = 0.58; RTVue, P = 0.53; 3D OCT, P = 0.42). For example, for Cirrus, the AUC of average cpRNFL thickness (0.969) did not differ with the AUC of average GCL/IL (0.946), which exhibited the highest AUC among macular parameters (P = 0.58). 

The sensitivities of each parameter were calculated with target specificities ≥95%. We constructed Venn diagrams of the average cpRNFL and GCC thicknesses for each SD-OCT instrument in the glaucoma groups to investigate whether macular parameters could diagnose glaucomatous abnormalities in eyes that were negative based on the cpRNFL thickness (Fig. 2). The agreement between the average cpRNFL thickness and the average GCC thickness in Cirrus, RTVue, and 3D OCT was 65% (n = 55/84); 77% (n = 65/84); and 73% (n = 61/84), respectively. 

Figure 2

 

Venn diagrams illustrate the percentage of eyes that were judged as glaucomatous based on the criteria of the average cpRNFL thickness and the average GCC from Cirrus, RTVue, and 3D OCT at a fixed specificity of 95% for highly myopic glaucoma.

Discussion

In this study, the ability to discriminate between highly myopic glaucoma and normal eyes with high myopia was evaluated in three SD-OCT instruments: Cirrus, RTVue, and 3D OCT. The abilities of these instruments to detect the average cpRNFL thicknesses and macular GCC thicknesses were not significantly different. RTVue exhibited a higher capability than Cirrus and 3D OCT in the measurement of the nasal quadrant of the cpRNFL thickness. These findings in cpRNFL measurements are in line with our previous report for detecting nonhighly myopic glaucoma. ¹¹ In the mRNFL thickness, the AUC in 3D OCT was higher than in Cirrus. The macular GCC showed a comparable AUC with the cpRNFL thickness for detecting highly myopic glaucoma in the three instruments. As far as we know, this report is the first to compare the diagnostic abilities of SD-OCTs in glaucomatous eyes with high myopia. 

It has already been shown that myopic eyes have thinner average RNFL measurements in OCT than emmetropic eyes. ^15–18 In addition, it was reported that myopia affected the distribution pattern of the RNFL thickness around the optic disc; with increasing myopia, the superotemporal and inferotemporal RNFL bundles tend to converge temporally. ^19,20 In this study, the cpRNFL thicknesses in highly myopic normal eyes were thinner than those in nonhighly myopic normal eyes except for the temporal quadrant, although the highly myopic normal group included a younger population on average. It is established that the cpRNFL thickness decreases with aging. Contrarily, the temporal cpRNFL thickness in highly myopic eyes was thicker. These findings agreed with previous studies. ^18,21,24 The internal normative database in each instrument consisted of nonhighly myopic normal eyes. We need to be aware that the measured cpRNFL thickness, except for the nasal quadrant, inevitably overestimates the glaucoma diagnostic ability of highly myopic eyes when the internal database is used. Therefore, we needed the normal database with high myopia when the cpRNFL measurements by OCT were used to diagnose highly myopic glaucoma. To strictly construct normative data in high myopia, the normal eyes in this study were enrolled as the hospital-based control with the exclusion criteria for suspicious eyes with GON. 

As for the effect of axial length or myopia on the thicknesses of macular parameters measured by OCT, the conclusions were still controversial. Some studies found a correlation between total macular thickness and axial length, ^31,32 but others did not. ^33,34 One study evaluating inner retinal layers showed that a negative correlation existed between axial lengths, ³⁵ but another did not. ³⁶ The projection artifact of the scanning area should be considered, especially in highly myopic eyes. Because a larger area was scanned in highly myopic eyes than in normal eyes, the measured thickness may be underestimated in macular measurements. The similar artifact could be occurred in cpRNFL measurements; cpRNFL thickness is measured more distally to the disc in highly myopic eyes. ^37,38 To minimize an effect of ocular magnification on the scanning area by axial length, several methods are available for correcting distances measured in the back of the eye for magnification due to the optical characteristics of the eye. To correct for magnification based on corneal curvature, refractive error, and axial length, the Littmann formula was adapted in 3D OCT measurements but not in the Cirrus and RTVue measurements. ³⁹ Most macular parameters measured by Cirrus and RTVue in highly myopic normal eyes were thinner than in nonhighly myopic normal eyes in this study as shown in Table 3. However, the effects of axial length and refraction on macula parameters should be investigated in further studies that focus on such purpose. This study showing similar AUCs among the three instruments suggested that the correction for ocular magnification on both the cpRNFL and macular parameters could be ignored for the detection with high myopia. 

Accumulating evidence is beginning to suggest that measurements of the inner retinal layers in the macular region may be additional parameters that can be used for glaucoma detection. ^2,3,5,40,41 The recent versions of Cirrus and 3D OCT allow for the separation of the RNFL from GCL at the macula. ^7,35,42 We also compared the mRNFL and GCL/IPL measurements between Cirrus and 3D OCT. Unfortunately, RTVue was not included in this analysis because the software of this instrument does not separate the RNFL and GCL. In results, 3D OCT exhibited higher AUCs for the mRNFL measurements than Cirrus. This was in agreement with our previous reports evaluating the abilities for nonhighly myopic glaucomatous eyes. ¹¹ Differences in the segmentation algorithm to demarcate the mRNFL from GCL and the scan area may underlie the discrepancy, but the reason was not fully understood. 

Past studies comparing diagnostic abilities of the GCC thickness with cpRNFL thickness in glaucomatous eyes with high myopia suggest conflicting results. Shoji et al. have shown that the GCC obtained by RTVue was a significantly better performance than the cpRNFL thickness for detecting glaucoma with high myopia. That study enrolled 51 perimetric glaucomatous eyes with high myopia (mean deviation [MD] = −8.1 dB) and 31 highly myopic eyes (≤ −5.0 dB) without glaucomatous visual field loss with the definition as no glaucomatous eyes. ²² They also demonstrated a similar conclusion in an additional study with 38 highly myopic eyes without glaucomatous visual field and 53 perimetric glaucomatous eyes with high myopia (mean MD = −5.6dB). ²³ These two studies did not describe the optic disc appearance and also included a number of eyes after cataract surgery. However, Kim et al. limited the enrollment criteria for subjects without any ocular surgeries and studied 24 highly myopic eyes (≤−6.00D) with normal appearing disc and 21 glaucomatous eyes with high myopia (mean MD = −8.56dB). They concluded that the GCC obtained by RTVue showed higher AUC than the cpRNFL thickness, but the difference was not statistically significant. ²¹ In the recent study, enrolling 49 highly myopic glaucomatous eyes (mean MD = −7.44dB) and 22 highly myopic normal eyes, macular parameters obtained by Cirrus showed comparable abilities with the cpRNFL for detecting highly myopic glaucomatous eyes. This study, which enrolled a larger sample size than those past studies, revealed that all of the instruments did not show significant differences between the best macular parameter and cpRNFL parameter in the AUCs. Our study demonstrated that the three instruments had similar abilities between the cpRNFL and GCC. In addition, as far as we know, this study is the first report to demonstrate that the mRNFL thickness had a comparable detecting ability as the cpRNFL thickness for highly myopic glaucoma. These result for the equivalent abilities between cpRNFL and macular parameters in each instrument was in a line with our past report in non-highly glaucomatous myopic eyes.¹¹  

Our study has some limitations. First, the HFA and OCT measurements were not performed on the same day in most subjects. However, these examinations were performed within 3 months. All glaucomatous eyes exhibited stable intraocular pressure, and the treatment modality was not altered during the present study. Therefore, the reduction in examined parameters during the examination periods was likely negligible. Second, our study's design was a case-control study including patients with well-established glaucoma, and using a separate group of normal subjects as hospital based-controls could substantially overestimate the diagnostic performance. ^43,44 Finally, we included some highly myopic eyes with mild or severe glaucomatous visual field loss. Clinically, the OCT analysis for diagnosing early or preperimetric glaucoma plays a more important role than in advanced glaucomatous eyes. However, the preliminary sample calculation suggested that we would not meet a large enough sample size when we limited the subjects to compare the AUCs between the instruments. 

In conclusion, we demonstrated that the diagnostic performances of the average cpRNFL thickness and average GCC thickness for the identification of highly myopic glaucoma were similar between Cirrus, RTVue, and 3D OCT in our studied population. In addition, the cpRNFL thickness had as high of a diagnostic performance as the GCC measurements for glaucoma with high myopia. Although the usage of SD-OCT in glaucoma was not commercially recommended in highly myopic eyes, these instruments would be useful for diagnosing highly myopic glaucoma if the normative database in highly myopic eyes was available. 

Acknowledgments

Supported by JSPS KAKENHI Grant Number 25462715 (AK) from the Japanese Government; the Suda Memorial Foundation (AK); the Mishima Memorial Foundation (AK); and the Santan Pharmaceutical Founder Commemoration Ophthalmic Research Fund (AK). 

Disclosure: A. Akashi, None; A. Kanamori, None; M. Nakamura, None; M. Fujihara, None; Y. Yamada, None; A. Negi, None