Quigley et al.1 reported that up to 40 to 50% of the retinal nerve fiber layer (RNFL) could be lost before visual field defects are detected by means of Goldmann kinetic perimetry. Therefore, the evaluation of RNFL thickness is important for early glaucoma detection. Many imaging technologies have been introduced for the objective assessment of the RNFL, such as scanning laser polarimetry (GDx),2-4 which estimates RNFL thickness by measuring the joint retardation of a polarized scanning laser beam. The GDx VCC machine (Laser Diagnostic Technologies, Inc) shows moderate to good discrimination ability in the diagnosis of glaucoma.5-7 In clinical glaucoma practice, it is also important to understand the severity of glaucoma.8 There is increasing evidence that shows that both structural and functional tests are important when it comes to assessing early damage and progression.9-11 Therefore, it is necessary to objectively assess the glaucomatous damage occurred. Here, in the present work, we conducted a comparative study to evaluate RNFL thickness in normal eyes and in glaucomatous eyes in differing stages of the disease. To our knowledge, there are no reports on RNFL thickness measured with GDx VCC as a function of the glaucoma severity grades that focus on a Chinese population. We aimed to evaluate the diagnostic ability of GDx VCC to discriminate between the different stages of the glaucoma disease. The RNFL thickness measured by GDx VCC will be compared across the groups described above. The correlation between the resulting parameters and the outcome of visual field testing was evaluated for the different subgroups of glaucomatous eyes.

This cross-sectional study included 50 normal eyes from 50 normal subjects and 123 glaucomatous eyes from 123 glaucoma patients. The glaucoma patients are among those people who had sought treatment at the Department of Ophthalmology, China Medical University Hospital (CMUH). Subjects with normal eyes were volunteers from the staff or their family members at the CMUH. This research follows the tenets of the Declaration of Helsinki. Informed consent was obtained from all participants, and the study was approved by the Institutional Review Board of the CMUH.

Each subject underwent a detailed ophthalmic examination, including slit-lamp biomicroscopy, gonioscopy, pachymetry, Goldmann applanation tonometry, stereoscopic examination of the optic disc, and standard automated perimetry (SAP, 30-2 mode, Humphrey Field Analyzer, model 750, HFA; Carl Zeiss Meditec, Inc.). All subjects underwent were assessed by means of the GDx VCC imaging technique within 3 months of the clinical examination and visual field testing. GDx VCC imaging was performed by one experienced operator. Subjects with a best-corrected visual acuity of less than 20/40, a spherical equivalent outside the ±6.0 diopter range, and a cylinder correction >3.0 diopters were excluded. Patients whose eyes suffered from coexisting retinal diseases, uveitis, or unknown optic neuropathy were also excluded from this study.

Normal control eyes had normal findings in clinical examination, an intraocular pressure (IOP) not exceeding 21 mmHg, no history of increased IOP, normal-looking optic nerve12 (intact rim, no evidence of hemorrhage, focal rim thinning, notching, glaucomatous excavation or RNFL defects) and normal visual-field results. A normal visual field was defined as a mean deviation (MD) and a pattern standard deviation (PSD) within 95% confidence limits, and a Glaucoma Hemifield Test (GHT) result within normal limits.

Inclusion criteria for the glaucoma groups included an intraocular pressure above 22 mmHg, an open angle, and a reproducible glaucomatous visual field defect in the absence of any other abnormalities that could explain that defect.

The important clinical challenges in glaucoma are the detection of early glaucomatous damage and its progression over time.9 Therefore, it is also important to know the severity of glaucoma. As far as the classification of glaucoma severity is concerned, several methods have been proposed to grade glaucoma on the basis of visual field changes detected on standard white-on-white perimetry.8,9 However, the objective evaluation of peripapillary RNFL and of optic disc changes constitutes an important part of glaucoma diagnosis, especially when standard automated perimetry is not a feasible option, as it is the case in the late stages of the disease.16 SLP provides a good and objective method for RNFL thickness evaluation. It has been demonstrated that SLP-measured RNFL thickness shows a good correlation with histological measurements.17,18 It is found to be highly reproducible in a long-term test-retest situation, supporting the use of this instrument for longitudinal assessment of the RNFL.19 In addition, low-to-high myopia is not related to clinically relevant variations of the parameters measured with SLP, as assessed with the GDx VCC device.20 Several studies have evaluated the diagnostic capability of GDx VCC in glaucoma detection, especially at its early stages.4-7 Weinreb et al.4 showed that the parameter with the largest AROC (0.83) for discriminating 54 glaucomatous eyes (average MD: 6.49 dB) from 40 healthy eyes was the average RNFL thickness within the superior quadrant. In the study conducted by Resu and Lemij,5 they reported that the sensitivity of the NFI to identify glaucoma patients with mild, moderate, and severe damage was equal to 83.8%, 92.9%, and 90.1%, respectively, at the cutoff level of ≥40. Besides, they also reported that GDx VCC had a high discriminating power for the detection of primary open-angle glaucoma (POAG) (average MD: 8.45 dB). Medeiros et al.6 reported that the GDx VCC-related index showing the highest AROC for glaucoma detection (average MD: –4.87 dB) was NFI (AROC, 0.91). Da Pozzo et al.7 reported that GDx VCC showed a moderate-to-good discriminating ability. Our previous report also showed that GDx VCC shows only a moderate ability to distinguish normal eyes from eyes with early glaucoma (mean MD: -3.32±2.20 dB).21 In the present study, we report that NFI is still the best discriminating parameter when it comes to differentiating normal from glaucomatous eyes in the various stages of the disease (AROC values: normal vs. early, 0.942; normal vs. moderate, 0.985, normal vs. severe glaucoma, 1.000). In addition, we also evaluated the discriminating power of GDx VCC for the various glaucoma subgroups. The results are similar to the ones to those found in previous studies, even though different populations and different visual field severities had been chosen. Many studies had previously reported that the NFI is the best parameter to differentiate glaucomatous eyes from normal ones.4-7,21 The NFI is based on an advanced form of neural network analysis. It has been trained on a large sample of representative healthy and glaucomatous eyes and utilizes information from the entire RNFL thickness map to optimize the discrimination between the two groups. The output of the NFI is a single value that ranges from 1 to 100 and which indicates the overall integrity of the RNFL. It is scientifically reasonable to think that the NFI might always be the best parameter for differentiating glaucomatous eyes from normal ones. Besides, the results of the present work are in good agreement with previous reports, where GDx VCC demonstrated excellent diagnostic capability in glaucoma in a purely Chinese population, even in its early stages. However, when it comes to differentiating the severity of glaucoma, GDx VCC demonstrates only a moderate ability. (early vs. moderate glaucoma: the best AROC (0.732) was obtained for the NFI; for moderate vs. severe, the best AROC (0.652) was for the TSNIT average; for early vs. severe, the best AROC (0.852) was obtained for the NFI). This might be due to a considerable overlap between the RNFL-thickness plots (as measured with GDx VCC) corresponding to the various glaucoma subgroups. Consequently, we believe that the functional results of each glaucoma patient should be analyzed jointly with the structural results yielded by GDx VCC when evaluating the severity of their glaucoma.

Some studies have evaluated the correlations between visual field indices and structural parameters.22-27 Weinreb et al.22 reported a significant correlation between retardation ratio and both the full-field and the sensitivity sectors using the Nerve Fiber Analyzer (NFA) I. Tjon-Fo-Sang and Lemij23 found a significant correlation between retardation and visual field sensitivity. Chen et al.24 reported a significant correlation between visual field index and maximum modulation, regional modulation, and ratio using NFA II. Know et al.25 found that the number had the highest correlation with Humphrey's mean deviation, using GDx. Reus and Lemij26 found a significant correlation between GDx VCC retardation and the sensitivity as measured with the HFA. Iester et al.27 reported that some of the GDx VCC indices significantly correlated with visual field indices in glaucomatous patients (NFI, r=-0.35; superior average, r=0.28; TSNIT average, r=-0.24; inferior average, r=0.21). Our current result is in agreement with previous reports that stated that, in glaucomatous eyes, there was a significant correlation between a visual field index (MD and PSD) and some of the parameters yielded by GDx VCC. However, an interesting result that was pointed out is that the correlation seemed to be stronger for the moderateglaucoma subgroup, compared to that obtained for the severe- or the early-glaucoma subgroup. The most plausible reason for this is that the linear model for the structural/functional relationship is really stronger in the moderate stages of glaucoma. In a previous study by Leung et al's,28 they reported that the structure/function relationship was better explained with nonlinear models (logarithmic regression) when visual sensitivity's MD (dB) was plotted against RNFL thickness. Another possible reason might originate from cataract changes, which might influence the result of visual field performance. However, we should not draw any conclusion from this tentative result, since we will address this topic (the functional/structural relationship in the different stages of glaucoma) in the near future. However, we agree that it is mandatory to follow-up the RNFL thickness of each glaucoma patient using GDx VCC in addition to undertaking conventional perimetry, when it comes to evaluating the disease progression of glaucoma.

There were some limitations in our study. First, a major limitation was the selection bias in the subject inclusion/exclusion process. Although we chose our glaucoma group according to an experienced doctor, it is inevitable that these patients were identified on the basis of particular patterns of structural and functional abnormalities that met preconceived notions, which may result in a biased outcome of the comparison.13 For example, inclusion criteria for normal subjects may be “superb” normal under the strict criteria for normal (normal optic nerve appearance and visual field defect).

Moreover, there were similar problems during the image selection process.29 To increase accuracy and to obtain goodquality scans, we excluded those patients for whom only poor images were obtained, especially the atypical birefringence patterns. However, it is inevitable to suffer from an image selection bias in the context of a technical imaging study.29 Therefore, these could overestimate the diagnostic accuracy of GDx VCC. Second, the difference in age between the normal healthy subjects and the glaucoma patients may also constitute a source of bias that may increase the discriminating power of GDx VCC for glaucoma detection in this study. Third, the sample size is not large enough. Further studies with larger sample sizes are necessary to calculate likelihood ratios that can be used more confidently in the clinical practice when it comes to interpreting GDx VCC data in glaucoma management. However, our tentative result can serve as a basis for further longitudinal and large-scale studies, especially those that would evaluate the long-term RNFL thickness changes in glaucomatous eyes in the different stages of the disease.

In conclusion, RNFL thickness measured by GDx VCC may serve as a useful tool to accurately and more objectively distinguish normal from glaucomatous eyes. It may help to differentiate various severity levels of glaucoma within the Taiwan Chinese population. The NFI was the best discriminating parameter when it comes to differentiating normal from glaucomatous eyes.