July 2012
Volume 53, Issue 8
Free
Glaucoma  |   July 2012
Detection of Localized Retinal Nerve Fiber Layer Defects with Posterior Pole Asymmetry Analysis of Spectral Domain Optical Coherence Tomography
Author Affiliations & Notes
  • Je Hyun Seo
    From the Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea; the
    Seoul National University Bundang Hospital, Seongnam, Korea; the
    Department of Ophthalmology, Pusan National University Yangsan Hospital, Pusan, Korea; the
  • Tae-Woo Kim
    From the Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea; the
    Seoul National University Bundang Hospital, Seongnam, Korea; the
  • Robert N. Weinreb
    Hamilton Glaucoma Center and Department of Ophthalmology, University of California San Diego, La Jolla, California; and the
  • Ki Ho Park
    From the Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea; the
  • Seok Hwan Kim
    From the Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea; the
    Seoul National University Boramae Hospital, Seoul, Korea.
  • Dong Myung Kim
    From the Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea; the
  • Corresponding author: Tae-Woo Kim, Department of Ophthalmology, Seoul National University Bundang Hospital, Seoul National University College of Medicine, 166 Gumi-dong, Bundang-gu, Seongnam, Gyeonggi-do 463-707, Korea; twkim7@snu.ac.kr
Investigative Ophthalmology & Visual Science July 2012, Vol.53, 4347-4353. doi:10.1167/iovs.12-9673
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Je Hyun Seo, Tae-Woo Kim, Robert N. Weinreb, Ki Ho Park, Seok Hwan Kim, Dong Myung Kim; Detection of Localized Retinal Nerve Fiber Layer Defects with Posterior Pole Asymmetry Analysis of Spectral Domain Optical Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2012;53(8):4347-4353. doi: 10.1167/iovs.12-9673.

      Download citation file:


      © 2016 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

Purpose.: To investigate the diagnostic ability of posterior pole asymmetry analysis (PPAA) with spectral domain optical coherence tomography (SD-OCT) for detecting localized retinal nerve fiber layer defects (RNFLD).

Methods.: Eighty-four open-angle glaucoma subjects (84 eyes) with localized, wedge-shape RNFLDs by red-free RNFL photography and 122 eyes of healthy subjects were enrolled. The subjects were examined by SD-OCT to obtain circumpapillary RNFL (cpRNFL) thickness as well as PPAA. The PPAA provides a corresponding cell to cell comparison between hemispheres within the central 20° and presents the difference using gray scale. Area under the receiver operating characteristic curve (AUROC) was calculated. The sensitivity and specificity of PPAA to detect glaucoma with the localized defects were also examined using various criteria; two consecutive black cells (criterion A), three consecutive black cells (criterion B), and four consecutive black cells (criterion C).

Results.: An AUROC of PPAA based on the number of black cells was 0.958 ± 0.013. The sensitivity and specificity of PPAA were 95.2% and 81.1% for criterion A, 83.3% and 92.6% for criterion B, and 69.0% and 98.4% for criterion C, respectively. The sensitivity and specificity of the cpRNFL thickness were 85.7% and 95.1%, respectively.

Conclusions.: SD-OCT PPAA detects localized RNFLDs with high sensitivity and specificity, which are comparable to the cpRNFL thickness profile of OCT. These findings suggest that the PPAA can be complementary to other tests for diagnosing glaucoma in patients with localized RNFLDs.

Introduction
Several studies have shown that assessment of macular thickness could be a surrogate method in evaluating glaucomatous damage. 15 Since the ganglion cell layer (GCL) is thickest at the macula (up to 6 to 12 cells thick) and, along with retinal nerve fiber layer (RNFL), constitutes 30% to 35% of the retinal thickness in this region, it is considered that the damage in the GCL can be more readily detected over the macular lesion than the peripheral retina. 1 However, previous studies on measuring macular retinal thickness using optical coherence tomography (OCT) have shown only moderate sensitivity and specificity for detecting glaucoma, and these values were inferior to those with circumpapillary RNFL (cpRNFL) thickness. 27 Such studies used time domain (TD) OCT and were reliant on conventional macular map parameters (concentric rings; fovea, inner macular, and outer macular with diameters of 1, 3, and 6 mm; 9 sectors), 27 which are not customized for detecting glaucoma. A recent study using spectral domain (SD) OCT also showed that the conventional macular thickness map had lower diagnostic ability for detecting glaucoma than measuring cpRNFL thickness. 8  
Early glaucomatous defects are often localized to either side of the horizontal meridian. This opens the possibility that evaluating the symmetry between upper and lower hemifield may help detect early glaucomatous damage. Bagga et al. 9 has shown that macular thickness symmetry, which seeks significant differences between the inferior and superior regions, had an area under the receiving operating characteristic curve (AUROC) of 0.76 to 0.84. Although they showed that a macular thickness symmetry test has potential to detect glaucomatous damage, such an AUROC may be less than desirable for clinical use. Moreover, their macular symmetry test was limited by the TD-OCT technology available at the time they acquired their data. Of particular note, it was conducted using only four 5-mm radial scans centered on the fovea, which would lead to low sensitivities. Furthermore, their normative data were generated from 20 normal eyes, which were not age matched with the glaucoma group. 
Recently, a symmetry test, posterior pole asymmetry analysis (PPAA), has been newly installed in the Spectralis (Heidelberg Engineering, Heidelberg, Germany) SD-OCT. This SD-OCT has higher axial resolution (5 to 7 μm) than the most widely used TD-OCT. 10 The PPAA uses a thickness value of 64 cells obtained from a macular area equivalent to a central 20° visual field 11 ; thus, offering the potential of improved diagnostic capability compared to the previously reported symmetry test. 
The purpose of this study was to investigate the ability of SD-OCT PPAA to detect localized RNFL defects (RNFLD). 
Methods
This study was conducted on consecutive open-angle glaucoma (OAG) patients and healthy subjects attending the Seoul National University Bundang Hospital Glaucoma Clinic from March 2011 to August 2011, who satisfied the inclusion and exclusion criteria. The study was approved by the Seoul National University Bundang Hospital Institutional Review Board and conformed to the Declaration of Helsinki. Informed consent was obtained from all subjects. Prior to the study, participants underwent a comprehensive ophthalmologic examination including refraction, IOP measurement by Goldmann applanation tonometry, central corneal thickness measurement (Orbscan; Orbtek Inc, Bausch & Lomb, Rochester, NY), stereoscopic disc photography, red-free RNFL photography, standard automated perimetry using a Swedish Interactive Threshold Algorithm (SITA Standard 24-2, Humphrey Field Analyzer 750 SITA Standard 24-2; Carl Zeiss Meditec, Dublin, CA), and cpRNFL thickness measurement and PPAA by Spectralis OCT. 
Only OAG patients with a localized RNFLD, 11 as shown in the red-free RNFL photography, which was limited to one hemifield, were enrolled. Subjects were excluded if they had best-corrected visual acuity less than 20/40, spherical equivalent refractive error beyond ± 5.0 diopter (D), cylinder correction greater than ± 3.0 D, history of ocular surgery other than uncomplicated cataract surgery, history of ocular trauma and uveitis, other diseases affecting the visual field (e.g., diabetic retinopathy, retinal vein occlusion, ischemic optic neuropathy), or other diseases affecting retinal thickness (e.g., epiretinal membrane, macular edema, drusen). Subjects with an unacceptable quality of OCT image (see below) were also excluded. 
OAG was defined as the presence of glaucomatous optic neuropathy such as rim thinning, notching, and disc hemorrhage with associated visual field defect. A glaucomatous visual field change was defined as (1) outside the normal limit on glaucoma hemifield test, (2) three abnormal points with P less than 5% probability of being normal, one with P less than 1% by pattern deviation, and (3) a pattern standard deviation (PSD) of 5% if the visual field was otherwise normal, confirmed on two consecutive tests. Only reliable visual fields (fixation loss rate ≤ 20%, false positive and false negative error rates ≤ 25%) tests were included in the analysis. 
Healthy subjects were enrolled by advertisement. Healthy control eyes had an IOP less than or equal to 21 mmHg with no history of increased IOP, an absence of glaucomatous disc appearance, no visible RNFLD according to red-free RNFL photography, no retinal pathology, and normal visual field by standard automated perimetry. 
When both eyes were eligible for the study in one subject, a randomly selected eye was included. 
Red-Free Retinal Nerve Fiber Layer Photography
Red-free fundus photographs of the RNFL were taken using a Canon fundus camera (EOS D60 instrument; Canon, Tochigiken, Japan) with pupil dilation. The red-free RNFL photographs were evaluated on an LCD monitor (Samsung Electronics Inc., Seoul, Korea). A localized RNFLD was defined on the red-free RNFL photographs when their width at a one disc diameter distance from the edge of the disc was larger than a major retinal vessel, diverging in an arcuate or wedge shape from the edge of the disc. The presence of a localized RNFLD on a red-free photograph was made by agreement between the two independent readers (JHS, TWK). Angular width was measured from lines (line a, line b) from the center of the disc to a disc margin where the RNFLD meets the disc center (c) (Fig. 1A). 
Figure 1. 
 
Topographic measurement of angular width of RNFLDs by red-free RNFL photograph. (A) Angular width consisted of lines (line a, line b) from the center of the disc to a disc margin where the RNFLD meets the disc center (c). (B) Localized RNFLDs detected by optical coherence tomography. A localized OCT RNFLD was defined by two criteria (1) more than or equal to one sector abnormal at the less than 1% level, and (2) at the less than 5% level. (C) Localized RNFLDs detected by PPAA. Black cells indicate mean difference over 30 μm. This subject has seven black cells on PPAA.
Figure 1. 
 
Topographic measurement of angular width of RNFLDs by red-free RNFL photograph. (A) Angular width consisted of lines (line a, line b) from the center of the disc to a disc margin where the RNFLD meets the disc center (c). (B) Localized RNFLDs detected by optical coherence tomography. A localized OCT RNFLD was defined by two criteria (1) more than or equal to one sector abnormal at the less than 1% level, and (2) at the less than 5% level. (C) Localized RNFLDs detected by PPAA. Black cells indicate mean difference over 30 μm. This subject has seven black cells on PPAA.
RNFL Thickness Analysis and Posterior Pole Asymmetry Analysis Using OCT
Images were acquired with the SD-OCT in near-infrared mode (820 nm) after pupillary dilation. The OCT was taken on the same day as the fundus photography was performed. The cpRNFL thickness was measured using a circular scan (12° in diameter). For cpRNFL measurements, Spectralis OCT provides the average measurement value for four quadrants and six sectors, and the global average. If the eye had abnormality in any quadrant or sector, it was considered an abnormal OCT (OCT RNFLD). A localized OCT RNFLD was defined by two criteria (1) greater than or equal to one sector abnormal at the less than 1% level, and (2) at the less than 5% level (Fig. 1B). 
The volume scan in the central 20° area for each eye was performed for PPAA using a 30° × 25° OCT scan. The quality of the scans was indicated by a color scale at the bottom of the scan, where it had to be in the green range to be considered a good quality scan. In addition, the SD-OCT macular map provides a color scale representation of topographic retinal thickness, which helps to evaluate the image quality. The authors examined all the B-scan images in each eye to determine whether there were any segmentation errors in the images. The following criteria were used for identifying segmentation failures in the B-scan: obvious disruption of the detected border and/or border wandering (jumping of the detected border) for 5% consecutive or 20% cumulative of the entire image. 12 The posterior pole map provides a retinal thickness value of 64 (8 × 8) cells within each cell. The 8 × 8 grid was positioned symmetrically to the fovea-disc axis with the central point of the grid to the fovea. PPAA provides the data derived from the cell to cell comparison between corresponding cells across the hemisphere within each eye. For the inter-hemisphere comparison, the fovea-disc axis was used as the symmetry line (Fig. 1C). 
For the PPAA, a difference greater than or equal to 30 μm was used as the cutoff value for the cell to cell comparison. This criterion was selected based on findings reported by previous studies. 9,13 Bagga et al. 9 reported that mean macular thickness was 30-μm thinner in patients with glaucoma compared to healthy subjects (259.7 ± 9.9 μm vs. 229.0 ± 13.3 μm). Inuzuka et al. 13 has also reported that the difference between outer macular retinal thickness of apparently normal hemifield and defective hemifield within an eye was 29.4 μm (266.3 ± 16.7 μm vs. 236.9 ± 14.3 μm). In addition, when the PPAA retinal thickness difference is represented by gray scale, it is difficult to clearly discern the values using colors other than black (black indicates that differences of retinal thickness is over 30 μm). For these two reasons, the authors used the black as the cutoff for meaningful difference in PPAA. 
Data Analysis
The AUROC was calculated to assess the ability of the overall numbers of black cells. Based on the AUROC analysis, criteria (e.g., the number of black cells) that might be clinically meaningful were selected, and the sensitivity and specificity of such criteria were compared with those of cpRNFL thickness measurement using McNemar's test with continuity correction. Data analysis was performed using the Statistical Package for Social Sciences (version 17.0; SPSS Inc., Chicago, IL) and R program (R Development Core Team, Vienna, Austria). A P value less than 0.05 was considered statistically significant. 
Results
Subject Characteristics
This study initially involved 213 OAG patients and 150 healthy subjects. Of the OAG patients, 78 subjects were excluded due to ambiguous information in RNFL photography (e.g., diffuse RNFLD), 20 subjects due to a retinal lesion, and 31 subjects due to unacceptable quality of OCT image, leaving a sample of 84 OAG patients. Of the control subjects, 126 subjects who were age matched with glaucoma patients were selected. Of these, four subjects were excluded due to unacceptable OCT image, leaving a sample of 122 eyes of 122 subjects. The mean age was 57.0 ± 13.3 years in the OAG group and 56.4 ± 13.2 years for the healthy control group. Sex, refraction, and initial IOP were similar between the two groups. Mean deviation (MD) of standard automated perimetry was −3.2 ± 3.3 dB for the localized RNFLD group and −0.29 ± 1.42 dB for the healthy control group (Table 1). 
Table 1. 
 
Demographic Characteristics of the Study Subjects
Table 1. 
 
Demographic Characteristics of the Study Subjects
Glaucoma (n = 84) Normal control (n = 122) P
Age (y) 57.0 ± 13.3 56.4 ± 13.2 0.73*
Sex (male %) 36 (42.8) 47 (38.5) 0.57†
Systemic hypertension n (%) 29 (34.5) 28 (23.0) 0.08†
Diabetic mellitus n (%) 12 (14.3) 13 (10.6) 0.52†
Spherical equivalent (diopters) –1.13 ± 2.76 –0.82 ± 2.1 0.36*
Humphrey C24-2 threshold visual field
 MD (dB) –3.2 ± 3.3 –0.29 ± 1.42 <0.0001*
 PSD (dB) 4.7 ± 3.6 1.82 ± 0.45 <0.0001*
 Central cornea thickness (μm) 555.9 ± 36.7 561.1 ± 42.7 0.47*
 IOP without medication (mmHg) 14.8 ± 3.3 14.5 ± 3.0 0.41*
Location of RNFLDs
 Superior 26
 Inferior 58
Diagnostic Ability According to the Number of Black Cells
The AUROC of the PPAA (numbers of consecutive black cells) was 0.958 ± 0.013 (Fig. 2). Table 2 presents the sensitivity and specificity for various cutoff values based on cell numbers. From the AUROC analysis, cutoff values of two, three, and four consecutive black cells were considered to be potentially useful in the clinical practice. The number of consecutive black cells beyond this range showed either poor sensitivity or specificity. The sensitivity and specificity of the PPAA for detecting localized RNFLDs using the two, three, and four consecutive black cell criterion (criterion A, B, and C in order) were 95.2% and 81.1%, 83.3%, and 92.6%, and 69.0% and 98.4%, respectively. 
Figure 2. 
 
ROC curves of posterior pole asymmetry analysis (numbers of black cell). The AUROC of PPAA (numbers of black cell) was 0.958 ± 0.013.
Figure 2. 
 
ROC curves of posterior pole asymmetry analysis (numbers of black cell). The AUROC of PPAA (numbers of black cell) was 0.958 ± 0.013.
Table 2. 
 
Sensitivity and Specificity of Posterior Pole Asymmetry Analysis: According to the Black Cell (difference > 30 μm) Numbers
Table 2. 
 
Sensitivity and Specificity of Posterior Pole Asymmetry Analysis: According to the Black Cell (difference > 30 μm) Numbers
Cell Number Sensitivity Specificity
1 1 0.385
2 0.952 0.811
3 0.833 0.926
4 0.690 0.984
5 0.583 1.000
6 0.524 1.000
7 0.500 1.000
8 0.417 1.000
9 0.333 1.000
10 0.286 1.000
11 0.262 1.000
12 0.250 1.000
13 0.214 1.000
14 0.155 1.000
15 0.131 1.000
16 0.083 1.000
17 0.060 1.000
18 0.036 1.000
19 0.024 1.000
20 0.012 1.000
21 0.000 1.000
Comparison of the Sensitivity and Specificity of PPAA and cpRNFL Thickness Analysis
The sensitivity and specificity of the cpRNFL thickness using OCT with the criterion more than or equal to one sector abnormality at the less than 1% level for detecting localized RNFLDs were 85.7% and 95.1%, respectively. The sensitivity was comparable with that of PPAA criterion B (P = 0.824), lower than that of PPAA criterion A (P = 0.039), and higher than that of PPAA criterion C (P = 0.007). The specificity was comparable with that of PPAA criterion B (P = 0.607) and C (P = 0.289), and higher than that of criterion A (P = 0.001). 
The sensitivity and specificity of the cpRNFL thickness using OCT with the criterion more than one or equal to sector at the less than 5% level were 90.5% and 81.1%, respectively. The sensitivity of this criterion was comparable with that of PPAA criterion A (P = 0.289) and B (P = 0.210), and higher than that of PPAA criterion C (P < 0.0001). The specificity was comparable with that of PPAA criterion A (P = 1.000), and lower than those of criterion B and C (P = 0.016, P < 0.0001, respectively) (Table 3). Figure 3 illustrates a case included in the study. A localized defect was observed in the superotemporal quadrant on the red-free photograph. This defect was also identified by PPAA. However, the defect was not detected by cpRNFL thickness. 
Table 3. 
 
Comparison of Sensitivity, Specificity between cpRNFL Thickness Protocol, and PPAA Using Optical Coherence Tomography
Table 3. 
 
Comparison of Sensitivity, Specificity between cpRNFL Thickness Protocol, and PPAA Using Optical Coherence Tomography
Sensitivity Specificity
cpRNFL OCT Abnormality < 1% level 85.7% 95.1%
Abnormality < 5% level 90.5% 81.1%
PPAA OCT Criteria A 95.2% 81.1%
P* = 0.039, P† = 0.289 P* = 0.001, P† = 1.000
Criteria B 83.3% 92.6%
P* = 0.824, P† = 0.210 P* = 0.607, P† = 0.016
Criteria C 69.0% 95.1%
P* = 0.007, P† < 0.0001 P* = 0.289, P† < 0.0001
Figure 3. 
 
A case with a localized RNFLD. (A) A localized RNFLD is observed by red-free RNFL photography (white arrows). (B) The defect was also identified by PPAA (blue arrow). (C) The defect was not detected by cpRNFL measurement (black arrow).
Figure 3. 
 
A case with a localized RNFLD. (A) A localized RNFLD is observed by red-free RNFL photography (white arrows). (B) The defect was also identified by PPAA (blue arrow). (C) The defect was not detected by cpRNFL measurement (black arrow).
Angular Width of Red-Free RNFLD and the Sensitivity of PPAA and cpRNFL Thickness Measurement
For the photographic RNFLD with width less than 10°, PPAA using criterion A and B showed apparently higher sensitivity than cpRNFL thickness measurement (99% confidence interval (CI) and 95 % CI). For the photographic RNFLDs with width more than 30°, cpRNFL thickness showed apparently higher sensitivity than PPAA. However, those differences were not statistically significant (Table 4). 
Table 4. 
 
Sensitivity of Retinal Nerve Fiber Layer Thickness and Posterior Pole: Asymmetry Analysis According to the Angular Width of Retinal Nerve Fiber Layer Defect using Optical Coherence Tomography
Table 4. 
 
Sensitivity of Retinal Nerve Fiber Layer Thickness and Posterior Pole: Asymmetry Analysis According to the Angular Width of Retinal Nerve Fiber Layer Defect using Optical Coherence Tomography
Angular Width of RNFLD on Fundus Photograph Number of Eyes Sensitivity of cpRNFL Number (Sensitivity) of PPAA
99% CI 95% CI Criteria A Criteria B Criteria C
<10° 12 50.0% 66.7% 10 (83.3%) P* = 0.22 P† = 0.63 9 (75.0%) P* = 0.38 P† = 1.00 3 (25.0%) P* = 0.38 P† = 0.063
11°–20° 20 75.0% 80.0% 19 (95.0%) P* = 0.13 P† = 0.25 14 (70.0%) P* = 1.00 P† = 0.73 13 (65%) P* = 0.75 P† = 0.51
21°–30° 14 92.8% 100% 14 (100%) P*† = 1.00 13 (92.8%) P*† = 1.00 11 (78.5%) P* = 0.50 P† = 0.25
31°–40° 12 100% 100% 11 (91.6%) P*† = 1.00 9 (75.0%) P*† = 0.25 7 (58.3%) P*† = 0.07
40°–50° 15 100% 100% 15 (100%) P*† = 1.00 14 (93.3%) P*† = 1.00 13 (86.6%) P*† = 0.48
>50° 11 100% 100% 11 (100%) P*† = 1.00 11 (100%) P*† = 1.00 11 (100%) P*† = 1.00
Total 84 85.7% 90.5% 80 (95.2%) 70 (83.3%) 58 (69.0%)
Discussion
A previous study on a macular symmetry test using TD-OCT 9 reported that the AUROC value range was 0.76 to 0.84, which may be less than desirable. However, Asrani et al. 11 recently suggested that the macular symmetry analysis using SD-OCT may be an effective approach to detect and potentially monitor patients with glaucoma. They have customized the retinal thickness protocol to acquire detailed retinal thickness measurements of the central 20° of the posterior pole. Although their evaluation was based on only four case samples, they showed that the macular symmetry test may accurately detect both early and moderate glaucoma. 
In the present study, the authors demonstrate that the PPAA of SD-OCT can detect localized RNFLDs with comparable sensitivity and specificity to those obtained with cpRNFL thickness measurement. Their results showed that PPAA was a higher diagnostic value (AUROC = 0.958) than results of previous studies measuring the macular retinal thickness. Previously, the AUROC of the macular retinal thickness measurement has been reported to be between 0.60 and 0.81. 2,7,1416 In the authors' study samples, the best AUROC for the macular retinal thickness measurement was 0.791, which was obtained at the outer inferior sector (data not presented). They speculate that the higher diagnostic value of PPAA is attributed to several factors. First, the PPAA tests the symmetry of the macular thickness, while the previous studies compared the measured macular thickness between healthy subjects and glaucoma patients. The authors' data suggest that examining the macular symmetry may be a more effective approach to detect the glaucomatous abnormality than the comparison with healthy subjects. Second, previous studies used conventional macular parameters obtained from a 6-mm diameter region, which corresponds to central 10°. Their use of an area of approximately 10° 15,17 may be too small to detect glaucomatous damage, which often starts more peripheral to 10°. 18,19 The PPAA have 64 cells on a 25° × 30° extent, which may cover a large area of approximately 20°. 11 Last, previous studies were performed by TD-OCT. 27 SD-OCT allows imaging of intraretinal microstructure at a much faster scan rate with higher resolution than TD-OCT. 20 Moreover, eye tracking capability of Spectralis OCT enables reduction of artifact produced by microsaccades. 21  
The authors' data are in line with Um et al.'s study, 22 which evaluated diagnostic capability of the evaluation of asymmetry in hemifield macular thickness using the same OCT device. They arbitrarily divided both the superior and inferior macular area into five zones, simulating the glaucoma hemifield test in the standard automated perimetry and compared the corresponding pairs. Using this approach, they found that the sensitivity of the macular hemifield test was significantly greater than that of cpRNFL measurement in the early glaucoma group. In the present study, the authors used the cell to cell comparison map presented in the PPAA result print out and the abnormality was defined based on the number of black cells. Although the two studies used different approaches, results of both studies suggest that detecting asymmetry in the macular hemifield thickness may be a useful diagnostic aid in the detection of glaucoma. 
Previous studies have shown the cpRNFL measurement had unsatisfactory diagnostic capability to detect such an early RNFLD. 23,24 Thus, it was a specific aim of this study to see whether the PPAA is superior to cpRNFL measurement in detecting very early stage localized RNFLD. Although the PPAA showed apparently higher sensitivity than cpRNFL measurement for detecting an RNFLD with width less than 10°, the difference was not statistically significant. It appears that detecting a localized defect in the very early stage is also challenging with the use of PPAA. 
Measuring the ganglion cell complex (GCC) is another strategy to evaluate the macular region to diagnose glaucoma using SD-OCT. The AUROC value of GCC thickness analysis has been reported to be from 0.87 to 0.922. 14,16 Although direct comparison is not feasible due to the different study populations, the authors' result (AUROC = 0.958) suggests that PPAA may have better, or comparable, diagnostic ability than GCC analysis. They considered that a larger scan area of PPAA (8 × 8 grid, one grid = 3° × 3°, approximately a square millimeter), 11 than that of GCC (up to 6 × 6-mm diameter circle), may provide an advantage to detect glaucomatous damage, which often starts within the superotemporal and inferotemporal quadrant. 
In the present study, only eyes with localized defects were included in the glaucoma group. This was because one of the primary aims of the present study was to evaluate whether PPAA has an advantage over cpRNFL to detect early glaucomatous damage as described above. Although localized RNFLDs are easily identified in red-free photography, even in the early stage, it is often difficult to discriminate between normal eyes and those with mild diffuse atrophy. 25 Faulty classification of mild diffuse atrophy to normal should result in erroneous interpretation. By including only eyes with localized RNFLDs, in which the border of the defect could be clearly defined, the authors were able to separately compare the performance of PPAA and cpRNFL thickness measurement to detect early stage RNFL damage (e.g., defect with narrow angular width).  
Due to this study design, the present study resulted only in information regarding the ability of PPAA to detect localized RNFLDs. The merit of PPAA for detecting diffuse atrophy has yet to be determined. 
The authors' study has several limitations. First, all of the participants were Korean. Therefore, further studies with other ethnic groups are needed. Second, they used fixed cutoff value retinal thickness differences (≥30 μm), regardless of retinal topographic considerations. To resolve this problem, more detailed age matched normative data that reflect the regional variation are needed. Last, they included only patients with a localized RNFLD, which involved one hemifield. Thus, the result may not be applied to the general glaucoma patient population. 
In conclusion, SD-OCT PPAA can detect localized RNFLDs with high sensitivity and specificity, and is comparable with the RNFL map of OCT. These findings suggest that the PPAA is complementary to other tests for diagnosing glaucoma in patients with a localized RNFLD. 
References
Zeimer R Asrani S Zou S Quigley H Jampel H . Quantitative detection of glaucomatous damage at the posterior pole by retinal thickness mapping. A pilot study. Ophthalmology . 1998;105:224–231. [CrossRef] [PubMed]
Guedes V Schuman JS Hertzmark E Optical coherence tomography measurement of macular and nerve fiber layer thickness in normal and glaucomatous human eyes. Ophthalmology . 2003;110:177–189. [CrossRef] [PubMed]
Lederer DE Schuman JS Hertzmark E Analysis of macular volume in normal and glaucomatous eyes using optical coherence tomography. Am J Ophthalmol . 2003;135:838–843. [CrossRef] [PubMed]
Hess DB Asrani SG Bhide MG Enyedi LB Stinnett SS Freedman SF . Macular and retinal nerve fiber layer analysis of normal and glaucomatous eyes in children using optical coherence tomography. Am J Ophthalmol . 2005;139:509–517. [CrossRef] [PubMed]
Leung CK Chan WM Yung WH Comparison of macular and peripapillary measurements for the detection of glaucoma: an optical coherence tomography study. Ophthalmology . 2005;112:391–400. [CrossRef] [PubMed]
Maier P Funk J . Diagnostic accuracy of the Retinal Thickness Analyser: differentiation between normal eyes and eyes with glaucoma or macular pathologies. Graefe's Arch Clinical Exp Ophthalmol . 2006;244:1113–1118. [CrossRef]
Parikh RS Parikh SR Thomas R . Diagnostic capability of macular parameters of Stratus OCT 3 in detection of early glaucoma. Br J Ophthalmol . 2010;94:197–201. [CrossRef] [PubMed]
Na JH Sung KR Baek S Sun JH Macular Lee Y . and retinal nerve fiber layer thickness: which is more helpful in the diagnosis of glaucoma? Invest Ophthalmol Vis Sci . 2011;52:8094–8101. [CrossRef] [PubMed]
Bagga H Greenfield DS Knighton RW . Macular symmetry testing for glaucoma detection. J Glaucoma . 2005;14:358–363. [CrossRef] [PubMed]
Sung KR Kim JS Wollstein G Folio L Kook MS Schuman JS . Imaging of the retinal nerve fibre layer with spectral domain optical coherence tomography for glaucoma diagnosis. Br J Ophthalmol . 2011;95:909–914. [CrossRef] [PubMed]
Asrani S Rosdahl JA Allingham RR . Novel software strategy for glaucoma diagnosis: asymmetry analysis of retinal thickness. Arch Ophthalmol . 2011;129:1205–1211. [CrossRef] [PubMed]
Ishikawa H Stein DM Wollstein G Beaton S Fujimoto JG Schuman JS . Macular segmentation with optical coherence tomography. Invest Ophthalmol Vis Sci . 2005;46:2012–2017. [CrossRef] [PubMed]
Inuzuka H Kawase K Sawada A Aoyama Y Yamamoto T . Macular Retinal Thickness in Glaucoma With Superior or Inferior Visual Hemifield Defects [published online ahead of print August 29, 2011]. J Glaucoma . 2011:doi: 10.1097/IJG.0b013e318225472a .
Rao HL Zangwill LM Weinreb RN Sample PA Alencar LM Medeiros FA . Comparison of different spectral domain optical coherence tomography scanning areas for glaucoma diagnosis. Ophthalmology . 2010;117:1692–1699. [CrossRef] [PubMed]
Tan O Chopra V Lu AT Detection of macular ganglion cell loss in glaucoma by Fourier-domain optical coherence tomography. Ophthalmology . 2009;116:2305–2314. [CrossRef] [PubMed]
Mori S Hangai M Sakamoto A Yoshimura N . Spectral-domain optical coherence tomography measurement of macular volume for diagnosing glaucoma. J Glaucoma . 2010;19:528–534. [CrossRef] [PubMed]
Sakamoto A Hangai M Nukada M Three-dimensional imaging of the macular retinal nerve fiber layer in glaucoma with spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2010;51:5062–5070. [CrossRef] [PubMed]
Harrington DO . The Bjerrum Scotoma. Am J Ophthalmol . 1965;59:646–656. [CrossRef] [PubMed]
Mikelberg FS Drance SM . The mode of progression of visual field defects in glaucoma. Am J Ophthalmol . 1984;98:443–445. [CrossRef] [PubMed]
Tan O Li G Lu AT Varma R Huang D . Mapping of macular substructures with optical coherence tomography for glaucoma diagnosis. Ophthalmology . 2008;115:949–956. [CrossRef] [PubMed]
Han IC Jaffe GJ . Evaluation of artifacts associated with macular spectral-domain optical coherence tomography. Ophthalmology . 2010;117:1177–1189. [CrossRef] [PubMed]
Um TW Sung KR Wollstein G Yun SC Na JH Schuman JS . Asymmetry in hemifield macular thickness as an early indicator of glaucomatous change. Invest Ophthalmol Vis Sci . 2012;53:1139–1144. [CrossRef] [PubMed]
Kim TW Park UC Park KH Kim DM . Ability of Stratus OCT to identify localized retinal nerve fiber layer defects in patients with normal standard automated perimetry results. Invest Ophthalmol Vis Sci . 2007;48:1635–1641. [CrossRef] [PubMed]
Jeoung JW Park KH Kim TW Khwarg SI Kim DM . Diagnostic ability of optical coherence tomography with a normative database to detect localized retinal nerve fiber layer defects. Ophthalmology . 2005;112:2157–2163. [CrossRef] [PubMed]
Airaksinen PJTA Werner EB . Clinical evaluation of the optic disc and retinal nerve fiber layer. In: Ritch RSM Krupin T , ed. The Glaucomas . St. Louis: Mosby; 1996:617–657.
Footnotes
 Supported by grants from the National Research Foundation of Korea funded by Korean Government (2010-0004210).
Footnotes
 Disclosure: J.H. Seo, None; T.-W. Kim, None; R.N. Weinreb, Heidelberg Engineering (F), Topcon (F), Zeiss-Meditec (F, C), Optovue (F, C), Nidek (F); K.H. Park, None; S.H. Kim, None; D.M. Kim, None
Figure 1. 
 
Topographic measurement of angular width of RNFLDs by red-free RNFL photograph. (A) Angular width consisted of lines (line a, line b) from the center of the disc to a disc margin where the RNFLD meets the disc center (c). (B) Localized RNFLDs detected by optical coherence tomography. A localized OCT RNFLD was defined by two criteria (1) more than or equal to one sector abnormal at the less than 1% level, and (2) at the less than 5% level. (C) Localized RNFLDs detected by PPAA. Black cells indicate mean difference over 30 μm. This subject has seven black cells on PPAA.
Figure 1. 
 
Topographic measurement of angular width of RNFLDs by red-free RNFL photograph. (A) Angular width consisted of lines (line a, line b) from the center of the disc to a disc margin where the RNFLD meets the disc center (c). (B) Localized RNFLDs detected by optical coherence tomography. A localized OCT RNFLD was defined by two criteria (1) more than or equal to one sector abnormal at the less than 1% level, and (2) at the less than 5% level. (C) Localized RNFLDs detected by PPAA. Black cells indicate mean difference over 30 μm. This subject has seven black cells on PPAA.
Figure 2. 
 
ROC curves of posterior pole asymmetry analysis (numbers of black cell). The AUROC of PPAA (numbers of black cell) was 0.958 ± 0.013.
Figure 2. 
 
ROC curves of posterior pole asymmetry analysis (numbers of black cell). The AUROC of PPAA (numbers of black cell) was 0.958 ± 0.013.
Figure 3. 
 
A case with a localized RNFLD. (A) A localized RNFLD is observed by red-free RNFL photography (white arrows). (B) The defect was also identified by PPAA (blue arrow). (C) The defect was not detected by cpRNFL measurement (black arrow).
Figure 3. 
 
A case with a localized RNFLD. (A) A localized RNFLD is observed by red-free RNFL photography (white arrows). (B) The defect was also identified by PPAA (blue arrow). (C) The defect was not detected by cpRNFL measurement (black arrow).
Table 1. 
 
Demographic Characteristics of the Study Subjects
Table 1. 
 
Demographic Characteristics of the Study Subjects
Glaucoma (n = 84) Normal control (n = 122) P
Age (y) 57.0 ± 13.3 56.4 ± 13.2 0.73*
Sex (male %) 36 (42.8) 47 (38.5) 0.57†
Systemic hypertension n (%) 29 (34.5) 28 (23.0) 0.08†
Diabetic mellitus n (%) 12 (14.3) 13 (10.6) 0.52†
Spherical equivalent (diopters) –1.13 ± 2.76 –0.82 ± 2.1 0.36*
Humphrey C24-2 threshold visual field
 MD (dB) –3.2 ± 3.3 –0.29 ± 1.42 <0.0001*
 PSD (dB) 4.7 ± 3.6 1.82 ± 0.45 <0.0001*
 Central cornea thickness (μm) 555.9 ± 36.7 561.1 ± 42.7 0.47*
 IOP without medication (mmHg) 14.8 ± 3.3 14.5 ± 3.0 0.41*
Location of RNFLDs
 Superior 26
 Inferior 58
Table 2. 
 
Sensitivity and Specificity of Posterior Pole Asymmetry Analysis: According to the Black Cell (difference > 30 μm) Numbers
Table 2. 
 
Sensitivity and Specificity of Posterior Pole Asymmetry Analysis: According to the Black Cell (difference > 30 μm) Numbers
Cell Number Sensitivity Specificity
1 1 0.385
2 0.952 0.811
3 0.833 0.926
4 0.690 0.984
5 0.583 1.000
6 0.524 1.000
7 0.500 1.000
8 0.417 1.000
9 0.333 1.000
10 0.286 1.000
11 0.262 1.000
12 0.250 1.000
13 0.214 1.000
14 0.155 1.000
15 0.131 1.000
16 0.083 1.000
17 0.060 1.000
18 0.036 1.000
19 0.024 1.000
20 0.012 1.000
21 0.000 1.000
Table 3. 
 
Comparison of Sensitivity, Specificity between cpRNFL Thickness Protocol, and PPAA Using Optical Coherence Tomography
Table 3. 
 
Comparison of Sensitivity, Specificity between cpRNFL Thickness Protocol, and PPAA Using Optical Coherence Tomography
Sensitivity Specificity
cpRNFL OCT Abnormality < 1% level 85.7% 95.1%
Abnormality < 5% level 90.5% 81.1%
PPAA OCT Criteria A 95.2% 81.1%
P* = 0.039, P† = 0.289 P* = 0.001, P† = 1.000
Criteria B 83.3% 92.6%
P* = 0.824, P† = 0.210 P* = 0.607, P† = 0.016
Criteria C 69.0% 95.1%
P* = 0.007, P† < 0.0001 P* = 0.289, P† < 0.0001
Table 4. 
 
Sensitivity of Retinal Nerve Fiber Layer Thickness and Posterior Pole: Asymmetry Analysis According to the Angular Width of Retinal Nerve Fiber Layer Defect using Optical Coherence Tomography
Table 4. 
 
Sensitivity of Retinal Nerve Fiber Layer Thickness and Posterior Pole: Asymmetry Analysis According to the Angular Width of Retinal Nerve Fiber Layer Defect using Optical Coherence Tomography
Angular Width of RNFLD on Fundus Photograph Number of Eyes Sensitivity of cpRNFL Number (Sensitivity) of PPAA
99% CI 95% CI Criteria A Criteria B Criteria C
<10° 12 50.0% 66.7% 10 (83.3%) P* = 0.22 P† = 0.63 9 (75.0%) P* = 0.38 P† = 1.00 3 (25.0%) P* = 0.38 P† = 0.063
11°–20° 20 75.0% 80.0% 19 (95.0%) P* = 0.13 P† = 0.25 14 (70.0%) P* = 1.00 P† = 0.73 13 (65%) P* = 0.75 P† = 0.51
21°–30° 14 92.8% 100% 14 (100%) P*† = 1.00 13 (92.8%) P*† = 1.00 11 (78.5%) P* = 0.50 P† = 0.25
31°–40° 12 100% 100% 11 (91.6%) P*† = 1.00 9 (75.0%) P*† = 0.25 7 (58.3%) P*† = 0.07
40°–50° 15 100% 100% 15 (100%) P*† = 1.00 14 (93.3%) P*† = 1.00 13 (86.6%) P*† = 0.48
>50° 11 100% 100% 11 (100%) P*† = 1.00 11 (100%) P*† = 1.00 11 (100%) P*† = 1.00
Total 84 85.7% 90.5% 80 (95.2%) 70 (83.3%) 58 (69.0%)
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×