Free
Glaucoma  |   April 2014
Topographic Profiles of Retinal Nerve Fiber Layer Defects Affect the Diagnostic Performance of Macular Scans in Preperimetric Glaucoma
Author Affiliations & Notes
  • Mi Jeung Kim
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
  • Jin Wook Jeoung
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
  • Ki Ho Park
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
  • Yun Jeong Choi
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
  • Dong Myung Kim
    Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
    Department of Ophthalmology, Seoul National University College of Medicine, Seoul, Korea
  • Correspondence: Jin Wook Jeoung, Department of Ophthalmology, Seoul National University Hospital, Seoul National University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul 110-744, Korea; [email protected]
Investigative Ophthalmology & Visual Science April 2014, Vol.55, 2079-2087. doi:https://doi.org/10.1167/iovs.13-13506
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mi Jeung Kim, Jin Wook Jeoung, Ki Ho Park, Yun Jeong Choi, Dong Myung Kim; Topographic Profiles of Retinal Nerve Fiber Layer Defects Affect the Diagnostic Performance of Macular Scans in Preperimetric Glaucoma. Invest. Ophthalmol. Vis. Sci. 2014;55(4):2079-2087. https://doi.org/10.1167/iovs.13-13506.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To evaluate the influence of topographic profiles (i.e., inner directional angle and angular width) of localized retinal nerve fiber layer (RNFL) defects on the diagnostic performance of macular ganglion cell–inner plexiform layer (GCIPL) thickness in discriminating preperimetric glaucoma (PPG) eyes from normal control eyes.

Methods.: The ganglion cell analysis algorithm in Cirrus OCT was performed to determine the macular GCIPL thickness. Areas under the receiver operating characteristic curves (AUROCs) and the sensitivities/specificities based on an internal normative database were evaluated. The effect of the inner directional angle and the angular width of localized RNFL defects on the diagnostic performance of macular GCIPL parameters were evaluated by using linear-by-linear association analysis and logistic regression analysis.

Results.: Ninety-two patients with PPG and 92 age-matched healthy control subjects were enrolled in this study. The AUROC of the best parameters in macular GCIPL was 0.823 (inferotemporal sector), which showed no significant difference in comparison to the best parameters of peripapillary RNFL (7 o'clock sector, 0.764) and optic nerve head (rim area, 0.767) (for all comparisons, P > 0.05). A significant linear association was observed between the inner directional angle of RNFL defects and the sensitivity of macular GCIPL parameters for detecting RNFL defects. The angular width of RNFL defects was not significantly associated with the sensitivity of macular GCIPL parameters.

Conclusions.: The diagnostic ability of macular GCIPL parameters was comparable to that of peripapillary RNFL and ONH parameters in PPG. The inner directional angle of RNFL defects, but not the angular width, affects the diagnostic sensitivity of macular GCIPL parameters.

Introduction
Glaucoma is a progressive optic neuropathy that is characterized by the loss of retinal ganglion cells (RGCs) and their axons. It presents as structural changes in the optic nerve head (ONH) and in the retinal nerve fiber layer (RNFL). 1 Because structural changes precede clinically detectable visual field (VF) defects, 24 it is important to identify the structural changes associated with RGC loss as early as possible. For preperimetric glaucoma (PPG), which shows glaucomatous optic disc and RNFL abnormalities with normal VF, these changes are an especially important issue because early detection of structural changes provides the best opportunities for preventing permanent visual function loss. 
Cirrus optical coherence tomography (OCT) (software version 6.0; Carl Zeiss Meditec, Dublin, CA, USA) has introduced an inbuilt ganglion cell analysis (GCA) algorithm that allows successful and reproducible segmentation of the inner macular layers (i.e., a combination of the ganglion cell layer and the inner plexiform layer [GCIPL]). 5,6 To date, several studies investigating the diagnostic performance of macular GCIPL analysis have shown promising results. 711  
The peripapillary location and the angular width of RNFL defects reportedly can affect the diagnostic detection of these defects. 12 Because the scan area of the macular GCIPL differs from that of the peripapillary RNFL (pRNFL), we assumed that the relationship between the diagnostic performance of macular GCPIL parameters and the topographic profiles (i.e., inner directional angle and angular width) of localized RNFL defects might show different patterns. However, to our knowledge, there have been no studies investigating the influence of the topographic profile of localized RNFL defects on the diagnostic performance of GCIPL parameters in PPG. 
Therefore, this study was designed (1) to evaluate the diagnostic ability of macular GCIPL parameters in PPG patients, and (2) to investigate whether the topographic profiles of RNFL defects affect the diagnostic performance of macular GCIPL parameters. 
Methods
This study is based on the Macular Ganglion Cell Imaging Study, which is an ongoing study of glaucoma patients and healthy individuals at the Glaucoma Clinic of Seoul National University Hospital (Seoul, South Korea). The study followed the tenets of the Declaration of Helsinki and was approved by the institutional review board of the Seoul National University Hospital. 
Study Subjects
Eyes with PPG and normal control eyes meeting the eligibility criteria were enrolled from the Glaucoma Clinic of Seoul National University Hospital. If both eyes of a patient were eligible for the study, only one eye was randomly chosen for inclusion. 
All patients showed a best-corrected visual acuity (BCVA) of 20/40 or better, spherical equivalent refractive error within ±5.00 diopter (D), astigmatism within ±3.00 D, an open anterior chamber angle, and good-quality red-free fundus photographs. All patients included in the study were familiar with standard automated perimetry (SAP) (Humphrey field analyzer II; Carl Zeiss Meditec) from the earlier experience of having undergone at least two VF tests. The results of the Humphrey visual field (HVF) test with the Swedish interactive threshold algorithm 30-2 standard in this study were reliable in all subjects (i.e., fixation loss < 20%, false-positive and false-negative rates < 33%). Excluded from the study were eyes with a history of amblyopia, uveitis, intraocular surgery (excepting uncomplicated cataract surgery), diabetes, ocular diseases possibly affecting the peripapillary area (e.g., large peripapillary atrophy), or macular area (e.g., epiretinal membrane), and any other ocular or systemic diseases affecting the VF (e.g., retinal vein occlusion, ischemic optic neuropathy). 
Eyes with PPG were defined as eyes with normal VF results and one or more localized RNFL defects (on red-free fundus photographs) that were associated with a glaucomatous disc appearance (e.g., notching or thinning of neuroretinal rim), which have documented evidence of progression (e.g., focal or diffuse narrowing of neuroretinal rim, increased excavation, increased width or depth of RNFL defects) through stereoscopic disc photography (SDP) or red-free fundus photography performed at least 6 months before enrollment. A normal VF was defined as the pattern standard deviation (PSD) with a P value of more than 5% and the glaucoma hemifield test result within normal limits. 13  
Normal control eyes were defined as eyes having an IOP ≤ 21 mm Hg with no history of increased IOP, an absence of glaucomatous disc appearance, no visible RNFL defect on red-free fundus photography, and a normal VF result in at least two VF tests. Matching for age was performed by randomly selecting one healthy control subject within 2 years of age for each subject with PPG. 
Ophthalmologic Examinations
All subjects underwent a complete ophthalmic examination that included a medical history review, measurement of BCVA, slit-lamp biomicroscopy, IOP measurements by Goldmann applanation tonometry, gonioscopy, dilated fundoscopic examination using a 90-D lens, SDP, red-free fundus photography (VX-10; Kowa Optimed, Tokyo, Japan), SAP, and Cirrus OCT. 
After maximum pupil dilation, red-free fundus photographs were acquired through 60-degree, wide-angle views of the optic disc using a digital camera focused on the retina. Then, they were stored in the JPEG format and displayed on a liquid crystal display monitor. Two glaucoma specialists (JWJ, MJK) independently evaluated the red-free fundus photographs by adjusting the brightness and contrast to accurately define the border of RNFL defect. This was performed without knowledge of the subject's clinical information, such as the OCT or VF test results. If there was a discrepancy between the two observers, it was resolved by the adjudication of a third glaucoma specialist (KHP). 
On red-free fundus photography, a lesion was defined as a localized RNFL defect when its width, at a 1 disc-diameter distance from the edge of the disc, was larger than the width of a major retinal vessel, and it diverged and reached the edge of the disc. 14  
To assess the peripapillary location and the angular width of localized RNFL defects, a method devised by Hwang et al. 15 and Yoo and Park 16 was applied with modifications. In brief, clock-face circles were drawn around the ONH on red-free fundus photographs, the diameter and location of the circles corresponded as closely as possible to the circle displayed in the video mode of the pRNFL thickness analysis report by Cirrus OCT. The two edge points where the RNFL defect met this circle at the inner (i.e., temporal) and outer (i.e., nasal) borders were determined. The inner and outer directional lines that passed through the center of the circle and each edge point were drawn. The direction angles were measured between the directional line and the temporal equator (which was set at 0°) in a clockwise direction in right eyes and in a counterclockwise direction in left eyes. The peripapillary angular width of the localized RNFL defect was determined by subtracting the inner direction angle (α) from the outer direction angle (β) (Fig. 1). 
Figure 1
 
The procedure to determine the peripapillary angular width and the directional angle of the RNFL defect. On red-free fundus photography, clock-face circles are drawn. The diameter and location of the circles are positioned as closely as possible to the scan circle by Cirrus OCT (Carl Zeiss Meditec). The point at which the RNFL defect intersects this circle at the inner and outer borders determines the two edge points. The inner (i.e., temporal) directional lines (a) and the outer (i.e., nasal) directional lines (b), which pass through the center of circle and each edge point, respectively, are drawn. The inner directional angle (α) and the outer directional angle (β) are then measured between the each directional line and the temporal equator (which was set as 0°) in a clockwise direction in right eyes and in a counterclockwise direction in left eyes. The peripapillary angular width of the RNFL defect is defined by subtracting the inner directional angle (α) from the outer directional angle (β) (i.e., βα).
Figure 1
 
The procedure to determine the peripapillary angular width and the directional angle of the RNFL defect. On red-free fundus photography, clock-face circles are drawn. The diameter and location of the circles are positioned as closely as possible to the scan circle by Cirrus OCT (Carl Zeiss Meditec). The point at which the RNFL defect intersects this circle at the inner and outer borders determines the two edge points. The inner (i.e., temporal) directional lines (a) and the outer (i.e., nasal) directional lines (b), which pass through the center of circle and each edge point, respectively, are drawn. The inner directional angle (α) and the outer directional angle (β) are then measured between the each directional line and the temporal equator (which was set as 0°) in a clockwise direction in right eyes and in a counterclockwise direction in left eyes. The peripapillary angular width of the RNFL defect is defined by subtracting the inner directional angle (α) from the outer directional angle (β) (i.e., βα).
OCT Measurements
The OCT images were obtained by using the Cirrus OCT device with software version 6.0. Only images that were well-centered on the optic disc or fovea with signal strength of 6 or greater were included in the analyses. The macular GCIPL thickness within a 6 × 6 × 2-mm (14.13 mm2) elliptical annulus centered on the fovea was measured and computed by the GCA algorithm, which was embedded in the Cirrus OCT device. This annulus cube has an inner vertical diameter of 1 mm, an outer vertical diameter of 4 mm, an inner horizontal diameter of 1.2 mm, and an outer horizontal diameter of 4.8 mm. The GCIPL thickness was then analyzed by eight parameters, which included the average thickness, minimum thickness, and six sectors (Fig. 2D). Several previous reports have described this computational method in detail. 5,6  
Figure 2
 
A case of preperimetric glaucoma with a localized RNFL defect. (A) Red-free fundus photography shows the inferior temporal localized RNFL defect (arrowheads). (B) The HVF test indicates normal visual field. (C) The pRNFL map does not detect the RNFL defect. (D) However, the RNFL defect is detected by the macular GCIPL parameters, the GCIPL thickness map (arrow), the deviation-from-normal map (asterisk), the macular sector map (arrow), and the minimal GCIPL thickness (asterisk).
Figure 2
 
A case of preperimetric glaucoma with a localized RNFL defect. (A) Red-free fundus photography shows the inferior temporal localized RNFL defect (arrowheads). (B) The HVF test indicates normal visual field. (C) The pRNFL map does not detect the RNFL defect. (D) However, the RNFL defect is detected by the macular GCIPL parameters, the GCIPL thickness map (arrow), the deviation-from-normal map (asterisk), the macular sector map (arrow), and the minimal GCIPL thickness (asterisk).
The pRNFL thickness within a 3.46-mm diameter circle, which was automatically positioned around the optic disc, was measured by the optic disc cube 200 × 200 protocol. The pRNFL thickness was then analyzed by 17 parameters that included the average thickness, four quadrant sectors, and 12 clock-hour sectors (Fig. 2C). 17  
In this study, clock-hour sectors were assessed in a clockwise direction for the right eye and in a counterclockwise direction for left eye, with the temporal sector set at 9 o'clock. For example, the 12 o'clock sector corresponds to the superior sector, 3 o'clock sector to the nasal, 6 o'clock sector to the inferior, and 9 o'clock sector to the temporal, in both eyes. ONH parameters, such as the rim area, the cup-to-disc area ratio (CDR), and the vertical cup-to-disc diameter ratio (VCDR) were measured by an internal ONH analysis algorithm (Fig. 2C). 17  
For the six-sector map of macular GCIPL and for the four quadrants and 12 clock-hour map of pRNFL, the yellow color code or red color code was defined as the detection of RNFL defects at the 5% abnormality level and 1% abnormality level, respectively, when it corresponded to the location of the RNFL defect on the red-free fundus photography. For the deviation-from-normal map, the detection of RNFL defects was defined as when the cluster of five or more contiguous yellow or red superpixels corresponded to the location of the RNFL defect on the red-free fundus photograph. If the cluster was composed of both yellow and red superpixels, it was defined as the detection of RNFL defects abnormal at the 1% level based on an internal normative database. 
Statistical Analysis
Statistical analyses were performed by using SPSS version 19.0 software (SPSS, Inc., Chicago, IL, USA) and MedCalc 12.3.0 software (MedCalc Software, Mariakerke, Belgium). The diagnostic ability of each GCIPL parameter was determined by computing the areas under the receiver operating characteristic curves (AUROCs) and the sensitivities/specificities calculated based on the internal normative database. The AUROCs of the GCIPL parameters were compared with the AUROCs of other Cirrus OCT parameters by using the methods described by DeLong et al. 18 A linear-by-linear association test and a logistic regression test were performed to assess the relationship between the diagnostic ability of GCIPL parameters to detect RNFL defects and the topographic profiles of RNFL defects. A P value less than 0.05 was considered statistically significant. 
Results
During the enrollment period, 209 eyes were initially involved (117 eyes with PPG and 92 normal control eyes). Of the 117 eyes with PPG, four eyes were excluded from further analysis because of ambiguous RNFL defects. Therefore, the study sample involved 205 eyes (113 eyes with PPG and 92 normal control eyes). After age-matching between the two groups, this study finally included 184 eyes of 184 subjects (92 eyes of 92 subjects with PPG and 92 eyes of 92 age-matched healthy control subjects). 
Basic Characteristics of the Study Subjects
Table 1 presents the demographic and background characteristics. All study subjects were Korean. We accomplished age-matching; therefore, there was no significant difference in the mean age between the healthy control subjects (57.6 ± 11.3 years; range, 26–81 years) and the patients with PPG (57.8 ± 11.4 years; range, 26–83 years) (P = 0.92). The sex, refraction, and BCVA of the patients were also not significantly different between the two groups (for all, P > 0.05). Because we defined PPG as eyes having normal VF results, there was no significant difference in the global VF indices between the two groups (P > 0.05). 
Table 1
 
Demographic Characteristics of the Study Subjects
Table 1
 
Demographic Characteristics of the Study Subjects
Healthy Control, n = 92 Preperimetric Glaucoma, n = 92 P Value
Age, y 57.6 ± 11.3 57.8 ± 11.4 0.920*
Sex, male:female 50:42 45:47 0.461†
BCVA, logMAR 0.04 ± 0.64 0.03 ± 0.64 0.634*
Spherical equivalents, D −0.80 ± 1.62 −0.37 ± 1.62 0.214*
HVF C30-2 threshold
 Mean deviation, dB −0.45 ± 1.86 −0.16 ± 1.61 0.287*
 PSD, dB 1.92 ± 0.66 1.99 ± 0.86 0.552*
Diagnostic Performance of Macular GCIPL, pRNFL, and ONH Parameters
A comparison of OCT parameters between two groups showed that all parameters of three categories were significantly reduced in the PPG group than in the healthy control group (for all, P < 0.05), except at the 9 o'clock sector of pRNFL (P = 0.531) (Supplementary Table S1). Table 2 shows the AUROCs of all Cirrus OCT parameters for diagnosing PPG. Of the eight GCIPL parameters, the inferotemporal sector (AUROC = 0.823), and the minimum (AUROC = 0.821) and the average (AUROC = 0.744) thickness showed the best diagnostic performances. There was no significant difference in the AUROCs of these three best GCIPL parameters (for all, P > 0.05). 
Table 2
 
Discriminating Ability of Macular GCIPL, pRNFL, and ONH Parameter for Preperimetric Glaucoma
Table 2
 
Discriminating Ability of Macular GCIPL, pRNFL, and ONH Parameter for Preperimetric Glaucoma
AUROC (95% CI) Specificity ≥ 80%
Sensitivity Specificity
GCIPL, μm
 Average 0.744 (0.674–0.814) 40.2 87.0
 Minimum 0.821 (0.761–0.880) 65.2 83.7
 Superotemporal 0.707 (0.632–0.781) 42.4 80.4
 Superior 0.642 (0.562–0.722) 21.7 82.6
 Superonasal 0.631 (0.551–0.711) 29.3 81.5
 Inferonasal 0.653 (0.575–0.732) 35.9 80.4
 Inferior 0.727 (0.655–0.799) 45.7 81.5
 Inferotemporal 0.823 (0.763–0.884) 58.7 81.5
RNFL, μm
 Average 0.752 (0.683–0.821) 46.7 83.7
Quadrant
 Superior 0.689 (0.613–0.765) 35.9 81.5
 Nasal 0.675 (0.597–0.752) 47.8 82.6
 Inferior 0.744 (0.542–0.815) 87.0 89.0
 Temporal 0.623 (0.542–0.704) 23.9 81.5
Clock-hours*
 12 Superior 0.584 (0.501–0.667) 31.5 80.4
 11 0.712 (0.639–0.785) 46.7 82.6
 10 0.619 (0.538–0.700) 23.9 80.4
 9 Temporal 0.549 (0.466–0.633) 20.7 80.4
 8 0.607 (0.526–0.689) 29.3 80.4
 7 0.764 (0.694–0.833) 46.7 81.5
 6 Inferior 0.668 (0.589–0.747) 26.1 80.4
 5 0.604 (0.522–0.686) 21.7 80.4
 4 0.641 (0.562–0.720) 39.1 81.5
 3 Nasal 0.612 (0.530–0.694) 25.0 80.4
 2 0.636 (0.555–0.716) 31.5 80.4
 1 0.650 (0.571–0.729) 35.9 81.5
ONH measurements
 CDR 0.686 (0.609–0.764) 47.8 81.5
 VCDR 0.714 (0.639–0.789) 53.3 80.4
 Rim area 0.767 (0.698–0.835) 50.0 80.4
There was no significant difference in the AUROCs among the GCIPL, pRNFL, and ONH parameters (which were above 0.7) (for all, P > 0.05); except, the minimum and the inferotemporal sectors of GCIPL thickness showed significantly higher AUROCs than did the inferior quadrant and the 11 o'clock sector of pRNFL, and the VCDR (for all, P < 0.05). There was no significant difference in the AUROCs for the single best diagnostic parameters of GCIPL (inferotemporal sector, 0.823), pRNFL (7 o'clock sector, 0.764), and ONH (rim area, 0.767) (for all, P > 0.05) (Table 3). 
Table 3
 
Comparison of the AUROC Curves of the Various OCT Parameters
Table 3
 
Comparison of the AUROC Curves of the Various OCT Parameters
P Value* GCIPL Parameters
Average Minimum Superotemporal Inferior Inferotemporal
pRNFL parameters
 Average 0.816 0.056 0.277 0.491 0.059
 Inferior quadrant 0.994 0.044 0.444 0.669 0.025
 7-o'clock sectors 0.643 0.155 0.250 0.380 0.094
 11-o'clock sectors 0.535 0.022 0.925 0.774 0.022
ONH parameters
 Rim area 0.564 0.164 0.177 0.338 0.151
 VCDR 0.568 0.027 0.889 0.802 0.026
Sensitivity and Specificity, Based on an Internal Normative Database
On red-free fundus photography, 73 eyes of the PPG group showed a single localized RNFL defect that was confined to one hemiretina (i.e., superior or inferior hemiretina). The other 19 eyes showed dual lesions of localized RNFL defects. Therefore, there were 111 cases of localized RNFL defects in total. Table 4 presents the sensitivity and specificity of the GCIPL and pRNFL parameters in detecting these 111 cases of RNFL defects. Based on an internal normative database, the sensitivity of the GCIPL parameters was generally higher than the sensitivity of pRNFL parameters, with the exception of the average GCIPL thickness at the 5% level and the GCIPL deviation-from-normal map at the 1% level. The deviation-from-normal maps at the 5% level (79.3%) and at the 1% level (60.4%) showed the highest sensitivity among the overall GCIPL parameters. 
Table 4
 
Overall Sensitivity and Specificity for the Macular GCIPL and pRNFL Parameters, Based on an Internal Normative Database
Table 4
 
Overall Sensitivity and Specificity for the Macular GCIPL and pRNFL Parameters, Based on an Internal Normative Database
Sensitivity, % (95% CI) Specificity, % (95% CI)
Macular GCIPL
 Macular sectors
  Abnormal at the 5% level 59.5 (49.7–68.5) 96.7 (90.1–99.2)
  Abnormal at the 1% level 44.1 (34.8–53.9) 98.9 (93.2–99.9)
 Average GCIPL thickness
  Abnormal at the 5% level 27.9 (20.0–37.4) 100 (95.0–100.0)
  Abnormal at the 1% level 19.8 (13.1–28.7) 100 (95.0–100.0)
 Deviation-from-normal map
  Abnormal at the 5% level 79.3 (70.3–86.2) 84.8 (75.4–91.1)
  Abnormal at the 1% level 60.4 (50.6–69.4) 93.5 (85.8–97.3)
pRNFL
 ≥1 clock-hour
  Abnormal at the 5% level 47.7 (38.3–57.4) 88.0 (79.2–93.6)
  Abnormal at the 1% level 23.4 (16.1–32.6) 97.8 (91.6–99.6)
 ≥1 quadrant
  Abnormal at the 5% level 37.8 (29.0–47.6) 98.9 (93.2–99.9)
  Abnormal at the 1% level 21.6 (14.6–30.6) 98.9 (93.2–99.9)
 Average RNFL thickness
  Abnormal at the 5% level 30.6 (22.4–40.2) 100 (95.0–100.0)
  Abnormal at the 1% level 12.6 (7.3–20.6) 100 (95.0–100.0)
 Deviation-from-normal map
  Abnormal at the 5% level 74.8 (65.5–82.3) 93.5 (85.8–97.3)
  Abnormal at the 1% level 62.2 (52.4–71.0) 96.7 (90.1–99.2)
With the criterion of “abnormal” set at the 5% level, the specificity of the GCIPL parameters for detecting localized RNFL defects ranged from 84.8% to 100% and the specificity of the pRNFL parameters ranged from 88.0% to 100%. Figure 2 shows an example of a PPG patient with a localized RNFL defect. In this patient, the RNFL defect was detected on the macular GCIPL map, but not on the pRNFL map. 
The Influence of Topographic Profiles on Diagnostic Performance
We selected single localized RNFL defects for each PPG eye to measure their topographic profiles. For the 19 eyes with dual lesions of localized RNFL defects, the more dominant lesion of both was selected. Therefore, we measured the inner directional angle (α) and the peripapillary angular width (βα) in a total of 92 localized RNFL defects. In calculating the peripapillary angular width (βα), 29 localized RNFL defects were excluded because of an ambiguous outer direction line, which overlapped with the retinal vessels. 
The sensitivity of the macular GCIPL map in detecting localized RNFL defects was negatively associated with the inner directional angle (α) of RNFL defects (deviation-from-normal map, P = 0.036; sector map, P = 0.012). For example, the sensitivities of the GCIPL deviation-from-normal map with an inner directional angle (α) of less than 30°, 30° to 59.9°, and greater than 60° were 100.0%, 84.1%, and 68.2%, respectively. However, the sensitivity of the pRNFL map was not significantly associated with the inner directional angle (α) (for all, P > 0.05) (Table 5). 
Table 5
 
Inner Directional Angle of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Table 5
 
Inner Directional Angle of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Inner Directional Angle, Degrees Number of RNFL Defects Deviation-From-Normal Map Topographic Territory Map
Sector Map Clock-Hour Map§ Quadrant Map||
Sensitivity of GCIPL, % Sensitivity of pRNFL, % Sensitivity of GCIPL, % Sensitivity of pRNFL, %
<30 7 100.0 85.7 85.7 57.1 14.3
30–59.9 63 84.1 76.2 66.7 50.8 38.1
>60 22 68.2 63.6 40.9 36.4 18.2
Total 92 81.5 73.9 62.0 47.8 31.5
P value* 0.036 0.175 0.012 0.22 0.473
For the peripapillary angular widths (βα) of RNFL defects, there was a trend toward greater angular widths corresponding to more sensitive detection of RNFL defects in both parameters. However, the trend for the linear association was not statistically significant (for all, P > 0.05), except for the clock-hour pRNFL map (P = 0.017) (Table 6). 
Table 6
 
Peripapillary Angular Width of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Table 6
 
Peripapillary Angular Width of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Inner Directional Angle, Degrees Number of RNFL Defects Deviation-From-Normal Map Topographic Territory Map
Macular Sectors Map Clock-Hour Map§ Quadrant Map||
Sensitivity of GCIPL, % Sensitivity of pRNFL, % Sensitivity of GCIPL, % Sensitivity of pRNFL, %
<20 17 64.7 58.8 47.1 29.4 17.6
20–39.9 34 79.4 73.5 61.8 50.0 29.4
>40 12 91.7 75.0 66.7 75.0 41.7
Total 63 77.8 69.8 58.7 49.2 28.6
P value* 0.083 0.314 0.269 0.017 0.159
Multiple logistic regression analysis was performed to assess the independent influence of the inner direction angle (α) and the angular width (βα) on the diagnostic performance of the GCIPL parameters. We found a significant association between the inner directional angle (α) and the sensitivity of the GCIPL parameters (deviation-from-normal map, P = 0.023; sector map, P = 0.009). By contrast, the angular width (βα) did not show a significant association with the sensitivity of the GCIPL parameters (for both, P > 0.05) (Table 7). Figure 3 shows the cases of preperimetric localized RNFL defects with different inner directional angle. 
Figure 3
 
Preperimetric RNFL defects with different inner directional angles. (A) A superior temporal RNFL defect on red-free fundus photography (arrowheads) with a large inner directional angle (α = 67°) was not detected by the macular GCIPL deviation-from-normal map and the macular GCIPL sector map. This defect was detected only by the pRNFL deviation-from-normal map (asterisk). This suggests that if the temporal margin of RNFL defect is far from the fovea, the sensitivity of the macular GCIPL map is low. In this circumstance, the deviation-from-normal map of the pRNFL is a sensitive parameter for identifying preperimetric RNFL defects. (B) An inferior temporal RNFL defect on red-free fundus photography (arrowheads) with a small inner direction angle (α = 37°) is detected in the deviation-from-normal map (asterisk) and in the sector map (arrow) of the macular GCIPL. However, the pRNFL map shows unreliable results that do not correspond to the RNFL defect on red-free fundus photography. This circumstance suggests that if the temporal margin of RNFL defect is close to the fovea, the macular GCIPL parameters are more sensitive than the pRNFL parameters.
Figure 3
 
Preperimetric RNFL defects with different inner directional angles. (A) A superior temporal RNFL defect on red-free fundus photography (arrowheads) with a large inner directional angle (α = 67°) was not detected by the macular GCIPL deviation-from-normal map and the macular GCIPL sector map. This defect was detected only by the pRNFL deviation-from-normal map (asterisk). This suggests that if the temporal margin of RNFL defect is far from the fovea, the sensitivity of the macular GCIPL map is low. In this circumstance, the deviation-from-normal map of the pRNFL is a sensitive parameter for identifying preperimetric RNFL defects. (B) An inferior temporal RNFL defect on red-free fundus photography (arrowheads) with a small inner direction angle (α = 37°) is detected in the deviation-from-normal map (asterisk) and in the sector map (arrow) of the macular GCIPL. However, the pRNFL map shows unreliable results that do not correspond to the RNFL defect on red-free fundus photography. This circumstance suggests that if the temporal margin of RNFL defect is close to the fovea, the macular GCIPL parameters are more sensitive than the pRNFL parameters.
Table 7
 
Results of Binary Logistic Regression Analysis for the Topographic Profiles of RNFL Defects and the Diagnostic Performance of the Macular GCIPL Parameters
Table 7
 
Results of Binary Logistic Regression Analysis for the Topographic Profiles of RNFL Defects and the Diagnostic Performance of the Macular GCIPL Parameters
Variables GCIPL Deviation-From-Normal Map* GCIPL Sector Map
β SE Wald Statistics OR (95% CI) P Value β SE Wald Statistics OR (95% CI) P Value
Inner direction angle, deg −0.1 0.04 5.178 0.91 (0.84–0.99) 0.023 −0.09 0.03 6.85 0.92 (0.86–0.98) 0.009
Angular width, deg 0.01 0.03 0.073 1.01 (0.95–1.06) 0.787 −0.02 0.02 0.38 0.99 (0.94–1.03) 0.539
Discussion
In the present study, we confirmed that the diagnostic ability of macular GCIPL parameters was comparable to the diagnostic ability of pRNFL and ONH parameters in patients with PPG. We also found that the sensitivity of the macular GCIPL parameters for detecting RNFL defects was significantly influenced by the inner directional angle of RNFL defects, but not by the peripapillary angular width of RNLF defects. 
There have been numerous studies that have introduced the automatic segmentation of inner retinal layers by using various types of spectral domain OCT (SD-OCT). 5,1929 It has been reported that SD-OCT can successfully segment and measure macular GCIPL thickness with excellent reproducibility. 5 Na et al. 9 reported a significant difference in the macular GCIPL thickness between healthy control eyes and eyes with PPG, which suggested that macular GCIPL thickness might serve as an early indicator of glaucomatous structural damage. Furthermore, Mwanza et al. 7 reported that the ability of GCIPL parameters to diagnose early glaucoma was high and comparable to that of pRNFL and ONH parameters. In addition, Takayama and colleagues 10 reported that the AUROCs of minimum GCIPL thickness were comparable to the average pRNFL thickness for early glaucoma. Nouri-Mahdavi et al. 11 also reported that the best regional measures from each algorithm showed comparable performances for detection of early glaucoma. 
In our results, the AUROCs of GCIPL parameters were comparable to those of the pRNFL parameters and the ONH parameters. This result was somewhat similar to that of previous studies performed mostly in early glaucoma. However, the values of AUROCs in this study were relatively lower than those of previous studies. 7,10,11 We speculate that this was because we included only patients with PPG. Because the diagnostic performance of imaging devices is highly affected by the severity of disease, 30 the relatively low diagnostic performance may be related to the very early stage of glaucomatous damage in study subjects. 
Because GCIPL parameters were computed from the data of a restricted scan area (i.e., an elliptical annulus centered on the fovea), the sensitivity for detecting RGC loss might be affected by where the RGC loss is mostly located. As a clue to this hypothesis, Shin and colleagues 31 recently reported that the average and the minimum GCIPL thickness demonstrated superior diagnostic performance in comparison with the average pRNFL thickness, in eyes with parafoveal VF loss. However, the opposite was observed in eyes with peripheral VF loss. They suggest that, because RNFL defects are located closer to the fovea in eyes with parafoveal VF loss than in eyes with peripheral VF loss, the GCIPL thickness is more remarkably decreased in eyes with parafoveal VF loss. However, in eyes with PPG, it is impossible to consider the location of VF loss. Therefore, we investigated whether there was a direct correlation between the location of RNFL defects and the diagnostic performance of GCIPL parameters. In the present study, the sensitivity of the GCIPL parameters was negatively associated with the inner directional angle of RNFL defects. This finding suggests that if the RNFL defects are located closer to the fovea, there are more chances for a macular GCIPL scan to detect them. This was consistent with the findings in a previous report by Shin et al. 31  
By contrast, the peripapillary angular width of RNFL defects was not significantly associated with the sensitivity of the GCIPL parameters. This was opposite of the results of previous reports that showed the angular width had a significant influence on the ability of pRNFL parameters to detect RNFL defects. 12 We speculate that these conflicting results are caused by a discrepancy in the scan area of the macular GCIPL and the pRNFL. Because the scan area of the macular GCIPL was an annulus centered on the fovea, the distance of an RNFL defect from the fovea might be a primary determining factor for detecting these defects by the GCIPL scan. By contrast, the scan area of pRNFL was a circle centered on the optic disc; therefore, the angular width of RNFL defects might be a primary determining factor for detecting these defects by the pRNFL scan. 
In this study, the angular width of RNFL defects was significantly associated with the sensitivity of the pRNFL clock-hour map, but not with the sensitivity of the pRNFL deviation-from-normal map and the pRNFL quadrant map. It appears that the relatively insufficient number of cases for the statistical analysis causes these conflicting results. Our observation nevertheless suggests that greater angular widths correspond to more sensitive detection of RNFL defects by the pRNFL scan. To clarify this matter, further stratified investigations with a large number of patients are needed. 
This study has several limitations. First, in the enrollment of the healthy control group, some subjects who have glaucomalike discs may have been included. The reason is we recruited the study subjects from the glaucoma clinic, rather than from the general population. Therefore, it is possible that the calculated AUROCs, sensitivity, and specificity may have been underestimated. Second, as mentioned previously, a fair number of localized RNFL defects were excluded from the analysis because of the ambiguity in defining their boundary. This might result in an insufficient statistical power. Further verification with a large number of subjects is required. Third, all subjects were enrolled from the Korean population. Considering the RNFL thickness can vary according to race/ethnicity, 32,33 the RNFL thickness in Korean individuals may differ from that in other ethnic groups. Because the internal normative database of Cirrus OCT is primarily compiled from Caucasian subjects, the sensitivity and specificity computed on the basis of a normative database might be somewhat affected by this racial difference. To exclude this effect, a race-specific normative database may be necessary. Fourth, we excluded patients with high ametropia. To represent the general population, further studies, including a wide range of refractive errors and varying races, will be necessary. 
In summary, this study demonstrates that the diagnostic ability of macular GCIPL parameters was comparable to the diagnostic ability of pRNFL and ONH parameters in patients with PPG. Our results suggest that the inner directional angles of RNFL defects affect the diagnostic sensitivity of macular GCIPL parameters. Because the location of glaucomatous damage has a significant influence on the diagnostic performance of imaging devices, clinicians should be aware of these issues when interpreting OCT results in the relevant population. 
Supplementary Materials
Acknowledgments
Supported by Grant A121615 from the Korea Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea. The authors alone are responsible for the content and writing of the paper. 
Disclosure: M.J. Kim, None; J.W. Jeoung, None; K.H. Park, None; Y.J. Choi, None; D.M. Kim, None 
References
Quigley HA Dunkelberger GR Green WR. Retinal ganglion cell atrophy correlated with automated perimetry in human eyes with glaucoma. Am J Ophthalmol . 1989; 107: 453–464. [CrossRef] [PubMed]
Sommer A Katz J Quigley HA Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol . 1991; 109: 77–83. [CrossRef] [PubMed]
Tuulonen A Airaksinen PJ. Initial glaucomatous optic disk and retinal nerve fiber layer abnormalities and their progression. Am J Ophthalmol . 1991; 111: 485–490. [CrossRef] [PubMed]
Tuulonen A Lehtola J Airaksinen PJ. Nerve fiber layer defects with normal visual fields. Do normal optic disc and normal visual field indicate absence of glaucomatous abnormality? Ophthalmology . 1993; 100: 587–597.
Mwanza JC Oakley JD Budenz DL Macular ganglion cell-inner plexiform layer: automated detection and thickness reproducibility with spectral domain-optical coherence tomography in glaucoma. Invest Ophthalmol Vis Sci . 2011; 52: 8323–8329. [CrossRef] [PubMed]
Mwanza JC Durbin MK Budenz DL Profile and predictors of normal ganglion cell-inner plexiform layer thickness measured with frequency-domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2011; 52: 7872–7879. [CrossRef] [PubMed]
Mwanza JC Durbin MK Budenz DL Glaucoma diagnostic accuracy of ganglion cell-inner plexiform layer thickness: comparison with nerve fiber layer and optic nerve head. Ophthalmology . 2012; 119: 1151–1158. [CrossRef] [PubMed]
Kotowski J Folio LS Wollstein G Glaucoma discrimination of segmented cirrus spectral domain optical coherence tomography (SD-OCT) macular scans. Br J Ophthalmol . 2012; 96: 1420–1425. [CrossRef] [PubMed]
Na JH Sung KR Baek S Detection of glaucoma progression by assessment of segmented macular thickness data obtained using spectral domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2012; 53: 3817–3826. [CrossRef] [PubMed]
Takayama K Hangai M Durbin M A novel method to detect local ganglion cell loss in early glaucoma using spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2012; 53: 6904–6913. [CrossRef] [PubMed]
Nouri-Mahdavi K Nowroozizadeh S Nassiri N Macular ganglion cell/inner plexiform layer measurements by spectral domain optical coherence tomography for detection of early glaucoma and comparison to retinal nerve fiber layer measurements. Am J Ophthalmol . 2013; 156: 1297–1307. [CrossRef] [PubMed]
Hwang YH Kim YY Kim HK Ability of cirrus high-definition spectral-domain optical coherence tomography clock-hour, deviation, and thickness maps in detecting photographic retinal nerve fiber layer abnormalities. Ophthalmology . 2013; 120: 1380–1387. [CrossRef] [PubMed]
Anderson DR Patella VM. Automated Static Perimetry . 2nd ed. St. Louis, MO: Mosby; 1999: 93, 121–136.
Hoyt WF Frisen L Newman NM. Fundoscopy of nerve fiber layer defects in glaucoma. Invest Ophthalmol . 1973; 12: 814–829. [PubMed]
Hwang JM Kim TW Park KH Correlation between topographic profiles of localized retinal nerve fiber layer defects as determined by optical coherence tomography and red-free fundus photography. J Glaucoma . 2006; 15: 223–228. [CrossRef] [PubMed]
Yoo YC Park KH. Influence of angular width and peripapillary position of localized retinal nerve fiber layer defects on their detection by time-domain optical coherence tomography. Jpn J Ophthalmol . 2011; 55: 115–122. [CrossRef] [PubMed]
Savini G Carbonelli M Barboni P. Spectral-domain optical coherence tomography for the diagnosis and follow-up of glaucoma. Curr Opin Ophthalmol . 2011; 22: 115–123. [CrossRef] [PubMed]
DeLong ER DeLong DM Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics . 1988; 44: 837–845. [CrossRef] [PubMed]
Arintawati P Sone T Akita T The applicability of ganglion cell complex parameters determined from SD-OCT images to detect glaucomatous eyes. J Glaucoma . 2013; 22: 713–718. [CrossRef] [PubMed]
Takagi ST Kita Y Yagi F Macular retinal ganglion cell complex damage in the apparently normal visual field of glaucomatous eyes with hemifield defects. J Glaucoma . 2012; 21: 318–325. [CrossRef] [PubMed]
Rolle T Briamonte C Curto D Ganglion cell complex and retinal nerve fiber layer measured by fourier-domain optical coherence tomography for early detection of structural damage in patients with preperimetric glaucoma. Clin Ophthalmol . 2011; 5: 961–969. [CrossRef] [PubMed]
Rao HL Babu JG Addepalli UK Retinal nerve fiber layer and macular inner retina measurements by spectral domain optical coherence tomograph in Indian eyes with early glaucoma. Eye (Lond) . 2012; 26: 133–139. [CrossRef] [PubMed]
Seong M Sung KR Choi EH Macular and peripapillary retinal nerve fiber layer measurements by spectral domain optical coherence tomography in normal-tension glaucoma. Invest Ophthalmol Vis Sci . 2010; 51: 1446–1452. [CrossRef] [PubMed]
Mori S Hangai M Sakamoto A Spectral-domain optical coherence tomography measurement of macular volume for diagnosing glaucoma. J Glaucoma . 2010; 19: 528–534. [CrossRef] [PubMed]
Sevim MS Buttanri B Acar BT Ability of fourier-domain optical coherence tomography to detect retinal ganglion cell complex atrophy in glaucoma patients. J Glaucoma . 2013; 22: 542–549. [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]
Chen J Huang H Wang M Fourier domain OCT measurement of macular, macular ganglion cell complex, and peripapillary RNFL thickness in glaucomatous Chinese eyes. Eur J Ophthalmol . 2012; 22: 972–979. [CrossRef] [PubMed]
Kotera Y Hangai M Hirose F Three-dimensional imaging of macular inner structures in glaucoma by using spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2011; 52: 1412–1421. [CrossRef] [PubMed]
Lisboa R Paranhos A Jr Weinreb RN Comparison of different spectral domain OCT scanning protocols for diagnosing preperimetric glaucoma. Invest Ophthalmol Vis Sci . 2013; 54: 3417–3425. [CrossRef] [PubMed]
Leite MT Zangwill LM Weinreb RN Effect of disease severity on the performance of Cirrus spectral-domain OCT for glaucoma diagnosis. Invest Ophthalmol Vis Sci . 2010; 51: 4104–4109. [CrossRef] [PubMed]
Shin HY Park HY Jung KI Glaucoma diagnostic ability of ganglion cell-inner plexiform layer thickness differs according to the location of visual field loss. Ophthalmology . 2013; 17: 578–572.
Knight OJ Girkin CA Budenz DL Effect of race, age, and axial length on optic nerve head parameters and retinal nerve fiber layer thickness measured by Cirrus HD-OCT. Arch Ophthalmol . 2012; 130: 312–318. [CrossRef] [PubMed]
Alasil T Wang K Keane PA Analysis of normal retinal nerve fiber layer thickness by age, sex, and race using spectral domain optical coherence tomography. J Glaucoma . 2013; 22: 532–541. [CrossRef] [PubMed]
Figure 1
 
The procedure to determine the peripapillary angular width and the directional angle of the RNFL defect. On red-free fundus photography, clock-face circles are drawn. The diameter and location of the circles are positioned as closely as possible to the scan circle by Cirrus OCT (Carl Zeiss Meditec). The point at which the RNFL defect intersects this circle at the inner and outer borders determines the two edge points. The inner (i.e., temporal) directional lines (a) and the outer (i.e., nasal) directional lines (b), which pass through the center of circle and each edge point, respectively, are drawn. The inner directional angle (α) and the outer directional angle (β) are then measured between the each directional line and the temporal equator (which was set as 0°) in a clockwise direction in right eyes and in a counterclockwise direction in left eyes. The peripapillary angular width of the RNFL defect is defined by subtracting the inner directional angle (α) from the outer directional angle (β) (i.e., βα).
Figure 1
 
The procedure to determine the peripapillary angular width and the directional angle of the RNFL defect. On red-free fundus photography, clock-face circles are drawn. The diameter and location of the circles are positioned as closely as possible to the scan circle by Cirrus OCT (Carl Zeiss Meditec). The point at which the RNFL defect intersects this circle at the inner and outer borders determines the two edge points. The inner (i.e., temporal) directional lines (a) and the outer (i.e., nasal) directional lines (b), which pass through the center of circle and each edge point, respectively, are drawn. The inner directional angle (α) and the outer directional angle (β) are then measured between the each directional line and the temporal equator (which was set as 0°) in a clockwise direction in right eyes and in a counterclockwise direction in left eyes. The peripapillary angular width of the RNFL defect is defined by subtracting the inner directional angle (α) from the outer directional angle (β) (i.e., βα).
Figure 2
 
A case of preperimetric glaucoma with a localized RNFL defect. (A) Red-free fundus photography shows the inferior temporal localized RNFL defect (arrowheads). (B) The HVF test indicates normal visual field. (C) The pRNFL map does not detect the RNFL defect. (D) However, the RNFL defect is detected by the macular GCIPL parameters, the GCIPL thickness map (arrow), the deviation-from-normal map (asterisk), the macular sector map (arrow), and the minimal GCIPL thickness (asterisk).
Figure 2
 
A case of preperimetric glaucoma with a localized RNFL defect. (A) Red-free fundus photography shows the inferior temporal localized RNFL defect (arrowheads). (B) The HVF test indicates normal visual field. (C) The pRNFL map does not detect the RNFL defect. (D) However, the RNFL defect is detected by the macular GCIPL parameters, the GCIPL thickness map (arrow), the deviation-from-normal map (asterisk), the macular sector map (arrow), and the minimal GCIPL thickness (asterisk).
Figure 3
 
Preperimetric RNFL defects with different inner directional angles. (A) A superior temporal RNFL defect on red-free fundus photography (arrowheads) with a large inner directional angle (α = 67°) was not detected by the macular GCIPL deviation-from-normal map and the macular GCIPL sector map. This defect was detected only by the pRNFL deviation-from-normal map (asterisk). This suggests that if the temporal margin of RNFL defect is far from the fovea, the sensitivity of the macular GCIPL map is low. In this circumstance, the deviation-from-normal map of the pRNFL is a sensitive parameter for identifying preperimetric RNFL defects. (B) An inferior temporal RNFL defect on red-free fundus photography (arrowheads) with a small inner direction angle (α = 37°) is detected in the deviation-from-normal map (asterisk) and in the sector map (arrow) of the macular GCIPL. However, the pRNFL map shows unreliable results that do not correspond to the RNFL defect on red-free fundus photography. This circumstance suggests that if the temporal margin of RNFL defect is close to the fovea, the macular GCIPL parameters are more sensitive than the pRNFL parameters.
Figure 3
 
Preperimetric RNFL defects with different inner directional angles. (A) A superior temporal RNFL defect on red-free fundus photography (arrowheads) with a large inner directional angle (α = 67°) was not detected by the macular GCIPL deviation-from-normal map and the macular GCIPL sector map. This defect was detected only by the pRNFL deviation-from-normal map (asterisk). This suggests that if the temporal margin of RNFL defect is far from the fovea, the sensitivity of the macular GCIPL map is low. In this circumstance, the deviation-from-normal map of the pRNFL is a sensitive parameter for identifying preperimetric RNFL defects. (B) An inferior temporal RNFL defect on red-free fundus photography (arrowheads) with a small inner direction angle (α = 37°) is detected in the deviation-from-normal map (asterisk) and in the sector map (arrow) of the macular GCIPL. However, the pRNFL map shows unreliable results that do not correspond to the RNFL defect on red-free fundus photography. This circumstance suggests that if the temporal margin of RNFL defect is close to the fovea, the macular GCIPL parameters are more sensitive than the pRNFL parameters.
Table 1
 
Demographic Characteristics of the Study Subjects
Table 1
 
Demographic Characteristics of the Study Subjects
Healthy Control, n = 92 Preperimetric Glaucoma, n = 92 P Value
Age, y 57.6 ± 11.3 57.8 ± 11.4 0.920*
Sex, male:female 50:42 45:47 0.461†
BCVA, logMAR 0.04 ± 0.64 0.03 ± 0.64 0.634*
Spherical equivalents, D −0.80 ± 1.62 −0.37 ± 1.62 0.214*
HVF C30-2 threshold
 Mean deviation, dB −0.45 ± 1.86 −0.16 ± 1.61 0.287*
 PSD, dB 1.92 ± 0.66 1.99 ± 0.86 0.552*
Table 2
 
Discriminating Ability of Macular GCIPL, pRNFL, and ONH Parameter for Preperimetric Glaucoma
Table 2
 
Discriminating Ability of Macular GCIPL, pRNFL, and ONH Parameter for Preperimetric Glaucoma
AUROC (95% CI) Specificity ≥ 80%
Sensitivity Specificity
GCIPL, μm
 Average 0.744 (0.674–0.814) 40.2 87.0
 Minimum 0.821 (0.761–0.880) 65.2 83.7
 Superotemporal 0.707 (0.632–0.781) 42.4 80.4
 Superior 0.642 (0.562–0.722) 21.7 82.6
 Superonasal 0.631 (0.551–0.711) 29.3 81.5
 Inferonasal 0.653 (0.575–0.732) 35.9 80.4
 Inferior 0.727 (0.655–0.799) 45.7 81.5
 Inferotemporal 0.823 (0.763–0.884) 58.7 81.5
RNFL, μm
 Average 0.752 (0.683–0.821) 46.7 83.7
Quadrant
 Superior 0.689 (0.613–0.765) 35.9 81.5
 Nasal 0.675 (0.597–0.752) 47.8 82.6
 Inferior 0.744 (0.542–0.815) 87.0 89.0
 Temporal 0.623 (0.542–0.704) 23.9 81.5
Clock-hours*
 12 Superior 0.584 (0.501–0.667) 31.5 80.4
 11 0.712 (0.639–0.785) 46.7 82.6
 10 0.619 (0.538–0.700) 23.9 80.4
 9 Temporal 0.549 (0.466–0.633) 20.7 80.4
 8 0.607 (0.526–0.689) 29.3 80.4
 7 0.764 (0.694–0.833) 46.7 81.5
 6 Inferior 0.668 (0.589–0.747) 26.1 80.4
 5 0.604 (0.522–0.686) 21.7 80.4
 4 0.641 (0.562–0.720) 39.1 81.5
 3 Nasal 0.612 (0.530–0.694) 25.0 80.4
 2 0.636 (0.555–0.716) 31.5 80.4
 1 0.650 (0.571–0.729) 35.9 81.5
ONH measurements
 CDR 0.686 (0.609–0.764) 47.8 81.5
 VCDR 0.714 (0.639–0.789) 53.3 80.4
 Rim area 0.767 (0.698–0.835) 50.0 80.4
Table 3
 
Comparison of the AUROC Curves of the Various OCT Parameters
Table 3
 
Comparison of the AUROC Curves of the Various OCT Parameters
P Value* GCIPL Parameters
Average Minimum Superotemporal Inferior Inferotemporal
pRNFL parameters
 Average 0.816 0.056 0.277 0.491 0.059
 Inferior quadrant 0.994 0.044 0.444 0.669 0.025
 7-o'clock sectors 0.643 0.155 0.250 0.380 0.094
 11-o'clock sectors 0.535 0.022 0.925 0.774 0.022
ONH parameters
 Rim area 0.564 0.164 0.177 0.338 0.151
 VCDR 0.568 0.027 0.889 0.802 0.026
Table 4
 
Overall Sensitivity and Specificity for the Macular GCIPL and pRNFL Parameters, Based on an Internal Normative Database
Table 4
 
Overall Sensitivity and Specificity for the Macular GCIPL and pRNFL Parameters, Based on an Internal Normative Database
Sensitivity, % (95% CI) Specificity, % (95% CI)
Macular GCIPL
 Macular sectors
  Abnormal at the 5% level 59.5 (49.7–68.5) 96.7 (90.1–99.2)
  Abnormal at the 1% level 44.1 (34.8–53.9) 98.9 (93.2–99.9)
 Average GCIPL thickness
  Abnormal at the 5% level 27.9 (20.0–37.4) 100 (95.0–100.0)
  Abnormal at the 1% level 19.8 (13.1–28.7) 100 (95.0–100.0)
 Deviation-from-normal map
  Abnormal at the 5% level 79.3 (70.3–86.2) 84.8 (75.4–91.1)
  Abnormal at the 1% level 60.4 (50.6–69.4) 93.5 (85.8–97.3)
pRNFL
 ≥1 clock-hour
  Abnormal at the 5% level 47.7 (38.3–57.4) 88.0 (79.2–93.6)
  Abnormal at the 1% level 23.4 (16.1–32.6) 97.8 (91.6–99.6)
 ≥1 quadrant
  Abnormal at the 5% level 37.8 (29.0–47.6) 98.9 (93.2–99.9)
  Abnormal at the 1% level 21.6 (14.6–30.6) 98.9 (93.2–99.9)
 Average RNFL thickness
  Abnormal at the 5% level 30.6 (22.4–40.2) 100 (95.0–100.0)
  Abnormal at the 1% level 12.6 (7.3–20.6) 100 (95.0–100.0)
 Deviation-from-normal map
  Abnormal at the 5% level 74.8 (65.5–82.3) 93.5 (85.8–97.3)
  Abnormal at the 1% level 62.2 (52.4–71.0) 96.7 (90.1–99.2)
Table 5
 
Inner Directional Angle of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Table 5
 
Inner Directional Angle of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Inner Directional Angle, Degrees Number of RNFL Defects Deviation-From-Normal Map Topographic Territory Map
Sector Map Clock-Hour Map§ Quadrant Map||
Sensitivity of GCIPL, % Sensitivity of pRNFL, % Sensitivity of GCIPL, % Sensitivity of pRNFL, %
<30 7 100.0 85.7 85.7 57.1 14.3
30–59.9 63 84.1 76.2 66.7 50.8 38.1
>60 22 68.2 63.6 40.9 36.4 18.2
Total 92 81.5 73.9 62.0 47.8 31.5
P value* 0.036 0.175 0.012 0.22 0.473
Table 6
 
Peripapillary Angular Width of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Table 6
 
Peripapillary Angular Width of RNFL Defects and the Diagnostic Sensitivity of Macular GCIPL and pRNFL Parameters
Inner Directional Angle, Degrees Number of RNFL Defects Deviation-From-Normal Map Topographic Territory Map
Macular Sectors Map Clock-Hour Map§ Quadrant Map||
Sensitivity of GCIPL, % Sensitivity of pRNFL, % Sensitivity of GCIPL, % Sensitivity of pRNFL, %
<20 17 64.7 58.8 47.1 29.4 17.6
20–39.9 34 79.4 73.5 61.8 50.0 29.4
>40 12 91.7 75.0 66.7 75.0 41.7
Total 63 77.8 69.8 58.7 49.2 28.6
P value* 0.083 0.314 0.269 0.017 0.159
Table 7
 
Results of Binary Logistic Regression Analysis for the Topographic Profiles of RNFL Defects and the Diagnostic Performance of the Macular GCIPL Parameters
Table 7
 
Results of Binary Logistic Regression Analysis for the Topographic Profiles of RNFL Defects and the Diagnostic Performance of the Macular GCIPL Parameters
Variables GCIPL Deviation-From-Normal Map* GCIPL Sector Map
β SE Wald Statistics OR (95% CI) P Value β SE Wald Statistics OR (95% CI) P Value
Inner direction angle, deg −0.1 0.04 5.178 0.91 (0.84–0.99) 0.023 −0.09 0.03 6.85 0.92 (0.86–0.98) 0.009
Angular width, deg 0.01 0.03 0.073 1.01 (0.95–1.06) 0.787 −0.02 0.02 0.38 0.99 (0.94–1.03) 0.539
×
×

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.

×