Structure-Function Relationship of the Macular Visual Field Sensitivity and the Ganglion Cell Complex Thickness in Glaucoma

Abstract

Purpose.: We attempted to understand better the relationship between the macular visual field (VF) mean sensitivity (MS) assessed by standard automated perimetry (SAP) and the ganglion cell complex thickness (GCCT), and macular peripapillary retinal nerve fiber layer thickness (mpRNFLT) assessed by spectral domain optical coherence tomography (SD-OCT, RTVue-100) in open-angle glaucoma (OAG) patients.

Methods.: We enrolled in the study 217 OAG patients with baseline intraocular pressure (IOP) in the statistically normal range. GCCT and mpRNFLT measurements, using the ganglion cell complex (GCC) and the optic nerve head (ONH) modes of RTVue-100 OCT, were obtained for analysis. Macular VF sensitivity was recorded in the dB and 1/L scales. The relationship of the function (MS) and structure (GCCT, mpRNFLT) was sought globally and in two VF sectors (superior and inferior).

Results.: The relationship of the macular VF sensitivity (dB) to the GCC, and mpRNFL global (R ² = 0.111, 0.127) and sectoral (superior R ² = 0.358, 0.171; inferior R ² = 0.227, 0.263) thicknesses were statistically significant (all P < 0.05). The relationship of the macular VF sensitivity to the GCCT differed significantly from that of the macular VF sensitivity to the mpRNFL in the superior VF sector (R ² = 0.358 vs. 0.171, P < 0.05).

Conclusions.: GCCT determined by SD-OCT (RTVue-100) showed a statistically significant structure-function association with macular VF, and the strength of the association was greater than that of the mpRNFL with macular VF in the superior central VF area.

Introduction

The macular visual field (VF), including central vision, is of paramount importance if a glaucoma patient is to enjoy normal daily activities. In a recent study, glaucoma patients assigned the greatest importance to tasks, for example reading, that use perifoveal vision. ^1,2 Therefore, preservation of the macular VF is a key concern in glaucoma management. 

The ability to detect structural changes in the macular region may be valuable to identify or predict the central VF defects associated with glaucoma. Previous studies have shown that eyes with normal-tension glaucoma (NTG) show VF defects that are dense and more central than those seen in patients with primary open-angle glaucoma (POAG), as determined by computerized threshold perimetry. ^3–5 This central macular VF area can be aligned with the region containing the macular peripapillary retinal nerve fiber layer (mpRNFL), as shown on the maps of Weber et al. ⁶ and Garway-Heath et al. ⁷  

Newer versions of optical coherence tomography (OCT) incorporating spectral-domain (SD) technology, offer higher scan resolution and increased scan speed compared to conventional time-domain (TD)-OCT. The RTVue-100 OCT (Optovue, Inc., Fremont, CA) is one commercially available OCT device using SD technology. RTVue-100 OCT incorporates a ganglion cell complex (GCC) scan mode to measure the inner macular retinal layer thickness from the internal limiting membrane to the inner plexiform layer, centered around the fovea and covering the central macula. Previous studies have demonstrated that GCC thickness (GCCT) measurements derived from GCC scan data were significantly lower in glaucomatous eyes with VF defects than in healthy eyes ^8–14 ; the glaucoma discrimination ability was similar to that afforded by measurement of the peripapillary retinal nerve fiber layer thickness (pRNFLT). These data suggested that macular thickness measurement by SD-OCT was sufficiently sensitive to be useful in the early detection of glaucoma, as compared to pRNFLT measurements. 

If a measurement tool is to be useful for glaucoma diagnosis and for tracking disease progression, the glaucomatous structure-function relationship should be tested and validated in any VF area of interest as it provides a practical way to predict how VF will behave in reference to structural change during the course of disease progression. A number of studies have reported high correlations between global VF sensitivity and pRNFLT/GCCT in glaucomatous subjects with the use of TD- and SD-OCT. ^9,15–17 However, limited information is available on the precise nature of the structure-function relationship between the macular VF assessed by standard automated perimetry (SAP) and that obtained using SD-OCT-derived macular measurement. Therefore, in our study we evaluated and compared the strength of the relationship between macular VF retinal sensitivity, and GCCT and mpRNFLT using the GCC and optic nerve head (ONH) modes of RTVue in a large series of open-angle glaucoma (OAG) eyes with intraocular pressure (IOP) in the statistically normal range (IOP ≤21 mm Hg) as these eyes constitute the predominant form of glaucoma in our part of Asia. We also compared the structure-function relationship of the extramacular VF to that of the macular VF area in the same cohort group. 

Methods

Subjects

All glaucoma subjects were recruited prospectively and in a consecutive manner, from the Glaucoma Clinic of Asan Medical Center in Seoul, Korea, between July 2010 and July 2011. On initial evaluation, each subject underwent a complete ophthalmologic examination, including medical, ocular, and family history; visual acuity (VA) testing; the Humphrey field analyzer (HFA) Swedish Interactive Threshold Algorithm (SITA) 24-2 test (Carl Zeiss Meditec, Dublin, CA); multiple IOP measurements using Goldmann applanation tonometry (GAT); stereoscopic ONH photography; and RTVue-100 OCT scanning. All glaucoma patients underwent two reliable VF tests within one month. To minimize the learning effect, the second reliable HFA test was used in the analysis. The presence or absence of learning effect was assessed by detecting the improvement of mean deviation (MD) in the second VF test. ¹⁸ The second VF test was required in all study eyes before the second results were used in the analyses. All participants gave written informed consent before enrollment. All procedures conformed to the Declaration of Helsinki, and the study was approved by the Institutional Review Board of Asan Medical Center at the University of Ulsan in Seoul, Korea. 

Inclusion and Exclusion Criteria

For inclusion in the study, all participants had to have OAG with the following criteria: best-corrected VA of 20/30 or better, with a spherical equivalent within ±5 diopters (D) and a cylinder correction within +3 D; presence of a normal anterior chamber and open angle on slit-lamp and gonioscopic examinations; maximum untreated IOP less than 22 mm Hg using GAT in the physician's office; and two reliable HFA test results with a false-positive error <15%, a false-negative error <15%, and a fixation loss <20%. 

Glaucomatous eyes in our study were defined as having glaucomatous VF defects confirmed by at least two reliable VF examinations; and the presence of a compatible glaucomatous optic disc that showed increased cupping (a vertical and/or horizontal cup-disc [C/D] ratio >0.6), a difference in vertical C/D ratio of >0.2 between eyes, diffuse or focal neural rim thinning, disc hemorrhage or retinal nerve fiber layer (RNFL) defects. None of the study eyes had a structural defect based on vertical and/or horizontal C/D ratio >0.6 only. All study eyes had other glaucomatous optic nerve changes described above along with glaucomatous VF defect. Eyes with glaucomatous VF defects were defined as those with a cluster of three points with probabilities of <5% on the pattern deviation map in at least one hemifield, and including at least one point with a probability of <1%; or a cluster of two points with a probability of <1%, and a glaucoma hemifield test (GHT) result outside normal limits; or a pattern standard deviation (PSD) outside 95% of the normal limits. One eye was selected randomly if both eyes were eligible for inclusion in the study. 

We excluded subjects with a history of secondary glaucoma, for example steroid glaucoma, trauma or uveitis, or intraocular surgery, or with any other ophthalmic or neurologic condition that affects the VF, including diabetes mellitus. In addition, individuals taking medications known to affect VF sensitivity were excluded. 

OCT

SD-OCT was performed using RTVue software (version A4, 0, 5, 100). All study patients underwent ONH and GCC protocols on the same day. Briefly, the ONH map protocol involves the performance of 12 radial scans, each 3.4-mm in length, and 13 concentric ring scans ranging from 1.3 to 4.9 mm in diameter, and all centered on the optic disc. This scan configuration yields 14,141 A-scans in 0.55 seconds. After image processing, a polar pRNFLT map of the 3.45-mm diameter ring as well as parameters describing optic disc features are provided. The map provides the average RNFLT in the temporal (316–45 degrees), superior (46–135 degrees), nasal (136–225 degrees), and inferior (226–315 degrees) quadrants as well as the overall average along the entire measurement circle. In addition, each quadrant is divided into four sectors, and the software provides the RNFLT in each of these 16 sectors. 

The GCC protocol explores parameters within a ring-scan diameter of 6 mm; the center of the GCC scan is shifted approximately 1 mm temporal to the fovea so as to improve the sampling of temporal peripheral nerve fibers. The variables generated by the GCC analysis include the average, superior (0–180 degrees), and the inferior (180–360 degrees) inner retinal thickness. 

Images with signal strength indices (SSIs) less than 45 for GCC and ONH imaging or with overt decentration of the measurement circle or with segmentation error or movement artifacts were excluded from the analysis. ONH segmentation was identified by detecting the internal limiting membrane (ILM) and the RNFL lines at the retinal surface at the ILM and at the bottom of the RNFL layer. For GCC segmentation, ILM and inner plexiform layer (IPL) segmentation lines were identified at the retinal surface at the ILM and at the bottom of the IPL. SD-OCT movement artifacts were detected by a line going across the enface image and a shift in the image to the left or right. In addition, a break in the blood vessel patterns above and below this line also was looked for to detect SD-OCT movement error. All included scans had to have a clear and even illumination of the fundus image in which the foveal pit was visible clearly for the GCC image acquisition, and a clear optic disc and scan circle images for the ONH scan acquisition. All scan images were acquired by a single, experienced scanner operator. Pharmacologic dilation was performed to a minimum pupil diameter of 5 mm in all patients so as to ensure optimal image quality and careful retinal segmentation. 

Definition of the Macular and Extramacular Visual Fields

We defined the macular VF area as a central VF island that corresponds topographically to the GCC map. Visual field data in the macular VF area were derived based on the map proposed by Garway-Heath et al. ⁷ The GCC map yields a 6-mm map of the macular area (10 degrees in the superior and inferior directions, 7 degrees in the nasal direction, and 13 degrees in the temporal direction). Therefore, the macular VF covers 10 degrees in the inferior and superior areas, 7 degrees in the temporal area, and 13 degrees in the nasal area relative to fovea, covering 12 macular VF loci. The extramacular VF was defined as the remaining area lying outside the macular VF. 

Mapping Structure (GCCT, Macular pRNFLT, and Extramacular pRNFLT) to Function (SAP)

As the GCC scan yields a 6-mm map of the macular area (10 degrees in the superior and inferior directions, 7 degrees in the nasal direction, and 13 degrees in the temporal direction), the macular VF was projected slightly nasally on the Humphrey 24-2 SITA map, as shown in Figure 1. ^7,19 The global macular VF mean sensitivity (MS) in the Humphrey SAP was defined as the average value of the differential light sensitivity (DLS) obtained at each 12-test-point in the map. This average value then was correlated with the global average GCCT. The superior macular VF MS, corresponding to the inferior hemiretinal GCCT, was calculated from six test points in the superior macular hemifield, while the inferior macular VF MS corresponding to the superior hemiretinal GCCT, was calculated from the six test points in the inferior macular hemifield. 

Figure 1. 

 

Circle: the visual field (VF) sensitivity values of the area covered by the ganglion cell complex map. The extramacular VF sensitivity values are outside the circle except for the two test locations within the blind spot and the two test points located nasal to the blind spot (dashed squared box).

Based on the ONH maps, ^7,19 the six, temporal pRNFL sectors (TU1, TU2, ST2, TL1, TL2, and IT2) shown in the ONH scan mode of RTVue SD-OCT correspond topographically to the GCC scan area. This measurement is referred to as the mpRNFL (Fig. 2). Therefore, we calculated the average thickness of the six temporal sectors of the pRNFL, and this measurement is referred to as the average mpRNFLT. We also calculated the average of the three temporal sectors of the macular pRNFL to obtain the superior and inferior macular VF correlations (TL1, TL2, and IT2 sectors or TU1, TU2, and ST2 sectors, respectively). The superior macular VF MS associated with the inferior mpRNFLT measurements was calculated using data from six test points in the superior macular VF, and the inferior macular VF MS pertaining to the superior mpRNFLT test data was calculated from data obtained at six test points in the inferior macular VF. 

Figure 2. 

 

Among the 16 sectors of the ONH map, the average of the six temporal sectors (ST2, TU2, TU1, TL1, TL2, and IT2) of the pRNFL thicknesses, which correspond to the GCC scan areas, were defined as the mpRNFL thickness (the average of the ST2, TU2, and TU1 is the superior mpRNFL thickness, and the average of the TL1, TL2, and IT2 is the inferior mpRNFL thickness). The remaining 10 sectors are categorized as empRNFLT.

The global, extramacular retinal MS was defined as the average value of the DLS obtained at each of 38 points except for the 12 test points that correspond to the GCC map on the Humphery 24-2 SITA map and the two points nasal to the blind spot. We also calculated the average thickness of the 10 extramacular sectors (ST1, SN1, SN2, NU2, NU1, NL1, NL2, IN2, IN1, and IT1) of pRNFL and this measurement is referred to as the average extramacular pRNFLT (empRNFLT). The superior, extramacular retinal MS associated with the inferior hemiretinal empRNFLT (NL1, NL2, IN2, IN1, and IT1) was calculated from 19 test points in the superior extramacular hemifield, and the inferior extramacular retinal MS referring to the superior hemiretinal empRNFLT (ST1, SN1, SN2, NU2, and NU1) was calculated from 19 test points in the inferior, extramacular hemifield. 

The VF MS was expressed in two forms: on the decibel (dB) and on unlogged 1/L scales (L, luminance measured in lamberts). The DLS at each tested location can be simply written as DLS (dB) = 10 × log₁₀ (1/L). The nonlogarithmic 1/L value at each tested location was calculated by dividing the dB reading by 10 followed by derivation of the antilogarithm. Details of the MS calculation have been described previously. ²⁰  

Statistical Analysis

The macular VF MS was treated as the dependent variable and the corresponding GCCT and mpRNFLT as the independent variables in all regressions assessing the structure-function relationship. Likewise, the extramacular VF MS was treated as the dependent variable and the corresponding empRNFLT as the independent variable. VF MS was recorded in the dB and in the 1/L scales. 

Previous studies have demonstrated that second-order, polynomial regression models best describe the relationship between VF sensitivity expressed in the dB scale and RNFLT, while linear regression better defined the structure-function relationship when VF sensitivity was expressed in a linear scale. ^21–24 Therefore, relationships between various structural parameters and the corresponding retinal sensitivity were evaluated with second-order and linear regression analyses according to the retinal sensitivity scale. 

To compare the strength of the association between the thickness measurements obtained on GCC and ONH modes, and the VF MS in the macular and extramacular VF areas, we examined any coefficient of determination (R ²) to obtain a statistically significant difference using the bootstrapping method (500 replicates). ¹⁵ Bootstrapping was performed 500 times. Each R ² value was derived from 500 bootstrapped data and the difference between R ² values from two models was calculated for the purpose of comparison. We used the percentile bootstrapping method, which derives 95% confidence interval (CI) using 2.5% and 97.5% of the bootstrapped distribution. Assuming its bootstrapped values of the difference follow a normal distribution, a t-test was performed to test the null hypothesis that their R ² means between two models are equal. 

All reported P values are two-sided, and P < 0.05 was considered to indicate statistical significance. SPSS version 15.0 (SPSS Inc., Chicago, IL) and R version 2.12.0 (free software that can be downloaded from http://www.r-project.org) were used for our statistical analysis. 

Results

A total of 217 OAG eyes was enrolled in this cross-sectional study. Table 1 shows the patient demographics and biometric parameters. There was no statistically significant difference in SSI between ONH and GCC scanning modes. 

Table 1. 

 

Demographics and Clinical Characteristics of the Study Participants

	Glaucomatous Eyes
N (M/F)*	217 (103/114)
RE/LE	108/109
Age (y)*	55.49 (12.32)
VA (decimal)*	0.861 (0.19)
IOP (mm Hg)*	17.35 (2.73)
CCT (μm)*	528.90 (32.69)
Spherical error (D)*	−1.67 (2.03)
OCT parameters*
SSI of the ONH mode	58.68 (9.31)
AVG RNFLT (μm)	83.60 (10.93)
Superior RNFLT (μm)	85.75 (13.69)
Inferior RNFLT (μm)	80.82 (13.97)
Cup-to-disc ratio	0.76 (0.17)
SSI of the GCC mode	65.25 (9.02)
AVG GCC (μm)	78.91 (9.92)
Superior GCC (μm)	82.81 (10.11)
Inferior GCC (μm)	75.25 (11.12)
VF parameters
VFI (%)	85.06 (14.02)
MD (dB)*	−5.32 (4.76)
PSD (dB)*	7.25 (4.21)
Location of the VF defect
Macular/extramacular/both	39/31/147
Superior/inferior/both	109/55/53

 AVG, average; CCT, central corneal thickness; LE, left eye; RE, right eye; SSI, signal strength index; VFI, visual field index.

*  Mean (SD).

The strength of the relationship between GCCT and mpRNFLT, and VF MS expressed in two forms (dB and unlogged 1/L scale) are presented in Table 2. Our data demonstrated statistically significant associations between GCCT and mpRNFLT measured by two different modes (GCC and ONH), and corresponding VF MS by second-order polynomial and linear regression analysis (Table 2, P < 0.001). There were no significant differences in R ² values between the associations derived from GCCT and those from mpRNFLT, except in the superior macular VF sector (Table 2). Of note, there were no significant differences in the coefficients of determination, and thus indicating the strength of the relationship, between the structure-function relationship in the macular VF (versus GCCT and mpRNFLT) and that of the extramacular VF (versus empRNFLT), as measured by either mode (Tables 3, 4). 

Table 2. 

 

The Relationship between SD-OCT Thickness (GCCT, mpRNFLT) and VF Sensitivities in Two Forms (dB and Unlogged 1/L Scale) in the Macular Area

	GCCT	mpRNFLT	P Value*
AVG macular sensitivity vs. AVG GCCT and mpRNFLT
Linear (1/L)	0.093 (<0.001)	0.119 (<0.001)	0.695
Second-order polynomial (dB)	0.111 (0.035)	0.127 (0.009)	0.765
Superior macular sensitivity vs. inferior AVG GCCT and mpRNFLT
Linear (1/L)	0.333 (<0.001)	0.119 (<0.001)	0.041†
Second-order polynomial (dB)	0.358 (0.002)	0.171 (<0.001)	0.003†
Inferior macular sensitivity vs. superior AVG GCCT and mpRNFLT
Linear (1/L)	0.188 (<0.001)	0.144 (<0.001)	0.374
Second-order polynomial (dB)	0.227 (0.002)	0.263 (<0.001)	0.699

 The numbers represent the coefficients of determination (R ²) and P values (in parentheses). L, luminance measured in lamberts.

*  Comparison of the coefficient of determination (R ²) between the GCCT and mpRNFLT using the Bootstrap method.

†   P < 0.05.

Table 3. 

 

The Relationship between SD-OCT Thickness (empRNFLT) and VF Sensitivities in Two Forms (Decibel and Unlogged 1/L Scale) in the Extramacular Area

	R 2 (P Value)
AVG extramacular sensitivity vs. AVG empRNFLT
Linear (1/L)	0.106 (<0.001)
Second-order polynomial (dB)	0.231 (<0.001)
Superior extramacular sensitivity vs. inferior empRNFLT
Linear (1/L)	0.190 (<0.001)
Second-order polynomial (dB)	0.223 (<0.001)
Inferior extramacular sensitivity vs. superior empRNFLT
Linear (1/L)	0.083 (<0.001)
Second-order polynomial (dB)	0.298 (<0.001)

 The numbers represent the coefficients of determination (R ²) and P values (in parentheses).

Table 4. 

 

Comparison of the Coefficient of Determination (R ²) between the Extramacular and Macular Areas Using the Bootstrap Method

MS in 2 Forms	Extramacular	Macular
	empRNFLT	mpRNFLT	GCCT	P Value*	P Value†
AVG MS vs. AVG SD-OCT thickness (R 2)
Linear (1/L)	0.106	0.119	0.093	0.783	0.773
Second-order polynomial (dB)	0.231	0.127	0.111	0.242	0.278
Superior MS vs. inferior SD-OCT thickness (R 2)
Linear (1/L)	0.190	0.119	0.333	0.642	0.221
Second-order polynomial (dB)	0.223	0.171	0.358	0.567	0.116
Inferior MS vs. superior SD-OCT thickness (R 2)
Linear (1/L)	0.083	0.144	0.188	0.213	0.154
Second-order polynomial (dB)	0.298	0.263	0.267	0.663	0.591

*  Comparison of the coefficient of determination (R ²) between the empRNFLT and mpRNFLT.

†  Comparison of the coefficient of determination (R ²) between the empRNFL and macular area (GCC).

Figure 3 shows the structure-function relationship between the average GCCT and mpRNFLT, and the corresponding VF MS in the macular area, as expressed in dB and unlogged 1/L scales, respectively. Figure 4 shows the structure-function relationship between GCCT and empRNFLT, and the corresponding VF MS in the macular and extramacular VF areas, as expressed in dB and unlogged 1/L scales, respectively. 

Figure 3. 

 

Scatter plots showing the associations between the RTVue-100 thickness parameters, and the corresponding retinal sensitivities and expressed in dB. (A) Average thickness of the GCC and the mpRNFL versus the global macular MS. (B) Inferior average thickness of the GCC and mpRNFL versus the superior macular MS. (C) Superior average thickness of the GCC and mpRNFL versus the inferior macular MS. Scatter plots showing the associations between the RTVue-100 thickness parameters, and the corresponding retinal sensitivities and expressed in 1/L. (D) Average thickness of the GCC and mpRNFL versus that of the global macular MS. (E) Inferior average thickness of the GCC and mpRNFL versus the superior macular MS. (F) Superior average thickness of the GCC and mpRNFL versus the inferior macular MS.

Figure 4. 

 

Scatter plots showing the associations between the RTVue-100 thickness parameters and the corresponding retinal sensitivities and expressed in dB. (A) Average thickness of the GCC versus the global macular MS, and the average thickness of the empRNFL versus the global peripheral MS. (B) Inferior average thickness of the GCC versus the superior macular MS, and the inferior average thickness of the empRNFL versus the superior extramacular MS. (C) Superior average thickness of the GCC versus the inferior macular MS, and the superior average thickness of the empRNFL versus the inferior extramacular MS. Scatter plots showing the associations between the RTVue-100 thickness parameters and the corresponding retinal sensitivities, and expressed in 1/L. (D) Average thickness of the GCC versus the global macular MS, and the average thickness of the empRNFL versus the global peripheral MS. (E) Inferior average thickness of the GCC versus the superior macular MS, and the inferior average thickness of the empRNFL versus the superior extramacular MS. (F) Superior average thickness of the GCC versus the inferior macular MS, and the superior average thickness of the empRNFL versus the inferior extramacular MS.

Discussion

As the strength of the structure-function relationship may vary as a function of the retinal and corresponding VF areas in which the association is assessed, the best way to evaluate the structure-function association in glaucoma is to compare local sensitivity to local structural measurements. ²⁵ Therefore, in our current study, central macular VF was selected as a local functional target because of its clinical relevance to OAG, and its structure-function relationship, therefore, was assessed using the GCC thickness. 

The use of experimental primate models of glaucoma has shown that ganglion cells in the foveal region seem to be vulnerable to glaucomatous injury, and that ganglion cell loss occurred even if glaucomatous changes were mild. ²⁶ In humans, loss of ganglion cells and reduced nerve fiber thickness also have been observed in the posterior pole region of glaucomatous eyes. ²⁷ However, accurate measurements of temporal RNFL corresponding to posterior pole using TD-OCT have been suboptimal due to a low signal-to-noise ratio as well as a lack of adjustment for spatial summation in the macular VF. ^21,28–31 In our study, average macular VF MS was related better to the temporal mpRNFLT as well as to the GCCT compared to the previous TD-OCT studies ^23,28 (R ² = 0.119, 0.093 according to the linear scale, respectively). Improved scan resolution and segmentation algorithms of the updated SD-OCT might have influenced the strength of the correlations between the two temporal thickness parameters and the corresponding macular VF, thereby resulting in a better structure-function relationship. Therefore, GCCT and mpRNFL from RTVue SD OCT may serve better as a structural parameter to predict or monitor macular VF change over time. Further longitudinal studies will be needed to validate this assumption. 

Overall, we found a significant correlation in the GCCT and mpRNFLT, and the macular VF MS regardless of the location (superior versus inferior VF) in our series of OAG eyes. However, in agreement with the previous findings of Caprioli et al., ³ eyes with localized RNFL defects near the inferior fovea were seen more often in our consecutive series of OAG eyes (54 superior versus 109 inferior RNFL defects). This greater sample size might have resulted in greater statistical power with higher inferior GCC thickness correlation with corresponding superior macular VF MS compared to the correlation of superior or average GCC thicknesses with corresponding VF in the macula (R ² = 0.33 vs. 0.188, 0.093 according to the linear scale, respectively). 

In our current study, GCCT showed a significantly better structure-function correlation with macular VF MS compared to that between mpRNFLT and the corresponding macular VF MS in the superior macular hemifield. The large number of patients (162) with superior macular hemifield damage might have resulted in a more robust and better correlation between inferior GCCT and superior macular hemifield than that between inferior mpRNFLT and superior macular hemifield. Another explanation is that the sectors corresponding to the mpRNFLT (TL1, TL2, IT2, TU1, TU2, ST2) do not match exactly with the 12 selected VF locations (Fig. 1) based on the Garway-Heath map. ⁷ In addition, the stronger correlation calculated in our study also may be attributable to the high correlation between the thicknesses of specific multiple macular layers, that is RNFL, ganglion cell, and inner plexiform layer, provided by the GCC scan as opposed to the single mpRNFL thickness measured by the ONH mode. Despite improved scan resolution and segmentation algorithms of the updated SD-OCT in the circumpapillary region, anatomic variation surrounding the ONH along with targeting the single layer (RNFL) in the temporal sector might explain the reduced strength of the correlations between mpRNFLT and macular VF compared to the correlation with GCCT. 

The clinical implication of finding a stronger structure-function association between the GCCT and macular VF MS compared to that between mpRNFLT and macular VF MS is that the use of the GCC scan mode can aid better in the detection and follow-up of OAG patients who may present with an early stage of macular VF defect, as was the case in our current patient series (MD = −5.32 dB), than that of the ONH mode. 

Recent histopathologic studies in monkeys trained for perimetric research, indicate that the perimetry slope in relation to the ganglion cell count varies with the eccentricity of the region being investigated. ³² With decibel scaling, visual sensitivity losses and ganglion cell densities varied with eccentricity, demonstrating that the structure-function relationship may vary with eccentricity and was steeper for the peripheral visual field location than for locations close to the fixation. ³³ Although direct comparisons with other studies are problematic owing to variations in the study populations and research design, the observed correlation between the average empRNFLT and extramacular retinal MS values was comparable to that between average GCCT or mpRNFLT and the macular visual field according to the dB scale as well as to 1/L (R ² = 0.231 vs. 0.127, 0.111 according to the dB scale, respectively; P > 0.05). Similar findings also were noted at both the inferior and superior VF sectors. 

In our consecutive series of eyes, many OAG patients had a VF defect in the macular VF area defined by the GCC scan obtained at the time of their study inclusion (n = 186 for macular VF vs. 178 for extramacular VF). This might have had an effect on the strength of the structure-function relationship in the macular compared to the extramacular VF (R ² = 0.358 vs. 0.223, respectively, in the superior VF; P > 0.05) according to the dB scale as the central VF structure-function relationship can be influenced by the sample size and disease severity. An additional explanation is that the sectors (ST1, SN1, SN2, NU2, NU1 NL1, NL2, IN2, IN1, and IT1) of the empRNFLT map represent a greater retinal area that would not be represented by the 38 points used in the current study using the SITA 24-2 test. This also might have affected the strength of the structure-function relationship in the extramacular region. Further longitudinal studies are needed to clarify GCC structural change with its corresponding functional decline during the progression of glaucoma using SD OCT to validate our findings. 

In agreement with previous study, ²³ the structure-function relationship found in our current study was described best with a nonlinear curve when macular VF MS was expressed according to the dB scale against GCCT and mpRNFLT. By plotting the macular VF sensitivity in 1/L against GCCT and mpRNFLT, the relationship was linear, as seen in Figure 3. Similar relationship patterns also were seen in the extramacular VF area when VF MS was expressed according to dB and the 1/L scale (Fig. 4). 

There are a few limitations to our study. More detailed macular VF MS using the 10-2 program might represent a more optimal combination for evaluating the relationship between GCCT, and mpRNFLT and VF MS. However, the area covered by GCCT or mpRNFLT by RTVue does not quite match the 10-2 program topographically as well as 12 macular VF points based on 24-2 SAP due to individual differences in the anatomy of the eyes studied. Other limitations of our current study include recruiting a relatively confined patient population, for example OAG with baseline IOP ≤21 mm Hg, and the use of a homogeneous population (Asian) as data from a single ethnic group with normal baseline IOP may not be generalized to other races with high IOP. However, as our study is exploratory in nature, evaluating the structure-function relationship using the GCC mode of SD OCT (RTVue-100) in macular and extramacular VF areas, it may be of value for guiding further studies regarding the clinical use of GCCT assessment using SD OCT for detecting and monitoring macular VF defects associated with OAG. We did not examine the structure-function relationship in smaller regions of the extramacular VF area by dividing the extramacular VF into multiple sectors per the Garway-Heath map. ⁷ Greater division of the extramacular VF might have resulted in a higher degree of structure-function relationship at certain peripheral sectors compared to the central macular VF seen in our current study 

In conclusion, the GCCT and mpRNFLT measured by RTVue-100 SD OCT were correlated significantly with the macular retinal MS assessed by HFA. The strength of the structure-function relationships with the GCCT was greater than that of mpRNFLT in the superior macular VF. There was a similar degree of strength in the structure-function relationship in the central macular VF compared to that of the extramacular VF area in our series of OAG eyes. GCCT as measured by RTVue-100 SD OCT may be a better target than mpRNFLT for examining the correlation of macular structural change with functional loss and for understanding glaucoma progression in the macular region. 

Footnotes

 Disclosure: J.H. Na, None; M.S. Kook, None; Y. Lee, None; S. Baek, None