April 2007
Volume 48, Issue 4
Free
Glaucoma  |   April 2007
Structure-Function Relationship Is Stronger with Enhanced Corneal Compensation than with Variable Corneal Compensation in Scanning Laser Polarimetry
Author Affiliations
  • Thế Anh Mai
    From The Rotterdam Eye Hospital, Rotterdam, The Netherlands.
  • Nicolaas J. Reus
    From The Rotterdam Eye Hospital, Rotterdam, The Netherlands.
  • Hans G. Lemij
    From The Rotterdam Eye Hospital, Rotterdam, The Netherlands.
Investigative Ophthalmology & Visual Science April 2007, Vol.48, 1651-1658. doi:10.1167/iovs.06-1003
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Thế Anh Mai, Nicolaas J. Reus, Hans G. Lemij; Structure-Function Relationship Is Stronger with Enhanced Corneal Compensation than with Variable Corneal Compensation in Scanning Laser Polarimetry. Invest. Ophthalmol. Vis. Sci. 2007;48(4):1651-1658. doi: 10.1167/iovs.06-1003.

      Download citation file:


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

      ×
  • Supplements
Abstract

purpose. To compare the structure-function relationship between peripapillary retinal nerve fiber layer (RNFL) retardation, measured with scanning laser polarimetry (SLP) with both enhanced and variable corneal compensation (ECC [enhanced corneal compensation] and VCC [variable corneal compensation], respectively; features of the GDx Nerve Fiber Analyzer; Carl Zeiss Meditec, Inc., Dublin, CA), and visual field (VF) sensitivity, measured with standard automated perimetry (SAP) in normal and glaucomatous eyes and the effect of marked atypical birefringence patterns (ABPs) on this relationship.

methods. Thirty-three healthy subjects, and 68 patients with primary open-angle glaucoma (POAG) took part in the study. ECC and VCC images were taken in one randomly selected eye of each subject. VF tests were also obtained in the same eyes. The structure-function relationship was assessed in six peripapillary sectors and their matching VF areas and was reassessed after eliminating eyes with marked ABPs.

results. Correlations (Spearman’s correlation coefficients, r s) in the structure-function relationship were generally stronger in images taken with ECC than in those taken with VCC. With ECC, the relationship was significantly more curvilinear when VF sensitivity was expressed in the standard decibel scale and more linear when VF sensitivity was expressed in an antilog scale than with VCC. When eyes with marked ABP images were removed from the analysis, the structure-function relationship with VCC improved, and no statistically significantly differences were found in the relationships between VCC and ECC.

conclusions. The structure-function relationship between RNFL retardation and SAP VF sensitivity was stronger in images obtained with the GDx ECC than with the GDx VCC (Carl Zeiss Meditec, Inc., Dublin, CA). ABPs, which appeared more markedly with VCC than with ECC, weakened the structure-function relationship.

Primary open-angle glaucoma (POAG) has been recognized as a progressive optic neuropathy characterized by an accelerated degeneration of retinal ganglion cells (RGCs) and their axons. 1 2 3 4 These structural changes, which may be evident as a local and/or diffuse thinning of the retinal nerve fiber layer (RNFL) 5 6 7 and of the neuroretinal rim, may eventually lead to loss of visual function. 6 Assessment of both structural and functional changes is therefore important in both clinical practice and clinical trials to support the diagnosis of glaucoma, and also to follow up any disease progression over time. 1 The relationship between structure and function, however, is not entirely understood. 8  
The relationship between structure, observed by different imaging techniques, including scanning laser polarimetry (SLP), and function, determined by standard automated perimetry (SAP) has been investigated before. 9 10 11 12 13 14 15 16 17 With SLP with variable corneal compensation (VCC; commercially available in the GDx Nerve Fiber Analyzer; Carl Zeiss Meditec, Inc., Dublin, CA), the structure-function relationship has been shown to be curvilinear when VF sensitivity is expressed in a decibel scale. 9 10 11 12 However, when VF sensitivity is expressed in an antilog (1/Lambert) scale, this relationship appears to be linear. 10 11 12 SLP is a noninvasive, noncontact diagnostic technique that indirectly quantifies the RNFL thickness. It is based on the principle that polarized light passing through the presumed form birefringent RNFL undergoes a measurable phase shift, known as retardation, that is linearly related to histologically measured RNFL thickness. 18 Because the anterior segment (mostly the cornea) can also exhibit birefringence, VCC was developed to obtain the true RNFL retardation by subtracting the eye-specific anterior segment retardation from the total retardation. 19 However, atypical birefringence patterns (ABPs), as seen in a subset of normal and glaucomatous eyes, may confound the RNFL thickness measurement by VCC. VCC images with ABPs are characterized by an abnormal retardation map (i.e., with variable areas of high retardation arranged in a spokelike peripapillary pattern, or splotchy areas of high retardation nasally and temporally). 20 Quantitatively, images with a typical scan score (TSS) of 80 or less have been reported to be atypical. 21 ABPs may be related to age, myopia, and blond fundi. Hypothetically, ABPs are caused by a low signal-to-noise ratio resulting from loss or attenuated reflectivity of the retinal pigment epithelium. 20  
SLP with enhanced corneal compensation (ECC), the latest software change of the GDx, has been introduced to optimize SLP imaging by improving the signal-to-noise ratio, notably in areas with a low signal. 22 It was first described by Knighton and Zhou 22 who assumed that the susceptibility of SLP to error, both optical (e.g., stray light) and electronic (e.g., noise, digitization error), is relatively large when the sensitivity of SLP to small retardation differences is low (low retardance, depolarization, or reduction in reflected intensity). The sensitivity of SLP to the main signal is increased by adding a predetermined birefringence (bias retarder) during image acquisition, allowing the total retardation to be shifted into a more sensitive region of the device’s detector to the polarization signal amplitude. The RNFL retardation is then calculated by subtracting corneal plus bias retardance from the total retardance. ECC has been reported to reduce both frequency and severity of ABPs. 21 23 24 Little is known, however, about whether the structure-function relationship is also becomes better with ECC than with VCC. 
In the present study, we assessed the relationship between RNFL thickness, measured by SLP with both VCC and ECC, and VF sensitivity. Because RNFL morphology appears to be better imaged with ECC than with VCC, 21 23 24 due to an improved signal-to-noise ratio, we expected the structure-function relationship to be better with ECC than with VCC. 
Materials and Methods
Subjects
Thirty-three healthy subjects and 68 patients with glaucoma, all of white origin, took part in the present study. Of these individuals, 52 (51.5%) were men. The right eye was studied in 45 (44.6%) of the subjects. The mean age ± SD of the healthy individuals and patients with glaucoma was 62.2 ± 13.0 and 67.5 ± 9.1 years, respectively. The difference in mean age between the healthy and glaucoma groups was thought to be of no clinical significance, despite its statistical significance (5.0 ± 2.2 years, P = 0.03 by two-tailed independent sample t-tests). The differences in gender and eye side among the study groups, however, were not statistically significant (χ2). Before imaging with SLP, all subjects underwent a complete ophthalmic examination including a white-on-white 24-2 full-threshold SAP VF test on the commercially available Humphrey Field Analyzer (HFA II; Carl Zeiss Meditec, Inc.), slit lamp biomicroscopy, intraocular pressure (IOP) measurement by Goldmann applanation tonometry, and gonioscopy. None of the subjects had any history of ocular disease (except glaucoma in the glaucoma group), intraocular surgery (except uncomplicated cataract surgery), or significant coexisting systemic disease with possible ocular involvement, such as diabetes mellitus or arterial hypertension. Only one eye per subject was included by random selection if both were eligible. All selected eyes had best corrected visual acuity (BCVA) of 20/40 or better. The range of spherical equivalent refractive error in all subjects was between −7.0 and +3.0 D. Only VFs that met reliability criteria of fixation losses <25%, false-negative and -positive responses ≤20% for healthy individuals were included. For glaucoma subjects, the same criteria for inclusion of the VFs were applied, except that up to 33% false-negative responses were considered acceptable. All protocols and methods used in the present study adhered to the tenets of the Declaration of Helsinki and were approved by the Institutional Human Experimentation Committee. Informed consent was obtained after the participants were informed about possible consequences of the study. 
Healthy subjects were either consecutively recruited from an ongoing longitudinal follow-up study of the Rotterdam Eye Hospital or from staff members, their friends and spouses, partners of the patients, or volunteers. In both eyes, all healthy subjects had an IOP of 21 mm Hg or less, a normal VF test result by SAP, an unremarkable slit lamp examination, open angles on gonioscopy, a healthy-looking optic disc (no diffuse or local rim thinning, cupping, or optic disc hemorrhages), and no other ocular abnormalities. A normal VF test result was defined as a mean deviation (MD) and a pattern SD (PSD) within 95% confidence limits, and a glaucoma hemifield test (GHT) result within normal limits. None of our healthy subjects reported having first- and/or second-degree family members with glaucoma. Their MD and PSD (both mean ± SD) were 0.3 ± 1.1 and 1.7 ± 0.5 dB, respectively. 
The glaucoma patients had, in their selected eyes, a glaucomatous appearance of their optic disc (diffuse or local rim thinning, cupping, possibly with optic disc hemorrhages), a corresponding SAP VF defect confirmed on two consecutive occasions, open angles by gonioscopy and no evidence of secondary glaucoma. A VF defect in the present study was considered glaucomatous if it had two or more adjacent points with P < 1% or deeper or three or more adjacent points with P < 5% or deeper in the total deviation plot or a GHT result outside normal limits that was not attributable to causes other than glaucoma. Their MD and PSD were −11.9 ± 8.0 and 9.2 ± 3.7 dB respectively. The glaucoma eyes were classified based on VF defects severity described by Hodapp et al. 25 Twenty-one (30.9%) patients with glaucoma were considered to have mild and moderate and 47 (69.1%) to have severe VF defects. 
Image Acquisition
All subjects were imaged with a commercially available SLP (GDx VCC, software version 5.4.0; and GDx ECC, software version 5.5.0.11; Carl Zeiss Meditec, Inc., Dublin, CA). Details of the SLP instrument have been described elsewhere. 26 27  
In short, the GDx VCC is a modified SLP system with two linear retarders in rotating mounts, so that both the retardance and axis of the unit can be adjusted as required. A near-infrared laser (785 nm in wavelength) scans the ocular fundus in a raster pattern and captures an image with a field 40° horizontally and 20° vertically (which includes both the peripapillary and the macular regions). In contrast with the earlier version of the SLP with fixed corneal compensation, SLP VCC, in which birefringence of the anterior segment was compensated as though all individuals had a slow axis of corneal birefringence 15° nasally downward with a magnitude of 60 nm, 19 28 the VCC algorithm allows eye-specific anterior segment birefringence compensation, both in corneal polarization magnitude (CPM) and in corneal polarization axis (CPA), 19 based on the macular retardance profile. The RNFL retardation (in nanometers) is calculated based on the retarder-adjusted eye-specific CPA and CPM, and is then converted into thicknesses (in micrometers), based on a fixed conversion factor of 0.67 nm/μm. 28  
In the ECC mode, a large known bias retarder is introduced so that the combination of corneal plus bias retardance becomes close to 50 nm and with a slow axis close to vertical, so that the retardance measurement is shifted into a more sensitive region of the device’s detector to the polarization signal amplitude. RNFL retardation is then determined by mathematically subtracting the macular and bias retarder-induced birefringence from the total retardance. 22 24 The bias retarder-induced birefringence is then determined from the macular region. Finally, the RNFL retardance is mathematically recalculated. 22 24  
In the present study, GDx measurements of both eyes of all subjects were performed by two trained and experienced technicians according to a standard protocol. Images were scanned through undilated pupils while the room light was left on, and the subject was required to keep the head still during the whole session, with the face resting on the facemask, allowing the best alignment between the instrument’s anterior segment compensator and the position of the eye. The spherical equivalent refractive error of each eye was registered in the instrument, and an adjustment in 0.25-D steps was made manually, if necessary, to focus on the retina. The anterior segment birefringence was then determined. Next, images of the RNFL were obtained, first with VCC and then with ECC. A fixed size band of eight pixels wide (0.4-mm equivalent in an emmetropic eye), with the inner and outer diameter of 2.4 and 3.2 mm, respectively, was centered on the ONH, allowing the retardation values to be calculated within the band to yield 256 values. These values were subsequently grouped into 64 peripapillary points by the software, which were regrouped into six sectors similar to those first described for the optic nerve head (ONH) by Garway-Health et al., 29 and later for the RNFL by Reus and Lemij. 12 These sectors were named according to their locations as temporal (T; 311–40°), superotemporal (ST; 41–80°), inferotemporal (IT; 81–120°), nasal (N; 121–230°), inferonasal (IN; 231° – 270°), and inferotemporal (IN; 271° – 310°). We then correlated the mean RNFL retardation values of these sectors to the mean VF sensitivities in their corresponding VF locations (Fig. 1) . Only images of high quality (i.e., those with a centered optic disc, well-focused, and evenly and illuminated throughout the image, without any motion artifacts, and with an inbuilt proprietary quality score of ≥7) were selected. 
Visual Field Testing
White-on-white SAP with the 24-2 full-threshold paradigm was used to test VF sensitivity of all subjects. VF sensitivities of the 52 test points, expressed in the standard decibel scale were also transformed into an exponential (or antilog) scale with the formula: DLS (in decibels) = 10 log10 (L max/L), where DLS is the differential light sensitivity, L max is the maximum luminance (for HFA, L max is set to be 104 asb or 0 dB), and L is the differential luminance at threshold, yielding the calculation of DLS in antilog (1/Lambert) scale: DLS (1/L) = 10[(DLS-dB/10)−4]. The 52 VF location sensitivities (both in the dB and antilog scales) were then regrouped into the six areas as depicted in Figure 1 , after Garway-Health et al. 29  
Statistical Analysis
In the present study, we examined the correlations between both global and sectoral RNFL retardation values and the VF sensitivities in matching areas (Fig. 1) . The Spearman’s rank correlation coefficient (r s) was used to assess the strength of any relationship. We also used Williams’ formula, 30 as described by Steiger 31 to assess whether any differences between these nonindependent correlations were statistically significant. In short, this method uses a t-transformation of two nonindependent correlation coefficients, allowing an assessment of whether variable B (ECC RNFL) is significantly more strongly correlated with variable A (DLS in either the dB or antilog scale) than is variable C (VCC RNFL), taking into account the correlation coefficients between both A and B (rAB), and A and C (rAC), as well as the correlation coefficient between B and C (rBC). A t value was calculated by the following equation:  
\[\{eq\}t{=}(\mathrm{rAB}{-}\mathrm{rAC}){\cdot}{\surd}{[}(N{-}1){\cdot}(1{+}\mathrm{rBC}){]}/{\{}{[}2(N{-}1)/(N{-}3){]}{\cdot}{\vert}R{\vert}{+}{[}(\mathrm{rAB}{+}\mathrm{rAC})/2{]}^{2}{\cdot}{[}(1{-}\mathrm{rBC})^{3}{]}{\}}\]
where R = (1 − rAB2 − rAC2 − rBC2) + (2 · rAB · rAC · rBC), and n is the sample size. 
We performed both linear and logarithmic regression analyses by using the formulas: y = a + bx and y = a + b logx, respectively. Coefficient of determination (R 2) values of each regression model were determined by this method, which then allowed the two association models between structure and function to be compared with one another. The best fit of the regression models for each area was then tested with a Wilcoxon signed ranks test for two related samples, with the null hypothesis that the absolute prediction errors (residuals) have the same mean for both logarithmic and linear models. 
Williams’ formula 30 31 was also used to assess whether there were any statistically significant differences in R 2 between ECC and VCC, generated from logarithmic regression with the VF sensitivity in the decibel scale, and from linear regression with the VF sensitivity in the antilog scale. To do so, we took the square root of the R 2 (correlation coefficients), and applied William’s formula. 
In all regression models, the VF sensitivity was treated as the dependent variable, and RNFL retardation measured by GDx VCC and ECC was regarded as the independent variable. Corrections of the probability for multiple comparisons were performed with the false discovery rate (FDR) approach described by Benjamini and Hochberg. 32 In short, this approach involves finding the largest integer k such that p(k) · m/k ≤ α, where m is the number of comparisons, 1 ≤ k (=1, 2, 3 … ) ≤ m, p(k) is the probability at the kth comparison, and α is the level of statistical significance, or 0.05. 
Finally, the analyses were performed again in subjects with both ECC and VCC images showing no atypical birefringence patterns (an ABP image was defined as having a TSS of < 80). 21 Examining all images showed that this was an appropriate cutoff point for investigating any effect of ABPs on the strength and curvilinearity/linearity of the structure-function relationships. 
All statistical analyses were performed with commercial software (SPSS ver. 12.0.1 for Windows; SPSS, Inc., Chicago, IL, and Excel 2000, SR-1; Microsoft, Redmond, WA). 
Results
The analyses performed in all eyes showed a significantly stronger relationship between RNFL retardation and VF sensitivity (decibel scale) for most studied areas with ECC than with VCC (Table 1) . The n sector, however, showed a stronger relationship with VCC than with ECC. In addition, in the T sector, there was no statistically significant structure-function relationship with either ECC or VCC (Table 1)
We suspected that the structure-function correlation was affected by any ABPs, notably present in the VCC images. We therefore recalculated r s after excluding eyes with either VCC or ECC atypical images. All images with a TSS of <80 with ECC or VCC showed atypical birefringence patterns, whereas those with a TSS of 80 or greater did not. With ECC, only two eyes had atypical images, whereas with VCC, as many as 41 were considered atypical. Of these, 33 were glaucomatous (48.5% of glaucomatous eyes), and 8 healthy eyes (24.2% of healthy eyes). When only eyes without ABPs were used for the analysis, the structure-function relationships with both VCC and ECC improved, particularly with VCC (Table 2) . There were no statistically significant differences in the r s anymore between ECC and VCC. Also, there now was a statistically significant correlation in the T sector with ECC (P = 0.02). 
Next, we investigated the differences in the curvilinearity and linearity of the structure-function relationships between ECC and VCC. Figure 2graphically shows the structure-function relationships (scatterplots), together with the best fit linear and curvilinear (logarithmic) regression lines. In Figure 2 , the structure-function relationship generally appeared curvilinear when VF sensitivity was expressed in the decibel scale and linear when it was expressed in the antilog scale. Tables 3 and 4summarize the coefficient of determination (R 2) of the linear and curvilinear regression models for each area. Table 3relates to VF sensitivity expressed in the decibel scale and Table 4to those in the antilog scale. Probabilities relate to differences in means of residuals between the two models, tested with the Wilcoxon signed-ranks test. Both tables therefore also show which regression model (linear/curvilinear) was the best fit model for the structure-function relationship. In the decibel scale, ECC data showed more sectors (ST, SN, and IN) with a statistically significantly curvilinear structure-function relationship than did VCC data (only ST). With VF sensitivity was expressed in the antilog scale, this relationship was more often significantly linear with ECC (globally, and in ST, and IT sectors) than with VCC (in ST and SN sectors). Compared with VCC, the structure-function correlations were generally stronger, both curvilinearly in the decibel scale, and linearly in the antilog scale (Table 5 , all eyes). Since there appeared no significant relationship in the T sector with VCC, we did not perform analysis of this sector. 
To assess the effect of atypical scans on the curvilinearity and linearity of the structure-function relationship, we performed all the related analyses again after eliminating eyes with atypical scans. As Tables 3 and 4show, the structure-function relationships improved with both ECC and VCC, most notably with VCC. With VCC, R 2 values were now greater than before the exclusion of ABP images. There were now also more areas with a statistically significantly curvilinear relationship in the decibel scale (not only in the ST sector, but also globally, and in SN sectors), as well as more areas with a significantly linear relationship in the antilog scale (not only in the ST, and SN sectors, but also globally, and in the IN sectors). 
Furthermore, the statistically significant differences in curvilinearity in the decibel scale, and linearity in the antilog scale between ECC and VCC disappeared (Table 5 , eyes without ABP images). 
Discussion
In this study, ECC yielded stronger correlations between RNFL retardation and VF sensitivity than did VCC, because of less atypical birefringence patterns in images taken with ECC than in those obtained with VCC. When only images without ABPs were used for the analysis, these differences in structure-function correlations between ECC and VCC were no longer observed. 
ECC has been shown to reduce the amount of atypical scan patterns in SLP. 21 23 Typical scans appear to reflect the true RNFL morphology better than do atypical scans. 21 23 24 With fewer ABPs (either reduced by ECC, or by selection of typical scans), the structure-function correlations became stronger, which may suggest that these correlations are more biological than those in atypical scans. Put differently, atypical scans appear to be noisier than typical ones. 
It has been suggested earlier that ABPs weaken the overall correlation between average RNFL thickness determined by GDx VCC and SAP VF sensitivity (MD) in a sample of 20 healthy and 60 glaucomatous eyes. 20 The weakening effect of ABPs was, however, not statistically significant (Pearson correlation coefficient, r = 0.71 in NBP eyes (n = 39), compared with 0.42 in ABP eyes (n = 26; P = 0.09). Some reasons why the reportedly weaker correlations in ABP images failed to reach statistical significance may be a relatively small sample size, the use of Pearson’s correlation coefficient, which is only designed for linear correlations, and the use of overall MD values. To our knowledge, our current paper is the first to report that ABPs degrade the correlations between RNFL thickness in various peripapillary areas and the corresponding VF sensitivity, regardless of whether the structure was measured by GDx VCC or by GDx ECC. 
With SLP and SAP, the relationship between RNFL retardation and VF sensitivity in this study was more strongly curvilinear with the VF sensitivity expressed in the decibel scale, and linear with the VF sensitivity expressed in the antilog scale. This further reaffirmed the findings of earlier studies 10 11 12 17 on the nature of the structure-function relationship. Furthermore, we think that structure-function correlations between SLP and any functional method, not necessarily restricted to SAP, should be studied with typical scans only, either by using ECC, or by eliminating atypical scans. Because the decibel scale tends to stretch the functional sensitivity at the lower end of its range, and the antilog scale at its higher end, we think that one should select the type of scale that best meets the requirements of the study. Clinically, noise and the measurement variability may be more likely to weaken the linearity of the structure-function relationship at the higher end of the antilog scale, and only studies using various stimuli scales may be able to investigate this issue. 
Although it has been linked to age, myopia, and blond fundi, ABPs in the present study appeared to be more apparent in glaucomatous eyes (48%) than in healthy eyes (24%). A possible explanation for this higher prevalence in glaucomatous eyes may be that the signal-to-noise ratio was further diminished by thinning of the RNFL in the glaucomatous eyes. 
In a subset of subjects with glaucoma, even when the VF sensitivity was nearly zero dB, there still appeared a measurable RNFL thickness of 20 μm or more with both ECC and VCC. Possible explanations that have been suggested include the presence of birefringence in the remaining nonfunctioning axons 12 or RNFL of supportive tissue 33 after most axons have been lost. It is also possible that there is a measurement offset (a floor effect). Furthermore, the structure-function relationship for the T sector was surprisingly negative with VCC in all eyes (Table 1) , which became more intuitively positive in eyes without ABPs. This may suggest that the erroneous relationship was due to spurious RNFL measurements in eyes with ABPs. 
In conclusion, the relationship between RNFL retardation and SAP VF sensitivity was stronger with ECC than with VCC. ABPs degrade the structure-function relationship. 
 
Figure 1.
 
Areas of HFA 24-2 VF test and corresponding sectors of the RNFL (right), after Garway-Health et al. 29 and Reus and Lemij. 12
Figure 1.
 
Areas of HFA 24-2 VF test and corresponding sectors of the RNFL (right), after Garway-Health et al. 29 and Reus and Lemij. 12
Table 1.
 
Structure-Function Correlation Coefficients (r s) for ECC and VCC in All Eyes with the VF Sensitivity Expressed in the Decibel Scale
Table 1.
 
Structure-Function Correlation Coefficients (r s) for ECC and VCC in All Eyes with the VF Sensitivity Expressed in the Decibel Scale
ONH RNFL ECC r s P VCC r s P P of Difference
Global 0.70 <0.001 0.54 <0.001 0.009, S
T 0.19 0.06 −0.11 0.30 0.0082, S
ST 0.78 <0.001 0.65 <0.001 0.0085, S
SN 0.66 <0.001 0.59 <0.001 0.0340, S
N 0.39 <0.001 0.43 <0.001 0.6534, NS
IN 0.62 <0.001 0.45 <0.001 0.0171, S
IT 0.66 <0.001 0.45 <0.001 0.0025, S
Table 2.
 
As Table 1but in Eyes without ABP
Table 2.
 
As Table 1but in Eyes without ABP
ONH RNFL ECC r s P VCC r s P P of Difference
Global 0.78 <0.001 0.77 <0.001 0.9685, NS
T 0.29 0.02 0.11 0.39 0.2007, NS
ST 0.79 <0.001 0.77 <0.001 0.8098, NS
SN 0.74 <0.001 0.73 <0.001 0.9421, NS
N 0.41 <0.001 0.48 <0.001 0.5587, NS
IN 0.65 <0.001 0.62 <0.001 0.7912, NS
IT 0.79 <0.001 0.68 <0.001 0.1364, NS
Figure 2.
 
Scatterplots of global and sectoral peripapillary RNFL retardation measured with ECC and VCC, against VF sensitivity (or differential light sensitivity, DLS), expressed in the decibel scale (left) (dB) and in the antilog (1/L) scale (right) in all eyes.
Figure 2.
 
Scatterplots of global and sectoral peripapillary RNFL retardation measured with ECC and VCC, against VF sensitivity (or differential light sensitivity, DLS), expressed in the decibel scale (left) (dB) and in the antilog (1/L) scale (right) in all eyes.
Table 3.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Decibel Scale, in Various Areas, for ECC and VCC
Table 3.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Decibel Scale, in Various Areas, for ECC and VCC
All Eyes Eyes without ABP Images
ECC VCC ECC VCC
R 2 log/R 2 lin P R 2 log/R 2 lin P R 2 log/R 2 lin P R 2 log/R 2 lin P
Global 0.50/0.47 0.064, NS 0.31/0.31 0.559, NS 0.54/0.50 0.047, NS 0.54/0.48 0.025, S
ST 0.44/0.43 0.001, S 0.31/0.30 0.007, S 0.46/0.45 0.008, S 0.41/0.38 0.008, S
SN 0.32/0.30 0.008, S 0.26/0.24 0.045, NS 0.43/0.40 0.016, S 0.43/0.40 0.021, S
T
N 0.16/0.13 0.127, NS 0.15/0.14 0.891, NS 0.22/0.18 0.016, S 0.17/0.16 1.000, NS
IN 0.44/0.40 0.027, S 0.20/0.19 0.573, NS 0.46/0.42 0.054, NS 0.33/0.30 0.174, NS
IT 0.36/0.35 0.049, NS 0.14/0.13 0.953, NS 0.57/0.52 0.004, S 0.40/0.39 0.927, NS
Table 4.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Antilog (1/L) Scale, in Various Areas for ECC and VCC
Table 4.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Antilog (1/L) Scale, in Various Areas for ECC and VCC
All Eyes Eyes without ABP Images
ECC VCC ECC VCC
R 2 lin/R 2 log P R 2 lin/R 2 log P R 2 lin/R 2 log P R 2 lin/R 2 log P
Global 0.50/0.47 <0.001, S 0.27/0.22 0.062, NS 0.55/0.51 0.007, S 0.53/0.51 0.011, S
ST 0.54/0.48 <0.001, S 0.35/0.32 0.001, S 0.56/0.51 0.018, S 0.52/0.47 0.027, S
SN 0.40/0.37 0.096, NS 0.32/0.25 0.008, S 0.46/0.45 0.218, NS 0.49/0.46 0.030, NS
T
N 0.14/0.14 0.598, NS 0.20/0.17 0.293, NS 0.13/0.12 0.457, NS 0.21/0.21 0.669, NS
IN 0.31/0.30 0.075, NS 0.16/0.14 0.053, NS 0.32/0.31 0.018, S 0.32/0.31 0.023, S
IT 0.41/0.39 0.004, S 0.17/0.13 0.111, NS 0.52/0.51 0.165, NS 0.44/0.42 0.385, NS
Table 5.
 
Coefficients of Determination for Logarithmic Regression Model with VF in the Decibel Scale, and Linear Regression Models with VF in the Antilog Scale, with ECC and VCC
Table 5.
 
Coefficients of Determination for Logarithmic Regression Model with VF in the Decibel Scale, and Linear Regression Models with VF in the Antilog Scale, with ECC and VCC
ECC and VCC R 2 for Logarithmic Regression Model with VF Sensitivity in the Decibel Scale ECC and VCC R 2 for Linear Regression Model with VF Sensitivity in the 1/L Scale
In All Eyes In Eyes without ABP Images In All Eyes In Eyes without ABP Images
ECC/VCC P * ECC/VCC P * ECC/VCC P * ECC/VCC P *
Global 0.50/0.31 0.0083, S 0.54/0.54 0.9916, NS 0.50/0.27 0.0016, S 0.55/0.53 0.8808, NS
ST 0.44/0.31 0.0967, NS 0.46/0.41 0.5892, NS 0.54/0.35 0.0134, S 0.56/0.52 0.6441, NS
SN 0.32/0.26 0.3483, NS 0.43/0.43 0.9668, NS 0.40/0.32 0.3673, NS 0.46/0.49 0.7815, NS
T
N 0.16/0.15 0.8447, NS 0.22/0.17 0.2525, NS 0.14/0.20 0.7222, NS 0.13/0.21 0.2591, NS
IN 0.44/0.20 0.0012, S 0.46/0.33 0.1315, NS 0.31/0.16 <0.0001, S 0.32/0.32 0.9554, NS
IT 0.36/0.14 0.0018, S 0.57/0.40 0.0399, S 0.41/0.17 <0.0001, S 0.52/0.44 0.3269, NS
American Academy of Ophthalmology. Preferred Practice Pattern: Primary Open Angle Glaucoma. 2005;3.American Academy of Ophthalmology San Francisco, CA.
Kerrigan-BaumrindLA, QuigleyHA, PeaseME, KerriganDF, MitchellRS. Number of ganglion cells in glaucoma eyes compared with threshold visual field tests in the same persons. Invest Ophthalmol Vis Sci. 2000;41:741–748. [PubMed]
QuigleyHA. Neuronal death in glaucoma. Prog Retin Eye Res. 1999;18:39–57. [CrossRef] [PubMed]
QuigleyHA, AddicksEM, GreenWR. Optic nerve damage in human glaucoma. III. Quantitative correlation of nerve fiber loss and visual field defect in glaucoma, ischemic neuropathy, papilledema, and toxic neuropathy. Arch Ophthalmol. 1982;100:135–146. [CrossRef] [PubMed]
QuigleyHA, MillerNR, GeorgeT. Clinical evaluation of nerve fiber layer atrophy as an indicator of glaucomatous optic nerve damage. Arch Ophthalmol. 1980;98:1564–1571. [CrossRef] [PubMed]
TuulonenA, AiraksinenPJ. Initial glaucomatous optic disk and retinal nerve fiber layer abnormalities and their progression. Am J Ophthalmol. 1991;111:485–490. [CrossRef] [PubMed]
JonasJB, SchiroD. Localized wedge shaped defects of the retinal nerve fiber layer in glaucoma. Br J Ophthalmol. 1994;78:285–290. [CrossRef] [PubMed]
ArtesPH, ChauhanBC. Longitudinal changes in the visual field and optic disc in glaucoma. Prog Retin Eye Res. 2005;24:333–354. [CrossRef] [PubMed]
BowdC, ZangwillLM, WeinrebRN. Association between scanning laser polarimetry measurements using variable corneal polarization compensation and visual field sensitivity in glaucomatous eyes. Arch Ophthalmol. 2003;121:961–966. [CrossRef] [PubMed]
SchlottmannPG, De CillaS, GreenfieldDS, CaprioliJ, Garway-HeathDF. Relationship between visual field sensitivity and retinal nerve fiber layer thickness as measured by scanning laser polarimetry. Invest Ophthalmol Vis Sci. 2004;45:1823–1829. [CrossRef] [PubMed]
ReusNJ, LemijHG. Relationships between standard automated perimetry, HRT confocal scanning laser ophthalmoscopy, and GDx VCC scanning laser polarimetry. Invest Ophthalmol Vis Sci. 2005;46:4182–4188. [CrossRef] [PubMed]
ReusNJ, LemijHG. The relationship between standard automated perimetry and GDx VCC measurements. Invest Ophthalmol Vis Sci. 2004;45:840–845. [CrossRef] [PubMed]
JonasJB, GrundlerAE. Correlation between mean visual field loss and morphometric optic disk variables in the open-angle glaucomas. Am J Ophthalmol. 1997;124:488–497. [CrossRef] [PubMed]
Garway-HeathDF, HolderGE, FitzkeFW, HitchingsRA. Relationship between electrophysiological, psychophysical, and anatomical measurements in glaucoma. Invest Ophthalmol Vis Sci. 2002;43:2213–2220. [PubMed]
Garway-HeathDF, ViswanathanA, WestcottM, KamalD, FritzkeFW, HitchingsRA. Relationship between perimetric light sensitivity and optic disc neuroretinal rim area.WallM WildJM eds. Perimetry Update. 1998/1999;381–389.Kugler Publications The Hague, The Netherlands.
HoffmannEM, MedeirosFA, SamplePA, et al. Relationship between patterns of visual field loss and retinal nerve fiber layer thickness measurements. Am J Ophthalmol. 2006;141:463–471. [CrossRef] [PubMed]
BowdC, ZangwillLM, MedeirosFA, et al. Structure-function relationships using confocal scanning laser ophthalmoscopy, optical coherence tomography, and scanning laser polarimetry. Invest Ophthalmol Vis Sci. 2006;47:2889–2895. [CrossRef] [PubMed]
WeinrebRN, DreherAW, ColemanA, et al. Histopathologic validation of Fourier-ellipsometry measurements of retinal nerve fiber layer thickness. Arch Ophthalmol. 1990;108:557–560. [CrossRef] [PubMed]
ZhouQ, WeinrebRN. Individualized compensation of anterior segment birefringence during scanning laser polarimetry. Invest Ophthalmol Vis Sci. 2002;43:2221–2228. [PubMed]
BaggaH, GreenfieldDS, FeuerWJ. Quantitative assessment of atypical birefringence images using scanning laser polarimetry with variable corneal compensation. Am J Ophthalmol. 2005;139:437–446. [CrossRef] [PubMed]
TóthM, HollóG. Enhanced corneal compensation for scanning laser polarimetry on eyes with atypical polarisation pattern. Br J Ophthalmol. 2005;89:1139–1142. [CrossRef] [PubMed]
KnightonR, ZhouQ. New techniques: Nerve Fiber Analyzer GDx.IesterM Garway-HealthDF LemijHG eds. Optic Nerve Head and Retinal Nerve Fiber Analysis. 2005;117–119.Dogma Savona, Italy.
SehiM, GuaquetaDC, GreenfieldDS. An enhancement module to improve the atypical birefringence pattern using scanning laser polarimetry with variable corneal compensation. Br J Ophthalmol. 2006;90:749–753. [CrossRef] [PubMed]
ReusNJ, ZhouQ, LemijHG. Enhanced imaging algorithm for scanning laser polarimetry with variable corneal compenstation. Invest Ophthalmol Vis Sci. 2006;47:3870–3877. [CrossRef] [PubMed]
HodappE, ParrishRK, AndersonDR. Clinical decision in glaucoma. 1993;52–61.CV Mosby St. Louis.
WeinrebRN, ShakibaS, ZangwillL. Scanning laser polarimetry to measure the nerve fiber layer of normal and glaucomatous eyes. Am J Ophthalmol. 1995;119:627–636. [CrossRef] [PubMed]
WeinrebRN, BowdC, ZangwillLM. Glaucoma detection using scanning laser polarimetry with variable corneal polarization compensation. Arch Ophthalmol. 2003;121:218–224. [CrossRef] [PubMed]
ZhouQ. Detection of glaucomatous retinal nerve fiber layer damage by scanning laser polarimetry with custom corneal compensation. Proc SPIE Int Soc Opt Eng. 2003;4951:32–41.
Garway-HeathDF, PoinoosawmyD, FitzkeFW, HitchingsRA. Mapping the visual field to the optic disc in normal tension glaucoma eyes. Ophthalmology. 2000;107:1809–1815. [CrossRef] [PubMed]
WilliamsEJ. The comparison of regression variables. J R Stat Soc Series B. 1959;21:396–399.
SteigerJH. Tests for comparing elements of a correlation matrix. Psychol Bull. 1980;87:245–251. [CrossRef]
BenjaminiY, HochbergY. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B. 1995;57:289–300.
QuigleyHA, AddicksEM. Quantitative studies of retinal nerve fiber layer defects. Arch Ophthalmol. 1982;100:807–814. [CrossRef] [PubMed]
Figure 1.
 
Areas of HFA 24-2 VF test and corresponding sectors of the RNFL (right), after Garway-Health et al. 29 and Reus and Lemij. 12
Figure 1.
 
Areas of HFA 24-2 VF test and corresponding sectors of the RNFL (right), after Garway-Health et al. 29 and Reus and Lemij. 12
Figure 2.
 
Scatterplots of global and sectoral peripapillary RNFL retardation measured with ECC and VCC, against VF sensitivity (or differential light sensitivity, DLS), expressed in the decibel scale (left) (dB) and in the antilog (1/L) scale (right) in all eyes.
Figure 2.
 
Scatterplots of global and sectoral peripapillary RNFL retardation measured with ECC and VCC, against VF sensitivity (or differential light sensitivity, DLS), expressed in the decibel scale (left) (dB) and in the antilog (1/L) scale (right) in all eyes.
Table 1.
 
Structure-Function Correlation Coefficients (r s) for ECC and VCC in All Eyes with the VF Sensitivity Expressed in the Decibel Scale
Table 1.
 
Structure-Function Correlation Coefficients (r s) for ECC and VCC in All Eyes with the VF Sensitivity Expressed in the Decibel Scale
ONH RNFL ECC r s P VCC r s P P of Difference
Global 0.70 <0.001 0.54 <0.001 0.009, S
T 0.19 0.06 −0.11 0.30 0.0082, S
ST 0.78 <0.001 0.65 <0.001 0.0085, S
SN 0.66 <0.001 0.59 <0.001 0.0340, S
N 0.39 <0.001 0.43 <0.001 0.6534, NS
IN 0.62 <0.001 0.45 <0.001 0.0171, S
IT 0.66 <0.001 0.45 <0.001 0.0025, S
Table 2.
 
As Table 1but in Eyes without ABP
Table 2.
 
As Table 1but in Eyes without ABP
ONH RNFL ECC r s P VCC r s P P of Difference
Global 0.78 <0.001 0.77 <0.001 0.9685, NS
T 0.29 0.02 0.11 0.39 0.2007, NS
ST 0.79 <0.001 0.77 <0.001 0.8098, NS
SN 0.74 <0.001 0.73 <0.001 0.9421, NS
N 0.41 <0.001 0.48 <0.001 0.5587, NS
IN 0.65 <0.001 0.62 <0.001 0.7912, NS
IT 0.79 <0.001 0.68 <0.001 0.1364, NS
Table 3.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Decibel Scale, in Various Areas, for ECC and VCC
Table 3.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Decibel Scale, in Various Areas, for ECC and VCC
All Eyes Eyes without ABP Images
ECC VCC ECC VCC
R 2 log/R 2 lin P R 2 log/R 2 lin P R 2 log/R 2 lin P R 2 log/R 2 lin P
Global 0.50/0.47 0.064, NS 0.31/0.31 0.559, NS 0.54/0.50 0.047, NS 0.54/0.48 0.025, S
ST 0.44/0.43 0.001, S 0.31/0.30 0.007, S 0.46/0.45 0.008, S 0.41/0.38 0.008, S
SN 0.32/0.30 0.008, S 0.26/0.24 0.045, NS 0.43/0.40 0.016, S 0.43/0.40 0.021, S
T
N 0.16/0.13 0.127, NS 0.15/0.14 0.891, NS 0.22/0.18 0.016, S 0.17/0.16 1.000, NS
IN 0.44/0.40 0.027, S 0.20/0.19 0.573, NS 0.46/0.42 0.054, NS 0.33/0.30 0.174, NS
IT 0.36/0.35 0.049, NS 0.14/0.13 0.953, NS 0.57/0.52 0.004, S 0.40/0.39 0.927, NS
Table 4.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Antilog (1/L) Scale, in Various Areas for ECC and VCC
Table 4.
 
Coefficients of Determination (R 2) for Logarithmic and Linear Regression Models with VF Sensitivity Expressed in the Antilog (1/L) Scale, in Various Areas for ECC and VCC
All Eyes Eyes without ABP Images
ECC VCC ECC VCC
R 2 lin/R 2 log P R 2 lin/R 2 log P R 2 lin/R 2 log P R 2 lin/R 2 log P
Global 0.50/0.47 <0.001, S 0.27/0.22 0.062, NS 0.55/0.51 0.007, S 0.53/0.51 0.011, S
ST 0.54/0.48 <0.001, S 0.35/0.32 0.001, S 0.56/0.51 0.018, S 0.52/0.47 0.027, S
SN 0.40/0.37 0.096, NS 0.32/0.25 0.008, S 0.46/0.45 0.218, NS 0.49/0.46 0.030, NS
T
N 0.14/0.14 0.598, NS 0.20/0.17 0.293, NS 0.13/0.12 0.457, NS 0.21/0.21 0.669, NS
IN 0.31/0.30 0.075, NS 0.16/0.14 0.053, NS 0.32/0.31 0.018, S 0.32/0.31 0.023, S
IT 0.41/0.39 0.004, S 0.17/0.13 0.111, NS 0.52/0.51 0.165, NS 0.44/0.42 0.385, NS
Table 5.
 
Coefficients of Determination for Logarithmic Regression Model with VF in the Decibel Scale, and Linear Regression Models with VF in the Antilog Scale, with ECC and VCC
Table 5.
 
Coefficients of Determination for Logarithmic Regression Model with VF in the Decibel Scale, and Linear Regression Models with VF in the Antilog Scale, with ECC and VCC
ECC and VCC R 2 for Logarithmic Regression Model with VF Sensitivity in the Decibel Scale ECC and VCC R 2 for Linear Regression Model with VF Sensitivity in the 1/L Scale
In All Eyes In Eyes without ABP Images In All Eyes In Eyes without ABP Images
ECC/VCC P * ECC/VCC P * ECC/VCC P * ECC/VCC P *
Global 0.50/0.31 0.0083, S 0.54/0.54 0.9916, NS 0.50/0.27 0.0016, S 0.55/0.53 0.8808, NS
ST 0.44/0.31 0.0967, NS 0.46/0.41 0.5892, NS 0.54/0.35 0.0134, S 0.56/0.52 0.6441, NS
SN 0.32/0.26 0.3483, NS 0.43/0.43 0.9668, NS 0.40/0.32 0.3673, NS 0.46/0.49 0.7815, NS
T
N 0.16/0.15 0.8447, NS 0.22/0.17 0.2525, NS 0.14/0.20 0.7222, NS 0.13/0.21 0.2591, NS
IN 0.44/0.20 0.0012, S 0.46/0.33 0.1315, NS 0.31/0.16 <0.0001, S 0.32/0.32 0.9554, NS
IT 0.36/0.14 0.0018, S 0.57/0.40 0.0399, S 0.41/0.17 <0.0001, S 0.52/0.44 0.3269, NS
×
×

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.

×