March 2006
Volume 47, Issue 3
Free
Glaucoma  |   March 2006
Influence of Disease Severity and Optic Disc Size on the Diagnostic Performance of Imaging Instruments in Glaucoma
Author Affiliations
  • Felipe A. Medeiros
    From the Hamilton Glaucoma Center and Department of Ophthalmology, University of California, San Diego, California.
  • Linda M. Zangwill
    From the Hamilton Glaucoma Center and Department of Ophthalmology, University of California, San Diego, California.
  • Christopher Bowd
    From the Hamilton Glaucoma Center and Department of Ophthalmology, University of California, San Diego, California.
  • Pamela A. Sample
    From the Hamilton Glaucoma Center and Department of Ophthalmology, University of California, San Diego, California.
  • Robert N. Weinreb
    From the Hamilton Glaucoma Center and Department of Ophthalmology, University of California, San Diego, California.
Investigative Ophthalmology & Visual Science March 2006, Vol.47, 1008-1015. doi:10.1167/iovs.05-1133
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Felipe A. Medeiros, Linda M. Zangwill, Christopher Bowd, Pamela A. Sample, Robert N. Weinreb; Influence of Disease Severity and Optic Disc Size on the Diagnostic Performance of Imaging Instruments in Glaucoma. Invest. Ophthalmol. Vis. Sci. 2006;47(3):1008-1015. doi: 10.1167/iovs.05-1133.

      Download citation file:


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

      ×
  • Supplements
Abstract

purpose. To evaluate the influence of disease severity and optic disc size on the diagnostic accuracy of three imaging technologies for structural assessment in glaucoma: confocal scanning laser ophthalmoscopy, scanning laser polarimetry, and optical coherence tomography.

methods. One hundred five patients with glaucoma and 61 normal subjects were recruited from the Diagnostic Innovations in Glaucoma Study (DIGS). All individuals underwent imaging with the GDx Variable Corneal and Lens Compensator (VCC; Carl Zeiss Meditec, Inc., Dublin, CA), the Heidelberg Retina Tomograph II (HRT II; Heidelberg Engineering, GmbH, Dossenheim, Germany), and the Stratus OCT (Stratus OCT; Carl-Zeiss Meditec, Inc.) within a 6-month period. Severity of disease was based on the AGIS (Advanced Glaucoma Intervention Study) visual field score. To evaluate the influence of severity of glaucoma and optic disc size on the diagnostic accuracy of the imaging instruments, the sensitivities of the tests were fitted as a function of the AGIS score and disc area, by using logistic marginal regression models.

results. The severity of visual field loss had a significant influence on the sensitivity of all imaging instruments. More severe disease was associated with increased sensitivity. This influence was similar among the three instruments. With regard to optic disc area, larger optic discs were associated with decreased sensitivity for the Stratus OCT parameter Average Thickness and the GDx VCC parameter Nerve Fiber Indicator, whereas small optic discs were associated with increased sensitivity. For the HRT II parameter Moorfields regression analysis classification, an inverse effect was observed.

conclusion. The diagnostic performances of the GDx VCC, HRT II, and Stratus OCT were significantly influenced by the severity of the disease and optic disc size. These covariates should be taken into account when comparing the performances of these tests for glaucoma diagnosis.

The performance of diagnostic tests in medicine can vary among subgroups of patients according to the severity and clinical presentation of the disease. Because of the pioneering work of Ransohoff and Feinstein 1 more than 20 years ago, it became clear that a single value of sensitivity or specificity, two commonly used diagnostic measures, may not be applicable to all subgroups of patients with the disease. 
Diagnostic tests tend to be more sensitive in advanced stages of the disease, and measures of diagnostic accuracy obtained from studies that include only patients with moderate or severe disease may not be applicable to patients in the early stage or suspected of having the disease. 2 3 Therefore, a study evaluating a diagnostic test should include a broad spectrum of patients with the disease so that the performance of the test can be evaluated at different stages of severity, providing valid estimates of accuracy that can be generalized to other populations of patients. 4 5 Moreover, it is possible that the comparison of the diagnostic abilities of different tests is influenced by the severity of the disease. For example, it is possible that a particular test is more sensitive at early stages of the disease, whereas another test may be more sensitive at moderate or advanced stages. Therefore, it is important to characterize the relationship between the performance of the diagnostic tests and the severity of disease and to evaluate how this relationship affects the comparison between different tests. 
Recently, several imaging tests have become available for structural evaluation of the optic disc and retinal nerve fiber layer in glaucoma and to assist in the diagnosis of the disease. Confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph II [HRT II]; Heidelberg Engineering, GmbH, Dossenheim, Germany), scanning laser polarimetry (GDx Variable Corneal and Lens Compensator [VCC]; Carl-Zeiss Meditec, Dublin, CA), and optical coherence tomography (Stratus OCT; Carl-Zeiss Meditec, Inc., Dublin, CA) are three technologies that use different properties of the light to obtain their measurements. 6 7 Although previous studies have evaluated the diagnostic abilities of these instruments, no investigation has yet been performed on the possible influence of disease severity as a variable influencing the performance of these tests and the comparison among them. 
Another covariate that can potentially affect the diagnostic performance of imaging tests in glaucoma is the size of the optic disc. Previous histologic and clinical studies have shown that RNFL thickness and optic disc topographic measurements may be affected by optic disc size. 8 9 10 It is possible, therefore, that the diagnostic accuracy of imaging tests would vary depending on the size of the optic disc. When evaluating the influence of optic disc size, it is important to adjust for the confounding effect of disease severity. It is likely that the influence of optic disc size would be greater at early stages of the disease, as the amount of neural loss in advanced cases would facilitate disease detection, regardless of the size of the optic disc. 
In the present study, we evaluated the influence of disease severity and optic disc size on the diagnostic accuracy of the three mentioned imaging technologies for structural assessment in glaucoma. To accomplish this, we used statistical models that allowed simultaneous evaluation of the effect of these covariates on the comparisons among the three instruments. 
Methods
This was an observational case–control study. Patients from this study were included in a prospective longitudinal study designed to evaluate optic nerve structure and visual function in glaucoma (DIGS; Diagnostic Innovations in Glaucoma Study) conducted at the Hamilton Glaucoma Center (University of California, San Diego). Informed consent was obtained from all participants. The University of California San Diego Human Subjects Committee approved all protocols and the methods described adhered to the tenets of the Declaration of Helsinki. 
Each subject underwent a comprehensive ophthalmologic examination including review of medical history, best corrected visual acuity, slit lamp biomicroscopy, intraocular pressure (IOP) measurement using Goldmann applanation tonometry, gonioscopy, dilated funduscopic examination with a 78-D lens, stereoscopic optic disc photography, and automated perimetry with the 24-2 Swedish Interactive Threshold Algorithm (SITA; Carl Zeiss Meditec, Inc.). To be included, subjects had to have best corrected visual acuity of 20/40 or better, spherical refraction within ± 5.0 D, cylinder correction within ± 3.0 D, and open angles on gonioscopy. Eyes with coexisting retinal disease, uveitis, or nonglaucomatous optic neuropathy were also excluded from the investigation. 
Eyes were classified as glaucomatous if they had repeatable (at least two consecutive) abnormal visual field test results, defined as a pattern standard deviation (PSD) outside the 95% normal confidence limits and/or a Glaucoma Hemifield Test result outside normal limits, regardless of the appearance of the optic disc. Eyes were also classified as glaucomatous if they had documented evidence of progressive glaucomatous change in the appearance of the optic disc as assessed by simultaneous stereoscopic optic disc photographs (TRC-SS; Topcon Instrument Corp. of America, Paramus, NJ), regardless of visual field test results. The evidence of progressive glaucomatous damage had to be present before the imaging test date. The use of this composite reference standard for glaucoma diagnosis allowed us to evaluate the accuracy of diagnostic tests in a broad spectrum of patients with the disease—as patients with visual field loss and patients with normal visual fields, but confirmed progressive glaucomatous optic nerve damage, were included. 
For evaluation of progressive optic disc damage, stereoscopic sets of slides were examined with a stereoscopic viewer (Pentax; Asahi Optical Co., Tokyo, Japan). The photographs were evaluated by two experienced graders, each of whom was masked to the subject’s identity and to the other test results. For inclusion, photographs had to be graded as being of adequate quality or better. To identify a subgroup of patients with progressive glaucomatous optic disc change for this study, our research database was reviewed for all patients who had been imaged using the three instruments and who had been observed for at least 1 year before the imaging test date. A total of 284 patients had stereophotographs graded for progression. For each patient, the most recent stereophotograph was compared to the oldest available one, to maximize the chance of detecting progressive optic disc change. Each observer was masked to the temporal sequence of the photographs. Definition of change was based on focal or diffuse thinning of the neuroretinal rim, increased excavation, or enlargement of RNFL defects. Changes in rim color, presence of disc hemorrhage or progressive parapapillary atrophy were not sufficient for characterization of progression. Discrepancies between the two graders were either resolved by consensus or by adjudication of a third experienced grader. Initial agreement between graders was obtained in 83% of the cases. 
Visual field results were not used as a criterion for inclusion in the study in the population with progressive optic disc change. However, results of the visual field test closest to the imaging date were analyzed as part of the study results in all patients. The AGIS (Advanced Glaucoma Intervention Study) score, described in detail elsewhere, 11 was used to evaluate the severity of visual field loss. It is based on the extent of depression at different locations of the visual field test in the total deviation plot and can range from 0 (no field loss) to 20 (end stage). A computer program was developed in commercial software (MatLab ver.7.0; The MathWorks, Inc., Natick, MA) to allow automatic calculation of the AGIS score from data exported from the Humphrey perimeter (Carl Zeiss Meditec, Inc.). 
Normal control eyes had intraocular pressures of 21 mm Hg or less with no history of increased IOP, and a normal visual field result. A normal visual field was defined as a mean deviation (MD) and PSD within 95% confidence limits, and GHT results within normal limits. Normal control eyes also had a healthy appearance of the optic disc and RNFL (no diffuse or focal rim thinning, optic disc hemorrhage, or RNFL defects), as evaluated by clinical examination. 
All patients had visual field and imaging tests within 6 months. When both eyes of the same patient fulfilled inclusion criteria, one eye was randomly selected for inclusion in the analysis. 
Imaging Instruments
GDx VCC Scanning Laser Polarimeter.
All patients were imaged with a commercially available scanning laser polarimeter with variable corneal compensation (GDx VCC, software ver. 5.5.1; Carl Zeiss Meditec). The general principles of scanning laser polarimetry and the algorithm used for variable corneal compensation have been described in detail elsewhere. 12 13 14 Assessment of GDx VCC image quality was performed by an experienced examiner masked to the subject’s identity and results of the other tests. The assessment was based on the appearance of the reflectance image, the presence of residual anterior segment retardation, and the presence of an atypical pattern of retardation. To be classified as good quality, an image had to have focused and evenly illuminated reflectance with a centered optic disc. To be acceptable, the mean image also had to have residual anterior segment retardation of 15 nm or less and a typical scan score more than 25. The typical scan score is a measure provided by the GDx VCC standard software that indicates the presence of atypical patterns of retardation that can generate spurious RNFL thickness measurements. 
GDx VCC parameters provided in the standard printout of the instrument and investigated in this study were superior average, inferior average, TSNIT (temporal-superior-nasal-inferior-temporal) average, TSNIT SD and the nerve fiber indicator (NFI). The NFI is calculated using a support vector machine algorithm based on several RNFL measures and assigns a number from 0 to 100 to each eye. The higher the NFI, the greater the likelihood the patient has glaucoma. 
HRT II Confocal Scanning Laser Ophthalmoscope.
The HRT II (software ver. 1.4.1.5; Heidelberg Engineering GmbH) uses confocal scanning laser principles to obtain a three-dimensional topographic image of the optic nerve. Its principles of working have been described in detail elsewhere. 15 For each patient, three topographical images were obtained and were combined and automatically aligned to make a single mean topography used for analysis. Magnification errors were corrected using patients’ corneal curvature measurements. An experienced examiner outlined the optic disc margin on the mean topographic image while viewing stereoscopic photographs of the optic disc. Good-quality images had to have a focused reflectance image with an SD not greater than 50 μm. 
For investigation of the diagnostic performance of the HRT II, the present study used the results provided by the Moorfields regression analysis (MRA). The MRA is a comparison of the subject’s rim area with a predicted rim area for a given disc area and age, based on confidence limits of a regression analysis derived from 112 normal eyes of white subjects. 16 Each sector is classified as normal if the measurement falls within 95% confidence interval (CI), borderline if the measurement falls between the 95% and 99.9% CI and outside normal limits if the measurement falls below the 99.9% CI. The MRA also provides results for the global rim area (MRA global), as well as a final classification (MRA classification). A normal MRA classification requires the MRA analysis of all sectors and the global rim area to be within normal limits. A borderline MRA classification occurs when at least one of the sectors or the global rim area is borderline, and an outside-normal-limits result occurs when test results of at least one sector or the global rim area are outside normal limits. 
The HRT II also provides measurements of optic disc area, which were used in the present study to evaluate the influence of optic disc size on the diagnostic performance of the instruments. 
Stratus OCT.
The commercially available optical coherence tomograph, Stratus OCT (software version 4.0; Carl Zeiss Meditec, Inc.), was used to assess parapapillary RNFL thickness measurements. Optical coherence tomography employs the principles of low coherence interferometry and is analogous to ultrasound B-mode imaging, but utilizes light instead of sound to acquire high-resolution images of ocular structures. 17 More detail of its principles of operation can be found in several publications. 6 7 17 The fast RNFL algorithm was used to obtain RNFL thickness measurements with the Stratus OCT. Three images were acquired from each subject, with each image consisting of 256 A-scans along a 3.4-mm-diameter circular ring around the optic disc. A mean image was automatically created by the Stratus OCT software. Quality assessment of Stratus OCT scans was evaluated by an experienced examiner masked to the subject’s identity and to the results of the other tests. Good-quality scans had to have a centered circular ring around the optic disc, an adequate signal strength (≥6) and no sign of failure of the algorithm for detection of the RNFL boundaries. 
Parapapillary RNFL thickness parameters evaluated in this study were average (360° measure), temporal quadrant (316–45°), superior quadrant (46–135°), nasal quadrant (136–225°), and inferior quadrant (226°–315°). 
Statistical Analysis
Descriptive statistics included mean and SD for normally distributed variables, and median, first-quartile and third-quartile values for non-normally distributed variables. Student’s t-tests or Mann-Whitney tests were used to evaluate demographic and clinical differences between glaucomatous and normal subjects. For categorical variables, χ2 tests or Fisher exact tests were used. 
To evaluate the influence of severity of glaucoma on the diagnostic accuracy of the imaging instruments, the sensitivities of the tests were fitted as a function of the AGIS score with a logistic marginal regression modeling approach, as proposed by Leisenring et al. 18 and others, 19 20 Logistic models have long been used to evaluate influence of covariates on diagnostic tests. 21 Previous methods, however, were limited by the fact that they could not accommodate correlated data—that is, it was not possible to compare directly the effect of covariates on several tests performed on the same group of patients. In contrast, the methodology proposed by Leisenring et al. 18 allows the comparison of the sensitivities of different tests performed in the same group of patients, adjusting for the severity of disease. Inference regarding model parameters is easily implemented with the use of generalized estimating equations to take into account the correlations between observations. These models also allow the evaluation of other covariates potentially affecting diagnostic performance, such as the area of the optic disc. For inclusion in the logistic model, the diagnostic tests have to be dichotomized so that their sensitivities can be compared at the same level of specificity. The result is a binary variable for sensitivity that is 1 or 0, according to the result of the diagnostic test (i.e., above or below the cutoff) in each one of the glaucomatous subjects. This binary variable for sensitivity is entered as the dependent variable in the logistic regression model, whereas the covariates (i.e., test type, disease severity, and optic disc size) are entered as independent categorical or continuous variables. As sensitivity is the parameter being studied, only the information from glaucomatous subjects is included in the model, although normal subjects are used to obtain cutoffs according to the specificity level. 
The general form for the logistic regression model is  
\[\mathrm{logit}(S)\ {=}\ {\beta}_{0}\ {+}\ {\beta}_{1}X_{1}\ {+}\ {\beta}_{2}X_{2}\ {+}\ {\ldots}{+}{\beta}_{n}X_{n},\]
where S is the sensitivity, X is the covariate, and β is the coefficient of the model. 
The sensitivity can then be calculated by  
\[\frac{\mathrm{exp}({\beta}_{0}\ {+}\ {\beta}_{1}X_{1}\ {+}\ {\beta}_{2}X_{2}\ {+}\ {\ldots}\ {+}\ {\beta}_{n}X_{n})}{1\ {+}\ \mathrm{exp}({\beta}_{0}\ {+}\ {\beta}_{1}X_{1}\ {+}\ {\beta}_{2}X_{2}\ {+}\ {\ldots}\ {+}\ {\beta}_{n}X_{n})}\ {\cdot}\]
 
By making one of the covariates, let’s say X 1, a dummy variable indicating test type, the sensitivities of each test can be calculated separately. In the present study, interactions terms were also included in the logistic regression model to evaluate whether the degree of severity of visual field loss or the size of the optic disc influenced the sensitivities of each of the tests in a similar or different fashion. 
P < 0.05 was considered statistically significant. Statistical analyses were performed on computer (SPSS ver.13.0; SPSS Inc., Chicago, IL; and S-PLUS ver. 6.0; Mathsoft Inc., Seattle, WA). 
Results
The study included 105 patients with glaucoma and 61 normal subjects. Table 1shows demographic and clinical characteristics of the subjects included in the study. Of the 105 patients with glaucoma, 84 (80%) had visual field loss in the visual field test closest to the imaging date, whereas 21 (20%) patients had normal visual fields, but progressive optic disc damage was demonstrated on stereophotographs. 
The average AGIS score of the 105 patients with glaucoma was 3.04 (median, 1.00; first quartile, 0; third quartile, 4.50). Figure 1shows the distribution of AGIS scores in the patients with glaucoma included in the study. 
To estimate levels of sensitivity and specificity for the HRT II MRA parameters, borderline MRA results were included as outside normal limits. This approach resulted in moderate sensitivities and specificities for the best MRA parameter, MRA classification (Table 2) . Although some of the other MRA parameters had higher specificities, their sensitivities were much lower. Based on the specificity of 83% found for the MRA classification, the parameters of the GDx VCC and Stratus OCT were dichotomized to match this specificity as closely as possible so that the tests could be compared at the same level of specificity in the control group. Indeed, the specificities of these three parameters were matched at exactly the same value of 83%. Table 2shows sensitivities for fixed specificities for the parameters evaluated in the study, regardless of disease severity. For each imaging instrument, the parameter with highest sensitivity was selected for inclusion in the logistic regression model: Average Thickness for the Stratus OCT, NFI for the GDx VCC, and MRA classification for the HRT II. The logistic model of sensitivity, incorporating AGIS score and optic disc size as independent variables, had the following form:  
\[\mathrm{logit}(S)\ {=}\ {\beta}_{0}\ {+}\ {\beta}_{1}GDx\ {+}\ {\beta}_{2}HRT\ {+}\ {\beta}_{3}AGIS\ {+}\ {\beta}_{4}AGIS\ {\times}\ GDx\ {+}\ {\beta}_{5}AGIS\ {\times}\ HRT\ {+}\ {\beta}_{6}Disc\ {+}\ {\beta}_{7}Disc\ {\times}\ GDx\ {+}\ {\beta}_{8}Disc\ {\times}\ HRT,\]
where S is to the sensitivity of the test (at fixed specificity), GDx and HRT are variables coding for the type of test (with Stratus OCT used as the reference test), AGIS represents the AGIS scores for severity of visual field defect, Disc represents the optic disc area, and the variables AGIS × GDx, AGIS × HRT, Disc × GDx, and Disc × HRT represent interactions between severity of disease or optic disc area and type of test. 
The estimates of the coefficients of the regression model are shown in Table 3 . The severity of visual field loss had a significant influence on the sensitivity of the imaging instruments, as indicated by the statistically significant coefficient representing the AGIS score (P = 0.036). This influence was similar among the three instruments—that is, the increase in sensitivity with increasing severity of visual field loss occurred in a similar way for all three instruments, as indicated by the nonsignificant coefficients for the interaction terms AGIS × GDx (P = 0.769) and AGIS × HRT (P = 0.156). Table 4shows a simplified regression model after exclusion of the interaction terms between disease severity and type of test. 
With regard to optic disc area, larger optic discs were associated with decreasing sensitivity for the Stratus OCT parameter average thickness, as observed by the negative and significant value of the coefficient (β6) representing optic disc area (Table 4 , Fig. 2 ). A similar effect was also observed for the GDx VCC parameter NFI. Larger optic discs were associated with decreasing sensitivity, whereas small optic discs were associated with increasing sensitivity (Fig. 3) . The magnitude of this effect was smaller than with the Stratus OCT parameter average thickness, although not significantly different, as evident by the lack of statistical significance of the coefficient (β7) representing the interaction term Disc × GDx in the multivariate model (P = 0.303). For the HRT II parameter MRA classification, an inverse effect was observed. Larger optic discs were associated with increasing sensitivity, whereas smaller optic discs were associated with decreasing sensitivity (Fig. 4) . This result is indicated by the positivity and statistical significance of the interaction term Disc × HRT in the logistic regression model (P = 0.015). Note that, because Stratus OCT was used as the reference test, the evaluation of the effect of optic disc area on GDx VCC or HRT II required that the coefficients of the interaction terms involving these tests be added to the coefficient representing the main effect of this covariate. Therefore, the final effect of disc area on GDx VCC is represented by the sum of the coefficients β6 and β7, whereas the effect of disc area on the HRT II is represented by the sum of the coefficients β6 and β8. In contrast, for the Stratus OCT, the effect of optic disc area on this test is represented only by the coefficient β6
Table 5shows sensitivities for different AGIS scores and different optic disc sizes with all three imaging instruments. For smaller optic discs, the sensitivities of the Stratus OCT parameter average thickness and the GDx VCC parameter NFI were higher than that of the HRT II parameter MRA classification. However, for larger optic discs, the relationship between sensitivities of the different instruments changed. The sensitivity of the HRT II MRA classification tended to be similar to that of the GDx VCC NFI and higher than that of the Stratus OCT parameter average thickness. 
Discussion
In the present study, we demonstrated that the severity of visual field loss and size of the optic disc had significant influence on the performance of three imaging devices used for glaucoma diagnosis—the GDx VCC scanning laser polarimeter, the HRT II confocal scanning laser ophthalmoscope, and the Stratus OCT. Our results showed the importance of taking into account these variables when evaluating and comparing the performances of these instruments. They could also, at least partially, explain previous disagreements found among these technologies when used for glaucoma detection. 6 22 23  
The severity of visual field loss, as measured by the AGIS score, had a significant influence on the sensitivity of all the three evaluated imaging devices. The increase in sensitivity with increasing severity of visual field loss occurred in a similar manner for the GDx VCC, the HRT II, and the Stratus OCT parameters. The inclusion of patients with glaucoma without visual field loss along with patients with early, moderate, and advanced visual field defects allowed us to evaluate the performance of these tests in a broad spectrum of patients with the disease. It should be emphasized that the diagnosis of glaucoma in patients without visual field loss was performed based on documented evidence of progressive optic disc damage on stereophotographs and, therefore, was more robust than a diagnosis of glaucomatous optic neuropathy based on an optic disc examination at a single time point, as used in other studies. This approach was originally proposed by Medeiros et al. 24 for evaluation of diagnostic tests in glaucoma. In the absence of visual field loss, a diagnosis of certainty of glaucoma usually requires demonstrating a previous history of progressive glaucomatous changes to the optic nerve. It should be noted, however, that this inclusion criterion is still limited in its ability to identify glaucoma. It is possible that patients with glaucomatous damage to the optic nerve but no visual field loss were not included in the study simply because they did not progress during the follow-up period covered by stereophotographs. This would limit the inclusion of patients with early disease who did not show progressive disc damage. However, this is a current limitation of all diagnostic studies in glaucoma, as no perfect reference standard exists for this disease. 
Previous studies have also evaluated the influence of optic disc size on the diagnostic accuracy of imaging instruments in glaucoma. 25 26 27 28 Most of these studies have divided their patient populations in subgroups according to the size of the optic disc and compared measures of diagnostic accuracy between the different subgroups. This approach, however, can lead to an important loss of power to detect significant differences. 29 Another important limitation of most of the previous studies is the lack of adjustment for the severity of disease when comparing diagnostic measures in subgroups of patients with different optic disc sizes. It is likely that variables influencing the diagnostic performance of imaging tests would have more influence in early stages of the disease than in advanced ones. In patients with advanced glaucoma, the large amount of neural loss would presumably allow detection of abnormality, regardless of optic disc size. Moreover, it is possible that subgroups of patients divided arbitrarily on the basis of optic disc size would present different stages of disease severity, introducing a confounding factor into the comparison. In the present study, we used statistical models to analyze the effect of severity of disease and optic disc size on the diagnostic tests. This approach is more powerful than the analysis of stratified subgroups, which would be more susceptible to variability due to the small number of subjects in each subgroup. 
We demonstrated that the size of the optic disc had a significant influence on the sensitivity for glaucoma diagnosis of all three imaging devices. This influence, however, occurred in different ways for the GDx VCC and Stratus OCT than for the HRT II. For the HRT II, larger optic discs were associated with increased sensitivity for glaucoma detection using the Moorfields regression analysis, whereas smaller discs were associated with decreased sensitivity. In contrast, for the GDx VCC parameter NFI and Stratus OCT parameter average thickness, larger optic discs were associated with decreased sensitivity for glaucoma detection, whereas smaller discs were associated with increased sensitivity. As expected, this effect was more pronounced in the early and moderate stages of the disease compared with the advanced stages, as illustrated by the Figures 2 3 and 4 . In these graphs, it was clear that the relationship between sensitivity of the test and optic disc size was steeper in the areas of lower values of the AGIS score, compared with areas with higher scores and more advanced disease. The results presented in Table 5clearly show the importance of taking into account the size of the optic disc when comparing the diagnostic performance of these instruments for glaucoma detection. Furthermore, the estimates of sensitivity in patients with AGIS score equal to zero provide an indication and comparison of the performance of these instruments in the detection of glaucomatous damage before apparent visual field loss on standard automated perimetry, a finding that has not been reported before. 
Previous studies have found similar results regarding the relationship between optic disc size and the diagnostic accuracy of HRT parameters. 25 27 28 30 In early glaucomatous damage, the loss of neuroretinal rim in a small disc is presumably more difficult to detect than in a larger disc, as the former would still present a proportion between rim and cup similar to that of a normal disc. Although the Moorfields regression analysis was developed to take into account the size of the optic disc when evaluating rim area, our results suggest that this parameter is still influenced by the size of the optic disc. Ford et al. 28 found similar results when studying 104 patients with glaucoma and 48 normal subjects stratified in three groups according to the optic disc size. In their study, higher sensitivities of the Moorfields regression analysis parameters were found in patients with larger discs than in those with smaller discs. 
Previous studies demonstrated a positive correlation between the size of the optic disc and RNFL thickness measurements obtained by OCT. 31 32 In those studies, subjects with larger discs were found to have higher RNFL thickness measurements, whereas individuals with smaller discs had lower measurements. This correlation is supported by histologic studies demonstrating that larger optic discs show a larger number of nerve fibers. 8 These findings could explain the decrease in sensitivity of the Stratus OCT for glaucoma detection in patients with large discs, as patients with early glaucomatous damage and large optic discs would still present thicknesses considered within normal limits. A similar reasoning could be applicable to the GDx VCC. However, the reasons for the decrease in performance of the GDx VCC and Stratus OCT for glaucoma detection in patients with large discs remain speculative, and further research is necessary to evaluate this question. 
To estimate levels of specificity and sensitivity for the parameters included in the study, we first analyzed the results of the HRT II MRA analysis. Because the results of the MRA analysis are categorical, only two levels of specificity are possible for this parameter, one after dichotomizing the test by including borderline results as outside normal limits and the other by including borderline results as within normal limits. As the first approach generates higher levels of sensitivity, we have opted to use it when comparing sensitivities of the different tests in the present study. Although the second approach resulted in higher levels of specificity (≥95%), the sensitivities were low, ranging from 3% to 53% for the different MRA parameters. When we reanalyzed the data using specificity close to 95% for the parameters of GDx VCC, HRT II, and Stratus OCT, we found similar results in terms of the relationship between optic disc size and disease severity on the sensitivity of the tests, as those obtained using specificity close to 80%. 
The parameters from HRT II, GDx VCC, and Stratus OCT included in the regression model were also the ones that demonstrated best diagnostic performance in other studies evaluating these instruments. 6 24 33 34 It should be noted that differences on how these parameters are calculated could also be related to some of the differences among the instruments found in our study. For example the HRT II MRA takes into account disc size in its analysis, whereas GDx VCC and Stratus OCT analyses do not. Also, the GDx VCC NFI is calculated from several measures of RNFL thickness, including not only those obtained in the circle of fixed diameter around the optic disc, but also measurements performed far from the disc, which could be less influenced by the size of the optic disc. This could explain, at least partially, the relatively smaller effect of optic disc size on the diagnostic accuracy of this instrument compared with the Stratus OCT, as OCT RNFL thickness measurements are acquired only in the circle of fixed diameter around the optic disc. 
In the present study, we used only RNFL parameters to evaluate the performance of the Stratus OCT. However, this instrument provides optic nerve head topographic parameters and macular thickness parameters that have been demonstrated to be useful in glaucoma. 35 36 37 Although a more comprehensive evaluation of this instrument would also include the above-mentioned scanning areas, previous studies comparing the several Stratus OCT analysis algorithms have demonstrated that the RNFL measures perform as well as the optic nerve head parameters and significantly better than macular thickness parameters. 35 36 37  
In conclusion, the diagnostic performances of the GDx VCC, HRT II, and Stratus OCT were demonstrated to be significantly influenced by the severity of the disease and the size of the optic disc. These covariates should be taken into account when comparing the performances of these tests for glaucoma diagnosis. 
 
Table 1.
 
Demographic and Clinical Characteristics of the Subjects Included in the Study
Table 1.
 
Demographic and Clinical Characteristics of the Subjects Included in the Study
Parameter Glaucoma (n = 105) Normal (N = 61) P
Age, y (mean ± SD) 68 ± 12 67 ± 9 0.761
Gender, female, n (%) 57 (54) 48 (79) 0.002
Race, n (%) 0.888
 White 78 (74) 45 (74)
 African American 18 (17) 12 (20)
 Asian 6 (6) 2 (3)
 Hispanic 3 (3) 2 (3)
MD (dB) −3.00 (−6.12; −1.82) −0.80 (−1.55; −0.14) <0.001
PSD (dB) 3.11 (2.10; 7.29) 1.66 (1.44; 1.82) <0.001
Figure 1.
 
Distribution of AGIS scores among the 105 patients with glaucoma included in the study.
Figure 1.
 
Distribution of AGIS scores among the 105 patients with glaucoma included in the study.
Table 2.
 
Sensitivities at Specificities ≥80% for the Parameters Evaluated in the Study
Table 2.
 
Sensitivities at Specificities ≥80% for the Parameters Evaluated in the Study
GDx VCC Sn/Sp (%) Stratus OCT Sn/Sp (%) HRT II* Sn/Sp (%)
NFI 86/83 Average thickness 79/83 MRA classification 70/83
TSNIT average 77/80 Superior thickness 69/80 MRA global 42/97
Superior average 79/80 Inferior thickness 79/83 MRA temporal 15/98
Inferior average 72/80 Temporal thickness 39/83 MRA temp/superior 33/95
TSNIT SD 66/80 Nasal thickness 47/80 MRA temp/inferior 47/93
MRA nasal 32/98
MRA nasal superior 44/92
MRA nasal inferior 42/97
Table 3.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates
Table 3.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates
Variable Coefficient Estimate 95% CI P
Constant β0 0.738 (0.067–1.409)
Type of test
 GDx VCC NFI β1 0.470 (−0.400–1.342) 0.289
 HRT II MRA classification β2 −0.189 (−0.948–0.570) 0.625
AGIS score β3 0.291 (0.019–0.526) 0.036
AGIS × GDx VCC NFI β4 0.055 (−0.314–0.426) 0.769
AGIS × HRT II MRA classification β5 −0.184 (−0.438–0.070) 0.156
Optic disc area β6 −1.140 (−2.238–−0.042) 0.042
Optic disc area × GDx VCC NFI β7 0.663 (−0.514–1.840) 0.270
Optic disc area × HRT II MRA classification β8 1.613 (0.378–2.850) 0.010
Table 4.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates, but Excluding Interaction Terms between Severity of the Disease and Test Type
Table 4.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates, but Excluding Interaction Terms between Severity of the Disease and Test Type
Variable Coefficient Estimate 95% CI P
Constant β0 0.897 (0.342–1.452)
Type of test
 GDx VCC NFI β1 0.526 (−0.100–1.153) 0.100
 HRT II MRA classification β2 −0.551 (−1.152–0.048) 0.071
AGIS score β3 0.198 (0.075–0.321) 0.002
Optic disc area β6 −1.040 (−2.049–−0.033) 0.043
Optic disc area × GDx VCC NFI β7 0.646 (−0.584–1.875) 0.303
Optic disc area × HRT II MRA classification β8 1.471 (0.291–2.651) 0.015
Figure 2.
 
Relationship between sensitivities for the Stratus OCT parameter average thickness and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 2.
 
Relationship between sensitivities for the Stratus OCT parameter average thickness and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 3.
 
Relationship between sensitivity for the GDx VCC parameter nerve fiber indicator and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 3.
 
Relationship between sensitivity for the GDx VCC parameter nerve fiber indicator and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 4.
 
Relationship between sensitivities for the HRT II Moorfields regression analysis classification and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with decreasing optic disc area, and this effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 4.
 
Relationship between sensitivities for the HRT II Moorfields regression analysis classification and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with decreasing optic disc area, and this effect was more pronounced at the early stages of the disease (lower AGIS scores).
Table 5.
 
Sensitivities at Fixed Specificity at 83%, According to the Regression Model for Different AGIS Scores and Optic Disc Sizes
Table 5.
 
Sensitivities at Fixed Specificity at 83%, According to the Regression Model for Different AGIS Scores and Optic Disc Sizes
Parameter Optic Disc Area (mm2) AGIS Score
0 3 6 9
HRT II MRA classification 1.00 47 62 74 84
2.00 58 71 82 89
3.00 68 79 87 93
Stratus OCT average thickness 1.00 88 93 96 98
2.00 73 83 90 94
3.00 50 63 76 85
GDx VCC NFI 1.00 86 92 95 97
2.00 81 88 93 96
3.00 74 84 90 94
RansohoffDF, FeinsteinAR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. N Engl J Med. 1978;299:926–930. [CrossRef] [PubMed]
LachsMS, NachamkinI, EdelsteinPH, GoldmanJ, FeinsteinAR, SchwartzJS. Spectrum bias in the evaluation of diagnostic tests: lessons from the rapid dipstick test for urinary tract infection. Ann Intern Med. 1992;117:135–140. [CrossRef] [PubMed]
MoonsKG, van EsGA, DeckersJW, HabbemaJD, GrobbeeDE. Limitations of sensitivity, specificity, likelihood ratio, and Bayes’ theorem in assessing diagnostic probabilities: a clinical example. Epidemiology. 1997;8:12–17. [CrossRef] [PubMed]
PhilbrickJT, HorwitzRI, FeinsteinAR, LangouRA, ChandlerJP. The limited spectrum of patients studied in exercise test research: analyzing the tip of the iceberg. JAMA. 1982;248:2467–2470. [CrossRef] [PubMed]
MulherinSA, MillerWC. Spectrum bias or spectrum effect?—subgroup variation in diagnostic test evaluation. Ann Intern Med. 2002;137:598–602. [CrossRef] [PubMed]
MedeirosFA, ZangwillLM, BowdC, WeinrebRN. Comparison of the GDx VCC scanning laser polarimeter, HRT II confocal scanning laser ophthalmoscope, and stratus OCT optical coherence tomograph for the detection of glaucoma. Arch Ophthalmol. 2004;122:827–837. [CrossRef] [PubMed]
ZangwillLM, MedeirosFA, BowdC, WeinrebRN. Optic nerve imaging: recent advances.KrieglsteinGK WeinrebRN eds. Essentials in Ophthalmology: Glaucoma. 2004;63–91.Springer New York.
JonasJB, SchmidtAM, Muller-BerghJA, Schlotzer-SchrehardtUM, NaumannGO. Human optic nerve fiber count and optic disc size. Invest Ophthalmol Vis Sci. 1992;33:2012–2018. [PubMed]
VarmaR, SkafM, BarronE. Retinal nerve fiber layer thickness in normal human eyes. Ophthalmology. 1996;103:2114–2119. [CrossRef] [PubMed]
RadiusRL. Thickness of the retinal nerve fiber layer in primate eyes. Arch Ophthalmol. 1980;98:1625–1629. [CrossRef] [PubMed]
Advanced Glaucoma Intervention Study. 2. Visual field test scoring and reliability. Ophthalmology. 1994;101:1445–1455. [CrossRef] [PubMed]
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, DreherAW, ColemanA, QuigleyH, ShawB, ReiterK. 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]
WeinrebRN. Laser scanning tomography to diagnose and monitor glaucoma. Curr Opin Ophthalmol. 1993;4:3–6. [CrossRef]
WollsteinG, Garway-HeathDF, HitchingsRA. Identification of early glaucoma cases with the scanning laser ophthalmoscope. Ophthalmology. 1998;105:1557–1563. [CrossRef] [PubMed]
HuangD, SwansonEA, LinCP, et al. Optical coherence tomography. Science. 1991;254:1178–1181. [CrossRef] [PubMed]
LeisenringW, PepeMS, LongtonG. A marginal regression modelling framework for evaluating medical diagnostic tests. Stat Med. 1997;16:1263–1281. [CrossRef] [PubMed]
MartusP, StrouxA, JunemannAM, et al. GEE approaches to marginal regression models for medical diagnostic tests. Stat Med. 2004;23:1377–1398. [CrossRef] [PubMed]
LiangKY, ZegerSL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22. [CrossRef]
CoughlinSS, TrockB, CriquiMH, PickleLW, BrownerD, TefftMC. The logistic modeling of sensitivity, specificity, and predictive value of a diagnostic test. J Clin Epidemiol. 1992;45:1–7. [CrossRef] [PubMed]
ZangwillLM, BowdC, BerryCC, et al. Discriminating between normal and glaucomatous eyes using the Heidelberg Retina Tomograph, GDx Nerve Fiber Analyzer, and Optical Coherence Tomograph. Arch Ophthalmol. 2001;119:985–993. [CrossRef] [PubMed]
GreaneyMJ, HoffmanDC, Garway-HeathDF, NaklaM, ColemanAL, CaprioliJ. Comparison of optic nerve imaging methods to distinguish normal eyes from those with glaucoma. Invest Ophthalmol Vis Sci. 2002;43:140–145. [PubMed]
MedeirosFA, ZangwillLM, BowdC, SamplePA, WeinrebRN. Use of progressive glaucomatous optic disk change as the reference standard for evaluation of diagnostic tests in glaucoma. Am J Ophthalmol. 2005;139:1010–1018. [CrossRef] [PubMed]
IesterM, MikelbergFS, DranceSM. The effect of optic disc size on diagnostic precision with the Heidelberg retina tomograph. Ophthalmology. 1997;104:545–548. [CrossRef] [PubMed]
IesterM, JonasJB, MardinCY, BuddeWM. Discriminant analysis models for early detection of glaucomatous optic disc changes. Br J Ophthalmol. 2000;84:464–468. [CrossRef] [PubMed]
BathijaR, ZangwillL, BerryCC, SamplePA, WeinrebRN. Detection of early glaucomatous structural damage with confocal scanning laser tomography. J Glaucoma. 1998;7:121–127. [PubMed]
FordBA, ArtesPH, McCormickTA, NicolelaMT, LeBlancRP, ChauhanBC. Comparison of data analysis tools for detection of glaucoma with the Heidelberg Retina Tomograph. Ophthalmology. 2003;110:1145–1150. [CrossRef] [PubMed]
StreinerDL. Breaking up is hard to do: the heartbreak of dichotomizing continuous data. Can J Psychiatry. 2002;47:262–266. [PubMed]
MardinCY, HornFK. Influence of optic disc size on the sensitivity of the Heidelberg Retina Tomograph. Graefes Arch Clin Exp Ophthalmol. 1998;236:641–645. [CrossRef] [PubMed]
SaviniG, ZaniniM, CarelliV, SadunAA, Ross-CisnerosFN, BarboniP. Correlation between retinal nerve fibre layer thickness and optic nerve head size: an optical coherence tomography study. Br J Ophthalmol. 2005;89:489–492. [CrossRef] [PubMed]
BowdC, ZangwillLM, BlumenthalEZ, et al. Imaging of the optic disc and retinal nerve fiber layer: the effects of age, optic disc area, refractive error, and gender. J Opt Soc Am A Opt Image Sci Vis. 2002;19:197–207. [CrossRef] [PubMed]
ReusNJ, LemijHG. Diagnostic accuracy of the GDx VCC for glaucoma. Ophthalmology. 2004;111:1860–1865. [CrossRef] [PubMed]
MedeirosFA, ZangwillLM, BowdC, MohammadiK, WeinrebRN. Comparison of scanning laser polarimetry using variable corneal compensation and retinal nerve fiber layer photography for detection of glaucoma. Arch Ophthalmol. 2004;122:698–704. [CrossRef] [PubMed]
MedeirosFA, ZangwillLM, BowdC, VessaniRM, SusannaR, Jr, WeinrebRN. Evaluation of retinal nerve fiber layer, optic nerve head, and macular thickness measurements for glaucoma detection using optical coherence tomography. Am J Ophthalmol. 2005;139:44–55. [CrossRef] [PubMed]
WollsteinG, IshikawaH, WangJ, BeatonSA, SchumanJS. Comparison of three optical coherence tomography scanning areas for detection of glaucomatous damage. Am J Ophthalmol. 2005;139:39–43. [CrossRef] [PubMed]
LeungCK, ChanWM, HuiYL, et al. Analysis of retinal nerve fiber layer and optic nerve head in glaucoma with different reference plane offsets, using optical coherence tomography. Invest Ophthalmol Vis Sci. 2005;46:891–899. [CrossRef] [PubMed]
Figure 1.
 
Distribution of AGIS scores among the 105 patients with glaucoma included in the study.
Figure 1.
 
Distribution of AGIS scores among the 105 patients with glaucoma included in the study.
Figure 2.
 
Relationship between sensitivities for the Stratus OCT parameter average thickness and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 2.
 
Relationship between sensitivities for the Stratus OCT parameter average thickness and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 3.
 
Relationship between sensitivity for the GDx VCC parameter nerve fiber indicator and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 3.
 
Relationship between sensitivity for the GDx VCC parameter nerve fiber indicator and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with increasing optic disc area, and the effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 4.
 
Relationship between sensitivities for the HRT II Moorfields regression analysis classification and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with decreasing optic disc area, and this effect was more pronounced at the early stages of the disease (lower AGIS scores).
Figure 4.
 
Relationship between sensitivities for the HRT II Moorfields regression analysis classification and optic disc area, according to the level of visual field damage measured by the AGIS score. Sensitivity for glaucoma detection decreased with decreasing optic disc area, and this effect was more pronounced at the early stages of the disease (lower AGIS scores).
Table 1.
 
Demographic and Clinical Characteristics of the Subjects Included in the Study
Table 1.
 
Demographic and Clinical Characteristics of the Subjects Included in the Study
Parameter Glaucoma (n = 105) Normal (N = 61) P
Age, y (mean ± SD) 68 ± 12 67 ± 9 0.761
Gender, female, n (%) 57 (54) 48 (79) 0.002
Race, n (%) 0.888
 White 78 (74) 45 (74)
 African American 18 (17) 12 (20)
 Asian 6 (6) 2 (3)
 Hispanic 3 (3) 2 (3)
MD (dB) −3.00 (−6.12; −1.82) −0.80 (−1.55; −0.14) <0.001
PSD (dB) 3.11 (2.10; 7.29) 1.66 (1.44; 1.82) <0.001
Table 2.
 
Sensitivities at Specificities ≥80% for the Parameters Evaluated in the Study
Table 2.
 
Sensitivities at Specificities ≥80% for the Parameters Evaluated in the Study
GDx VCC Sn/Sp (%) Stratus OCT Sn/Sp (%) HRT II* Sn/Sp (%)
NFI 86/83 Average thickness 79/83 MRA classification 70/83
TSNIT average 77/80 Superior thickness 69/80 MRA global 42/97
Superior average 79/80 Inferior thickness 79/83 MRA temporal 15/98
Inferior average 72/80 Temporal thickness 39/83 MRA temp/superior 33/95
TSNIT SD 66/80 Nasal thickness 47/80 MRA temp/inferior 47/93
MRA nasal 32/98
MRA nasal superior 44/92
MRA nasal inferior 42/97
Table 3.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates
Table 3.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates
Variable Coefficient Estimate 95% CI P
Constant β0 0.738 (0.067–1.409)
Type of test
 GDx VCC NFI β1 0.470 (−0.400–1.342) 0.289
 HRT II MRA classification β2 −0.189 (−0.948–0.570) 0.625
AGIS score β3 0.291 (0.019–0.526) 0.036
AGIS × GDx VCC NFI β4 0.055 (−0.314–0.426) 0.769
AGIS × HRT II MRA classification β5 −0.184 (−0.438–0.070) 0.156
Optic disc area β6 −1.140 (−2.238–−0.042) 0.042
Optic disc area × GDx VCC NFI β7 0.663 (−0.514–1.840) 0.270
Optic disc area × HRT II MRA classification β8 1.613 (0.378–2.850) 0.010
Table 4.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates, but Excluding Interaction Terms between Severity of the Disease and Test Type
Table 4.
 
Results of the Logistic Regression Model of Sensitivity Incorporating Disease Severity and Optic Disc Area as Covariates, but Excluding Interaction Terms between Severity of the Disease and Test Type
Variable Coefficient Estimate 95% CI P
Constant β0 0.897 (0.342–1.452)
Type of test
 GDx VCC NFI β1 0.526 (−0.100–1.153) 0.100
 HRT II MRA classification β2 −0.551 (−1.152–0.048) 0.071
AGIS score β3 0.198 (0.075–0.321) 0.002
Optic disc area β6 −1.040 (−2.049–−0.033) 0.043
Optic disc area × GDx VCC NFI β7 0.646 (−0.584–1.875) 0.303
Optic disc area × HRT II MRA classification β8 1.471 (0.291–2.651) 0.015
Table 5.
 
Sensitivities at Fixed Specificity at 83%, According to the Regression Model for Different AGIS Scores and Optic Disc Sizes
Table 5.
 
Sensitivities at Fixed Specificity at 83%, According to the Regression Model for Different AGIS Scores and Optic Disc Sizes
Parameter Optic Disc Area (mm2) AGIS Score
0 3 6 9
HRT II MRA classification 1.00 47 62 74 84
2.00 58 71 82 89
3.00 68 79 87 93
Stratus OCT average thickness 1.00 88 93 96 98
2.00 73 83 90 94
3.00 50 63 76 85
GDx VCC NFI 1.00 86 92 95 97
2.00 81 88 93 96
3.00 74 84 90 94
×
×

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.

×