July 2016
Volume 57, Issue 9
Open Access
Articles  |   August 2016
Comparison of Bruch's Membrane Opening Minimum Rim Width and Peripapillary Retinal Nerve Fiber Layer Thickness in Early Glaucoma Assessment
Author Affiliations & Notes
  • Jonas M. D. Gmeiner
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Wolfgang A. Schrems
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Christian Y. Mardin
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Robert Laemmer
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Friedrich E. Kruse
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Laura M. Schrems-Hoesl
    Department of Ophthalmology University of Erlangen-Nuremberg, Erlangen, Germany
  • Correspondence: Wolfgang A. Schrems, Department of Ophthalmology, University of Erlangen-Nuremberg, Schwabachanlage 6, 91054 Erlangen, Germany; [email protected]
Investigative Ophthalmology & Visual Science August 2016, Vol.57, OCT575-OCT584. doi:https://doi.org/10.1167/iovs.15-18906
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jonas M. D. Gmeiner, Wolfgang A. Schrems, Christian Y. Mardin, Robert Laemmer, Friedrich E. Kruse, Laura M. Schrems-Hoesl; Comparison of Bruch's Membrane Opening Minimum Rim Width and Peripapillary Retinal Nerve Fiber Layer Thickness in Early Glaucoma Assessment. Invest. Ophthalmol. Vis. Sci. 2016;57(9):OCT575-OCT584. https://doi.org/10.1167/iovs.15-18906.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: To compare the diagnostic value of Bruch's membrane opening minimum rim width (BMO-MRW) and retinal nerve fiber layer thickness (RNFLT) in patients with ocular hypertension, preperimetric, and perimetric glaucoma.

Methods: One hundred eighty-one eyes consisting of 40 healthy controls, 41 ocular hypertensive subjects, 50 subjects with preperimetric glaucoma and 50 with perimetric glaucoma were included. One randomly selected eye was included. All patients underwent slit-lamp examination, funduscopy, achromatic perimetry, and 24-hour IOP profile. Bruch's membrane opening–MRW and RNFLT (3 peripapillary circle scans, 12°/14°/16°) data were obtained using spectral domain optical coherence tomography. Areas under the receiver operating characteristics curves (AUROC) as well as sensitivity at fixed specificity were computed globally and for six vertical split sectors. Venn diagrams were plotted to identify patients that were diagnosed by one of the two parameters only.

Results: For RNFLT the smallest circle diameter showed highest diagnostic accuracy and was used for comparison with BMO-MRW. Distinguishing perimetric glaucoma, RNFLT and BMO-MRW showed comparable AUROCs in global (AUROC, 95% confidence interval: 0.954, 0.911–0.996 and 0.929, 0.872–0.986) and sectoral (0.929, 0.877–0.981 and 0.946, 0.904–0.996) analysis. For preperimetric glaucoma BMO-MRW and RNFLT also demonstrated comparable performance in global (0.839, 0.757–0.921 and 0.821, 0.731–0.912) and sectoral (0.860, 0.782–0.938 and 0.835, 0.750–0.920) analysis. When identifying ocular hypertensive eyes AUROCs were lower for global RNFLT and BMO-MRW (0.493, 0.365–0.621 and 0.562, 0.433–0.691). A combined parameter showed an AUROC of 0.959, 0.921 to 0.996 for perimetric and 0.849, 0.770 to 0.929 for preperimetric glaucoma.

Conclusions: Bruch's membrane opening–MRW and RNFLT are comparably useful parameters for discrimination of preperimetric and perimetric glaucomatous eyes and show potential to assist each other in glaucoma diagnosis. (www.ClinicalTrials.gov number, NTC00494923; Erlangen Glaucoma Registry.)

Open-angle glaucoma is a major cause of blindness affecting millions worldwide and considering todays demographic trends both prevalence and incidence are expected to increase.1 Early glaucoma diagnosis followed by sufficient reduction of IOP significantly improves the patient's outcome, emphasizing the importance of a reliable and objective detection, and monitoring of early structural defects.2,3 Many patients develop characteristic changes such as retinal nerve fiber layer thickness (RNFLT) loss, neuroretinal rim thinning, or increasing excavation cup depth prior to experiencing visual field defects.4,5 However, large interindividual variability of the optic nerve head complicates the assessment of these defects.5 Standard clinical ophthalmoscopy of the neuroretinal rim demonstrates only moderate classification accuracy6 and is highly depending on the skill level of the ophthalmologist. The high subjectivity of the examination also leads to moderate interobserver agreement in glaucoma diagnosis.710 
Optical coherence tomography (OCT) was first described by Huang et al.11 in 1991 and has become an essential tool in ocular imaging since then. It provides an objective quantitative method of detecting nerve fiber layer thickness using low-coherence interferometry. In glaucoma assessment OCT imaging has proven itself as a valuable imaging modality for scanning peripapillary RNFLT.12 Early attempts to introduce optic nerve head parameters into clinical OCT examination foundered on the lower reproducibility and the effort needed to correct missegmentation of the optic disc margin, even though several studies indicated comparable diagnostic competence.13,14 However, akin to ophthalmoscopic examination of the neuroretinal rim the new parameter Bruch's membrane opening minimum rim width (BMO-MRW) was proposed as an assessment of the optic nerve head. It measures the minimum distance from the inner opening of the BMO to the internal limiting membrane (ILM). In contrast to ophthalmoscopy, it presents stable borders and a geometrically more accurate evaluation of neuroretinal rim tissue.1518 Recently, Chauhan et al.19 showed that BMO-MRW provides better performance at discriminating perimetric glaucoma than established confocal scanning laser tomography parameters, OCT measurements of RNFLT and BMO horizontal rim width, which is a variant of BMO-MRW but uses the horizontal instead of the minimum distance between BMO and ILM. Pollet-Villard et al.20 found a stronger association of visual field sensitivity with BMO-MRW than other optic nerve head parameters or RNFLT. 
To our knowledge, this study is the first to investigate and compare the diagnostic capability of RNFLT and BMO-MRW in differentiating healthy eyes not only from advanced perimetric glaucoma but also from those with ocular hypertension (OH) and preperimetric glaucoma. This is especially important, as detecting early stages of the disease allows early treatment and a better outcome of the patient.2,3 
Methods
This prospective, controlled, single-center study was approved by the Ethics committee of the University of Erlangen-Nuremberg (Erlangen, Germany) and it adhered to the tenets of the Declaration of Helsinki. 
From each of the 40 healthy controls, 41 patients with ocular hypertension, 50 with preperimetric, and 50 with perimetric glaucoma, one eye was randomly selected and included in this study. Informed consent was obtained from all patients, which were recruited from the Erlangen Glaucoma Registry, a clinical registry founded in 1991 for cross-sectional and longitudinal observation study of patients diagnosed with open-angle glaucoma or ocular hypertension (clinical trials identifier: NCT00494923). 
All patients and controls were of Caucasian descent and underwent standard ophthalmic examination including slit-lamp inspection, funduscopy, achromatic perimetry, IOP measurement (Goldmann applanation tonometry), visual acuity testing, refractometry, and spectral-domain optical coherence tomography (SD-OCT; Spectralis, Heidelberg Engineering, Germany). A 24-hour IOP profile was performed for all patients. 
Perimetric and preperimetric glaucoma had a pathologic appearance of the optic disc, such as neuroretinal rim thinning, increased excavation of the optic disc, higher cup to disc ratio vertically than horizontally, or retinal nerve fiber loss. Perimetric glaucoma showed abnormal visual field as defined in the perimetry section while preperimetric glaucoma did not. 
Ocular hypertensive subjects had an IOP above 21 mm Hg measured in the 24-hour IOP profile but no signs of morphologic or functional defects in slit-lamp examination, optic disc inspection, and perimetry. All subjects identified as glaucoma or ocular hypertension were then treated with IOP-lowering medication at the discretion of a glaucoma expert. 
Inclusion criteria for healthy controls were a single measurement of IOP less than 21 mm Hg, no pathologic findings in slit-lamp examination or optic disc inspection, refractive error of less than 7 diopters (D) equivalent sphere or less than 3 D astigmatism, and normal visual field in achromatic perimetry. 
Optical Coherence Tomography
Spectral-domain OCT imaging was performed using the Glaucoma Module Premium Edition for Spectralis SD-OCT (Software Version 6.0f; Heidelberg Engineering). This software enables measuring the BMO-MRW, a parameter proposed to represent the neuroretinal rim “in an anatomically and geometrically accurate manner.”15 It marks the minimal distance between the BMO and the ILM. For BMO-MRW, 24 radial B-scans (length: 15°), for RNFLT three circle B-scans (12°/14°/16° or 3.5/4.1/4.6 mm for the emmetropic eye with an axial length of 23 mm), inner, middle, and outer RNFLT, were acquired. All scans were centered on the optic nerve head and aligned to the BMO-fovea axis using the internal positioning system. The internal eye tracking software TruTrack (Heidelberg Engineering, Heidelberg, Germany.21 For each B-scan 25 (BMO-MRW) or 100 (RNFLT) B-scans were averaged. Bruch's membrane opening and RNFLT segmentation was computed and reviewed by an experienced ophthalmologist to rule out segmentation errors. In cases of blood vessels shadowing we optimized image contrast from standard setting 12 to 16 to allow better visibility of BMO endpoints. In cases of bending of the BMO endpoint or Bruch's membrane doubling phenomenon, we verified the endpoint by considering neighboring scans. Scans showing poor quality scores of less than 15 or any OCT-related artifacts such as mirror artifacts, out of register artifacts, or blink artifacts were excluded.22 Bruch's membrane opening–MRW and RNFLT values were exported and subsequently calculated globally as well as for six sectors according to the legacy distribution of the Spectralis OCT (nasal [N], nasal superior [NS], temporal superior [TS], temporal [T], temporal inferior [TI], and nasal inferior [NI], T and N twice the size of the other sectors, Fig. 1).23 
Figure 1
 
Distribution of sectors. Distribution of the six sectors around the optic nerve head: temporal (T), temporal superior (TS), nasal superior (NS), nasal (N), nasal inferior (NI), and temporal inferior (TI). The dashed red line indicates the fovea to BMO axis, which was used to align the sectoral distribution.
Figure 1
 
Distribution of sectors. Distribution of the six sectors around the optic nerve head: temporal (T), temporal superior (TS), nasal superior (NS), nasal (N), nasal inferior (NI), and temporal inferior (TI). The dashed red line indicates the fovea to BMO axis, which was used to align the sectoral distribution.
Perimetry
Visual field tests were performed on all subjects using standard static white-on-white perimetry (Octopus 500, program G1; Interzeag, Schlieren, Switzerland) and Peridata software (version 3.3; in the public domain, http://www.peridata.org/index_e.htm) for evaluation of the pattern deviation map. The first test was excluded to avoid learning effects. False-negative and false-positive rates over 12% were considered insufficient and repeated. Tests showing three adjacent test points, one of which shows a reduction in sensitivity of greater than 10 dB while the other two show a reduction of greater than 5 dB, or two adjacent test points showing a reduction in sensitivity of greater than 10 dB were classified as pathologic. Findings had to be confirmed in two subsequent examinations. 
Analysis
The statistical analysis was performed using SPSS (version 22.0; Chicago, IL, USA). ANOVA and Scheffé-Test were used for discrimination of continuous variables in all subgroups, whereas χ2 test was used for differences in dichotomous variables. P less than 0.05 was considered statistically significant in all tests. The Bonferroni method was applied to correct probability values when performing multiple statistical tests.24 
For BMO-MRW and RNFLT, areas under the receiver operating characteristics curves (AUROCs) as well as sensitivity at fixed specificity (90%, 95%) were calculated globally and sectorally to determine their discrimination ability. Based on the highest global and sectoral AUROC, one of the three peripapillary scan circle sizes was selected. Areas under the receiver operating characteristics curves of BMO-MRW and the best RNFLT were compared using the method described by Delong et al.25 
Area-proportional Venn diagrams were computed using Euler 3 Applet,26 to depict how patients were diagnosed by the different parameters at fixed specificity of 90% and 95%. We used them to visualize how the two parameters correspond and to what extend they could complement each other in glaucoma diagnosis. We also created combined parameters for global and each sector by adding BMO-MRW to RNFLT values. To ensure equal contribution of both measurements we adjusted RNFLT values, multiplying them by a factor based on the ratio of mean BMO-MRW and RNFLT among healthy controls (combined parameter = BMO-MRW + RNFLT × (mean_BMO-MRW_control / mean_RNFLT_control). We then analyzed the AUROCs and sensitivity of these combined parameters. 
To investigate the influence of the optic disc size on the diagnostic ability of BMO-MRW we separated healthy controls, preperimetric, and perimetric glaucomatous eyes into two subgroups each, based upon the size of the BMO-area. We adjusted the corresponding subgroups regarding age. The cutoff value between small and large BMO-areas was 1.84 mm2. We also matched the two perimetric glaucomatous groups respecting disease severity using the enhanced glaucoma staging system.27 
Results
The participants' demographic and clinical characteristics at the time of OCT measurement are shown in Table 1. There were no significant differences in age (P = 0.594) or sex (P = 0.586) between all groups. Bruch's membrane opening area and global values of BMO-MRW as well as RNFLT are depicted in Table 2. Global values were significant thinner in perimetric and preperimetric glaucoma. 
Table 1
 
Demographic and Clinical Characteristics of the Subjects
Table 1
 
Demographic and Clinical Characteristics of the Subjects
Table 2
 
Morphometric Characteristics of the Subjects
Table 2
 
Morphometric Characteristics of the Subjects
Mean BMO-MRW and RNFLT showed a resembling pattern with highest values in superior and inferior sectors (Fig. 2). The deviation from normal demonstrates thinning of BMO-MRW and RNFLT in patients with preperimetric and perimetric glaucoma. Most pronounced loss was found in superior and inferior regions. There was no significant loss of BMO-MRW or RNFLT in the OH group (Table 2). Of the three scan circle sizes, inner RNFLT demonstrated highest global and sectoral AUROCs and was selected for comparison with BMO-MRW as seen in the following. For reasons of comprehensibility, “inner RNFLT” is hereafter referred to as “RNFLT” only. 
Figure 2
 
Thickness distribution around the optic nerve head. Thickness of BMO-MRW and RNFLT as well as the thickness deviation from normal by position around the optic nerve head from left to right for normal, ocular hypertensive subjects, preperimetric glaucoma, and perimetric glaucoma patients. Values are depicted as measured around the optic nerve head beginning from temporal (T) to superior (S), nasal (N), and inferior (I). Graphs include the 95% CI.
Figure 2
 
Thickness distribution around the optic nerve head. Thickness of BMO-MRW and RNFLT as well as the thickness deviation from normal by position around the optic nerve head from left to right for normal, ocular hypertensive subjects, preperimetric glaucoma, and perimetric glaucoma patients. Values are depicted as measured around the optic nerve head beginning from temporal (T) to superior (S), nasal (N), and inferior (I). Graphs include the 95% CI.
In discriminating between perimetric glaucoma and healthy controls global RNFLT and BMO-MRW present comparable AUROCs (AUROC ± confidence interval [CI] 0.954 [0.911–0.996] and 0.929 [0.872–0.986]). There was no significant difference between the two ROC curves (P = 0.3662). Global RNFLT had higher sensitivity than RNFLT than BMO-MRW fixed specificity of 95% (84% vs. 52%), but lower at 90% specificity (84% compared with 88%). The best sector for all parameters was the temporal inferior sector showing an AUROC of 0.929 (0.877–0.981) for RNFLT and 0.946 (0.904–0.988) for BMO-MRW. No significant difference between the two ROC curves was found (P = 0.5030). Temporal inferior RNFLT showed higher sensitivity than BMO-MRW at fixed specificity of 95% (82% vs. 74%), but lower at 90% specificity (84% compared with 88%). See Tables 3 and 4 for all sectoral and global AUROC values and sensitivities. 
Table 3
 
Diagnostic Performance to Detect Perimetric Glaucoma
Table 3
 
Diagnostic Performance to Detect Perimetric Glaucoma
Table 4
 
Diagnostic Performance to Detect Preperimetric Glaucoma
Table 4
 
Diagnostic Performance to Detect Preperimetric Glaucoma
Distinguishing preperimetric glaucoma from healthy controls global RNFLT and BMO-MRW both showed lower AUROCs than regarding perimetric glaucoma (0.839 [0.757–0.921] and 0.821 [0.731–0.912]) with no significant difference between the ROCs (P = 0.6749). Global RNFLT demonstrated higher sensitivity than BMO-MRW at fixed specificity of 95% (44% compared with 28%), but lower at 90% specificity (50% vs. 52%). The highest AUROC for RNFLT was found in the nasal inferior sector (0.860 [0.782–0.938]), whereas BMO-MRW had highest AUROC in the temporal superior sector (0.835 [0.750–0.920]). There was no significant difference between the ROC curves (P = 0.5068). Nasal inferior RNFLT had higher sensitivity than BMO-MRW at specificity of 95% (60% vs. 44%) and 90% (66% vs. 50%). 
Discriminating ocular hypertensive subjects both RNFLT and BMO-MRW showed lower AUROCs in global analysis (0.493 [0.365–0.621] and 0.562 [0.433–0.691]) than for the best sector (nasal superior, 0.616 [0.492–0.740] and 0.591 [0.462–0.720]). 
Figure 3 shows the ROC curves of BMO-MRW and RNFLT when differentiating ocular hypertensive subjects, preperimetric, and perimetric glaucoma from healthy eyes. 
Figure 3
 
Diagnostic performance of optical coherence tomography parameters. Area under receiver operating characteristic of (dashed) as well as inner (red), middle (yellow), and outer (green) RNFLT for discriminating perimetric and preperimetric glaucoma patients as well as ocular hypertensive subjects from normal eyes decrease in patients with less advanced glaucoma stage. Dashed line indicates 95% specificity. Area under receiver operating characteristic values are depicted inside each diagram.
Figure 3
 
Diagnostic performance of optical coherence tomography parameters. Area under receiver operating characteristic of (dashed) as well as inner (red), middle (yellow), and outer (green) RNFLT for discriminating perimetric and preperimetric glaucoma patients as well as ocular hypertensive subjects from normal eyes decrease in patients with less advanced glaucoma stage. Dashed line indicates 95% specificity. Area under receiver operating characteristic values are depicted inside each diagram.
See Supplementary Figure S1 that demonstrates the AUROCs of global BMO-MRW and RNFLT in different positions around the optic nerve head. 
Venn diagrams (Fig. 4) show a higher number of correctly diagnosed patients for RNFLT than BMO-MRW among perimetric (McNemar's test, P < 0.001) and preperimetric glaucoma (P = 0.013) for 95% specificity. We observed no statistically significant difference among RNFLT and BMO-MRW for 90% specificity among perimetric and preperimetric glaucoma (P = 0.727). The combined parameters showed higher but not significantly different AUROCs for all sectors when examining perimetric glaucoma and for some sectors when applied to preperimetric glaucoma (Tables 3, 4). 
Figure 4
 
Venn diagrams. Venn diagrams for BMO-MRW (dashed circles) and RNFLT (continuous circles) for distinguishing perimetric and preperimetric glaucoma at 95% and 90% specificity show the absolute number of patients correctly diagnosed by one, both, or none of the two parameters. In the upper left diagram one patient was solely distinguished by BMO-MRW. For reasons of comprehensibility this number is not depicted in the graph. Asterisks indicate which tests showed significantly different sensitivity between BMO-MRW and RNFLT (McNemar's test, P < 0.05).
Figure 4
 
Venn diagrams. Venn diagrams for BMO-MRW (dashed circles) and RNFLT (continuous circles) for distinguishing perimetric and preperimetric glaucoma at 95% and 90% specificity show the absolute number of patients correctly diagnosed by one, both, or none of the two parameters. In the upper left diagram one patient was solely distinguished by BMO-MRW. For reasons of comprehensibility this number is not depicted in the graph. Asterisks indicate which tests showed significantly different sensitivity between BMO-MRW and RNFLT (McNemar's test, P < 0.05).
Preperimetric and perimetric glaucoma patients were divided into two subgroups based on the size of the BMO. The subgroups' characteristics can be seen in Table 5. The corresponding perimetric and preperimetric glaucomatous subgroups showed significant difference in BMO-area (P < 0.001) but no difference in age (P = 0.549, P = 0.372) or Enhanced Glaucoma Staging System (P = 0.537).27 Both RNFLT and BMO-MRW presented higher but not significantly different AUROCs when examining perimetric or preperimetric eyes with smaller BMO. The results can be seen in Table 5. There was larger difference in the AUROCs for BMO-MRW than for RNFLT (Δ = 0.108 and Δ = 0.141 vs. Δ = 0.053 and Δ = 0.040). Both RNFLT and BMO-MRW also showed higher sensitivity at 95% specificity in the small BMO-area groups. 
Table 5
 
Influence of Bruch's Membrane Opening Area
Table 5
 
Influence of Bruch's Membrane Opening Area
Discussion
There are various assumptions underlying the proposal of BMO-MRW as a new parameter in glaucoma diagnosis. Due to its improved algorithm measuring the minimum instead of the horizontal distance between the BMO and the ILM, it provides a geometrically more accurate examination of the optic nerve head than previous scanning methods.15 Furthermore, studies have shown that the disc margin as assessed in standard clinical funduscopy does not represent a consistent anatomic structure.28 Therefore, interindividual variability of the incorporated anatomic structures, such as clinically invisible extensions of Bruch's membrane, leads to erroneous evaluation of the neuroretinal rim thickness.29 Bruch's membrane opening–MRW, based solely on the BM as an impenetrable layer, supposedly leads to better determination of the borders of neuroretinal rim compared with clinical ophthalmoscopy, and thereby enhances the precision of rim assessment. Lastly, there is large interindividual diversity in the axis between the optic nerve head and the fovea.30 These differences must be taken into account for sectoral analyses of either RNFLT or BMO-MRW.15 
In our study, all groups were adjusted regarding age, sex, and eye side, to reduce the influence of age-related decline of retinal nerve fibers and other sources of bias.18,31 We found that RNFLT and BMO-MRW provide comparable diagnostic performance when differentiating perimetric and preperimetric glaucomatous subjects from healthy controls, whereas a previous study by Chauhan et al.19 did not. At very high levels of specificity (95%) that are needed for clinical practice due to the low prevalence of the disease, we even found significantly higher diagnostic performance for RNFLT than for BMO-MRW. The higher diagnostic performance of RNFLT compared with BMO-MRW could be due to RNFLT being less influenced by the shape of the optic disc and its anatomic variability than BMO-MRW. Chauhan et al.19 ascertained that BMO-MRW had consistently better performance over RNFLT in diagnosing perimetric glaucomatous patients. They included few more healthy controls and a larger number of perimetric glaucoma patients. However, Chauhan et al.19 examined patients that had already been diagnosed as glaucomatous by a positive Glaucoma Hemifield Test. The benefit of OCT as a diagnostic tool in such cases is questionable, because perimetric defects themselves provide a sufficient diagnosis. Studies have shown that OCT can detect glaucomatous changes up to 8 years before the onset of visual defects, providing lead time for the disease management that could positively influence the patients' functional outcome.32 
To investigate the accuracy of BMO-MRW and RNFLT in early glaucoma diagnosis, we included preperimetric glaucomatous and ocular hypertensive patients in our study. However, due to a missing difference in diagnostic performance and an overall weak performance among ocular hypertensive subjects we focused on preperimetric and perimetric glaucoma. 
To our knowledge, we were also the first to present data for BMO-MRW and RNFLT that was obtained and analyzed using a sectoral distribution, which is aligned on the fovea to BMO axis. Compensating for the large interindividual variation of this axis allows an anatomically more consistent comparison of sectoral data between individuals and should further enhance the diagnostic performance.15 
We presented comparable and to some extend higher AUROCs than recent similar studies.20,33 However, one can find many studies investigating RNFLT only, some of which show very high AUROC values.14,34 Differences in the control groups and the severity of the disease exacerbates exact comparison of studies. Further studies need to investigate if diagnostic accuracy of RNFLT and BMO-MRW are improved by aligning data on this axis when comparing results from the same study group. For RNFLT of normals however, Resch et al.35 could not find a reduction of intersubject variability after compensation for disc-fovea angle. 
Venn diagrams show that especially in early stages of glaucoma a considerable fraction of up to 34% of the patients was detected by one of the two tests only. In addition, a combined parameter including BMO-MRW and RNFLT showed higher performance in all sectors when distinguishing perimetric glaucoma and for some sectors when examining preperimetric glaucoma. This suggests that in our population patients often show focal defects. Defects that manifest in distal peripapillary regions may be detected more easily by the peripapillary RNFLT scan while BMO-MRW may show better performance in identifying defects in the optic nerve head. In cases of optic disc hemorrhage or RNFL edema, BMO may provide favorable information as the aforementioned might bias the peripapillary RNFLT measurement. In myopic configuration of the optic disc BMO might show better performance as RNFLT is influenced by axial length.36 Further studies need to investigate these correlations. However, solely expanding criteria for identifying glaucoma to multiple parameters comes at the expense of decreased specificity. But, as recently presented, different models of combining multiple parameters can improve the diagnostic capability in early glaucoma assessment.34,37,38 Even our basic attempt of creating a combination parameter showed improvement in the diagnostic capability. We believe incorporating BMO-MRW in more elaborated models could further enhance glaucoma diagnosis by detecting patients that show defects mainly in the optic nerve head. 
The Glaucoma Module Premium Edition also features three differently sized peripapillary RNFLT circle scans. For comparison with BMO-MRWs performance we chose the smallest circle, as it consistently showed the best AUROC for global analysis and for the best sector (TI). This 12°, or 3.4 mm in the emmetropic eye, circle size was introduced due to its high reliability as well as low overlapping with the optic nerve head in early time-domain OCT testing and has barely been questioned since then.39 Leung et al.40 found that in glaucoma progression a 4-mm scan circle, which is best comparable with the 14° scan protocol showed highest accuracy in detecting glaucomatous changes. They reasoned that smaller circles correspond to regions of the peripapillary RNFLT where nerve fibers lie more compactly, and thus hinder detection of early defects. However, as the peripapillary RNFLT is thicker in regions closer to the optic disc, there is a higher signal-to-noise ratio for smaller circle scans. This may lead to higher precision of the 12° scan protocol.41 
Global RNFLT yielded higher performance than any sectoral analyses for all but one exception, inner RNFLT in distinguishing preperimetric glaucoma. This finding accords with previous studies42 and is said to possibly derive from an overlapping arrangement of peripapillary nerve fibers, resulting in variable diagnostic performance among different sectors of RNFLT.19,43 The sectoral performance varies highly, featuring two peaks in superior and inferior analyses. In agreement with the above mentioned this resembling pattern is stronger pronounced for inner RNFLT than for BMO-MRW. This could be due to the distribution of peripapillary RNFLT, which features thicker regions in inferior and superior sectors, whereas BMO-MRW is more evenly arranged. As the range of found AUROCs was also larger for RNFLT than for BMO-MRW, one may speculate that BMO-MRWs performance may be more consistent throughout sectors. 
Retinal nerve fiber layer thickness and BMO-MRW both presented higher AUROCs when examining perimetric and preperimetric glaucoma with smaller BMO-area, implying that their diagnostic capability might be negatively affected by an increasing size of the optic disc. Though for RNFLT there was very little deterioration, which could originate from an overlapping of the 12° scan circle with very large optic discs. This finding is in line with previous studies that found no significant influence of disc areas smaller than 4 mm2 on the RNFLT measurements alongside higher influence of disc area on optic nerve head parameters acquired with confocal scanning laser ophthalmoscopy.44,45 Bruch's membrane opening–MRW showed greater difference in both patient groups. This may due to the overall thinner BMO-MRW in large optic discs, where the axons enter the optic nerve head distributed over a wider rim area. As large but healthy discs may be erroneously classified as glaucomatous, this may lead to lower sensitivity. This assumption could account for the lower sensitivity of BMO-MRW at 95% specificity that we found in our population compared with Chauhan et al.19 Bruch's membrane opening–MRW showed a strong decrease of sensitivity at specificity above 95%. Our study group showed not only larger BMO-area than the population of Chauhan et al.18 and the reference group of the device, but also high variability. However, the number of patients in the subgroups was too small to derive statistically significant results. This relationship needs to be further investigated in studies including larger case numbers, preferably with comparison to the aforementioned normative study group, to adapt the findings to the new internal reference of the device. 
It is also possible, that the accuracy of BMO-MRW is deteriorated by retinal blood vessels. For RNFLT measurements, Mardin et al. found no influence of blood vessels on the diagnostic accuracy (Mardin CY, et al. IOVS 2009;50:ARVO E-Abstract 3333). In the peripapillary circle scans vessels are mainly depicted transversely, which minimizes the influence on the RNFLT measurements. However, vessels often enter the optic nerve head more irregularly, potentially leading to impaired diagnostic accuracy of BMO-MRW. 
Lastly, our results show an unusual pattern for the sector analysis in the preperimetric glaucoma group. We found the highest sectoral sensitivity of RNFLT for preperimetric glaucoma to be in the nasal inferior sector. Other studies reported the inferior quadrant or temporal inferior sector to show the best diagnostic accuracy.4648 A close look at Figure 2 reveals that maximum deviation from normal is found nasal inferiorly as well as temporal inferiorly. Our preperimetric group seems to show uncommonly high defects in the nasal inferior sector. We thoroughly reviewed fundus photographs and OCT images of the patients with these defects and found an unusually high number of superior centrally exiting vessel trunks instead of the more common vessel trunks located in the superior nasal quadrant.49 According to a study by Jonas et al.50 parapapillary atrophy is higher in the quadrant most distant to the central retinal vessel trunk, which “may indicate that, in eyes with […] an abnormal position of the central retinal vessel trunk, early glaucomatous rim changes should be looked for in the sector that is most distant to the central retinal vessel trunk exit.” This could explain the RNFLT loss in the nasal inferior and temporal inferior sector of our preperimetric group. Our study agrees with the aforementioned studies on the highest diagnostic accuracy for global RNFLT. Our analysis focuses in large parts on global RNFLT. We therefore believe the validity of our results should not be impaired much by this finding. 
Our results have shown the comparable potential of BMO-MRW and RNFLT in glaucoma diagnosis. However, this was a cross-sectional study only, and therefore did not assess its performance in glaucoma progression monitoring. Further studies are needed to ascertain if BMO-MRW can detect glaucoma progression with efficiency comparable to established methods.51 For adequate clinical assessment of glaucomatous eyes or suspects, we do not recommend relying merely on OCT imaging, but integrating its additional information in the clinical practice. 
Acknowledgments
Supported by grants from the German Research Foundation (Bonn, Germany). 
Disclosure: J.M.D. Gmeiner, None; W.A. Schrems, None; C.Y. Mardin, None; R. Laemmer, None; F.E. Kruse, None; L.M. Schrems-Hoesl, None 
References
Quigley HA, Broman AT. The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. 2006; 90: 262–267.
The Advanced Glaucoma Intervention Study (AGIS): 7. The relationship between control of intraocular pressure and visual field deterioration. The AGIS Investigators. Am J Ophthalmol. 2000; 130: 429–440.
Kass MA, Heuer DK, Higginbotham EJ, et al. The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol. 2002; 120: 701–713. discussion 829–830.
Keltner JL, Johnson CA, Anderson DR, et al. The association between glaucomatous visual fields and optic nerve head features in the Ocular Hypertension Treatment Study. Ophthalmology. 2006; 113: 1603–1612.
Jonas JB, Budde WM, Panda-Jonas S. Ophthalmoscopic evaluation of the optic nerve head. Surv Ophthalmol. 1999; 43: 293–320.
Denniss J, Echendu D, Henson DB, Artes PH. Discus: investigating subjective judgment of optic disc damage. Optom Vis Sci. 2011; 88: E93–E101.
Wollstein G, Garway-Heath DF, Fontana L, Hitchings RA. Identifying early glaucomatous changes. Comparison between expert clinical assessment of optic disc photographs and confocal scanning ophthalmoscopy. Ophthalmology. 2000; 107: 2272–2277.
Reus NJ, Lemij HG, Garway-Heath DF, et al. Clinical assessment of stereoscopic optic disc photographs for glaucoma: the European Optic Disc Assessment Trial. Ophthalmology. 2010; 117: 717–723.
Breusegem C, Fieuws S, Stalmans I, Zeyen T. Agreement and accuracy of non-expert ophthalmologists in assessing glaucomatous changes in serial stereo optic disc photographs. Ophthalmology. 2011; 118: 742–746.
Hadwin SE, Redmond T, Garway-Heath DF, et al. Assessment of optic disc photographs for glaucoma by UK optometrists: the Moorfields Optic Disc Assessment Study (MODAS). Ophthalmic Physiol Opt. 2013; 33: 618–624.
Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science. 1991; 254: 1178–1181.
Budenz DL, Michael A, Chang RT, McSoley J, Katz J. Sensitivity and specificity of the StratusOCT for perimetric glaucoma. Ophthalmology. 2005; 112: 3–9.
Mwanza JC, Oakley JD, Budenz DL, Anderson DR; for the Cirrus Optical Coherence Tomography Normative Database Study Group. Ability of cirrus HD-OCT optic nerve head parameters to discriminate normal from glaucomatous eyes. Ophthalmology. 2011; 118: 241–248. e241.
Bussel II, Wollstein G, Schuman JS. OCT for glaucoma diagnosis, screening and detection of glaucoma progression. Br J Ophthalmol. 2014; 98 (suppl 2): ii15–ii19.
Chauhan BC, Burgoyne CF. From clinical examination of the optic disc to clinical assessment of the optic nerve head: a paradigm change. Am J Ophthalmol. 2013; 156: 218–227. e212.
Povazay B, Hofer B, Hermann B, et al. Minimum distance mapping using three-dimensional optical coherence tomography for glaucoma diagnosis. J Biomed Opt. 2007; 12: 041204.
Chen TC. Spectral domain optical coherence tomography in glaucoma: qualitative and quantitative analysis of the optic nerve head and retinal nerve fiber layer (an AOS thesis). Trans Am Ophthalmol Soc. 2009; 107: 254–281.
Chauhan BC, Danthurebandara VM, Sharpe GP, et al. Bruch's membrane opening minimum rim width and retinal nerve fiber layer thickness in a normal white population: a multicenter study. Ophthalmology. 2015; 122: 1786–1794.
Chauhan BC, O'Leary N, Almobarak FA, et al. Enhanced detection of open-angle glaucoma with an anatomically accurate optical coherence tomography-derived neuroretinal rim parameter. Ophthalmology. 2013; 120: 535–543.
Pollet-Villard F, Chiquet C, Romanet JP, Noel C, Aptel F. Structure-function relationships with spectral-domain optical coherence tomography retinal nerve fiber layer and optic nerve head measurements. Invest Ophthalmol Vis Sci. 2014; 55: 2953–2962.
Langenegger SJ, Funk J, Toteberg-Harms M. Reproducibility of retinal nerve fiber layer thickness measurements using the eye tracker and the retest function of Spectralis SD-OCT in glaucomatous and healthy control eyes. Invest Ophthalmol Vis Sci. 2011; 52: 3338–3344.
Chhablani J, Krishnan T, Sethi V, Kozak I. Artifacts in optical coherence tomography. Saudi J Ophthalmol. 2014; 28: 81–87.
Garway-Heath DF, Poinoosawmy D, Fitzke FW, Hitchings RA. Mapping the visual field to the optic disc in normal tension glaucoma eyes. Ophthalmology. 2000; 107: 1809–1815.
Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ. 1995; 310: 170.
DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988; 44: 837–845.
Rodgers P, Stapleton G, Flower J, Howse J. Drawing area-proportional Euler diagrams representing up to three sets. IEEE Trans Vis Comput Graph. 2014; 20: 56–69.
Brusini P, Filacorda S. Enhanced Glaucoma Staging System (GSS 2) for classifying functional damage in glaucoma. J Glaucoma. 2006; 15: 40–46.
Reis AS, Sharpe GP, Yang H, Nicolela MT, Burgoyne CF, Chauhan BC. Optic disc margin anatomy in patients with glaucoma and normal controls with spectral domain optical coherence tomography. Ophthalmology. 2012; 119: 738–747.
Reis AS, O'Leary N, Yang H, et al. Influence of clinically invisible, but optical coherence tomography detected, optic disc margin anatomy on neuroretinal rim evaluation. Invest Ophthalmol Vis Sci. 2012; 53: 1852–1860.
He L, Ren R, Yang H, et al. Anatomic vs. acquired image frame discordance in spectral domain optical coherence tomography minimum rim measurements. PLoS One. 2014; 9: e92225.
Leung CK, Yu M, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: a prospective analysis of age-related loss. Ophthalmology. 2012; 119: 731–737.
Kuang TM, Zhang C, Zangwill LM, Weinreb RN, Medeiros FA. Estimating lead time gained by optical coherence tomography in detecting glaucoma before development of visual field defects. Ophthalmology. 2015; 122: 2002–2009.
El Chehab H, Delbarre M, Marechal M, et al. New neuroretinal rim analysis with spectral domain optical coherence tomography, Spectralis (Heidelberg Engineering, Germany). Preliminary study [in French]. J Fr Ophthalmol. 2015; 38: 46–52.
Mwanza JC, Warren JL, Budenz DL; for the Ganglion Cell Analysis Study Group. Combining spectral domain optical coherence tomography structural parameters for the diagnosis of glaucoma with early visual field loss. Invest Ophthalmol Vis Sci. 2013; 54: 8393–8400.
Resch H, Pereira I, Hienert J, et al. Influence of disc-fovea angle and retinal blood vessels on interindividual variability of circumpapillary retinal nerve fibre layer. Br J Ophthalmol. 2016; 100: 531–536.
Bendschneider D, Tornow RP, Horn FK, et al. Retinal nerve fiber layer thickness in normals measured by spectral domain OCT. J Glaucoma. 2010; 19: 475–482.
Choi YJ, Jeoung JW, Park KH, Kim DM. Clinical use of an optical coherence tomography linear discriminant function for differentiating glaucoma from normal eyes. J Glaucoma. 2016; 25: e162–e169.
Baskaran M, Ong EL, Li JL, et al. Classification algorithms enhance the discrimination of glaucoma from normal eyes using high-definition optical coherence tomography. Invest Ophthalmol Vis Sci. 2012; 53: 2314–2320.
Schuman JS, Pedut-Kloizman T, Hertzmark E, et al. Reproducibility of nerve fiber layer thickness measurements using optical coherence tomography. Ophthalmology. 1996; 103: 1889–1898.
Leung CK, Yu M, Weinreb RN, Lai G, Xu G, Lam DS. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: patterns of retinal nerve fiber layer progression. Ophthalmology. 2012; 119: 1858–1866.
Leung CK. Diagnosing glaucoma progression with optical coherence tomography. Curr Opin Ophthalmol. 2014; 25: 104–111.
Schrems WA, Mardin CY, Horn FK, Juenemann AG, Laemmer R. Comparison of scanning laser polarimetry and optical coherence tomography in quantitative retinal nerve fiber assessment. J Glaucoma. 2010; 19: 83–94.
Jeoung JW, Kim TW, Kang KB, Lee JJ, Park KH, Kim DM. Overlapping of retinal nerve fibers in the horizontal plane. Invest Ophthalmol Vis Sci. 2008; 49: 1753–1757.
Mansoori T, Balakrishna N, Viswanath K. Influence of disc area on retinal nerve fiber layer thickness measurement by spectral domain optical coherence tomography. Indian J Ophthalmol. 2014; 62: 615–618.
Oddone F, Centofanti M, Tanga L, et al. Influence of disc size on optic nerve head versus retinal nerve fiber layer assessment for diagnosing glaucoma. Ophthalmology. 2011; 118: 1340–1347.
Lu AT, Wang M, Varma R, et al. Combining nerve fiber layer parameters to optimize glaucoma diagnosis with optical coherence tomography. Ophthalmology. 2008; 115: 1352–1357.
Wu H, de Boer JF, Chen TC. Diagnostic capability of spectral-domain optical coherence tomography for glaucoma. Am J Ophthalmol. 2012; 153: 815–826. e812.
Chen HY, Huang ML. Discrimination between normal and glaucomatous eyes using Stratus optical coherence tomography in Taiwan Chinese subjects. Graefes Arch Clin Exp Ophthalmol. 2005; 243: 894–902.
Jonas JB, Fernandez MC. Shape of the neuroretinal rim and position of the central retinal vessels in glaucoma. Br J Ophthalmol. 1994; 78: 99–102.
Jonas JB, Budde WM, Nemeth J, Grundler AE, Mistlberger A, Hayler JK. Central retinal vessel trunk exit and location of glaucomatous parapapillary atrophy in glaucoma. Ophthalmology. 2001; 108: 1059–1064.
Wollstein G, Schuman JS, Price LL, et al. Optical coherence tomography longitudinal evaluation of retinal nerve fiber layer thickness in glaucoma. Arch Ophthalmol. 2005; 123: 464–470.
Figure 1
 
Distribution of sectors. Distribution of the six sectors around the optic nerve head: temporal (T), temporal superior (TS), nasal superior (NS), nasal (N), nasal inferior (NI), and temporal inferior (TI). The dashed red line indicates the fovea to BMO axis, which was used to align the sectoral distribution.
Figure 1
 
Distribution of sectors. Distribution of the six sectors around the optic nerve head: temporal (T), temporal superior (TS), nasal superior (NS), nasal (N), nasal inferior (NI), and temporal inferior (TI). The dashed red line indicates the fovea to BMO axis, which was used to align the sectoral distribution.
Figure 2
 
Thickness distribution around the optic nerve head. Thickness of BMO-MRW and RNFLT as well as the thickness deviation from normal by position around the optic nerve head from left to right for normal, ocular hypertensive subjects, preperimetric glaucoma, and perimetric glaucoma patients. Values are depicted as measured around the optic nerve head beginning from temporal (T) to superior (S), nasal (N), and inferior (I). Graphs include the 95% CI.
Figure 2
 
Thickness distribution around the optic nerve head. Thickness of BMO-MRW and RNFLT as well as the thickness deviation from normal by position around the optic nerve head from left to right for normal, ocular hypertensive subjects, preperimetric glaucoma, and perimetric glaucoma patients. Values are depicted as measured around the optic nerve head beginning from temporal (T) to superior (S), nasal (N), and inferior (I). Graphs include the 95% CI.
Figure 3
 
Diagnostic performance of optical coherence tomography parameters. Area under receiver operating characteristic of (dashed) as well as inner (red), middle (yellow), and outer (green) RNFLT for discriminating perimetric and preperimetric glaucoma patients as well as ocular hypertensive subjects from normal eyes decrease in patients with less advanced glaucoma stage. Dashed line indicates 95% specificity. Area under receiver operating characteristic values are depicted inside each diagram.
Figure 3
 
Diagnostic performance of optical coherence tomography parameters. Area under receiver operating characteristic of (dashed) as well as inner (red), middle (yellow), and outer (green) RNFLT for discriminating perimetric and preperimetric glaucoma patients as well as ocular hypertensive subjects from normal eyes decrease in patients with less advanced glaucoma stage. Dashed line indicates 95% specificity. Area under receiver operating characteristic values are depicted inside each diagram.
Figure 4
 
Venn diagrams. Venn diagrams for BMO-MRW (dashed circles) and RNFLT (continuous circles) for distinguishing perimetric and preperimetric glaucoma at 95% and 90% specificity show the absolute number of patients correctly diagnosed by one, both, or none of the two parameters. In the upper left diagram one patient was solely distinguished by BMO-MRW. For reasons of comprehensibility this number is not depicted in the graph. Asterisks indicate which tests showed significantly different sensitivity between BMO-MRW and RNFLT (McNemar's test, P < 0.05).
Figure 4
 
Venn diagrams. Venn diagrams for BMO-MRW (dashed circles) and RNFLT (continuous circles) for distinguishing perimetric and preperimetric glaucoma at 95% and 90% specificity show the absolute number of patients correctly diagnosed by one, both, or none of the two parameters. In the upper left diagram one patient was solely distinguished by BMO-MRW. For reasons of comprehensibility this number is not depicted in the graph. Asterisks indicate which tests showed significantly different sensitivity between BMO-MRW and RNFLT (McNemar's test, P < 0.05).
Table 1
 
Demographic and Clinical Characteristics of the Subjects
Table 1
 
Demographic and Clinical Characteristics of the Subjects
Table 2
 
Morphometric Characteristics of the Subjects
Table 2
 
Morphometric Characteristics of the Subjects
Table 3
 
Diagnostic Performance to Detect Perimetric Glaucoma
Table 3
 
Diagnostic Performance to Detect Perimetric Glaucoma
Table 4
 
Diagnostic Performance to Detect Preperimetric Glaucoma
Table 4
 
Diagnostic Performance to Detect Preperimetric Glaucoma
Table 5
 
Influence of Bruch's Membrane Opening Area
Table 5
 
Influence of Bruch's Membrane Opening Area
×
×

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.

×