Diagnostic Accuracy of Optical Coherence

Tomography and Scanning Laser

Tomography for Identifying Glaucoma in
Myopic Eyes
Rizwan Malik, MD, PhD,1,2 Anne C. Belliveau, BSc,1 Glen P. Sharpe, BSc,1 Lesya M. Shuba, MD, PhD,1
Balwantray C. Chauhan, PhD,1 Marcelo T. Nicolela, MD1

Purpose: Ruling out glaucoma in myopic eyes often poses a diagnostic challenge because of atypical optic disc
morphology and visual field defects that can mimic glaucoma. We determined whether neuroretinal rim assessment
based on Bruch’s membrane opening (BMO), rather than conventional optic disc margin (DM)-based assessment or
retinal nerve fiber layer (RNFL) thickness, yielded higher diagnostic accuracy in myopic patients with glaucoma.
Design: Case-control, cross-sectional study.
Participants: Myopic patients with glaucoma (n ¼ 56) and myopic normal controls (n ¼ 74).
Methods: Myopic subjects with refraction error greater than 2 diopters (D) (spherical equivalent) and typical
myopic optic disc morphology, with and without glaucoma, were recruited from a glaucoma clinic and a local
optometry practice. The final classification of myopic glaucoma or myopic control was based on consensus
assessment by 3 clinicians of visual fields and optic disc photographs. Participants underwent imaging with
confocal scanning laser tomography for measurement of DM rim area (DM-RA) and with spectral domain optical
coherence tomography (SD OCT) for quantification of a BMO-based neuroretinal rim parameter, minimum rim
width (BMO-MRW), and RNFL thickness.
Main Outcome Measures: Sensitivity of DM-RA, BMO-MRW, and RNFL thickness at a fixed specificity of
90% and partial area under the curves (pAUCs) for global and sectoral parameters for specificities 90%.
Results: Sensitivities at 90% specificity were 30% for DM-RA and 71% for both BMO-MRW and RNFL
thickness. The pAUC was higher for the BMO-MRW compared with DM-RA (P < 0.001), but similar to RNFL
thickness (P > 0.5). Sectoral values of BMO-MRW tended to have a higher, but nonsignificant, pAUC across all
sectors compared with RNFL thickness.
Conclusions: Bruch’s membrane opening MRW is more sensitive than DM-RA and similar to RNFL thickness
for the identification of glaucoma in myopic eyes and offers a valuable diagnostic tool for patients with glaucoma
with myopic optic discs. Ophthalmology 2016;123:1181-1189 ª 2016 by the American Academy of Ophthal-
mology.

Supplemental material is available at www.aaojournal.org.

The diagnosis of primary open-angle glaucoma relies on the
clinician’s ability to identify structural abnormality of the
optic disc.1 In myopia, identification of glaucomatous optic
disc changes often poses a diagnostic challenge2 because
atypical optic disc morphology, including substantial
peripapillary atrophy,3 varying degrees of disc tilt,4e6 and
abnormally large7,8 or small optic disc size, often is pre-
sent.5 Furthermore, individuals with myopia can have visual
field defects that mimic glaucomatous loss.6,9,10
Myopia is a significant risk factor for the development of
primary open-angle glaucoma,11e13 and there is strong
supporting evidence for an increase in the prevalence of
myopia, particularly in urban areas.14,15 As these myopic
individuals age, clinicians likely are to be increasingly faced
with the challenge of diagnosing glaucoma in myopic eyes,
requiring more sensitive and specific methods for diagnosis.
Automated imaging devices, such as confocal scanning
laser tomography (CSLT) and optical coherence tomogra-
phy, give an objective and reproducible measure of optic
nerve head or retinal nerve fiber layer (RNFL) structure.16
However, the utility of these devices has been limited in
myopic eyes because of low diagnostic accuracy.17e19 At
90% specificity, CSLT has been reported to have a sensi-
tivity of approximately 50% in these eyes.18
In clinical settings, CSLT has been used widely for neu-
roretinal rim measurement for quantifying the likelihood of
glaucoma20 and its progression.21 Neuroretinal rim
measurements from CSLT are based on the clinically

 2016 by the American Academy of Ophthalmology http://dx.doi.org/10.1016/j.ophtha.2016.01.052 1181

Published by Elsevier Inc. ISSN 0161-6420/16
 Ophthalmology Volume 123, Number 6, June 2016

Figure 1. Optic disc photographs of eyes in the control group, selected at random. A, Right eye of 72-year-old man with spherical equivalent (SE) of 3.25
diopters (D). B, Left eye of 53-year-old man with SE of 9.38 D. C, Left eye of 70-year-old man with SE of 6.25 D. D, Right eye of 59-year-old woman
with SE of 7.11 D. E, Right eye of 67-year-old woman with SE of 4.50 D. F, Right eye of 37-year-old man with SE of 7.50 D. G, Left eye of 78-year-old
man with SE of 7.25 D. H, Left eye of 37-year-old man with SE 7.12 D. I, Right eye of 62-year-old woman with SE of 9.00 D. J, Left eye of 63-year-old
man with SE of 3.50 D.

identifiable optic disc margin (DM). However, recent findings
with optical coherence tomography have challenged the
validity and accuracy of conventional DM-based mea-
sures.22 Alternative anatomically and geometrically accurate
rim parameters,23,24 based on the Bruch’s membrane open-
ing (BMO), have been proposed. A new parameter measuring
the minimum distance between the BMO and the internal
limiting membrane, termed “BMO-minimum rim width”
(BMO-MRW), seems to provide a better representation of the
amount of neuroretinal rim tissue than conventional DM-
based parameters. Previous studies have demonstrated better
diagnostic accuracy24 and greater correlation with other
structural and functional parameters25,26 of BMO-MRW
than conventional DM-based rim measures.
The primary aim of this study was to compare the
diagnostic accuracy of BMO-MRW with DM-based rim
assessment from CSLT and RNFL thickness measurements
to separate individuals with myopic optic disc morphology
with glaucoma (myopic glaucoma) from those without
glaucoma (myopic controls). Secondarily, we also explored
the sensitivity of these measures on a sectoral basis.
Methods
Participants
Study participants included myopic patients with glaucoma and
healthy controls with myopic optic disc morphology (defined
later). Patients with glaucoma and myopia were recruited pro-
spectively from the glaucoma clinic at the Eye Care Centre, Queen
Elizabeth II Health Sciences Centre, Halifax, Canada. Myopic
participants without glaucoma were recruited consecutively from a
local optometry practice. Typical myopic optic disc morphology
(Fig 1) was defined as the presence of significant beta type
of peripapillary atrophy, characterized by complete loss of
retinal pigment epithelium, adjacent to the optic disc.27 Beta
peripapillary atrophy usually was present sectorally, mostly
temporally, associated with the direction of the optic disc tilt, but
could also be present circumferentially around the optic disc.
Optic disc tilt also is part of the typical myopic optic disc
morphology, but its presence was not necessary for inclusion in
this study.
For both groups, additional inclusion criteria were best-
corrected visual acuity of 20/40, myopia greater than 2 di-
opters (D) (spherical equivalent refraction), cylinder correction
within 4 D, absence of retinal disease (including degenerative
myopia) or optic nerve disease other than glaucoma, and willing-
ness to participate in the study. If both eyes were eligible, 1 eye
was randomly selected for analysis.
The study was approved by the Ethics Review Board of Capital
Health and followed the tenants of the Declaration of Helsinki. All
subjects provided written informed consent.
Study Definition of Patients with Glaucoma and
Myopic Controls
The diagnosis of myopic glaucoma or myopic control was defined
by consensus among 3 fellowship-trained glaucoma subspecialists
who evaluated the visual fields and optic disc photographs from all
participants independently and were masked from all other de-
mographic and clinical information. To minimize bias and maintain
an independent reference standard for evaluating the diagnostic
accuracy of structural tests, visual field appearance was primarily
used for deciding the diagnostic group of the participants. In-
dividuals were included in the myopic control group if their visual
field was graded as normal or with abnormalities consistent with
myopia, but not glaucoma, independently by all 3 clinicians,
regardless of the grading given to their optic disc. If all 3 clinicians
graded the visual field as having glaucomatous abnormalities, the
participant was included in the myopic glaucoma group. In cases of
disagreement in visual field grading, the clinicians used the optic
disc evaluation to obtain a consensus decision to place the partic-
ipant in the glaucoma or control group (examples given in
“Results”).
Study Procedures
Study subjects had a variety of diagnostic tests (described later)
performed in 1 study visit. For individuals who were perimetrically

1182
 Malik et al 
Diagnosis of Glaucoma in Myopia
Figure 2. A, Infrared image showing the position of the radial section; B, Corresponding SD OCT image for this radial scan. Optical coherence tomography
landmarks and Bruch’s membrane opening minimum rim width (BMO-MRW) (blue arrow) for one of the images in the study. Red dots indicate the BMO.
The disc margin rim area (DM-RA), BMO-MRW, and retinal nerve fiber layer (RNFL) thickness profile for this patient is shown in Figure 8. ILM ¼ internal
limiting membrane.
naïve, a training visual field test was obtained initially, with the
second test used for the study. During the study visit, best-
corrected distance visual acuity, ocular biometry (IOLMaster,
Zeiss, Carl Zeiss Meditec, Dublin, CA), Goldmann applanation
tonometry, and a complete slit-lamp and fundus examination were
performed. Standard automated perimetry was performed with the
Humphrey Field Analyzer (Carl Zeiss Meditec) and the 24-2 SITA
strategy, with the appropriate near refractive correction. Only
reliable (15% false-positives, 30% fixation losses, and 30%
false-negatives) visual fields were included. All participants had
fundus imaging with stereo disc photography (Zeiss Visucam Pro
NM fundus nonmydriatic camera, Carl Zeiss Meditec Inc., Jena,
Germany), CSLT (Heidelberg Retina Tomograph, Heidelberg
Engineering GmbH, Heidelberg, Germany), and spectral domain
optical coherence tomography (SD OCT) (Spectralis, Heidelberg
Engineering GmbH).
For CSLT, mean topography and reflectance images were
automatically computed by the software from 3 individual 15
images centered on the optic disc. Only images with a mean
standard deviation mean pixel height <40 mm and without obvious
motion artifacts were accepted. The optic disc contour line was
drawn by an experienced technician and subsequently checked by a
clinician (RM). Rim area (RA) was computed using the standard
reference plane.28
Two scanning patterns were used with SD OCT. First, a radial
pattern comprising 24 angularly equidistant high-resolution 15 B-
scans was used to compute the neuroretinal rim parameters. Both
scan patterns were centered on the BMO center, which was aligned
with the fovea, because previous studies showed significant vari-
ation on the position of the fovea relative to the center of BMO,
which could have consequences in sectorial neuroretinal rim
analysis.29 Before scan acquisition, the operator manually marked
4 BMO points on 2 separate perpendicular radial scan lines. These
2 radial scans were selected on the basis of the clarity of the BMO.
The image acquisition was then completed, with the scan pattern
aligned with the center of the 4 points. After image acquisition,
the true BMO center was determined on the basis of all 48 BMO
points (24 radial scans). The true center (based on 48 points)
was then compared with the initial scan center (based on 4
points). If the discrepancy between these 2 central points was
greater than 100 mm, the whole process of image acquisition was
repeated. Data for each B-scan were averaged from 20 to 30
individual B-scans, with 1536 A-scans per B-scan. An automated
segmentation algorithm (Heidelberg Eye Explorer 1.7.1.0;
Heidelberg Engineering GmbH) was used to segment BMO and
internal limiting membrane.30 The segmentation was checked by
an experienced observer (GPS) and corrected when necessary.
The automated segmentation has been shown to be highly
comparable to manual segmentation in nonmyopic eyes.30
Second, a 12 circular scan centered on BMO center was used to
measure the circumpapillary RNFL thickness. The scan was
composed of 768 A-scans and data were averaged from 100
individual B-scans. For each SD OCT scan, the operator checked
the image quality, automated segmentation of RNFL, which was
corrected manually when necessary, and quality score (>20).
Neuroretinal Rim Parameters
For CSLT, the internal software (Heidelberg Eye Explorer) was
used to generate the DM-RA as well as the results of the Moor-
fields Regression Analysis (MRA).20 The MRA uses the
relationship between the optic disc area and the DM-RA and the
age of the subject in a normal population to derive the normal
prediction limits of neuroretinal RA. For a given optic disc, a
classification of within normal limits is made if the RA is within
the 95% prediction interval for the given disc area; borderline, if
the RA is within the 95% and 99.9% prediction intervals; or
outside normal limits if the RA is outside the 99.9% prediction
interval.28
The BMO-MRW parameter has been detailed by Reis et al.23
Briefly, 48 BMO points are yielded from the 24 radial scans
with 3-dimensional coordinates through which the software
automatically fitted a spline to derive a closed curve representing
the BMO around the ONH. From this, the BMO area was
computed and a best-fit plane representing the BMO reference
plane was computed. The BMO-MRW was defined as the mini-
mum distance between the BMO and the internal limiting mem-
brane (Fig 2) and determined automatically by the software. The
BMO-MRW was computed at the 48 equally spaced angular po-
sitions around the BMO center.23,31
Sample Size
It was estimated that for a power of 0.8, with type I error a set to
0.05, at least 55 participants would be needed in each group
(myopic control/myopic glaucoma) to detect a difference of 0.20
(taken to be clinically significant difference) between the area of
receiver operating characteristic (ROC) curves for DM-RA and
BMO-MRW, assuming an area under the curve (AUC) ROC of
0.70 for DM-RA.32
1183
 Ophthalmology Volume 123, Number 6, June 2016

Data Analysis Table 1. Mean (Standard Deviation) of Descriptive Parameters in

Myopic Controls and Myopic Patients with Glaucoma
The primary analysis involved the comparison of sensitivity of
global measures for identifying glaucoma in myopic eyes: BMO- Myopic Controls Myopic Glaucoma
MRW with DM-RA and BMO-MRW with RNFL thickness. In a n [ 74 n [ 56 P Value
secondary analysis, we also explored trends of sensitivity of these
parameters across 6 sectors. Age (yrs) 56.4 (11.3) 62.5 (9.46) 0.001
All rim parameters were computed globally and sectorally. For Sex (M:F) 33:41 34:22 0.08
SE (D) 5.81 (2.24) 6.06 (3.04) 0.61
CSLO, the sectors were 90 nasally and temporally and 45 for
Axial length (mm) 26.0 (1.26) 26.4 (1.61) 0.13
each of the other 4 sectors (superotemporal, superonasal, infer-
Disc size (mm2) 1.90 (0.64) 2.08 (0.98) 0.24
onasal, inferotemporal), with the temporal sector centered on the BMO area (mm2) 2.11 (0.66) 2.07 (0.76) 0.80
horizontal meridian. For SD OCT, there were four 40 sectors Mean deviation (dB) 0.52 (1.47) 7.08 (4.92) <0.001
(superior-temporal, inferior-temporal, superior-nasal, and inferior-
nasal), one 90 temporal sector, and one 110 nasal sector, with
the center of the temporal sector centered on the fovea. Raw values BMO ¼ Bruch’s membrane opening; D ¼ diopters; dB ¼ decibels; SE ¼
of both global and sectoral DM-RA were extracted from the CSLT spherical equivalent.
software, and raw values of global BMO-MRW and the circum-
papillary RNFL were extracted from the SD OCT software. The
AUC and partial AUC (pAUC) ROC curve for specificity 90% automated segmentation. The circular scans were all adequately
were computed and compared for these parameters. Partial areas centered, and none of these required manual adjustment.
were standardized by the method described by McClish.33 The The baseline characteristics of the study participants are shown
sensitivity of these parameters at fixed specificities of 90% and in Table 1. The spherical refractive error and axial length were
95% were reported. similar in the 2 groups (P > 0.12), as was the proportion of men
In addition to the DM-RA raw values, we also evaluated the (P ¼ 0.08, chi-square test). The patients with glaucoma were, on
diagnostic performance of the MRA. Because the MRA yields a average, 6 years older than the controls (P ¼ 0.001). Optic disc
borderline classification, 2 different MRA criteria were used: (1) size, as measured by CSLT, and the area enclosed by the Bruch’s
liberal, MRA1, where the borderline cases were classified as membrane (BMO area) were similar in the 2 groups (P > 0.23).
abnormal; and (2) conservative, MRA2, where borderline cases Visual field mean deviation, in decibels, was significantly lower
were classified as normal. Thus, there were 2 discrete points for the in the myopic glaucoma group compared with myopic controls
MRA analysis on the ROC curve corresponding to MRA1 and 2, (P < 0.001). The DM-RA, BMO-MRW, and RNFL thickness were
whereas DM-RA, MRW, and RNFL thickness were continuous all significantly lower in myopic patients with glaucoma compared
curves. with myopic controls (Table 2).
Categoric variables were compared using a chi-square test, and The AUC was higher (P < 0.01) for the BMO-MRW than the
continuous variables were compared using a t test. Receiver DM-RA (Fig 5 and Table 3). The pAUC for specificities 90%
operating characteristic analysis was performed with the pROC also was higher for the BMO-MRW compared with DM-RA
package34 in the open platform R software.35 Confidence intervals (P < 0.01) (Table 3). Both the AUC and pAUC for BMO-MRW
for AUC and pAUC were computed with a bootstrap method with were similar to RNFL thickness (P > 0.25). Sensitivities at a
2000 iterations. The AUCs were compared using function ROCtest fixed specificity of 90% were 30.0% for DM-RA and 71.4% for
in the package pROC, which uses a Wald statistic, dividing the both BMO-MRW and RNFL thickness. Respective sensitivity
observed difference by its standard error compared with the values for specificity of 95% are shown in Table 3. MRA1 yielded
standard normal distribution to report a P value.36 sensitivity of 64.3% and specificity of 71.6%, whereas the
respective values of MRA2 were 41.1% and 90.5% (not shown
in Table 3).
The ROC curves for 6 sectors are shown in Figure 6 (available
Results at www.aaojournal.org). The pAUCs for the specificity 90% for
the 6 sectors are shown in Table 4 (available at
A total of 131 eyes of 131 participants were included in this study. www.aaojournal.org). In general, the same trends were observed
Figure 3 (available at www.aaojournal.org) shows examples of with sectoral analysis as found with global parameters. The
visual fields that were classified as myopic normal and pAUC for BMO-MRW was higher than for DM-RA for each
glaucoma. Masked independent agreement for classifying visual sector (P  0.01). The pAUC for BMO-MRW tended to be higher
fields for the 3 observers is shown in Figure 4 (available at
www.aaojournal.org). All 3 clinicians agreed that 42 subjects
(32%) were glaucomatous and 72 subjects (55%) were
nonglaucomatous (i.e., complete agreement observed in 114 Table 2. Mean (Standard Deviation) of Diagnostic Parameters
subjects [88%]). Kappa agreement (L) ranged from 0.71 to 0.73 Evaluated in Myopic Controls and Myopic Patients with
among the 3 possible pairs of clinicians (good agreement). There Glaucoma
was incomplete initial agreement in visual field classification for
17 subjects (13%). From these, agreement among the 3 observers Myopic Controls Myopic Glaucoma
was reached by consensus in 16 subjects after reviewing their n [ 74 n [ 56 P Value
optic disc photographs; agreement could not be reached for 1 DM-RA (mm2) 1.38 (0.52) 1.09 (0.43) 0.0008
subject, who was excluded from the study. Therefore, the final BMO-MRW (mm) 295 (74.0) 182 (50.9) <0.0001
study sample consisted of 56 eyes from 56 subjects in the RNFL thickness (mm) 84.8 (10.8) 64.8 (10.9) <0.0001
myopic glaucoma group and 74 eyes from 74 subjects in the
myopic control group.
For the radial SD OCT images, the BMO had to be manually BMO ¼ Bruch’s membrane opening; DM-RA ¼ rim area (from CSLT);
MRW ¼ minimum rim width; RNFL ¼ retinal nerve fiber layer.
adjusted in at least 1 radial scan in 91 eyes (70%) after the

1184
 Malik et al 
Diagnosis of Glaucoma in Myopia

landmark that is accurately identified by the automated

Figure 5. Receiver operating characteristic (ROC) curve for global mea-
sures. See text for explanation of Moorfields regression analysis (MRA)1
and 2. The grey shaded box shows the maximum partial area under the curve
(pAUC) for specificities between 90% and 100%. BMO ¼ Bruch’s mem-
brane opening; MRW ¼ minimum rim width; RA ¼ (disc margin-based)
rim area (confocal scanning laser tomography [CSLT]); RNFL ¼ retinal
nerve fiber layer.
than for RNFL thickness, although statistical significance was
observed only in the inferonasal sector (P < 0.01). Across the 6
sectors, the sensitivity of MRA1 ranged from 40% to 60% and of
MRA2 ranged from 50% to 70% for MRA2.
Figures 7 and 8 show representative examples of a myopic
control and a myopic patient with glaucoma, respectively. In
both examples, the BMO-MRW and RNFL thickness results
seem to be more consistent with the status of the visual field than
the DM-RA results.
Discussion
Spectral domain OCT permits a better visualization of the
deep optic nerve and its exit through the scleral canal. The
termination of Bruch’s membrane, which delineates the
BMO, is, in most cases, an easily identifiable anatomic
delineation software.30 In addition to that, the axons
composing the optic nerve have to exit the eye by going
through the BMO, making it the logical choice as a plane
from which to measure neuroretinal rim.24 By measuring
the minimal distance between the BMO and the internal
limiting membrane (BMO-MRW), we can theoretically
have a better assessment of the amount of neuroretinal rim
in tilted optic nerves with an oblique insertion, commonly
seen in myopic individuals, as opposed to measuring
tissue on the plane of the BMO.
Our study showed that both global and sectoral BMO-
MRW had significantly better diagnostic performance in
identifying glaucoma in patients with myopia than the
conventional DM-RA measures, similar to what has been
reported in a study that limited the degree of myopia to
6 D.24 Likewise, the diagnostic performance of BMO-
MRW was comparable to that of RNFL thickness.24
However, the sensitivity of both BMO-MRW and RNFL
thickness was lower in myopic eyes (71% at 90% speci-
ficity). As a comparison, these parameters had higher
sensitivity at 90% sensitivity (85% and 81%, respectively)
in a previous study that excluded subjects with myopia
exceeding 6 D,24 although in the previous study patients
had an earlier stage of disease compared with the current
one (mean deviation of 4.0 and 7.1 decibels,
respectively). This finding is not surprising considering the
significant structural abnormalities in myopic eyes, which
adversely affect the diagnostic performance of imaging
devices in this population.
Several investigators have evaluated the diagnostic per-
formance of SD OCT in myopic eyes. The AUC for RNFL
thickness reported in myopic eyes ranged from 0.84 to
0.98,19,37,38 which was similar to what we observed in the
current study (0.90). However, it is difficult to compare
diagnostic performance across studies, because the inclu-
sion/exclusion criteria and the stage of disease vary. In the
current study, we selected only eyes with myopic optic disc
morphology in the control group, which is likely to reduce
sensitivity estimates.
Likewise, previous studies have found poor diagnostic
performance of DM-RA in myopic eyes.18,32 Mayama
et al18 found that the AUC was significantly lower for
myopic eyes compared with emmetropic eyes. The poor
diagnostic performance of DM-RA in these eyes can be
explained by a number of factors. First, the oblique insertion

Table 3. Areas and Partial Areas Under the Receiving Operator Characteristic Curve and Sensitivities at Fixed Specificities of 90% and
95%, with Corresponding 95% Confidence Intervals

Sensitivity
AUC pAUC 90% Specificity 95% Specificity
DM-RA 66.7 (57.7e76.2) 55.6 (50.5e63.5) 30.0 (8.94e48.2) 14.3 (3.63e36.6)
BMO-MRW 90.0 (84.4e94.4) 80.5 (73.8e87.8) 71.4 (55.48e85.7) 62.5 (48.2e78.6)
RNFL thickness 89.7 (83.9e94.7) 77.5 (67.8e87.1) 71.4 (57.1e84.6) 64.9 (28.8e78.8)

AUC ¼ total area under ROC curve; BMO ¼ Bruch’s membrane opening; DM-RA ¼ disc-margin rim area from CSLT; MRW ¼ minimum rim width;
pAUC¼ partial area under ROC curve for specificity range 90%e100%; RNFL ¼ retinal nerve fiber layer.

1185
 Ophthalmology Volume 123, Number 6, June 2016

Figure 7. Data for the right eye of a myopic control subject in the study: A, Visual field exam. B, Confocal scanning laser tomography (CSLT) image.
C, Bruch’s membrane opening minimum rim width (BMO-MRW) profile with corresponding infrared image D (red dots show BMO opening). E, Retinal
nerve fiber layer (RNFL) thickness profile with corresponding infrared image F. In this example, the Moorfields regression analysis (MRA) for CSLT was
“outside normal limits,” although both BMO-MRW and RNFL thickness, in addition to the visual field, showed normal results. BMO ¼ Bruch’s membrane
opening; INF ¼ inferior; NAS ¼ nasal; NS ¼ nasal superior; SUP ¼ superior; TMP ¼ temporal; TS ¼ temporal superior.

of the optic nerve fibers into the scleral canal is not taken
into consideration in conventional DM-RA estimates, where
measurements are made in the plane of the optic disc.
Second, the cup in these eyes often is shallow and sloping,
and the margin between cup and rim is difficult to identify.
Third, myopic discs often have a large amount of peri-
papillary atrophy, and the clinical DM can be difficult to
determine in these eyes. Fourth, myopic optic discs can be
abnormally small or abnormally large, and the diagnostic
performance of DM-RA has been shown to be negatively
affected by optic disc size, both small and large.39
In the sectoral analysis, BMO-MRW performed better
than RNFL thickness for all 6 sectors, although statistical
significance was observed only for the inferonasal sector.

1186
 Malik et al 
Diagnosis of Glaucoma in Myopia

Figure 8. Data from the left eye of a myopic subject with glaucoma in the study (same patient shown in Figure 2). A, Visual field exam. B, confocal scanning
laser tomography (CSLT) image. C, Minimum rim width (MRW) profile with Bruch’s membrane opening (BMO) opening with corresponding infrared
image (D). E, Retinal nerve fiber layer (RNFL) thickness profile with corresponding infrared image F. This patient had a superotemporal visual field
defect (Fig 7A). Dashed arrow in C shows a dip in the BMO-MRW, with thin RNFL in the same region (E). An RNFL defect is visible on the CSLT
image (B, blue arrow) and infrared optical coherence tomography (OCT) image (F, yellow arrow). Although the Moorfields regression analysis (MRA)
classification was only borderline inferonasally, the BMO-MRW was markedly thin in the same region (Fig 7C). INF ¼ inferior; NAS ¼ nasal; NS ¼ nasal
superior; SUP ¼ superior; TMP ¼ temporal; TS ¼ temporal superior.

Another interesting finding was that the inferotemporal Study Limitations

sector best differentiated between myopic controls and
myopic glaucomatous eyes. Further studies should better
evaluate the utility of sectoral analysis in myopic eyes, but
our results suggest that it might provide useful diagnostic
information.
There are a number of possible limitations of this study.
First, classification of glaucoma was primarily based on
visual field appearance with clinical judgment of the optic
disc in ambiguous cases. This approach permitted the

1187
 Ophthalmology Volume 123, Number 6, June 2016
inclusion of a range of both normal and glaucomatous
myopic optic disc appearances and was designed to reduce
potential spectrum bias (the selection of individuals without
diagnostic uncertainty), which could artificially increase
diagnostic accuracy.40,41 It is possible that some individuals
were misclassified; however, it is unlikely that this would
affect our overall conclusions because the relative, rather
than absolute, diagnostic capabilities of the studied param-
eters were being compared. Second, the control group was
younger than the patient group. Although the precise effect
of group age differences on sensitivity are difficult to
quantify, DM-RA,20 BMO-MRW,24 and RNFL thickness42
all decrease with age, and this age effect could influence our
results. Last, the inclusion in the study was partially based
on disc morphology, and so the outcomes may be biased
in terms of favoring rim, rather than RNFL thickness, for
identifying glaucoma. The diagnostic performance of
RNFL thickness may have been higher if the population
had been selected on the basis of established RNFL loss
rather than rim loss, although in the majority of subjects
inclusion was based on visual field assessment, and disc
morphology was used for inclusion only in 17 patients in
whom there was disagreement on the visual field
classification.
In conclusion, BMO-MRW and RNFL thickness yield a
higher diagnostic accuracy in myopic discs than DM-RA,
with all parameters performing less well compared with
nonmyopic eyes. It is possible that combining information
from BMO-MRW and RNFL thickness, as well as incor-
porating sectoral analysis and other structural information,
such as extent and type of peripapillary atrophy, may
yield to increased diagnostic performance in myopic eyes,
which is a subject of future study. We propose that
BMO-MRW should be incorporated into routine clinical
practice to assist clinicians in detecting glaucoma in myopic
patients.
Acknowledgments. The authors thank David Dobbelsteyn,
OD, Paul Gray, OD, Carl Davis, OD, Jeff Sangster, OD, Leah
Gallie, OD, and the staff at the Insight Optometry Group, Halifax,
Nova Scotia, for assistance with recruiting healthy volunteers for
this study.
References
1. Weinreb RN, Khaw PT. Primary open-angle glaucoma. Lancet
2004;363:1711–20.
2. Chang RT, Singh K. Myopia and glaucoma: diagnostic and
therapeutic challenges. Curr Opin Ophthalmol 2013;24:
96–101.
3. Chang L, Pan CW, Ohno-Matsui K, et al. Myopia-related
fundus changes in Singapore adults with high myopia. Am J
Ophthalmol 2013;155:991–999 e1.
4. How AC, Tan GS, Chan YH, et al. Population prevalence of
tilted and torted optic discs among an adult Chinese population
in Singapore: the Tanjong Pagar Study. Arch Ophthalmol
2009;127:894–9.
5. Samarawickrama C, Mitchell P, Tong L, et al. Myopia-related
optic disc and retinal changes in adolescent children from
Singapore. Ophthalmology 2011;118:2050–7.
1188
6. Tay E, Seah SK, Chan SP, et al. Optic disk ovality as an index
of tilt and its relationship to myopia and perimetry. Am J
Ophthalmol 2005;139:247–52.
7. Nagaoka N, Ohno-Matsui K, Saka N, et al. Clinical charac-
teristics of patients with congenital high myopia. Jpn J Oph-
thalmol 2011;55:7–10.
8. Jonas JB, Gusek GC, Naumann GO. Optic disk morphometry
in high myopia. Graefes Arch Clin Exp Ophthalmol 1988;226:
587–90.
9. Vuori ML, Mantyjarvi M. Tilted disc syndrome may mimic
false visual field deterioration. Acta Ophthalmol 2008;86:
622–5.
10. Hwang YH, Yoo C, Kim YY. Long-term development of
significant visual field defects in highly myopic eyes. Am J
Ophthalmol 2011;152:878–9. author reply 9e80.
11. Chon B, Qiu M, Lin SC. Myopia and glaucoma in the South
Korean population. Invest Ophthalmol Vis Sci 2013;54:
6570–7.
12. Vijaya L, Rashima A, Panday M, et al. Predictors for incidence
of primary open-angle glaucoma in a South Indian population:
the Chennai eye disease incidence study. Ophthalmology
2014;121:1370–6.
13. Marcus MW, de Vries MM, Junoy Montolio FG,
Jansonius NM. Myopia as a risk factor for open-angle glau-
coma: a systematic review and meta-analysis. Ophthalmology
2011;118:1989–1994 e2.
14. Lin LL, Shih YF, Hsiao CK, Chen CJ. Prevalence of myopia
in Taiwanese schoolchildren: 1983 to 2000. Ann Acad Med
Singapore 2004;33:27–33.
15. Vitale S, Sperduto RD, Ferris FL 3rd. Increased prevalence of
myopia in the United States between 1971-1972 and 1999-
2004. Arch Ophthalmol 2009;127:1632–9.
16. Greenfield DS, Weinreb RN. Role of optic nerve imaging in
glaucoma clinical practice and clinical trials. Am J Ophthalmol
2008;145:598–603.
17. Melo GB, Libera RD, Barbosa AS, et al. Comparison of optic
disk and retinal nerve fiber layer thickness in nonglaucomatous
and glaucomatous patients with high myopia. Am J Oph-
thalmol 2006;142:858–60.
18. Mayama C, Tsutsumi T, Saito H, et al. Glaucoma-induced
optic disc morphometric changes and glaucoma diagnostic
ability of Heidelberg Retina Tomograph II in highly myopic
eyes. PLoS One 2014;9:e86417.
19. Shoji T, Nagaoka Y, Sato H, Chihara E. Impact of high
myopia on the performance of SD-OCT parameters to detect
glaucoma. Graefes Arch Clin Exp Ophthalmol 2012;250:
1843–9.
20. Wollstein G, Garway-Heath DF, Hitchings RA. Identification
of early glaucoma cases with the scanning laser ophthalmo-
scope. Ophthalmology 1998;105:1557–63.
21. Bowd C, Balasubramanian M, Weinreb RN, et al. Performance
of confocal scanning laser tomograph Topographic Change
Analysis (TCA) for assessing glaucomatous progression.
Invest Ophthalmol Vis Sci 2009;50:691–701.
22. 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.e2.
23. 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;53:1852–60.
24. Chauhan BC, O’Leary N, Almobarak FA, et al. Enhanced
detection of open-angle glaucoma with an anatomically ac-
curate optical coherence tomography-derived neuroretinal rim
parameter. Ophthalmology 2013;120:535–43.
 Malik et al 
Diagnosis of Glaucoma in Myopia
25. Gardiner SK, Ren R, Yang H, et al. A method to estimate the
amount of neuroretinal rim tissue in glaucoma: comparison
with current methods for measuring rim area. Am J Oph-
thalmol 2014;157:540–549.e1e2.
26. Pollet-Villard F, Chiquet C, Romanet JP, et al. Structure-
function relationships with spectral-domain optical coherence
tomography retinal nerve fiber layer and optic nerve head
measurements. Invest Ophthalmol Vis Sci 2014;55:2953–62.
27. Jonas JB, Nguyen XN, Gusek GC, Naumann GO. Para-
papillary chorioretinal atrophy in normal and glaucoma eyes. I.
Morphometric data. Invest Ophthalmol Vis Sci 1989;30:
908–18.
28. Heidelberg-Engineering. Heidelberg Retinal Tomograph II
(manual). Dossenheim, Germany: Heidelberg Engineering
GmbH; 2001.
29. Danthurebandara VM, Sharpe GP, Hutchison DM, et al.
Enhanced structure-function relationship in glaucoma with
an anatomically and geometrically accurate neuroretinal
rim measurement. Invest Ophthalmol Vis Sci 2015;56:
98–105.
30. Almobarak FA, O’Leary N, Reis AS, et al. Automated seg-
mentation of optic nerve head structures with optical coher-
ence tomography. Invest Ophthalmol Vis Sci 2014;55:
1161–8.
31. Reis AS, Sharpe GP, Yang H, et al. Optic disc margin anatomy
in patients with glaucoma and normal controls with spectral
domain optical coherence tomography. Ophthalmology
2012;119:738–47.
32. Lee NY, Park HL, Park CK. Glaucoma detection in high
myopia with the Heidelberg Retina Tomograph 3. Semin
Ophthalmol 2015;1–6.
33. McClish DK. Analyzing a portion of the ROC curve. Med
Decis Making 1989;9:190–5.
Footnotes and Financial Disclosures
Originally received: November 18, 2015.
Final revision: January 19, 2016.
Accepted: January 29, 2016.
Available online: March 16, 2016. Manuscript no. 2015-2039.
1 Data collection: Malik, Belliveau, Sharpe, Shuba, Nicolela
Department of Ophthalmology and Visual Sciences, Dalhousie Univer-
sity, Halifax, Nova Scotia, Canada.
2
NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye
Hospital NHS Foundation Trust and UCL Institute of Ophthalmology,
London, United Kingdom.
Presented at the Association for Research in Vision and Ophthalmology,
May 4e8, 2014, Orlando, Florida.
Financial Disclosure(s):
The author(s) have made the following disclosure(s): R.M.: A part of the
author’s salary was paid by the National Institute for Health Research (UK)
during research fellowship.
B.C.C.: Unrestricted funding support  Heidelberg Engineering;
Consultant Allergan.
M.T.N.: Consultant  Alcon and Allergan.
R.M.: Supported by a Research Award from the Special Trustees of
Moorfields Eye Hospital (ST12 07E).
34. Robin X, Turck N, Hainard A, et al. pROC: an open-source
package for R and Sþ to analyze and compare ROC curves.
BMC Bioinformatics 2011;12:77.
35. R Development Core Team (2014). R: A language and
environment for statistical computing. Vienna, Austria: R
Foundation for Statistical Computing. Available at: https://
www.r-project.org/. Accessed June 1, 2013.
36. Pepe M, Longton G, Janes H. Estimation and comparison of
receiver operating characteristic curves. Stata J 2009;9:1.
37. Akashi A, Kanamori A, Nakamura M, et al. The ability of
macular parameters and circumpapillary retinal nerve fiber
layer by three SD-OCT instruments to diagnose highly myopic
glaucoma. Invest Ophthalmol Vis Sci 2013;54:6025–32.
38. Shoji T, Sato H, Ishida M, et al. Assessment of glaucomatous
changes in subjects with high myopia using spectral domain
optical coherence tomography. Invest Ophthalmol Vis Sci
2011;52:1098–102.
39. Hawker MJ, Vernon SA, Ainsworth G. Specificity of the
Heidelberg Retina Tomograph’s diagnostic algorithms in a
normal elderly population: the Bridlington Eye Assessment
Project. Ophthalmology 2006;113:778–85.
40. Rao HL, Kumbar T, Addepalli UK, et al. Effect of spectrum
bias on the diagnostic accuracy of spectral-domain optical
coherence tomography in glaucoma. Invest Ophthalmol Vis
Sci 2012;53:1058–65.
41. Medeiros FA, Ng D, Zangwill LM, et al. The effects of study
design and spectrum bias on the evaluation of diagnostic ac-
curacy of confocal scanning laser ophthalmoscopy in glau-
coma. Invest Ophthalmol Vis Sci 2007;48:214–22.
42. Alasil T, Wang K, Keane PA, et al. Analysis of normal retinal
nerve fiber layer thickness by age, sex, and race using spectral
domain optical coherence tomography. J Glaucoma 2013;22:
532–41.
Supported by a Canadian Glaucoma Clinical Research Council Grant.
Author Contributions:
Conception and design: Malik, Chauhan, Nicolela
Analysis and interpretation: Malik, Shuba, Chauhan, Nicolela
Obtained funding: Nicolela
Overall responsibility: Malik, Chauhan, Nicolela
Abbreviations and Acronyms:
AUC ¼ area under the curve; BMO ¼ Bruch’s membrane opening;
CSLT ¼ confocal scanning laser tomography; D ¼ diopters; DM ¼ disc
margin; MRA ¼ Moorfields Regression Analysis; MRW ¼ minimum rim
width; pAUC ¼ partial area under the curve; RA ¼ rim area;
RNFL ¼ retinal nerve fiber layer; ROC ¼ receiver operating characteristic;
SD OCT ¼ spectral domain optical coherence tomography; SE ¼ spherical
equivalent.
Correspondence:
Marcelo T. Nicolela, MD, Department of Ophthalmology and Visual Sci-
ences, Dalhousie University, 1276 South Park Street, 2W Victoria, Room
2035, Halifax, NS B3H 2Y9, Canada. E-mail: nicolela@dal.ca.
1189