Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

DOI 10.1007/s00417-010-1585-5

GLAUCOMA

Diagnostic ability of retinal ganglion cell complex, retinal

nerve fiber layer, and optic nerve head measurements
by Fourier-domain optical coherence tomography
Andreas Schulze & Julia Lamparter & Norbert Pfeiffer &
Fatmire Berisha & Irene Schmidtmann &
Esther M. Hoffmann

Received: 6 August 2010 / Revised: 16 November 2010 / Accepted: 18 November 2010 / Published online: 15 January 2011
# Springer-Verlag 2011

Abstract
Purpose To evaluate the diagnostic ability of Fourier-domain
optical coherence tomography (FD-OCT) measurements in
glaucoma patients, patients with ocular hypertension, and
normal subjects.
Methods Ninety-three participants with open-angle glaucoma
(OAG), 58 patients with ocular hypertension (OHT), and 60
healthy control subjects were included in the study. All study
participants underwent FD-OCT imaging. Retinal ganglion
cell complex (GCC), macular thickness, peripapillary retinal
nerve fiber layer thickness (RFNL), and optic nerve head
parameters (ONH) were measured in each participant. The
diagnostic ability was evaluated using area under the receiver
operating characteristics curves (AUROC).
Presented in part at the Annual Meeting of Association for Research in
Vision and Ophthalmology, Fort Lauderdale, USA, May 2009
Financial disclosure The FD-OCT RTVue-100 was provided by the
Optovue company (Fremont, USA) at no cost.
The authors have full control of all primary data, and they agree to
allow Graefes Archive for Clinical and Experimental Ophthalmology
to review their data upon request.
A. Schulze (*) : J. Lamparter : N. Pfeiffer : F. Berisha :
E. M. Hoffmann
Department of Ophthalmology,
University Medical Center Mainz,
Langenbeckstrae 1,
55131 Mainz, Germany
e-mail: andreas.schulze@unimedizin-mainz.de
I. Schmidtmann
Department of Medical Biometry,
Epidemiology and Informatics (IMBEI),
University Medical Center Mainz,
Obere Zahlbacher Str. 69,
55131 Mainz, Germany

Results Glaucoma patients showed a significant reduction in

GCC and macular retinal thickness compared to patients with
OHT and normal subjects. No differences in GCC were found
between the patients with OHT and normal subjects. The best
diagnostic ability in the comparison between glaucoma and
normal subjects after adjusting for age was found for cup-todisc ratio (AUROC = 0.848), RNFL average thickness
(AUROC=0.828), and GCC global loss volume (AUROC=
0.805). The diagnostic power of the best GCC, RNFL, and
ONH parameter did not show differences beyond random
variation (p>0.05).
Conclusions Imaging of the GCC using FD-OCT (RTVue100) has a comparable diagnostic ability to RNFL and
ONH measurements in distinguishing between glaucoma
patients and healthy subjects. No differences were found
between patients with OHT and normal subjects with regard
to ONH, RNFL, and GCC parameters.
Keywords Fourier-domain optical coherence tomography .
Glaucoma . Ocular hypertension . Retinal ganglion cell
complex . Diagnostic ability

Introduction
Optical coherence tomography (OCT) is a non-invasive
imaging method used to analyze the optic nerve head and
retinal layers. The ability of OCT to discriminate between
glaucomatous and normal eyes with measurements of the
optic nerve head (ONH), the retinal nerve fiber layer (RNFL),
and macular thickness (MT) has been demonstrated in various
studies [14]. Measurements of the ONH showed the highest
diagnostic accuracy for glaucoma detection (cup/disc area
ratio [2], RIM area [4]) and RNFL (inferior RNFL thickness

1040

[2, 5], mean RNFL thickness [4]). Segmentation discrimination and measurement of the thickness of retinal layers in
the macula region was not possible with time-domain OCT
devices. Tan et al. [6] measured the macular layers with a
complex semi-automatic segmentation algorithm for Stratus
OCT. The diagnostic power of the measurement of inner
retinal layers (nerve fiber layer, ganglion cell layer, and inner
plexiform layer, AUROC=0.91) was comparable to peripapillary nerve fiber layer measurements (average RNFL
thickness, AUROC=0.94). However, these measurements
require the use of manual and complex segmentation of the
retinal layers and may, in some cases, lead to a reduced
image quality. The development of Fourier-domain OCT
devices with higher imaging speed and higher resolution
than TD-OCT devices allows for an automatic segmentation
of the retinal layers. The RTVue-100 (Optovue, Fremont,
CA, USA) represents the first FD-OCT with an automatically integrated scan and analysis protocol for the measurement of the retinal ganglion cell complex (GCC). The GCC
consists of the retinal nerve fiber layer, the ganglion cell
layer, and the inner plexiform layer.
The aim of our study was to explore the diagnostic value
of ganglion cell complex, retinal nerve fiber layer, and optic
nerve head measurements using the FD-OCT RTVue-100 in
glaucoma patients, patients with ocular hypertension, and
normal subjects. The special focus of this investigation was
on the identification of the diagnostic ability of GCC
measurements to differentiate between glaucomatous
patients, patients with OHT, and healthy controls.

Materials and methods

Included in this observational case study were 211 eyes of 211
subjects. Ninety-three glaucoma patients, 58 ocular hypertensive patients, and 60 healthy normal subjects were enrolled.
All participants underwent slit-lamp examination, indirect dilated fundus examination with a 90D fundus lens,
Goldmann applanation tonometry, measurement of the
central corneal thickness (OCP, Heidelberg Engineering,
Heidelberg, Germany), visual field testing with a Humphrey
(Carl Zeiss Meditec, Jena, Germany) or Octopus (HAAGSTREIT, Wedel, Germany) visual field analyzer, and measurement of the ONH, RNFL, and GCC with Fourier-domain
OCT (RTVue-100, Optovue, Fremont, CA, USA).
All study participants had a best-corrected visual acuity of
20/60 or better, refraction error 5.0 diopters sphere and 2.0
diopters cylinder, no advanced lens opacities, no prior ocular
surgery or laser treatment, and no intraocular disease affecting
visual function or retinal structures (such as diabetic retinopathy or age-related macular degeneration). If both eyes of a
participant met the entry criteria, one eye was selected
randomly for this study.

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

Patients with primary open-angle, normal tension, and

pseudo exfoliation glaucoma with mild or moderate glaucomatous damage were included. Glaucoma was classified using
the enhanced visual field glaucoma staging system (GSS2)
according to Brusini [7], the GSS2 based on the three main
perimetric global indices (mean deviation, corrected pattern
standard deviation, and corrected loss variance) and plots on
an easy-to-use X-Y coordinate diagram. A further advantage
of the GSS2 is that it can be used with both the Humphrey
and the Octopus visual field system. Glaucoma patients with
a maximum GSS2 index stage 2 were included in this study.
GSS2 stage 2 is defined as a mean deviation 9 dB, and/or
pattern standard deviation 8%, and/or corrected loss
variance 64 dB2 [11]. Furthermore, both visual fields had
to have a false-positive error of less than 33%, a falsenegative error of less than 33%, and a fixation loss of less
than 20% to be defined as reliable.
Ocular hypertension was defined based on the presence
of an intraocular pressure of >21 mmHg with normal optic
nerve head appearance and normal visual field.
Normal subjects had to have at least one reliable normal
result on standard automated perimetry, normal disc appearance based on dilated fundus examination, and intraocular
pressure <21 mmHg.
The Fourier-domain OCT RTVue-100 was used for
GCC, RNFL, and ONH measurements. The RTVue-100
has an axial resolution of 5 m and acquires highresolution images with 26,000 axial scans per second.
The ganglion cell complex was measured using the scan
protocol GCC. The GCC was measured with one horizontal
line with a 7-mm scan length (467 axial scans per line,
centered 1 mm temporal to the fovea, and having a scan time
of 0.59 s) and 15 vertical lines with a 7-mm scan length (400
axial scans per line, 0.5-mm interval between two lines,
centered in the middle of the horizontal scan line).
For the measurement of optic nerve head parameters, a
three-dimensional image was initially achieved with the
scan protocol 3D Disc (101 lines with 512 axial scans,
66 mm volume centered in the middle of ONH, and a
scan time 2.2 s). Optic disc parameters were then measured
with the scan protocol ONH, using 12 radial line scans
3.4 mm in length (452 axial scans per line). The RNFL
thickness was measured with the same protocol using 13
concentric ring scans 1.3 to 4.9 mm in diameter (587 to 965
different axial scans per ring) centered in the optic disc.
The integrated signal strength index (SSI) was used to
control for image quality. The SSI represents the ratio
between the measuring beam and the reference beam. SSI
measurements of 50 and above were accepted (users
manual, Optovue Europe); scans with movement or
decentration artefacts were repeated.
The following GCC parameters were analyzed in this
study: average thickness, thickness in the superior and inferior

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

hemisphere, focal loss volume [FLV], and global loss volume

[GLV]). The FLV represents the total sum of significant GCC
loss in volume divided by the map area in percent; the GLV is
the sum of the pixels where the fractional deviation map value
is <0 divided by the total map area in percent.
Furthermore, the average thickness, thickness in the
superior and inferior hemisphere of the total central retinal
thickness, the thickness of the outer retina, and RNFL, and the
following ONH parameters: disc area, cup area and volume,
cup-to-disc ratio, RIM area, and volume were measured.
To describe the population, we computed means, standard
deviation, minimum, maximum, median, and quartiles for
quantitative variables, and they are displayed in box plots
where appropriate. For categorical variables, we computed
absolute and relative frequencies.
To compare parameters between patient groups, an
analysis of covariance (ANCOVA) with diagnosis group as
categorical factor and age as continuous covariate was used.
We report the estimated parameter mean at the mean age of the
total study population. For pairwise comparisons between
groups, the mean differences between group effects after
adjusting for age were determined. Mann-Whitney tests were
used to compare age between groups.
The discriminating ability of each parameter was described
by AUROC curves; given are point estimates and 95%
confidence intervals (CI). For the comparison of AUROC
curves, the method of DeLong et al. [8] was employed to test
for equality of the area under the curve (AUROC), which is
implemented in SAS PROC LOGISTIC.
For each RTVue-100 parameter, AUROC curves based
solely on the parameter and AUROC curves based on the
parameter in question, adjusting for gender, age, spherical
equivalent and/or for corneal thickness, were obtained.
Adjustment was performed by including gender, age, spherical equivalent, and corneal thickness into the logistic model
on which the AUROC curve was based. We further searched
for an optimal combination of parameters to predict glaucoma
by fitting a logistic regression model and applying forward
and forward stepwise variable selection.
Statistical analysis was performed using SPSS 17.0
(SPSS Inc. Chicago, IL, USA) and SAS 9.2 (SAS Institute
Inc., Cary, NC, USA). As mentioned above, the nature of
the present analysis is exploratory; no adjustment for
multiple testing was performed. In accordance with
standard statistical practice, p values equal to or less than
5% were described as significant, thereby verifying only the
local significance level.

Results
Ninety-three glaucoma patients (52 females, 41 males), 58
patients with ocular hypertension (29 females, 29 males),

1041

and 60 normal subjects (30 females, 30 males) were

enrolled in this study. OHT patients (mean age 58.5
12.8 years) and normal subjects (mean age 59.39.0 years)
were found to be significantly younger than glaucoma
patients (mean age 64.78.6 years).
The mean visual field loss was significantly higher in
the glaucoma group (-1.762.74 dB) compared to the
OHT (-0.351.86 dB) and the normal group (-0.051.42 dB).
The visual field test results were not statistically significantly
different between patients with ocular hypertension and
normal subjects.
Glaucoma patients further exhibited both thinner corneal
thickness (CCT) and lower intraocular pressure (IOP) than
OHT patients. Glaucoma patients and normal subjects had
similar CCT and IOP. Patients with ocular hypertension
showed a thicker CCT and higher IOP compared to healthy
subjects (Table 1).
The comparison of optic disc and RNFL parameters
between glaucoma and OHT or glaucoma and normal
subjects revealed significant differences for every parameter
excepting the disc area (Table 2). No differences were
found in any ONH and RNFL parameters of OHT patients
and normal subjects
Glaucoma patients showed a significantly thinner
GCC (average, superior, and inferior hemisphere) than
OHT patients or normal subjects (p<0.001). Focal loss
volume (FLV) and global loss volume (GLV) were
significantly higher in glaucoma than in OHT patients or
normal subjects (p<0.0001). The full retinal thickness was
also reduced in glaucoma patients compared with OHT
patients or healthy subjects. No differences in changes of
the outer retinal complex were found between the three
groups (Table 3).
Table 4 and Fig. 1 show the diagnostic ability of each
parameter using AUROC curves with adjustment for age to
distinguish between glaucoma and healthy eyes. After
adjusting for age, the parameter with the best discriminating
ability was cup-to-disc ratio (AUROC=0.848, 95% CI:
0.788-0.907), followed by RIM volume (AUROC=0.837,
95% CI: 0.775-0.899), RIM area (AUROC=0.834, 95%
CI: 0.771-0.896), RNFL average (AUROC=0.828, 95%
CI: 0.762-0.893), cup area (AUROC=0.823, 95% CI:
0.757-0.888) and RNFL inferior hemisphere (AUROC=
0.823, 95% CI: 0.756-0.890). Additional adjustment for
gender, corneal thickness, or spherical equivalent did not
change the sequence, the AUROC values were change only
marginally (all differences in AUROC were less than 0.01
and non-significant).
The best GCC parameters adjusted for age to
discriminate glaucoma from healthy subjects (Table 4)
included GCC global loss of volume (AUROC=0.805,
95% CI: 0.737-0.872) and GCC thickness in the inferior
hemisphere (AUROC = 0.802, 95% CI: 0.734-0.870).

1042

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

Table 1 Characteristics of the study population

Glaucoma

OHT

Normal

Number
Gender (F/M)
Age (MD SD)
Mean visual acuity (logMAR)
Mean visual acuity (logMAR)
at mean age
Spherical equivalent

93
52/41
64.78.6
0.090.09
0.08 (0.01)

58
29/29
58.512.8
0.080.14
0.08(0.02)

60
30/30
59.39.0
0.060.08
0.07 (0.01)

0.221.81

0.161.81

0.231.47

Mean spherical equivalent

at mean age
Mean CCT (m)
Mean CCT (m) at mean age
Mean IOP (mmHG)
Mean IOP (mmHG)
at mean age
MD visual field loss(dB)
Mean MD visual field loss(dB)
at mean age

0.43 (0.18)

0.04 (0.20)

0.38 (0.18)

532.135.7
533.2 (3.7)
16.06.3
15.8 (0.7)

559.833.7
558.9 (4.5)
19.94.9
20.1 (0.5)

540.431.1
539.7 (4.1)
15.43.4
15,6 (0.4)

1.762.74
1.66 (0.28)

0.351.86
0.45 (0.25)

0.051.42
0.11 (0.19)

Glaucoma vs.
normal (p)

Glaucoma vs.
OHT (p)

OHT vs.
normal (p)

0.0006

0.0053

0.9678

0.7109

0.8272

0.4781

0.0022

0.0873

0.2118

0.2479

<.0001

0.0016

0.7609

<.0001

<.0001

<.0001

0.0016

0.2774

F female, M male, MD mean deviation, SD standard deviation, CCT central corneal thickness, IOP intraocular pressure, p p value (obtained with
Mann-Whitney test or ANCOVA). Values in brackets following means at mean age are standard errors of estimates

Adjustment for gender, corneal thickness and spherical

equivalent yielded virtually the same results as adjusting
for age only.
The diagnostic power to distinguish between glaucoma
and healthy eyes of the best GCC parameter adjusted for
age (GCC global loss of volume, AUROC=0.805) was not
significantly different from the best OHN parameters (cupto-disc ratio, AUROC=0.848, p=0.2411; RIM volume,
AUROC=0.837, p=0.3581), and the best RNFL parameter
(RNFL average, AUROC=0.828, p=0.3647).
In the searching for an optimal model to predict
glaucoma, forward selection enabled the inclusion of cupto-disc ratio, RNFL average thickness, and GCC focal loss

of volume in addition to age. The area under the resulting

ROC curve was 0.878. When forward stepwise selection
was used, the resulting model included the same parameters, with the exception of RNFL average thickness. The
area under the corresponding AUROC curve was 0.871.
To distinguish patients with ocular hypertension from
normal subjects, the parameters with the highest AUROC
value were found for GCC focal loss of volume (AUROC=
0.593, 95% CI 0.554-0.754), followed by RNFL thickness
in the superior hemisphere (AUROC=0.548, 95% CI
0.441-0.654) and cup-to-disc ratio (AUROC=0.547, 95%
CI 0.439-0.655). Adjusting for gender or spherical equivalent did not increase the discrimination accuracy; adjusting

Table 2 Measurement of disc parameters and peripapillary retinal nerve fiber layer in glaucoma patients, OHT patients, and normal subjects

Disc area (mm2)

Cup area (mm2)
Cup volume (mm3)
Cup-to-disc ratio
RIM area (mm2)
RIM volume (mm3)
RNFL average (mm)
RNFL superior (mm)
RNFL inferior (mm)

Glaucoma mean
at mean age (SEE)

OHT mean
at mean age (SEE)

Normal mean
at mean age (SEE)

Glaucoma vs.
normal (p)

Glaucoma vs.
OHT (p)

OHT vs.
normal (p)

2.06
1.33
0.43
0.64
0.74
0.08
89.9
91.2
88.8

2.09 (0.06)
0.84 (0.07)
0.24 (0.04)
0.39 (0.03)
1.24 (0.06)
0.17 (0.02)
105.5 (1.5)
106.3 (1.6)
104.7 (1.7)

1.99 (0.04)
0.76 (0.06)
0.20 (0.03)
0.37 (0.02)
1.23 (0.04)
0.16 (0.01)
105.1 (1.3)
106.7 (1.4)
103.4 (1.4)

0.2044
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001

0.7958
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001

0.1969
0.3529
0.4388
0.5550
0.8614
0.7152
0.8544
0.8225
0.5152

(0.04)
(0.06)
(0.03)
(0.02)
(0.05)
(0.01)
(1.5)
(1.7)
(1.5)

SEE standard error of estimate, RNFL retinal nerve fiber layer, SSI signal strength index, p p value (obtained using ANCOVA)

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

1043

Table 3 Thickness of ganglion cell complex, outer retinal complex, and full retina layer in glaucoma, or OHT patients, and normal subjects

GCC average (mm)

GCC superior (mm)
GCC inferior (mm)
GCC FLV (%)
GCC GLV (%)
OUT ret avg (mm)
OUT ret sup (mm)
OUT ret inf (mm)
FULL ret avg (mm)
FULL ret sup (mm)
FULL ret inf (mm)

Glaucoma mean
at mean age (SEE)

OHT mean
at mean age (SEE)

Normal mean
at mean age (SEE)

Glaucoma vs.
normal (p)

Glaucoma vs.
OHT (p)

OHT vs.
normal (p)

86.7 (1.0)
87.4 (1.1)
85.9 (1.1)
2.8 (0.3)
11.1 (0.9)
170.9 (0.7)
172.2 (0.8)
50.8 (7.6)
257.6 (1.3)
258.4 (1.7)
253.9 (2.1)

94.6 (1.0)
94.0 (1.1)
95.2 (1.1)
1.3 (0.3)
5.1 (0.7)
172.0 (1.0)
173.2 (1.0)
46.6 (9.1)
265.9 (1.4)
266.7 (1.6)
263.1 (2.4)

94.7 (0.9)
94.0 (0.9)
95.4 (0.9)
0.8 (0.1)
4.4 (0.4)
172.3 (1.0)
173.9 (1.1)
66.1 (12.2)
266.7 (1.5)
265.9 (2.3)
262.3 (2.6)

<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
0.2633
0.4183
0.4201
<0.0001
0.0104
0.0125

<0.0001
<0.0001
<0.0001
0.0004
<0.0001
0.3680
0.1959
0.5472
<0.0001
0.0005
0.0047

0.8358
0.9584
0.8480
0.0574
0.3331
0.8047
0.6286
0.8139
0.6909
0.7700
0.8249

SEE standard error of estimated, GCC ganglion cell complex, GCC superior GCC in the superior hemisphere, GCC inferior GCC in the inferior
hemisphere, GCC FLV focal loss volume of GCC, GCC GLV global loss volume of GCC, OUT ret avg average of the outer retinal complex, OUT
ret sup outer retinal complex in the superior hemisphere, OUT ret inf outer retinal complex in inferior hemisphere, FULL ret avg average of the
full central retinal thickness, FULL ret sup full central retinal thickness in the superior hemisphere, FULL ret inf full central retinal thickness in the
inferior hemisphere, p p value (obtained using the ANCOVA)

for central corneal thickness, however, increased the

diagnostic ability. When adjusting for age and central
corneal thickness, the highest discrimination accuracy was
found for cup-to-disc ratio (AUROC=0.693, 95% CI
0.595-0.791), RNFL thickness in the superior hemisphere

(AUROC=0.693, 95% CI 0.595-0.791) and GCC global

loss of volume (AUROC=0.693, 95% CI 0.595-0.790).
In the search for an optimal model to predict ocular
hypertension, both forward selection and stepwise selection
lead to the inclusion of central corneal thickness in addition

1.0

Table 4 AUROC curves to distinguish between glaucomatous and

healthy eyes, adjusted for age
AUROC

95% Confidence interval

Cup-to-disc ratio
RIM volume (mm3)
RIM area (mm2)
RNFL average (mm)
Cup area (mm2)
RNFL inferior (mm)
RNFL superior (mm)

0.848
0.837
0.834
0.828
0.823
0.823
0.805

0.7880.907
0.7750.899
0.7710.896
0.7620.893
0.7570.889
0.7560.890
0.7360.874

GCC GLV (%)

GCC inferior (mm)
GCC average (mm)
Cup volume (mm3)
GCC FLV (%)
FULL ret average (mm)
FULL ret superior (mm)
FULL ret inferior (mm)
GCC superior (mm)
OUT ret superior (mm)
OUT ret average (mm)
OUT ret inferior (mm)
Disc area (mm2)

0.805
0.802
0.789
0.786
0.779
0.766
0.743
0.743
0.737
0.681
0.677
0.673
0.662

0.7370.872
0.7340.870
0.7190.860
0.7130.860
0.7070.852
0.6900.843
0.6620.824
0.6630.823
0.6590.815
0.5950.767
0.5910.763
0.5870.759
0.5740.749

Sensitivity

0.8

0.6

0.4

0.2

0.0
0.0

0.2

0.4

0.6

0.8

1.0

1 - Specificity
ROC Curve (Area)

RIM Volume (0.8370)

Cup Disc Ratio (0.8476)
RNFL Average (0.8277)
GCC Global loss volume (0.8045)
reference line

Fig. 1 AUROC curves for the best ONH, RNFL, and GCC
parameter, adjusted for age

1044

to age. The area under the corresponding AUROC curve in

both models was 0.687.

Discussion
The comparison of glaucoma patients with normal and
ocular hypertensive patients demonstrated significant differences for the ONH, RNFL, GCC parameters, and central
retinal thickness. When the retina in the ganglion cell
complex and outer retinal complex was separated, only the
GCC layer showed a significant reduction in thickness.
For the investigation of the diagnostic ability, AUROC
curves adjusted for age were used with a view to the agerelated reduction in retinal layers and the absence of agematched groups. The parameter with the best discriminating
ability adjusted for age was cup-to-disc ratio (AUROC=
0.848), closely followed by RIM volume (AUROC=
0.837), RIM area (AUROC=0.834), and RNFL average
(AUROC=0.828). The best GCC parameter for glaucoma
discrimination (global loss volume of GCC, AUROC=
0.805) had a slightly lower diagnostic ability than that of
RNFL thickness and OHN parameters, although this
difference was not statistically significant.
A more marked reduction in central retinal thickness found
in glaucoma patients compared with normal subjects using the
established time-domain OCT has been reported by a number
of authors [25]. Furthermore, the significantly higher
diagnostic ability for the measurement of RNFL parameters
compared to macular thickness measurement is well documented [14].
The segmentation of retinal layers using time-domain
OCT (Stratus OCT) is limited by the lower resolution
compared to Fourier-domain OCT. An automatic segmentation of retinal layers is not integrated into TD-OCT
devices. Ishikawa et al. 2005 [9] and Tan et al. 2007 [6]
developed a complex segmentation algorithm to identify
the boundaries between retinal layers for exported Stratus
OCT macular images. The best discrimination between
glaucoma patients (with or without visual field defects) and
normal subjects was found for the measurement of the
nerve fiber layer, ganglion cell layer, and inner plexiform
layer in the macular area. The combination of those three
layers with the inner retinal layer complex (now called
ganglion cell complex, GCC) had the highest repeatability
and best discrimination power compared with the measurement of macular retinal thickness for glaucoma diagnosis.
A good reproducibility of ONH and RNFL measurements
with FDT-OCT RTVue-100 has been reported by different
authors [10, 11]. Tan et al. [12] demonstrated a slightly
higher diagnostic ability for GCC parameters comparing
preperimetric and perimetric glaucoma patients with normal
subjects. In the perimetric glaucoma group, the authors

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

found the highest diagnostic accuracy for GCC-FLV and

GCC-GLV (AUROC 0.92), and in the preperimetric glaucoma group this was found for both GCC-GLV (AUROC
0.799) and the GCC average (AUROC 0.789). Seong et al.
[13] also demonstrated no significant differences between
AUROCs (0.945 for GCC, 0.973 for RNFL average) in
patients with normal tension glaucoma and healthy subjects.
In the present study, patients with early glaucoma damage
showed different GCC patterns as, e.g., point-shaped or
diffuse defects. This might offer an explanation for the
demonstration of GLV and FLV as the GCC parameters with
the best diagnostic accuracy.
No significant differences were observed in the RNFL,
ONH, and GCC parameters in the comparison of patients
with ocular hypertension and healthy subjects. The diagnostic ability (AUROCs) of the ONH, RNFL, and GCC
parameters adjusted for age and central corneal thickness
was moderate (cup-to-disc ratio, AUROC=0.693; RNFL
thickness in the superior hemisphere, AUROC=0.693;
GCC global loss of volume, AUROC=0.693). Results of
studies comparing FD-OCT measurements obtained in
OHT patients and healthy subjects have not been published
to date. Using the TD-OCT (Stratus-OCT, Carl Zeiss
Meditec, Jena, Germany), Anton et al. [14] found a thinner
RIM (volume and area), larger cup/disc area ratio, and
reduced RNFL comparing 95 OHT patients with 55 normal
subjects.
Limitations of the present study were the small number
of subjects in the OHT and the control group. Due to the
fact that we used a selected study population, our results
may not be representative of German/Caucasian population.
This is a problem with most diagnostic studies. Because our
patient groups (glaucoma patients, OHT patients, and
normal subjects) did not represent age-matched groups,
we used an analysis of covariance (ANCOVA) with age as
continuous covariate to compare ONH, RNFL, and GCC
parameter values and age-adjusted AUROC curves were
used for the investigation of diagnostic ability.
In accordance with the manufactures users manual, all
measurements were taken without pupil dilation. Results
reported by other authors [15, 16] and our own findings
(poster at ARVO 2010) showed no influence of pupil
dilation on RNFL measurements. Limitations of the automatic measurements without pupil dilation include constricted pupils, dry eye syndrome, as well as a high rate of
blink and corneal or lens opacities.
In conclusion, measurements of GCC without pupil
dilation using the Fourier-domain OCT RTVue-100 had a
slightly lower diagnostic power than RNFL or ONH
measurements for the discrimination between glaucoma
patients and normal subjects. The differences did not,
however, reach statistical significance. Follow-up studies
on patients with ocular hypertension and early glaucoma

Graefes Arch Clin Exp Ophthalmol (2011) 249:10391045

patients are required to determine the best parameter for the

diagnosis of early glaucoma and progression of glaucoma.
References
1. Anton A, Moreno-Montaes J, Blzquez F, Alvarez A, Martn B,
Molina B (2007) Usefulness of optical coherence tomography
parameters of the optic disc and the retinal nerve fiber layer to
differentiate glaucomatous, ocular hypertensive, and normal eyes.
BJ Glaucoma 16(1):18
2. Medeiros FA, Zangwill LM, Bowd C, Vessani RM, Susanna R Jr,
Weinreb RN (2005) Evaluation of retinal nerve fiber layer, optic
nerve head, and macular thickness measurements for glaucoma
detection using optical coherence tomography. Am J Ophthalmol
139(1):4455
3. Guedes V, Schuman JS, Hertzmark E, Wollstein G, Correnti A,
Mancini R, Lederer D, Voskanian S, Velazquez L, Pakter HM,
Pedut-Kloizman T, Fujimoto JG, Mattox C (2003) Optical
coherence tomography measurement of macular and nerve fiber
layer thickness in normal and glaucomatous human eyes.
Ophthalmology 110(1):177189
4. Wollstein G, Ishikawa H, Wang J, Beaton SA, Schuman JS (2005)
Comparison of three optical coherence tomography scanning areas
for detection of glaucomatous damage. Am J Ophthalmol 139
(1):3943
5. Leung CK, Chan WM, Yung WH, Ng AC, Woo J, Tsang MK, Tse
RK (2005) Comparison of macular and peripapillary measurements
for the detection of glaucoma: an optical coherence tomography
study. Ophthalmology 112(3):391400
6. Tan O, Li G, Lu AT, Varma R, Huang D, Advanced Imaging for
Glaucoma Study Group (2008) Mapping of macular substructures
with optical coherence tomography for glaucoma diagnosis.
Ophthalmology 115(6):949956
7. Brusini P, Filacorda S (2006) Enhanced Glaucoma Staging
System (GSS 2) for classifying functional damage in glaucoma.
J Glaucoma 15(1):4046

1045
8. DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing
the Areas under Two or More Correlated Receiver Operating
Characteristic Curves: A Nonparametric Approach. Biometrics
44:837845
9. Ishikawa H, Stein DM, Wollstein G, Beaton S, Fujimoto JG,
Schuman JS (2005) Macular segmentation with optical coherence
tomography. Invest Ophthalmol Vis Sci 46(6):20122017
10. Gonzlez-Garca AO, Vizzeri G, Bowd C, Medeiros FA, Zangwill
LM, Weinreb RN (2009) Reproducibility of RTVue retinal nerve
fiber layer thickness and optic disc measurements and agreement
with Stratus optical coherence tomography measurements. Am J
Ophthalmol 1067-1074, 1074.e1
11. Menke MN, Knecht P, Sturm V, Dabov S, Funk J (2008)
Reproducibility of nerve fiber layer thickness measurements using
3D Fourier-domain OCT. Invest Ophthalmol Vis Sci 49
(12):53865391
12. Tan O, Chopra V, Lu AT, Schuman JS, Ishikawa H, Wollstein G,
Varma R, Huang D (2009) Detection of macular ganglion cell loss
in glaucoma by Fourier-domain optical coherence tomography.
Ophthalmology 116(12):23052314, e1-2
13. Seong M, Sung KR, Choi EH, Kang SY, Cho JW, Um TW, Kim YJ,
Park SB, Hong HE, Kook M (2009) Diagnostic Comparison between
Macular and Peripapillary Retinal Nerve Fiber Layer Measurements
by Spectral Domain Optical Coherence Tomography in NormalTension Glaucoma. Invest Ophthalmol Vis Sci 51(3):14461452
14. Anton A, Moreno-Montaes J, Blzquez F, Alvarez A, Martn B,
Molina B (2007) Usefulness of optical coherence tomography
parameters of the optic disc and the retinal nerve fiber layer to
differentiate glaucomatous, ocular hypertensive, and normal eyes.
J Glaucoma 16(1):18
15. Garas A, Vargha P, Holl G (2010) Reproducibility of retinal
nerve fiber layer and macular thickness measurement with the
RTVue-100 optical coherence tomograph. Ophthalmology 117
(4):738746
16. Savini G, Carbonelli M, Parisi V, Barboni P (2009) Effect of pupil
dilation on retinal nerve fibre layer thickness measurements and
their repeatability with Cirrus HD-OCT. Eye (Lond) 24(9):1503
1508