[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.161.130.145. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Figure.
Retinal nerve fiber layer (RNFL) thickness determined by both Stratus OCT (Carl Zeiss Meditec, Inc, Dublin, California) and Cirrus HD-OCT (Carl Zeiss Meditec, Inc) in healthy subjects and in patients at 2 different glaucoma stages (group 1, early; group 2, moderate to advanced). OCT indicates optical coherence tomography; A, average; T, temporal quadrant; S, superior quadrant; N, nasal quadrant; I, inferior quadrant.

Retinal nerve fiber layer (RNFL) thickness determined by both Stratus OCT (Carl Zeiss Meditec, Inc, Dublin, California) and Cirrus HD-OCT (Carl Zeiss Meditec, Inc) in healthy subjects and in patients at 2 different glaucoma stages (group 1, early; group 2, moderate to advanced). OCT indicates optical coherence tomography; A, average; T, temporal quadrant; S, superior quadrant; N, nasal quadrant; I, inferior quadrant.

Table 1. 
Demographics of Healthy Subjects and Patients With Glaucoma
Demographics of Healthy Subjects and Patients With Glaucoma
Table 2. 
AUCs of RNFL Thicknesses Measured by Stratus OCT and Cirrus HD-OCT and Sensitivities at Similar Specificity Levels (80%-90%)a
AUCs of RNFL Thicknesses Measured by Stratus OCT and Cirrus HD-OCT and Sensitivities at Similar Specificity Levels (80%-90%)a
Table 3. 
Comparison of AUCs of RNFL Thicknesses Measured by Stratus OCT and Cirrus HD-OCT at Different Glaucoma Stagesa
Comparison of AUCs of RNFL Thicknesses Measured by Stratus OCT and Cirrus HD-OCT at Different Glaucoma Stagesa
Table 4. 
LRs and 95% CIs Using Normative Classifications With Either OCT Modality
LRs and 95% CIs Using Normative Classifications With Either OCT Modality
1.
Burgansky-Eliash  ZWollstein  GChu  T  et al.  Optical coherence tomography machine learning classifiers for glaucoma detection: a preliminary study. Invest Ophthalmol Vis Sci 2005;46 (11) 4147- 4152
PubMedArticle
2.
Huang  MLChen  HY Development and comparison of automated classifiers for glaucoma diagnosis using Stratus optical coherence tomography. Invest Ophthalmol Vis Sci 2005;46 (11) 4121- 4129
PubMedArticle
3.
Kanamori  ANagai-Kusuhara  AEscano  MFMaeda  HNakamura  MNegi  A Comparison of confocal scanning laser ophthalmoscopy, scanning laser polarimetry and optical coherence tomography to discriminate ocular hypertensionand glaucoma at an early stage. Graefes Arch Clin Exp Ophthalmol 2006;244 (1) 58- 68
PubMedArticle
4.
Lalezary  MMedeiros  FAWeinreb  RN  et al.  Baseline optical coherence tomography predicts the development of glaucomatous change in glaucoma suspects. Am J Ophthalmol 2006;142 (4) 576- 582
PubMedArticle
5.
Leung  CKChan  WMYung  WH  et al.  Comparison of macular and peripapillary measurements for the detection of glaucoma: an optical coherence tomography study. Ophthalmology 2005;112 (3) 391- 400
PubMedArticle
6.
Manassakorn  ANouri-Mahdavi  KCaprioli  J Comparison of retinal nerve fiber layer thickness and optic disk algorithms with optical coherence tomography to detect glaucoma. Am J Ophthalmol 2006;141 (1) 105- 115
PubMedArticle
7.
Medeiros  FAZangwill  LMBowd  CVessani  RMSusanna  R  JrWeinreb  RN Evaluation of retinal nerve fiber layer, optic nerve head, and macular thickness measurements for glaucoma detection using optical coherence tomography. Am J Ophthalmol 2005;139 (1) 44- 55
PubMedArticle
8.
Naithani  PSihota  RSony  P  et al.  Evaluation of optical coherence tomography and heidelberg retinal tomography parameters in detecting early and moderate glaucoma. Invest Ophthalmol Vis Sci 2007;48 (7) 3138- 3145
PubMedArticle
9.
Nouri-Mahdavi  KNikkhou  KHoffman  DCLaw  SKCaprioli  J Detection of early glaucoma with optical coherence tomography (StratusOCT). J Glaucoma 2008;17 (3) 183- 188
PubMedArticle
10.
Parikh  RSParikh  SSekhar  GC  et al.  Diagnostic capability of optical coherence tomography (Stratus OCT 3) in early glaucoma. Ophthalmology 2007;114 (12) 2238- 2243
PubMedArticle
11.
Sihota  RSony  PGupta  VDada  TSingh  R Comparing glaucomatous optic neuropathy in primary open angle and chronic primary angle closure glaucoma eyes by optical coherence tomography. Ophthalmic Physiol Opt 2005;25 (5) 408- 415
PubMedArticle
12.
Quigley  HAKatz  JDerick  RJGilbert  DSommer  A An evaluation of optic disc and nerve fiber layer examinations in monitoring progression of early glaucoma damage. Ophthalmology 1992;99 (1) 19- 28
PubMedArticle
13.
Sommer  AKatz  JQuigley  HA  et al.  Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol 1991;109 (1) 77- 83
PubMedArticle
14.
Zeyen  TGCaprioli  J Progression of disc and field damage in early glaucoma. Arch Ophthalmol 1993;111 (1) 62- 65
PubMedArticle
15.
Paunescu  LASchuman  JSPrice  LL  et al.  Reproducibility of nerve fiber thickness, macular thickness, and optic nerve head measurements using StratusOCT. Invest Ophthalmol Vis Sci 2004;45 (6) 1716- 1724
PubMedArticle
16.
Hodapp  EParrish  RKAnderson  DR Clinical Decisions in Glaucoma.  St Louis, MO Mosby1993;53
17.
Mills  RPBudenz  DLLee  PP  et al.  Categorizing the stage of glaucoma from pre-diagnosis to end-stage disease. Am J Ophthalmol 2006;141 (1) 24- 30
PubMedArticle
18.
DeLong  ERDeLong  DMClarke-Pearson  DL Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44 (3) 837- 845
PubMedArticle
19.
Jaeschke  RGuyatt  GHSackett  DL Users' guides to the medical literature, III: how to use an article about a diagnostic test. B: what are the results and will they help me in caring for my patients? the Evidence-Based Medicine Working Group. JAMA 1994;271 (9) 703- 707
PubMedArticle
20.
Budenz  DLMichael  AChang  RT McSoley  JKatz  J Sensitivity and specificity of the Stratus OCT for perimetric glaucoma. Ophthalmology 2005;112 (1) 3- 9
PubMedArticle
21.
Bowd  CZangwill  LMBerry  CC  et al.  Detecting early glaucoma by assessment of retinal nerve fiber layer thickness and visual function. Invest Ophthalmol Vis Sci 2001;42 (9) 1993- 2003
PubMed
22.
Budenz  DLAnderson  DRVarma  R  et al.  Determinants of normal retinal nerve fiber layer thickness measured by Stratus OCT. Ophthalmology 2007;114 (6) 1046- 1052
PubMedArticle
23.
Gurses-Ozden  RDurbin  MCallan  T  et al.  Distribution of retinal nerve fiber layer thickness using Cirrus HD-OCT spectral domain technology.  Paper presented at: ARVO 2008 Annual Meeting April 30, 2008 Fort Lauderdale, FLE-abstract 4632
Clinical Sciences
December 14, 2009

Comparison of Glaucoma Diagnostic Capabilities of Cirrus HD and Stratus Optical Coherence Tomography

Author Affiliations

Author Affiliations: Department of Ophthalmology, College of Medicine, University of Ulsan, Asan Medical Center, Seoul, Korea.

Arch Ophthalmol. 2009;127(12):1603-1609. doi:10.1001/archophthalmol.2009.296
Abstract

Objective  To compare the glaucoma diagnostic capabilities offered by Stratus and Cirrus spectral-domain optical coherence tomography (OCT).

Methods  One hundred subjects with glaucoma and 74 healthy subjects were tested by Stratus and Cirrus OCT. Areas under the receiver operating characteristic curves (AUCs) of average, 4-quadrant, and 12-sector retinal nerve fiber layer thicknesses and sensitivities at fixed specificities (80% and 90%) were compared when the 2 OCT modalities were used to evaluate patients with early or moderate to advanced glaucoma. Likelihood ratios using normative classifications were reported.

Results  Overall, both OCT instruments showed similar glaucoma discrimination capability in average retinal nerve fiber layer thickness (AUC, 0.953 [Cirrus] vs 0.934 [Stratus]; P = .15). Cirrus OCT displayed significantly higher AUCs in the average, inferior, temporal, and superior quadrants and 7-o’clock measurements in early stages of glaucoma. The between–OCT instrument AUCs did not differ significantly in moderate to advanced stages. Abnormal results for both OCT instruments, after comparison with their normative databases, were associated with high likelihood ratios.

Conclusions  In our series, the Cirrus OCT showed better glaucoma discrimination capability than Stratus OCT in early stages of glaucoma. Our findings suggest that spectral-domain technology of OCT may offer an improved capability of early-stage glaucoma detection.

The third-generation Stratus OCT (Carl Zeiss Meditec, Inc, Dublin, California) has shown good glaucoma diagnosis capabilities in many studies.111 Glaucomatous structural damage may precede functional loss associated with glaucoma progression.1214 Thus, quantitative and objective assessment of structural damage by optical coherence tomography (OCT) might enable detection of anatomical changes before occurrence of irreversible functional impairment.

The Cirrus HD-OCT (Carl Zeiss Meditec, Inc) uses spectral-domain technology and has recently become commercially available. This technology offers higher axial resolution and scan speed than the conventional time-domain Stratus OCT. The Cirrus HD-OCT has an axial resolution of 5 μm and a scan speed of 27 000 A-scans per second, whereas the figures for the conventional time-domain Stratus OCT instrument are 8 to 10 μm and 400 A-scans per second. The enhanced axial resolution of Cirrus HD-OCT might thus provide more detailed segmentation of the retinal nerve fiber layer (RNFL), thus leading to better data acquisition. Furthermore, the higher sampling rate of the new OCT instrument allows more data to be collected in shorter scan times. Therefore, we hypothesized that the Cirrus HD-OCT might be better than or at least comparable with the Stratus OCT when used to diagnose glaucomatous changes. The purpose of this study was to test this hypothesis by comparing the diagnostic capabilities of the 2 technologies and to compare their abilities to differentiate between healthy eyes and eyes with both early and moderate to advanced glaucomatous visual field (VF) loss.

METHODS
SUBJECTS

All study subjects were recruited prospectively, in a consecutive manner, from our glaucoma clinic and were examined between March 2008 and December 2008 at the Asan Medical Center, Seoul, Korea. At initial evaluation, all subjects underwent a complete ophthalmologic examination including medical, ocular, and family history; visual acuity testing; the Humphrey field analyzer (HFA) Swedish Interactive Threshold Algorithm 24-2 test (Carl Zeiss Meditec, Inc); multiple intraocular pressure (IOP) measurements using Goldmann applanation tonometry; stereoscopic optic nerve photography; and Stratus OCT and Cirrus HD-OCT. All patients with glaucoma had extensive experience with HFA testing. To minimize the learning effect, only the last 2 HFA test results were used for analysis. For inclusion in the study, all participants had to meet the following criteria: best-corrected visual acuity of 20/30 or better, with a spherical equivalent within ±5 diopters and a cylinder correction within +3 diopters; presence of a normal anterior chamber and open angle on slitlamp and gonioscopic examinations; and reliable HFA test results with a false-positive error less than 15%, a false-negative error less than 15%, and a fixation loss less than 20%. Subjects with any other ophthalmic disease that could result in HFA defects or with histories of diabetes mellitus were excluded. One eye was randomly selected if both eyes were found to be eligible. Age-matched, healthy eyes formed the control group. The control group consisted of staff, their family, spouses of patients, and volunteers from the eye clinic and hospital. The control group had no history of ocular symptoms or disease and no intraocular incisional or laser surgery. These control eyes had an IOP lower than 22 mm Hg, with no history of IOP elevation, and were normal by VF examination. Glaucomatous eyes were defined as those with glaucomatous VF defects confirmed by at least 2 reliable VF examination results and by the presence of a glaucomatous optic disc that showed increased cupping (vertical cup-disc ratio of >0.6), a difference in vertical cup-disc ratio of more than 0.2 between eyes, diffuse or focal neural rim thinning, hemorrhage, and nerve fiber layer defects. Eyes with glaucomatous VF defects were defined as those with a cluster of 3 points with probabilities of less than 5% on the pattern deviation map in at least 1 hemifield, including at least 1 point with a probability of less than 1%, or a cluster of 2 points with a probability of less than 1% and a glaucoma hemifield test result outside 99% of age-specific normal limits or a pattern standard deviation outside 95% of normal limits.

All participants gave written informed consent before enrollment. All procedures conformed to the Declaration of Helsinki and the study was approved by the institutional review board of the Asan Medical Center at the University of Ulsan, Seoul.

OPTICAL COHERENCE TOMOGRAPHY

The basic principles and technical characteristics of the Stratus OCT have been described elsewhere.15 We excluded all poor-quality scans, defined as those with signal strength less than 7; overt misalignment of the surface detection algorithm on at least 15% of consecutive A-scans or 20% of cumulative A-scans; or overt displacement of the measurement circle, as assessed subjectively.

With the Cirrus HD-OCT (software version 3.0.0.50), an optic disc cube obtains a 3-dimensional data set composed of 200 A-scans, from each of 200 B-scans, that cover an area of 6 mm2 centered on the optic disc. After creating an RNFL thickness map from this data set, the software automatically determines the disc center and then extracts a circumpapillary circle (1.73 mm in radius) from the cube data set for RNFL thickness measurement. We defined, and excluded, poor-quality scans, as described earlier. In addition, we excluded images where horizontal eye motion was observed within the measurement circle. Pharmacologic dilation was performed if the pupil was small.

All images were acquired by a single well-trained operator (K.R.K.) during the same visit. Both OCT technologies provide RNFL thickness maps for 4 quadrants (superior, inferior, nasal, and temporal) and 12 clock hours and include classifications derived by use of an internal normative database. For both instruments, 4 normative classifications were used. The 95th to 100th percentiles were hypernormal (shown in white on thickness maps); the fifth to 95th percentiles were normal (green); the first to fifth percentiles were borderline (yellow); and the less than first percentile was abnormal (red). White and green were regarded as normal in our analyses. All OCT data were aligned according to the orientation of the right eye. Thus, clock hour 9 of the circumpapillary scan represented the temporal optic disc side for both eyes.

STATISTICAL ANALYSIS

The Shapiro-Wilk test was used to test the distribution of numerical data. Normally distributed data were compared between healthy subjects and patients with glaucoma using unpaired t tests. To compare categorical data, the χ2 test was used. To test RNFL thickness discriminatory capability between healthy and glaucomatous eyes, the areas under the receiver operating characteristic curves (AUCs), including the overall average, the 4-quadrant, and average 12–clock hour RNFL thickness measurements, were calculated and compared. Sensitivities at fixed specificities of 80% and 90% were also calculated from the receiver operating characteristic curves. To compare between-instrument glaucoma discrimination capabilities at different stages of glaucoma, we divided subjects with glaucoma into 2 groups, an early group (group 1) and a moderate to advanced group (group 2), according to the Hodapp-Anderson-Parrish grading scale of VF severity. This staging system is described in detail elsewhere.16,17 By Hodapp-Anderson-Parrish criteria, 52 eyes of 52 patients had early VF defects (group 1) whereas 48 eyes of 48 patients had moderate to advanced VF defects (group 2). Between-instrument AUCs were compared in each group. Finally, the DeLong method was used to evaluate statistical differences between AUCs yielded by the 2 OCT technologies.18

Likelihood ratios (LRs) for glaucomatous change detection using a normative RNFL thickness classification were calculated. The use of LRs to predict posttest disease probability has been suggested by Jaeschke and colleagues.19 In their scheme, LRs higher than 10 or lower than 0.1 are associated with large effects on posttest probabilities; LRs from 5 to 10 or from 0.1 to 0.2, with moderate effects; LRs from 2 to 5 or from 0.2 to 0.5, with small effects; and LRs closer to 1 are insignificant. Likelihood ratios for each category (fifth to 95th percentile, green [within normal limits]; first to fifth percentile, yellow [borderline]; and <first percentile, red [abnormal classification]) were calculated separately. The 95% confidence intervals for LRs were also calculated. Statistical analysis was performed using SPSS version 15.0 (SPSS Inc, Chicago, Illinois) and MedCalc version 9.6 (MedCalc, Mariakerke, Belgium).

RESULTS

Of the 184 eyes that met our inclusion criteria, 8 eyes were excluded because of the unacceptable image quality provided by Cirrus HD-OCT. Six of these eyes had signal strengths less than 7, and 2 had horizontal eye motion within the measurement circle. Four eyes' images with Stratus OCT were excluded owing to poor quality. Two eyes were excluded because of poor quality based on both Cirrus HD-OCT and Stratus OCT images. The final study sample included 174 eyes from 174 subjects (100 patients with glaucoma and 74 healthy control subjects). Ninety-three were women, 81 were men, and all were Asian. The baseline demographic characteristics of healthy subjects and subjects with glaucoma are shown in Table 1.

RNFL THICKNESS COMPARISONS

In both Stratus OCT and Cirrus HD-OCT RNFL thickness measurements, all sectors showed significant differences between healthy subjects and group 1 patients except the 9-o’clock and 3-o’clock sectors. When healthy subjects and group 2 patients were compared, Stratus OCT measurements showed significant differences in all sectors except the 9-o’clock sector, while Cirrus HD-OCT measurements showed significant differences in all sectors except the 9-o’clock and 3-o’clock sectors (Figure).

AUC AND SENSITIVITY AT A FIXED SPECIFICITY FOR ALL GLAUCOMATOUS EYES

Overall, average RNFL thickness measurement by both Stratus OCT and Cirrus HD-OCT demonstrated good glaucoma discrimination capability, and there was no significant between–OCT instrument difference (AUCs, 0.934 vs 0.953; P = .15). However, for both OCT modalities, inferior-quadrant RNFL thickness measurements showed the highest AUCs and Cirrus HD-OCT was significantly better than Stratus OCT (0.935 [Stratus] vs 0.963 [Cirrus]; P = .04). Stratus OCT showed better glaucoma discrimination capability than Cirrus HD-OCT in the 3-o’clock RNFL thickness measurement (Table 2). Cirrus HD-OCT demonstrated higher sensitivities in more parametric measurements, at specificities of 90% or more, than did Stratus OCT (12 parameters vs 3 parameters, 2 parameters were measured with equal sensitivities by either OCT modality). At specificities of 80% or more, the number of parameters measured with higher sensitivity was similar for both OCT instruments (8 parameters vs 7 parameters, 2 parameters were measured with equal sensitivities by either OCT modality).

AUCs AT DIFFERENT GLAUCOMA STAGES

Cirrus HD-OCT demonstrated statistically better glaucoma discrimination capability in average, superior, inferior, and temporal quadrants and in the 7-o’clock sectors of group 1 patients as assessed by AUC calculations. In group 2 patients, the 2 OCT modalities showed no significant discrimination differences in most of the parameters while Cirrus HD-OCT showed statistically worse glaucoma discrimination capability than Stratus OCT in the nasal-quadrant and 3-o’clock measurements (Table 3).

LIKELIHOOD RATIOS

Table 4 shows the LRs after comparison with the instruments' normative database. Results outside normal limits for both OCT instruments using average and 4-quadrant RNFL thicknesses were associated with large effects on the posttest probability of glaucoma (infinite LRs). However, within–normal limits results were associated with small effects on the posttest probability of glaucoma for both OCT instruments (LRs, 0.35-0.92).

COMMENT
OVERALL DISCRIMINATION CAPABILITY

In this study (average mean deviation of −6.67 dB), both Stratus OCT and Cirrus HD-OCT demonstrated good glaucoma discrimination capabilities. Parikh and coworkers10 reported that the Stratus OCT AUC calculated using average RNFL thickness was 0.75 (95% confidence interval, 0.68-0.83) in patients with early-stage glaucoma (average mean deviation of −3.57 dB). Budenz and colleagues20 reported a higher Stratus OCT AUC (0.966; 95% confidence interval, 0.933-0.999), calculated with reference to mean RNFL thickness, in subjects with more advanced glaucoma (average mean deviation of −8.4 dB). Although direct comparison of AUC values between different reports is problematic, our calculations of a Stratus OCT AUC of 0.934 and a Cirrus HD-OCT AUC of 0.953, with reference to average RNFL thickness, showed that both OCT modalities offered good glaucoma diagnostic capabilities.

Our results indicated that inferior-quadrant RNFL thickness measurements by either Stratus OCT or Cirrus HD-OCT yielded the highest AUCs, in agreement with previous reports on Stratus OCT performance.10,20 This finding is also consistent with work using earlier OCT versions.3,21 Our finding may be explained by the generally acknowledged fact that VF defects tend to occur more frequently in the superior hemifield at early stages of glaucoma.

Overall, Cirrus HD-OCT showed better glaucoma discrimination ability in most of the parameters and a significant difference in the inferior quadrant. Stratus OCT showed a statistically better discrimination capability in the 3-o’clock sector, as assessed by AUC. The better overall glaucoma discrimination capability of Cirrus HD-OCT compared with Stratus OCT may be explained by differences in measurement techniques. Advances in technology incorporated in the Cirrus HD-OCT, and yielding both higher scan resolution and more accurate data registration, may contribute to improved glaucoma diagnostic capability.

Cirrus HD-OCT measured more parameters at higher sensitivities with specificities of 90% or more than did Stratus OCT, whereas both OCT modalities were similar in performance at specificities of 80% or more (Table 2). Because diagnostic criteria of high sensitivity with maximal specificity are desirable in clinical practice, our findings are encouraging and may suggest that Cirrus HD-OCT offers better glaucoma diagnostic potential than Stratus OCT in glaucomatous eyes with VF loss.

Interpreting diagnostic capabilities, including AUC values, sensitivity, and specificity obtained in case-control studies, using newer techniques such as Cirrus HD-OCT requires a careful consideration of the type of patient being studied. Currently, there is a lack of a gold standard for diagnosing glaucoma. Often, visualization of a change on optic disc photographs, which may take years to occur, or VF defects seen on achromatic automated perimetry, which may not show up until many retinal nerve fibers are already lost, are used to define glaucoma. The current study did not include so-called preperimetric glaucoma patients or glaucoma suspects, defined as patients who are suspected to have or have begun to lose nerve fibers in the absence of VF changes. This will require challenging studies to observe glaucoma suspects over a long period to determine if measurements on the Cirrus HD-OCT become abnormal before changes appear on achromatic perimetry or stereoscopic disc photographs. Therefore, it is important to keep in mind that our estimates of diagnostic capabilities associated with both technologies are unfairly increased and may not be directly applicable to the suspected patients evaluated in our clinical practice.

AUCs AT DIFFERENT STAGES OF GLAUCOMA

In a comparison of glaucoma discrimination capability at different disease stages, Cirrus HD-OCT showed significantly higher AUCs for patients with early-stage disease, in more sectors, compared with Stratus OCT. However, there were no significant measurement differences (including mean RNFL thicknesses) in most sectors in patients with moderate to advanced glaucoma. Because structural damage may precede functional decay in glaucoma, the diagnostic capabilities of structural tests in patients with later glaucoma stages may be of less clinical significance than early-stage detection in patients. Whether Cirrus HD-OCT may be valuable in early detection of structural glaucomatous damage remains to be explored in the future as this study was based on subjects with glaucoma with preexisting VF defects.

LIKELIHOOD RATIOS

In our study, in both superior and inferior sectors, LRs for outside–normal limits and borderline results from both Cirrus HD-OCT and Stratus OCT data were generally associated with large effects on posttest glaucoma probabilities. However, LRs for within–normal limits results were associated with insignificant changes with both instruments. In other words, such data are of limited use for exclusion of disease. Overall, LRs of within–normal limits results ranged from 0.35 to 0.97 for Cirrus HD-OCT and 0.50 to 0.95 for Stratus OCT.

In general, Cirrus HD-OCT showed higher LRs associated with outside–normal limits results as well as lower LRs with within–normal limits results in most sectors except for the nasal quadrant and 3-, 4-, and 5-o’clock sectors compared with Stratus OCT with borderline results. This may be because of the better sensitivity and specificity of Cirrus HD-OCT, as reviewed earlier. Another probable contributory factor is a difference between the internal normative databases of the 2 OCT instruments. In the Stratus OCT database, very few Asian individuals (3% of total) are included,22 whereas in construction of the Cirrus HD-OCT normative database, more than 20% of subjects were Asian.23 The substantially higher proportion of Asian ethnic data in the Cirrus HD-OCT database than in that of the Stratus OCT may contribute to glaucoma diagnostic accuracy as assessed by LR in Asian individuals.

At the nasal quadrant and 3-, 4-, and 5-o’clock sectors, the Cirrus HD-OCT performed more poorly than Stratus OCT, as assessed by LR calculations with borderline results and also when AUCs were compared. This deficiency of Cirrus HD-OCT in nasalside clock-hour measurements may be explained by the data acquisition mode. For example, large vessels in the optic nerve head dominate on the nasal side. Thus, data from this side are generally of low precision and suffer from inaccuracies.

Although the LRs provide an indication of how much the odds of disease change based on a positive or a negative result, they would be useful when obtained in a clinically more relevant population. In our study design, as well as many similar case-control studies, the normal control group may not be the ideal one in a real-life situation because it is a highly selected group of subjects with no suspicion for the disease. Furthermore, patients may have early glaucoma without achromatic perimetric abnormalities in real life. Our estimates of LRs, therefore, should be interpreted with caution.

Limitations of the current study include the use of a homogeneous population. Normative databases derived from a mixture of races might not be optimal for statistical comparisons of RNFL thickness data, although more Asian control eyes are included in the Cirrus HD-OCT database than in that of Stratus OCT. Data from a single Asian population, and including a large number of normal-tension glaucoma eyes, would be valuable to assist in sensitivity and specificity measurements. Also, we cannot exclude the possible effects of selection bias, in that most of our subjects with glaucoma had normal-tension glaucoma with an IOP lower than 22 mm Hg on multiple IOP measurements (82 of 100; 82%). Imaging comparisons of such glaucomatous eyes showing high IOPs might yield different outcomes in terms of RNFL thickness, sensitivity, and specificity.

In conclusion, both Cirrus HD-OCT and Stratus OCT RNFL thickness measurements showed good glaucoma diagnostic capabilities. Cirrus HD-OCT showed the better overall glaucoma discrimination capability in patients with early-stage glaucoma. Whereas abnormal and borderline results (as detected by comparison with normative databases) were associated with high LRs and large effects on posttest probabilities of glaucoma, normal results from either instrument were associated with small to insignificant effects on posttest probabilities. Finally, these findings may apply to Asian eyes based on our study design and they may not be generalized to other races.

Back to top
Article Information

Correspondence: Michael S. Kook, MD, Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, 388-1 Pungnap-2-dong, Songpa-gu, Seoul 138-736, Korea (mskook@amc.seoul.kr).

Submitted for Publication: February 11, 2009; final revision received May 21, 2009; accepted June 19, 2009.

Financial Disclosure: None reported.

References
1.
Burgansky-Eliash  ZWollstein  GChu  T  et al.  Optical coherence tomography machine learning classifiers for glaucoma detection: a preliminary study. Invest Ophthalmol Vis Sci 2005;46 (11) 4147- 4152
PubMedArticle
2.
Huang  MLChen  HY Development and comparison of automated classifiers for glaucoma diagnosis using Stratus optical coherence tomography. Invest Ophthalmol Vis Sci 2005;46 (11) 4121- 4129
PubMedArticle
3.
Kanamori  ANagai-Kusuhara  AEscano  MFMaeda  HNakamura  MNegi  A Comparison of confocal scanning laser ophthalmoscopy, scanning laser polarimetry and optical coherence tomography to discriminate ocular hypertensionand glaucoma at an early stage. Graefes Arch Clin Exp Ophthalmol 2006;244 (1) 58- 68
PubMedArticle
4.
Lalezary  MMedeiros  FAWeinreb  RN  et al.  Baseline optical coherence tomography predicts the development of glaucomatous change in glaucoma suspects. Am J Ophthalmol 2006;142 (4) 576- 582
PubMedArticle
5.
Leung  CKChan  WMYung  WH  et al.  Comparison of macular and peripapillary measurements for the detection of glaucoma: an optical coherence tomography study. Ophthalmology 2005;112 (3) 391- 400
PubMedArticle
6.
Manassakorn  ANouri-Mahdavi  KCaprioli  J Comparison of retinal nerve fiber layer thickness and optic disk algorithms with optical coherence tomography to detect glaucoma. Am J Ophthalmol 2006;141 (1) 105- 115
PubMedArticle
7.
Medeiros  FAZangwill  LMBowd  CVessani  RMSusanna  R  JrWeinreb  RN Evaluation of retinal nerve fiber layer, optic nerve head, and macular thickness measurements for glaucoma detection using optical coherence tomography. Am J Ophthalmol 2005;139 (1) 44- 55
PubMedArticle
8.
Naithani  PSihota  RSony  P  et al.  Evaluation of optical coherence tomography and heidelberg retinal tomography parameters in detecting early and moderate glaucoma. Invest Ophthalmol Vis Sci 2007;48 (7) 3138- 3145
PubMedArticle
9.
Nouri-Mahdavi  KNikkhou  KHoffman  DCLaw  SKCaprioli  J Detection of early glaucoma with optical coherence tomography (StratusOCT). J Glaucoma 2008;17 (3) 183- 188
PubMedArticle
10.
Parikh  RSParikh  SSekhar  GC  et al.  Diagnostic capability of optical coherence tomography (Stratus OCT 3) in early glaucoma. Ophthalmology 2007;114 (12) 2238- 2243
PubMedArticle
11.
Sihota  RSony  PGupta  VDada  TSingh  R Comparing glaucomatous optic neuropathy in primary open angle and chronic primary angle closure glaucoma eyes by optical coherence tomography. Ophthalmic Physiol Opt 2005;25 (5) 408- 415
PubMedArticle
12.
Quigley  HAKatz  JDerick  RJGilbert  DSommer  A An evaluation of optic disc and nerve fiber layer examinations in monitoring progression of early glaucoma damage. Ophthalmology 1992;99 (1) 19- 28
PubMedArticle
13.
Sommer  AKatz  JQuigley  HA  et al.  Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol 1991;109 (1) 77- 83
PubMedArticle
14.
Zeyen  TGCaprioli  J Progression of disc and field damage in early glaucoma. Arch Ophthalmol 1993;111 (1) 62- 65
PubMedArticle
15.
Paunescu  LASchuman  JSPrice  LL  et al.  Reproducibility of nerve fiber thickness, macular thickness, and optic nerve head measurements using StratusOCT. Invest Ophthalmol Vis Sci 2004;45 (6) 1716- 1724
PubMedArticle
16.
Hodapp  EParrish  RKAnderson  DR Clinical Decisions in Glaucoma.  St Louis, MO Mosby1993;53
17.
Mills  RPBudenz  DLLee  PP  et al.  Categorizing the stage of glaucoma from pre-diagnosis to end-stage disease. Am J Ophthalmol 2006;141 (1) 24- 30
PubMedArticle
18.
DeLong  ERDeLong  DMClarke-Pearson  DL Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44 (3) 837- 845
PubMedArticle
19.
Jaeschke  RGuyatt  GHSackett  DL Users' guides to the medical literature, III: how to use an article about a diagnostic test. B: what are the results and will they help me in caring for my patients? the Evidence-Based Medicine Working Group. JAMA 1994;271 (9) 703- 707
PubMedArticle
20.
Budenz  DLMichael  AChang  RT McSoley  JKatz  J Sensitivity and specificity of the Stratus OCT for perimetric glaucoma. Ophthalmology 2005;112 (1) 3- 9
PubMedArticle
21.
Bowd  CZangwill  LMBerry  CC  et al.  Detecting early glaucoma by assessment of retinal nerve fiber layer thickness and visual function. Invest Ophthalmol Vis Sci 2001;42 (9) 1993- 2003
PubMed
22.
Budenz  DLAnderson  DRVarma  R  et al.  Determinants of normal retinal nerve fiber layer thickness measured by Stratus OCT. Ophthalmology 2007;114 (6) 1046- 1052
PubMedArticle
23.
Gurses-Ozden  RDurbin  MCallan  T  et al.  Distribution of retinal nerve fiber layer thickness using Cirrus HD-OCT spectral domain technology.  Paper presented at: ARVO 2008 Annual Meeting April 30, 2008 Fort Lauderdale, FLE-abstract 4632
×