HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (7) Citing Articles via Google Scholar Google Scholar Articles by Naithani, P. Articles by Pandey, R. M. Search for Related Content PubMed PubMed Citation Articles by Naithani, P. Articles by Pandey, R. M.

Evaluation of Optical Coherence Tomography and Heidelberg Retinal Tomography Parameters in Detecting Early and Moderate Glaucoma

¹From the Glaucoma Research Facility, Dr. Rajendra Prasad Centre for Ophthalmic Sciences, and the ²Department of Biostatistics, All India Institute of Medical Sciences, New Delhi, India.

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
PURPOSE. To evaluate the relationship between optic nerve head (ONH) and peripapillary retinal nerve fiber layer (RNFL) parameters by optical coherence tomography (OCT) and confocal scanning laser ophthalmoscopy (Heidelberg retinal tomography; HRT; Heidelberg Engineering, Heidelberg, Germany) in early and moderate glaucoma and to compare several OCT-based automated classifiers with those inbuilt in HRT for detection of glaucomatous damage.

METHODS. This cross-sectional study included 60 eyes of 60 patients with glaucoma (30 early and 30 moderate visual field defects) and 60 eyes of 60 healthy subjects. All patients underwent Fast Optic Disc and Fast Peripapillary RNFL scans on the OCT and then HRT evaluation of the ONH during the same visit. Glaucoma variables obtained from OCT and HRT analyses were compared among the groups. Receiver operator characteristic (ROC) curves generated by performing linear discriminant analysis (LDA), artificial neural networks (ANNs), and classification and regression trees (CART) on OCT-based parameters were compared with the Moorfield regression analysis (MRA), R Bathija (RB), and FS Mickelberg (FSM) functions in the HRT, to classify eyes as either glaucomatous or normal.

RESULTS. No statistically significant difference was found in the disc area measured by the OCT and HRT analyses within each study group (P > 0.05). The areas under ROC curves were 0.9822 (LDF), 0.9791 (CART), and 0.9383 (ANN) as compared with 0.859 (FSM), 0.842 (RB) and 0.767 (MRA).

CONCLUSIONS. OCT-based automated classifiers performed better than HRT classifiers in distinguishing glaucomatous from healthy eyes. Such parameters should be integrated in the OCT to improve its diagnostic abilities.

As both CSLO and the OCT analyze structural parameters in the ONH for glaucoma diagnosis, the purpose of this study was to compare the performance of ONH and peripapillary RNFL parameters generated by OCT (StratusOCT 3000; software ver. 4.0; Carl Zeiss Meditec, Inc.) with those by CSLO with the HRT II (Heidelberg Retinal Tomograph; software version 2.0; Heidelberg Engineering GmbH, Heidelberg, Germany), for the detection of early to moderate glaucoma from control eyes.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
One hundred twenty eyes of 120 consecutive and eligible subjects (60 patients with glaucoma and 60 healthy control subjects) were included in this cross-sectional study. The study was approved by our institutional review board and complied with the tenets of Declaration of Helsinki. Informed consent was obtained. All subjects were 40 years of age or older, having a best-corrected visual acuity of 20/40 and a refractive error within ±6.0 D (spherical equivalent).

Eligible subjects underwent a complete ophthalmic evaluation, including review of medical history, manifest refraction, axial length (EchoScan US 3300; Nidek Corp., Gamagori, Japan), keratometry, central corneal pachymetry, slit lamp biomicroscopy, intraocular pressure (IOP) measurement using Goldmann applanation tonometry, gonioscopy, dilated fundus evaluation using a +90-D lens and automated perimetry (model 745 Humphrey visual field analyser [HFA], full-threshold program 30-2; Carl Zeiss Meditec, Inc.). OCT and HRT scanning was also performed within 1 week of the baseline examinations.

Normal control subjects had no ocular complaints or diseases. All had normal anterior segments on slit lamp biomicroscopy, open angles on gonioscopy and normal disc and macula. They also had IOP 21 mm Hg and reliable normal (mean deviation and pattern SD within 95% confidence limits and a Glaucoma Hemifield Test (GHT) result within normal limits full-threshold 30-2 Humphrey visual fields on more than two occasions. One eye of 60 such subjects was randomly selected for inclusion in the study.

Patients were categorized as having glaucoma if they had an IOP of >21 mm Hg and reliable, consistent glaucomatous visual field defects commensurate with optic nerve damage on more than two occasions. Glaucomatous visual field loss was defined as consistent presence of a cluster of three or more nonedge points on the pattern deviation plot in typical glaucomatous locations, CPSD with P < 5%, or a GHT outside normal limits. Sixty such eyes of 60 patients having early or moderate visual field defects according to the Hodapp-Anderson-Parrish²³ grading scale of visual field severity were included in the study.

Patients who had other intraocular or neurologic disease that affected the RNFL or optic disc, a secondary cause of raised intraocular pressure, or significant media opacity were excluded. Eyes with consistently unreliable visual field results (defined as false positives and negatives >33% and fixation losses >20%) were also excluded from the study.

All CSLO scans were performed with the HRT II (Heidelberg Retinal Tomograph; software ver. 2.0). A series of three good-quality scans for each eye were used for the ONH analysis. The following parameters were computed: disc area, cup area, rim area, cup-to-disc (C-D) area ratio, rim-to-disc (R-D) area ratio, vertical and horizontal C-D ratio and cup volume. Linear discriminant functions (LDF) developed by Mikelberg et al.²⁴ (FSM) and Bathija et al.²⁵ (RB), and the Moorfield regression analysis (MRA)²⁶ inbuilt in the HRT II were also evaluated in the study.

The OCT 3 (StratusOCT 3000; software version 4.0; Carl Zeiss Meditec, Inc.) was used to image the RNFL and the ONH. A masked operator performed imaging with the two algorithms at the same session. Disc area, cup area, rim area, C-D area ratio, horizontal and vertical C-D ratio, vertical integrated rim area (VIRA, estimate of total volume of rim tissue calculated by multiplying the average of individual rim areas by the circumference of the disc), and horizontal integrated rim width (HIRW, estimate of total area of rim tissue calculated by multiplying the average of individual rim widths by the circumference of the disc) were the ONH parameters evaluated.

The Fast RNFL Thickness protocol on OCT was used to yield three 3.4-mm-diameter circular scans for each eye. Presence of uniform signal intensity, strong reflectance signal from the RNFL and the retinal pigment epithelium resulting in clear demarcation of both layers without the absence of any part of image constituted a good-quality scan. The following parameters were calculated: average RNFL thickness, RNFL thickness in the superior and inferior hemifields, RNFL thickness in the four quadrants spanning 90° each and RNFL thickness in twelve 30° clock-hour sectors.

One-way ANOVA with the Bonferroni correction was used to compare the glaucoma parameters between the groups and between diagnostic modalities within each group. Receiver operator curves (ROC) were plotted for each parameter to evaluate its diagnostic ability. Linear discriminant analysis (LDA), artificial neural network (ANN), and classification and regression tree (CART) methods were used to develop three automated classifiers based on the glaucoma parameters measured by the OCT. Discriminant analysis was performed using all the ONH and RNFL variables to develop the best linear discriminant function (LDF). LDA has been used in various glaucoma studies^{27 28} for classifying patients according to disease severity. It assumes a Gaussian distribution of data and defines linear discrimination boundaries between the categories where it maximizes the variance between classes while minimizing the variance within classes. Neural network analysis mimics the brain’s problem-solving process. Just as humans apply knowledge gained from experience to new problems or situations, a neural network takes previously solved examples to build a system of "neurons" that make new decisions, classifications, and forecasts. It looks for patterns in training sets of data, learns these patterns, and develops the ability to classify new patterns correctly.²⁹ The CART method is unique in its methodology by making no previous assumptions in labeling a subject as normal or diseased.³⁰ The key elements of a CART analysis are a set of rules for splitting each node in a tree, deciding when a tree is complete, assigning each terminal node to a class outcome, and selecting the "right-sized" tree. All the ONH and RNFL parameters were simultaneously entered into the CART analysis software, to obtain the best classification tree based on minimum variables. Cross-validation of the LDA and ANN results was performed by randomly selecting 70% of the study population as a training set and the remaining 30% as the test set. Sensitivities and specificities of such a set were calculated. This process was repeated 10 times, and the average values were compared with the results initially obtained. Twenty-five-fold cross-validation was performed for CART analysis by omitting one twenty-fifth of the data for each series. Misclassification rates and ROCs were plotted to compare the classifiers’ performance with one another and with the HRT-based algorithms in discriminating glaucomatous from normal eyes.

	Results

Top Abstract Materials and Methods Results Discussion References

 
Demographic characteristics of the study population are summarized in Table 1 . The mean age in the control group was 60.17 years compared with 60.97 years in early glaucoma and 59.37 years in the moderate glaucoma group. There was no significant difference in the refractive status, axial length, and central corneal thickness measurements between the three groups (P > 0.05). Mean deviation (MD), pattern SD (PSD), and corrected pattern SD (CPSD) in the early and moderate glaucoma group were significantly worse compared with those parameters in normal subjects (P < 0.05).

On evaluating the percentage loss of average peripapillary RNFL thickness on OCT (Fig. 1) , a 28% decline was seen in eyes with early glaucoma when compared to the normal subjects (P < 0.001). A further decline of 4% in the average RNFL thickness occurred in the moderate glaucoma group compared with the early glaucoma group. A similar pattern was seen in the corresponding RNFL thickness across all four quadrants.

View larger version (7K): [in this window] [in a new window]   FIGURE 1. Distribution of the average RNFL thickness among the normal subjects (category 0) and patients with early (category 1) or moderate (category 2) glaucoma.

had a misclassification rate of 5% and an AUC of 0.9822. CART analysis was performed by simultaneously evaluating both optic disc and RNFL parameters to determine an optimal combination of parameters from both algorithms that had the highest AUCs in differentiating glaucoma from normal eyes (Fig. 2) . A combination of C-D vertical ratio and average RNFL thickness provided the best result, with an AUC of 0.9791 and a misclassification rate of 5.8%. The AUC for the ANN algorithm was 0.9383 with a misclassification rate of 6.5%. Figure 3 shows the ROCs of the three OCT-based classifiers and the MRA, RB discriminant function, and FSM function computed by the HRT II.

View larger version (14K): [in this window] [in a new window]   FIGURE 2. Classification and regression tree analysis flowchart for best combination of OCT-based ONH and peripapillary RNFL parameters.

View larger version (8K): [in this window] [in a new window]   FIGURE 3. Comparison of AUCs between the OCT-based classifiers (top) and HRT-generated functions (bottom) in distinguishing patients with glaucoma from normal subjects.

	Discussion

Top Abstract Materials and Methods Results Discussion References

Comparison of optic disc parameters measured by the HRT II and OCT 3 provided some interesting results in our study. Although there were no significant differences in the disc area measured by the HRT and the OCT among the three study groups, the cup area and related parameters were measured as significantly larger by the OCT than by the HRT in each group. Schuman et al.²¹ partly attributed the increase in the stereometric parameters to the significantly larger disc size measured by the OCT in their study. It seems, however, that the difference arises predominantly due to the anterior placement of the reference plane in the OCT compared with the HRT scan.

The AUC of HRT parameters in our study varied from 0.474 and 0.852 in the control versus the early glaucoma group and 0.535 to 0.894 in the control versus the moderate glaucoma group. These values are higher than those reported by Schuman et al.²¹ in whose study they varied from 0.48 to 0.73. The highest AUC among the HRT and OCT ONH parameters was for the vertical C-D ratio, both in differentiating control eyes from eyes with moderate or early glaucoma. Manassakorn et al.²² showed the vertical C-D ratio as the ONH parameter having the highest AUC in their study as well. The vertical C-D ratio is the most commonly evaluated parameter in clinically suspected glaucoma, and its high AUC in distinguishing early glaucoma from normal optic discs further reemphasizes its importance as a diagnostic sign. Average RNFL thickness has been documented to have the best discriminating ability among all OCT-based RNFL parameters in various studies,^{18 19 31 32} and this result was reproduced in our study as well. The discriminating ability of all the HRT as well as OCT parameters improved with an increase in glaucomatous damage among subjects.

The glaucomatous disease process leads to a progressive loss of retinal ganglion cells over time.³³ In our present study, there was a 28% decline in the average RNFL thickness in the eyes with early glaucoma compared with the normal eyes. This is in accordance with a previous study from our center³³ that showed a decline of 25% in the eyes with early glaucoma. Kerringnan-Baumrind et al.³⁴ have recorded a loss of 25% to 35% in retinal ganglion cells before visual defects were confirmed on automated perimetry. A further loss of only 4% in the average RNFL thickness was detected between the eyes with moderate and early glaucoma in our study. This result is similar to the observations made in a study by Galvao-Filho et al.,³⁵ in which no statistically significant difference was seen in the RNFL thickness between eyes with early and moderate glaucoma when measured using SLP (GDx VCC). Our results suggest that an initial loss of 28% of RNFL thickness occurs before visual field defects can be appreciated on automated achromatic perimetry. However, once this ganglion cell "buffer" is destroyed, further functional deterioration occurs out of proportion to the anatomic damage.

Extremes of disc sizes had a significant effect on the diagnostic ability of the MRA in our present study. Falsely labeling a disc abnormal due to its large size would subject such people to a barrage of glaucoma tests leading to unnecessary expense. Agarwal et al.³⁶ showed that a similar disc area biased the MRA misclassification of 14.5% among 275 normal eyes using HRT II in an Indian population. An equal number of glaucoma cases having significantly smaller disc areas were diagnosed as normal in the present study. Small discs are usually associated with small cups, and this makes clinical suspicion of glaucomatous cupping difficult in such patients. This, combined with the fact that these subjects had a higher rate of being labeled normal according to the MRA in our study, could lead to such patients’ escaping glaucoma detection until late in the disease process. The OCT-based classifiers, however, misclassified none of these eyes.

Among the three indigenously developed OCT classifiers developed in this study, LDF performed the best. Depending on the hemifield affected by the glaucomatous damage, superior or inferior RNFL tends to get thinned out of proportion to the adjoining nerve fiber bundles, and the incorporation of both superior and inferior quadrant RNFL thickness in the formula gives it a wider diagnostic application. The CART analysis provides for easily understood algorithms based on the minimum number of clinically useful variables. After the C-D vertical ratio was used for gross delineation of the easily discernible from the overlapping cases, the average RNFL thickness was used to correctly classify the latter further. The combination of these two parameters resulted in a high AUC and a low misclassification rate. ANNs too performed significantly better than the HRT classifiers in detection of patients with glaucoma.

Studies conducted recently on the discriminating capabilities of computerized classifiers have shown mixed results. In a study by Huang and Chen,³¹ automated classifiers performed significantly better than any isolated OCT parameter in distinguishing glaucomatous from normal eyes. However, Burgansky-Eliash et al.³² did not find any significant improvement in discriminating patients with glaucoma from healthy subjects by using indigenous classifiers in their study. Manassakorn et al.²² achieved 92% sensitivity and 95% specificity in their CART analysis based on the inferior quadrant RNFL thickness and vertical C-D ratio parameters of OCT 3. Our findings may help in incorporation of similar such algorithms in the OCT for the early quantification of glaucoma.

Our study has certain limitations. IOP >21 mm Hg was used as an inclusion criteria for patients with glaucoma, which could have lead to exclusion of a subset of patients demonstrating glaucomatous damage at lower pressures. Further investigation with a larger number of patients with glaucoma, irrespective of the underlying etiology, would tend to overcome this limitation and is currently under way. Glaucoma is a disease that demonstrates high variability in anatomic versus functional damage. However, an equal number of early and moderate cases were enrolled in our study, to develop classifiers to detect glaucoma progression at successive stages. Also, the particular analysis models we chose are unlikely to be the only ones that can be applied to such an evaluation. Given the good discriminating ability of many of the OCT parameters, there are bound to be other analytical combinations to distinguish glaucoma at an early stage.

In conclusion, the performance of OCT 3’s Fast RNFL and Fast Optic Disc protocols were superior to the HRT-based parameters in our study. AUCs evaluated by OCT showed a higher degree of accuracy compared with HRT in detection of early glaucoma. Disc area was a confounding variable in HRT- but not in OCT-based evaluation. Although the results are encouraging, studies on a larger scale are needed to fine-tune and integrate the OCT-based classifiers into the entire clinical picture for more accurate diagnosis of glaucoma at an early stage.

	Footnotes

 
Submitted for publication November 25, 2006; revised January 30, 2007; accepted May 3, 2007.

Disclosure: P. Naithani, None; R. Sihota, None; P. Sony, None; T. Dada, None; V. Gupta, None; D. Kondal, None; R.M. Pandey, None

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked "advertisement" in accordance with 18 U.S.C. 1734 solely to indicate this fact.

Corresponding author: Prashant Naithani, Glaucoma Services, Room No. 495, 4th Floor, Dr. R. P. Centre for Ophthalmic Sciences, All India Institute of Medical Sciences, New Delhi 110049, India; drprashantnaithani{at}yahoo.co.in.

	References

Top Abstract Materials and Methods Results Discussion References

 

1. Schuman JS, Hee HR, Arya AV, et al. Optical coherence tomography: a new tool for glaucoma diagnosis. Curr Opin Ophthalmol. 1995;6:89–95.[Medline][Order article via Infotrieve]
2. Rohschneider K, Burk RO, Kruse FE, et al. Reproducibility of optic nerve head topography with a new laser scanning tomographic device. Ophthalmology. 1994;101:1044–1049.[Web of Science][Medline][Order article via Infotrieve]
3. Zangwill L, Shakiba S, Caprioli J, et al. Agreement between clinicians and a confocal scanning laser ophthalmoscope in estimating C-D ratios. Am J Ophthalmol. 1995;119:415–421.[Web of Science][Medline][Order article via Infotrieve]
4. Kruse FE, Burk ROW, Volcker HE, et al. Reproducibility of topographic measurements of optic nerve head with topographic scanner. Ophthalmology. 1989;96:1320–1324.[Web of Science][Medline][Order article via Infotrieve]
5. Dreher AW, Tso PC, Weinreb RN. Reproducibility of optic nerve head measurements of normal and glaucomatous optic nerve head with laser tomographic scanner. Am J Ophthalmol. 1991;111:221–229.[Web of Science][Medline][Order article via Infotrieve]
6. Rohschneider K, Burk RO, Volcker HE. Reproducibility of topometric data acquisition in normal and glaucomatous optic nerve heads with the laser tomographic scanner. Graefes Arch Clin Exp Ophthalmol. 1993;231:457–464.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
7. Chauhan BC, LeBlanc RP, McCormick TA, et al. Test retest variability of topographic measurements with confocal laser scanning tomography in patients with glaucoma and control subjects. Am J Ophthalmol. 1994;118:9–15.[Web of Science][Medline][Order article via Infotrieve]
8. Janknecht P, Funk J. Optic nerve head analyser and Heidelberg retinal tomograph: accuracy and reproducibility of topographic measurements in a model eye and in volunteers. Br J Ophthalmol. 1994;78:760–768.[Abstract/Free Full Text]
9. Schuman JS, Hee MR, Puliafito CA, et al. Quantification of nerve fiber layer thickness in normal and glaucomatous eyes using optical coherence tomography. Arch Ophthalmol. 1995;113:586–596.[Abstract/Free Full Text]
10. Gurses-Ozden R, Ishikawa H, Hoh ST, et al. Increasing sampling density improves reproducibility of optical coherence tomography measurements. J Glaucoma. 1999;8:238–241.[Web of Science][Medline][Order article via Infotrieve]
11. Mistlberger A, Liebmann JM, Greenfield DS, et al. Heidelberg retinal tomography and optical coherence tomography in normal, ocular hypertensive and glaucomatous eyes. Ophthalmology. 1999;106:2027–2032.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
12. Pieroth L, Schuman JS, Hertzmark E, et al. Evaluation of focal defects of the nerve fiber layer using optical coherence tomography. Ophthalmology. 1999;106:570–579.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
13. Zangwill LM, Williams J, Berry CC, Knauser S, Weinreb RN. A comparison of optical coherence tomography and retinal nerve fiber layer photography for detection of retinal nerve fiber layer damage in glaucoma. Ophthalmology. 2000;107:1309–1315.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
14. Hoh ST, Greenfield DS, Mistlberger A, Liebmann JM, Ishikawa H, Ritch R. Optical coherence tomography and scanning laser polarimetry in normal, ocular hypertensive and glaucomatous eyes. Am J Ophthalmol. 2000;129:129–135.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
15. Bowd C, Weinreb RN, Williams JM, Zangwill LM. The retinal nerve fibre layer thickness in ocular hypertensive, normal and glaucomatous eyes with optical coherence tomography. Arch Ophthalmol. 2000;118:22–26.[Abstract/Free Full Text]
16. Wollstein G, Schuman JS, Price LL, et al. Optical coherence tomography (OCT) macular and peripapillary retinal nerve fibre layer measurements and automated visual fields. Am J Ophthalmol. 2004;138:218–225.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
17. Nouri-Mahdavi K, Hoffman D, Tannenbaum DP, Law SK, Caprioli J. Identifying early glaucoma with optical coherence tomography. Am J Ophthalmol. 2004;137:228–235.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
18. Wollstein G, Ishikawa H, Wang J, Beaton SA, Schuman JS. Comparison of three optical coherence tomography scanning areas for detection of glaucomatous damage. Am J Ophthalmol. 2005;139:39–43.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
19. Medeiros FA, Zangwill LM, Bowd C, Vessani RM, Susanna R, Jr, Weinreb 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;139:44–55.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
20. Lai E, Wollstein G, Price LL, et al. Optical coherence tomography disc assessment in optic nerves with peripapillary atrophy. Ophthalmic Surg Lasers Imaging. 2003;34:498–504.[Web of Science][Medline][Order article via Infotrieve]
21. Schuman JS, Wollstein G, Farra T, et al. Comparison of optic nerve head measurements obtained by optical coherence tomography and confocal scanning laser ophthalmoscopy. Am J Ophthalmol. 2003;135:504–512.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
22. Manassakorn A, Nouri-Mahdavi K, Caprioly J. Comparison of retinal nerve fiber layer thickness and optic disc algorithms with optical coherence tomography to detect glaucoma. Am J Ophthalmol. 2006;141:105–115.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
23. Hodapp E, Parrish RK, Anderson DR. Clinical Decisions in Glaucoma. 1993;84–125. CV Mosby St. Louis.
24. Mikelberg FS, Parfitt CM, Swindale NV, Graham SL, Drance SM, Gosine R. Ability of Heidelberg retinal tomograph to detect early glaucomatous visual field loss. J Glaucoma. 1995;4:242–247.[Medline][Order article via Infotrieve]
25. Bathija R, Zangwill L, Berry CC, Sample PA, Weinreb RN. Detection of early glaucomatous structural damage with confocal scanning laser tomography. J Glaucoma. 1998;7:121–127.[Web of Science][Medline][Order article via Infotrieve]
26. Wollstein G, Garway-Heath GF, Hitchings RA. Identification of early glaucoma cases with the scanning laser ophthalmoscope. Ophthalmology. 1998;105:1557–1563.[CrossRef][Web of Science][Medline][Order article via Infotrieve]
27. Bowd C, Chan K, Zangwill LM, et al. Comparing neural networks and discriminant functions for glaucoma detection using confocal scanning laser ophthalmoscopy of the optic disc. Invest Ophthalmol Vis Sci. 2002;43:3444–3454.[Abstract/Free Full Text]
28. Lester M, Jonas JB, Mardin CY, et al. Discrimination analysis models for early detection of glaucomatous optic disc changes. Br J Ophthalmol. 1997;115:725–728.
29. Bose NK, Liang P. Neural Network Fundamentals with Graphs, Algorithms and Applications. 1996; McGraw-Hill New York.
30. Cook EF, Goldman L. Asymmetric stratification: an outline for an efficient method for controlling confounding in cohort studies. Am J Ophthalmol. 1998;127:626–639.
31. Huang M, Chen H. Development and comparison of automated classifiers for glaucoma diagnosis using Stratus optical coherence tomography. Invest Ophthalmol Vis Sci. 2005;46:4121–4129.[Abstract/Free Full Text]
32. Burgansky-Eliash Z, Wollstein G, Chu T, et al. Optical Coherence Tomography learning classifiers for glaucoma detection- a preliminary study. Invest Ophthalmol Vis Sci. 2005;46:4147–4152.[Abstract/Free Full Text]
33. Sihota R, Sony P, Gupta V, Dada T, Singh R. Diagnostic capability of optical coherence tomography in evaluating the degree of glaucomatous retinal nerve fibre damage. Invest Ophthalmol Vis Sci. 2005;47:2006–2010.[Web of Science]
34. Kerringnan-Baumrind LA, Quigley HA, Pease ME, et al. Number of ganglion cells in glaucoma eyes compared with threshold visual field tests in same persons. Invest Ophthalmol Vis Sci. 2000;41:741–748.[Abstract/Free Full Text]
35. Galvao-Filho RP, Vessani RM, Sussana R. Comparison of retinal nerve fibre layer thickness and visual field loss between different glaucoma groups. Br J Ophthalmol. 2005;89:1004–1007.[Abstract/Free Full Text]
36. Agarwal HC, Gulati V, Sihota R. The normal optic nerve head on Heidelberg retinal tomograph II. Indian J Ophthalmol. 2003;51:25–33.[Medline][Order article via Infotrieve]

This article has been cited by other articles:

				S. B. Park, K. R. Sung, S. Y. Kang, K. R. Kim, and M. S. Kook Comparison of Glaucoma Diagnostic Capabilities of Cirrus HD and Stratus Optical Coherence Tomography Arch Ophthalmol, December 1, 2009; 127(12): 1603 - 1609. [Abstract] [Full Text] [PDF]

				S. Demirel, B. Fortune, J. Fan, R. A. Levine, R. Torres, H. Nguyen, S. L. Mansberger, S. K. Gardiner, G. A. Cioffi, and C. A. Johnson Predicting Progressive Glaucomatous Optic Neuropathy Using Baseline Standard Automated Perimetry Data Invest. Ophthalmol. Vis. Sci., February 1, 2009; 50(2): 674 - 680. [Abstract] [Full Text] [PDF]

				A. Ferreras, L. E. Pablo, A. B. Pajarin, J. M. Larrosa, V. Polo, and F. M. Honrubia Logistic Regression Analysis for Early Glaucoma Diagnosis Using Optical Coherence Tomography Arch Ophthalmol, April 1, 2008; 126(4): 465 - 470. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (7) Citing Articles via Google Scholar Google Scholar Articles by Naithani, P. Articles by Pandey, R. M. Search for Related Content PubMed PubMed Citation Articles by Naithani, P. Articles by Pandey, R. M.