Inter- and intraobserver variation in the analysis of optic disc images: comparison of the Heidelberg retina tomograph and computer assisted planimetry

Craniofacial anomalies in schizophrenia-relevant GFAP.HMOX10-12m mice

Subtle craniofacial dysmorphology has been reported in schizophrenia patients. This dysmorphology includes midline facial elongation, frontonasal anomalies and a sexually dimorphic deviation from normal directional asymmetry of the face, with male patients showing reduced and female patients showing enhanced facial asymmetry relative to healthy control subjects. GFAP.HMOX10-12m transgenic mice (Mus musculus) that overexpress heme oxygenase-1 in astrocytes recapitulate many schizophrenia-relevant neurochemical, neuropathological and behavioral features. As morphogenesis of the brain, skull and face are highly interrelated, we hypothesized that GFAP.HMOX10-12m mice may exhibit craniofacial anomalies similar to those reported in persons with schizophrenia. We examined craniofacial anatomy in male GFAP.HMOX10-12m mice and wild-type control mice at the early adulthood age of 6-8 months. We used computer vision techniques for the extraction and analysis of mouse head shape parameters from systematically acquired 2D digital images, and confirmed our results with landmark-based geometric morphometrics. We performed skull bone morphometry using digital calipers to take linear distance measurements between known landmarks. Relative to controls, adult male GFAP.HMOX10-12m mice manifested craniofacial dysmorphology including elongation of the nasal bones, alteration of head shape anisotropy and reduction of directional asymmetry in facial shape features. These findings demonstrate that GFAP.HMOX10-12m mice exhibit craniofacial anomalies resembling those described in schizophrenia patients, implicating heme oxygenase-1 in their development. As a preclinical mouse model, GFAP.HMOX10-12m mice provide a novel opportunity for the study of the etiopathogenesis of craniofacial and other anomalies in schizophrenia and related disorders.

Computed tomographic arthrography, gross anatomy and histology demonstrate a communication between synovial invaginations in the proximal aspect of the third interosseous muscle and the carpometacarpal joint in horses

This descriptive anatomical study investigates the relationship between the third interosseous muscle, also known as the suspensory ligament, and the carpometacarpal joint in forelimbs of horses, with the hypothesis that there was a direct synovial communication between these structures as shown by computed tomographic arthrography, histology, and gross anatomy sections. Computed tomography of the carpus and metacarpal region was performed on two groups. Group 1 consisted of eight cadaver limbs undergoing computed tomographic arthrography following injection of a mixture of positive contrast medium, saline, and color-pigmented fluid solution into the middle carpal joint. Group 2 consisted of eight forelimbs assessed using plain computed tomography. The images were interpreted subjectively for contrast medium distribution and objectively by comparing Hounsfield values of the proximal suspensory ligament at 0.5 cm intervals starting at the origin and extending 3 cm distal to the proximal subchondral bone plate of the third metacarpal bone. Of the 16 limbs, two were sectioned for gross anatomy and one was documented histologically. The proximal suspensory ligament was visualized with clear margins on computed tomography images. The positive contrast medium was found within the lateral and medial lobes of the suspensory ligament in all eight (100%) limbs. Hounsfield units within the suspensory ligament following contrast injection were significantly higher than in those in the plain CT group between 0.5 and 2.5 cm distal to the proximal subchondral bone plate of the third metacarpal bone (p < 0.05). The gross anatomy sections showed color pigmentation within the suspensory ligament correlating to the contrast medium distribution evident on computed tomography images. Histology confirmed a synovial lined cavity within the suspensory ligament. The demonstration of a direct synovial communication between the internal structure of the proximal suspensory ligament and the carpometacarpal joint in horses offers further explanation for commonly encountered interactions of diagnostic local anesthesia of the carpal and subcarpal regions. When performing diagnostic or therapeutic injections into the middle carpal joint, the likely effect on the proximal suspensory ligament should be considered. Furthermore, as the proximal suspensory ligament was identified clearly on CT images, further studies are needed to elucidate the utility of CT in clinical cases with suspected soft tissue pathology in the subcarpal region.

A combined convolutional and recurrent neural network for enhanced glaucoma detection

Glaucoma, a leading cause of blindness, is a multifaceted disease with several patho-physiological features manifesting in single fundus images (e.g., optic nerve cupping) as well as fundus videos (e.g., vascular pulsatility index). Current convolutional neural networks (CNNs) developed to detect glaucoma are all based on spatial features embedded in an image. We developed a combined CNN and recurrent neural network (RNN) that not only extracts the spatial features in a fundus image but also the temporal features embedded in a fundus video (i.e., sequential images). A total of 1810 fundus images and 295 fundus videos were used to train a CNN and a combined CNN and Long Short-Term Memory RNN. The combined CNN/RNN model reached an average F-measure of 96.2% in separating glaucoma from healthy eyes. In contrast, the base CNN model reached an average F-measure of only 79.2%. This proof-of-concept study demonstrates that extracting spatial and temporal features from fundus videos using a combined CNN and RNN, can markedly enhance the accuracy of glaucoma detection.

Application of Speckle-Tracking Echocardiography in an Experimental Model of Isolated Subendocardial Damage

BACKGROUND

The subendocardium is highly vulnerable to damage and is thus affected even in subclinical disease stages. Therefore, methods reflecting subendocardial status are of great clinical relevance for the early detection of cardiac damage and the prevention of functional impairment. The aim of this study was to investigate the potential ability of myocardial strain parameters to evaluate changes within the subendocardium.

METHODS

Male 129/Sv mice were injected with isoproterenol (ISO; n = 32) to induce isolated subendocardial fibrotic lesions or saline as appropriate control (n = 15). Transthoracic echocardiography was performed using a 30-MHz linear-frequency transducer coupled to a high-resolution imaging system, and acquired images were analyzed for conventional and strain parameters. The degree of collagen content within the different cardiac layers was quantified by histologic analysis and serum levels of tissue inhibitor of metalloproteinase-1, a biomarker for fibrosis, were assessed.

RESULTS

ISO treatment induced a marked increase in subendocardial collagen content in response to cell loss (control vs ISO, 0.6 ± 0.3% vs 5.8 ± 0.9%; P < .001) and resulted in a moderate increase in left ventricular wall thickness with preserved systolic function. Global longitudinal peak strain (LS) and longitudinal strain rate were significantly decreased in ISO-treated animals (LS, -15.49% vs -11.49% [P = .001]; longitudinal strain rate, -4.81 vs -3.88 sec^-1 [P < .05]), whereas radial and circumferential strain values remained unchanged. Global LS was associated with subendocardial collagen content (r = 0.46, P = .01) and tissue inhibitor of metalloproteinase-1 serum level (r = 0.52, P < .05). Further statistical analyses identified global LS as a superior predictor for the presence of subendocardial fibrosis (sensitivity, 84%; specificity, 80%; cutoff value, -14.4%).

CONCLUSION

Assessment of LS may provide a noninvasive method for the detection of subendocardial damage and may consequently improve early diagnosis of cardiac diseases.

Agreement among ophthalmologists in marking the optic disc and optic cup in fundus images

The purpose of this paper was to study the agreement between six ophthalmologists who manually marked the optic nerve head using fundus images. Four different parameters were considered from the manual marking images: (1) disc (area and centroid), (2) cup (area and centroid), (3) horizontal and vertical cup-to-disc ratios, and (4) including the previous two parameters for both horizontal and vertical cup-to-disc ratios, and investigated the comprehensive agreement and accuracy among all the ophthalmologists. The best agreement percentage for all the parameters combined was between ophthalmologists number one and three for 44 % of images, and the best accuracy was for ophthalmologist number one with 77.4 % of 315 total tested images. Our analysis shows that more than half of the images in the dataset were not agreed upon when considering all the parameters together.

Can Visual Field Progression be Predicted by Confocal Scanning Laser Ophthalmoscopic Imaging of the Optic Nerve Head in Glaucoma? (An American Ophthalmological Society Thesis)

PURPOSE

To determine whether confocal scanning laser ophthalmoscopic imaging (Heidelberg retinal tomography [HRT]) can predict visual field change in glaucoma.

METHODS

The study included 561 patients with glaucoma or ocular hypertension whose clinical course was followed at the Mount Sinai Faculty practice. Humphrey visual fields (HVFs) and HRT images were collected on one randomly selected eye per patient. Glaucoma progression was determined by the presence of two sequential statistically significant negative slopes in mean deviation (MD) or visual field index (VFI) at any point during the study period. Trend-based analysis on HRT parameters was used to determine progressive changes and whether these occurred before or after HVF change. Sensitivity and specificity of HRT to predict HVF change were calculated. HVF rate of change was correlated to the rate of change detected by HRT imaging.

RESULTS

Approximately 17% of patients progressed by either MD or VFI criteria. MD and VFI correlated highly and identified overlapping sets of patients as progressing. HRT global parameters had poor sensitivity (∼42%) and moderate specificity (∼67%) to predict HVF progression. Regional stereometric parameters were more sensitive (69%-78%) but significantly less specific (24%-27%). Sensitivity of global stereometric parameters in detecting HVF change was not significantly affected by the level of visual field damage (P=.3, Fisher exact test). HVF rate of change did not correlate with rate of change of HRT parameters.

CONCLUSIONS

Trend-based analysis of HRT parameters has poor sensitivity and specificity in predicting HVF change. This may be related specifically to HRT imaging or may reflect the fact that in some patients with glaucoma, functional changes precede structural alterations.

Can an inexperienced observer accurately plot disc contours using Heidelberg retinal Tomograph?

OBJECTIVE

To examine the reliability of inexperienced observers in plotting optic disc contours on Heidelberg retinal tomography images before and after training.

DESIGN

Observational study.

PARTICIPANTS

One hundred eyes that were randomly selected from the Singapore Indian Eye Study.

METHODS

Both eyes of subjects were imaged with Heidelberg Retina Tomograph 3 (HRT-3; Heidelberg Engineering, Heidelberg, Germany). Optic disc contours were plotted on the same images by 2 new observers on 2 separate occasions, before and after 2-hour standardized training on the skills and tools available to accurately identify and delineate optic disc contours. These plottings were compared with an experienced, trained glaucoma expert (gold standard). Agreement and variability were analyzed by interclass correlation tests and Bland-Altman plots.

RESULTS

A total of 182 images (18 excluded because of poor quality) from 89 Indian subjects were included. The mean age was 53.27 ± 7.25 years and 54.8% were male. There was moderate-to-high agreement between pretraining (both new observers) and experienced observer's results (interclass correlation values range, 0.76-0.99). The interclass correlation improved for all the HRT-3 parameters after the 2 new observers were adequately trained. Comparing the interclass correlation values before and after training, the differences for mean retinal nerve fibre layer thickness for Observer 1 and all the HRT-3 parameters for Observer 2 were statistically significant.

CONCLUSIONS

This study shows that it is easy to train a new inexperienced observer to plot optic disc contours on HRT images, which translates into improved and acceptable interobserver variability and agreement.

Topographical Analysis of Non-Glaucomatous Myopic Optic Discs Using a Confocal Scanning Laser Ophthalmoscope (TopSS)

OBJECTIVES

To show normative data of optic discs and the mechanism of glaucoma in people with myopia.

DESIGN

Cross-sectional study.

PARTICIPANTS

This study investigated 89 Korean adults with myopia but without glaucoma.

METHODS

Patients were divided into three groups according to the refractive error: low, moderate, and high; and axial length: normal or below normal length, moderately long, and extremely long. Optic disc variables were obtained by confocal scanning laser ophthalmoscope and compared among groups.

RESULTS

The optic disc parameters have a correlation between the refractive error and the optic disc parameters such as average depth, volume below, and half-depth volume. Those parameters also decreased as the axial length increased. The thickness of the volume above decreased significantly as the axial length increased, but a similar relationship was not evident with the refractive error change. In addition, the optic disc parameters were analyzed with respect to the 12 clockwise directions.

CONCLUSIONS

Analyses of optic disc parameters provided by TopSS™ revealed the height of the disc decreased as the myopic refractive error and/or axial length increased. The RNFL bundle became compacted in the thinner disc of the myopic population. This could be an explanation for the fragility of the RNFL in the myopic population. The 12 radial section analyses revealed the shallow cupping at the temporal side in the high-myopic, very-long-axis group. The neuroretinal rim (NRR) height significantly decreased at the superior and inferior sides. These findings suggest that the RNFL bundle should be under high mechanical strain in these sectors.

Laser scanning tomography in the EPIC-Norfolk Eye Study: principal components and associations

PURPOSE

To describe Heidelberg Retina Tomograph (HRT) measures, their principal components, and their associations in a British population.

METHODS

The European Prospective Investigation of Cancer (EPIC)-Norfolk Eye Study is nested within a multicenter cohort study. Measurements were taken with the HRT-2 and the software subsequently updated to yield HRT-3 parameters. Principal components analysis (PCA) was used to identify distinct components of the HRT variables. Generalized estimating equation models were used to examine associations of these components with age, sex, height, body mass index (BMI), blood pressure, social class, education, alcohol intake, smoking status, axial length, IOP, and lens status.

RESULTS

Complete data were available from 10,859 eyes of 6430 participants with a mean age of 68 years. Principal components analysis identified three components with an eigenvalue greater than 1, explaining 79.9% of the variance of all the HRT measures. These were named cup, retinal nerve fiber layer (RNFL), and rim based on the factor loadings they were most correlated with. Older age was significantly associated with a greater cup (P = 0.003), smaller RNFL (P < 0.001), and smaller rim (P < 0.001). Female sex (P = 0.001), higher education (P < 0.001), and shorter axial length (P < 0.001) were associated with a greater RNFL. Lower BMI and higher IOP were associated with a greater cup (both, P < 0.001) and a smaller rim (BMI, P = 0.001; IOP, P < 0.001).

CONCLUSIONS

Heidelberg Retina Tomograph measures in this cohort were largely explained by three principal components related to optic disc cup, RNFL, and rim. Associations with cup and rim were distinct to associations with RNFL, suggesting different underlying determinants.

Validity of a new optic disc grading software for use in clinical and epidemiological research

BACKGROUND

To determine the reliability and agreement of a new optic disc grading software program for use in clinical, epidemiological research.

DESIGN

Reliability and agreement study.

SAMPLES

328 monoscopic and 85 stereoscopic optic disc images.

METHODS

Optic disc parameters were measured using a new optic disc grading software (Singapore Optic Disc Assessment) that is based on polynomial curve-fitting algorithm. Two graders independently graded 328 monoscopic images to determine intergrader reliability. One grader regraded the images after 1 month to determine intragrader reliability. In addition, 85 stereo optic disc images were separately selected, and vertical cup-to-disc ratios were measured using both the new software and standardized Wisconsin manual stereo-grading method by the same grader 1 month apart. Intraclass correlation coefficient (ICC) and Bland-Altman plot analyses were performed.

MAIN OUTCOME MEASURES

Optic disc parameters.

RESULTS

The intragrader and intergrader reliability for optic disc measurements using Singapore Optic Disc Assessment was high (ICC ranging from 0.82 to 0.94). The mean differences (95% limits of agreement) for intergrader vertical cup-to-disc ratio measurements were 0.00 (-0.12 to 0.13) and 0.03 (-0.15 to 0.09), respectively. The vertical cup-to-disc ratio agreement between the software and Wisconsin grading method was extremely close (ICC = 0.94). The mean difference (95% limits of agreement) of vertical cup-to-disc ratio measurement between the two methods was 0.03 (-0.09 to 0.16).

CONCLUSIONS

Intragrader and intergrader reliability using Singapore Optic Disc Assessment was excellent. This software was highly comparable with standardized stereo-grading method. Singapore Optic Disc Assessment is useful for grading digital optic disc images in clinical, population-based studies.

Glaucoma Diagnosis and Monitoring Using Advanced Imaging Technologies

Advanced ocular imaging technologies facilitate objective and reproducible quantification of change in glaucoma but at the same time, impose new challenges on scientists and clinicians for separating true structural change from imaging noise. This review examines time-domain and spectral-domain optical coherence tomography, confocal scanning laser ophthalmoscopy and scanning laser polarimetry technologies and discusses the diagnostic accuracy and the ability of each technique for evaluation of glaucomatous progression. A broad review of the current literature reveals that objective assessment of retinal nerve fiber layer, ganglion cell complex and optic nerve head topography may improve glaucoma monitoring when used as a complementary tool in conjunction with the clinical judgment of an expert.

Test-retest variability in structural parameters measured with glaucoma imaging devices

In addition to classical stereo-disc photography, various glaucoma imaging devices were developed in the last two decades to quantitatively measure and record glaucoma-related structural parameters of the eye. In determining whether or not the glaucomatous damage progressed from baseline and in estimating the number of test results' optimal frequency needed to confirm disease progression, information relating to the test-retest variability of measurement results provided by each imaging device is indispensable. Such information enables the clinician to apply these devices in practice. The test-retest variability of a system is usually estimated using the Bland-Altman analysis and by calculating the coefficient of variation (CV), intraclass correlation coefficient (ICC), and minimum detectable changes (MDC). The reported CV, ICC, and MDC values for glaucoma-related structural parameter measurement results of stereo-disc photographs, confocal scanning laser ophthalmoscopes, scanning laser polarimeters, time-domain optical coherence tomography (OCT), spectral-domain OCT (SD-OCT), anterior-segment OCT, and ultrasound biomicroscope are systematically reviewed in this manuscript, which will enable the clinician to interpret measurement results provided by each glaucoma imaging devices and thus be useful in practice. Although SD-OCT systems may be currently prevailing because of the volume of information provided and the relatively better test-retest variability, these systems need improvement in their test-retest variability measurement capabilities.

Comparison of spectral-domain optical coherence tomography and Heidelberg retina tomograph III optic nerve head parameters in glaucoma

BACKGROUND

To evaluate the agreement between the optic nerve head (ONH) measurements obtained by spectral-domain optical coherence tomography/scanning laser ophthalmoscope (SD-OCT/SLO) and the Heidelberg retinal tomograph III (HRT-III), and to compare the ONH measurements of both devices in different glaucoma types.

METHODS

In a prospective study, 30 patients with primary open-angle glaucoma (POAG) and 30 patients with pseudoexfoliation glaucoma (PXG) were enrolled. All patients underwent SD-OCT/SLO and HRT-III evaluation of the ONH during the same visit. Agreement between measurements of SD-OCT/SLO and HRT-III were evaluated by determination of intraclass correlation coeficients. In addition, mean ONH measurements obtained with both devices were compared between POAG and PXG patient groups.

RESULTS

Mean SD-OCT/SLO measurements were greater than those of HRT-III, except for mean and maximum cup depth. Intraclass correlation coefficient values for disc area, rim area, cup area, cup/disc area ratio, mean cup depth and maximum cup depth were 0.367, 0.213, 0.632, 0.681, 0.775 and 0.661, respectively. No significant differences were found between ONH parameters of POAG and PXG patients as measured with both devices (p > 0.01).

CONCLUSION

ONH measurements with SD-OCT/SLO and HRT-III did not show clinically acceptable agreement in glaucoma patients. This precludes interchangeable use of these measurements in clinical practice. PXG and POAG patient groups displayed similar ONH measurements with both devices.

Heidelberg Retina Tomograph (HRT3) in population-based epidemiology: normative values and criteria for glaucomatous optic neuropathy

PURPOSE

To establish normative values for Heidelberg Retina Tomograph (HRT3) variables and to develop HRT3-based criteria for glaucomatous optic neuropathy for epidemiological research in a white population.

METHODS

Consecutive participants in the Rotterdam Study were examined with HRT and simultaneous stereoscopic fundus photography (ImageNet) in addition to other ophthalmic examinations including intraocular pressure (IOP) measurements and perimetry. Normative values for all HRT3 variables were determined in participants who met all the following criteria: no glaucomatous visual field loss (GVFL), an IOP of 21mmHg or less, no IOP lowering treatment, and a negative family history of glaucoma. Sensitivity was determined in participants with glaucomatous visual field loss at a fixed high specificity of 97.5% - a value commonly used in population-based epidemiology.

RESULTS

A total of 2516 participants were included in this study of whom 66 had glaucomatous visual field loss in at least one eye and 1680 fulfilled the criteria for contributing to the normative values. The HRT3 linear cup-disc ratio (LCDR) variable, adjusted for disc area, showed the highest sensitivity, 35%, at the required specificity of 97.5%. The 97.5th percentile of the LCDR was 0.67 for small discs (up to 1.5 mm(2)), 0.71 [corrected] for medium-sized discs and 0.76 [corrected] for large discs (above 2.0 mm(2)).The HRT3 Glaucoma Probability Score and previously published linear discriminant functions showed a lower sensitivity than LCDR at this specificity.

CONCLUSIONS

At the high specificity of 97.5% as is commonly used in population-based epidemiology, the sensitivity of the HRT3 is low - albeit not lower than that of the vertical cup-disc ratio as assessed with simultaneous stereoscopic fundus photography and analyzed with the ImageNet software. The LCDR variable, stratified for disc area, seems to be the most suitable variable to develop criteria for glaucomatous optic neuropathy for epidemiological purposes.

Digital versus film stereo-photography for assessment of the optic nerve head in glaucoma and glaucoma suspect patients

PURPOSE

One of the leading methods for optic nerve head assessment in glaucoma remains stereoscopic photography. This study compared conventional film and digital stereoscopy in the quantitative and qualitative assessment of the optic nerve head in glaucoma and glaucoma suspect patients.

METHODS

Fifty patients with glaucoma or suspected glaucoma underwent stereoscopic photography of the optic nerve head with a 35-mm color slide film and a digital camera. Photographs/images were presented in random order to 3 glaucoma specialists for independent analysis using a standardized assessment form. Findings for the following parameters were compared among assessors and between techniques: cup/disc (C/D) ratio, state of the optic rim, presence of peripapillary atrophy and appearance of the retinal nerve fiber layer, blood vessels, and lamina cribrosa. The film-based and image-based diagnoses (glaucoma yes/no) were compared as well.

RESULTS

Despite high level of agreement across graders using the same method for the horizontal and vertical C/D ratio, (intraclass correlations 0.80 to 0.83), the agreement across graders was much lower for the other parameters using the same method. Similarly the agreement between the findings of the same grader using either method was high for horizontal and vertical C/D ratio, but low for the other parameters. The latter differences were reflected in the disagreement regarding the final diagnosis: The diagnoses differed by technique for each grader in 18% to 46% of eyes, resulting in 38.5% of eyes diagnosed with glaucoma by film photography that "lost" their diagnosis on the digital images, whereas 18.7% of eyes diagnosed as nonglaucomatous by film photography were considered to have glaucoma on the digital images.

CONCLUSIONS

Although there is consistency between 35-mm film stereoscopy and digital stereoscopy in determining the cup/disc (C/D) ratio, in all other parameters large differences exist, leading to differences in diagnosis. Differences in capturing images between digital and film photography may lead to loss of information and misdiagnosis. Further studies are needed to determine the reliability of the new digital techniques.

Glaucomatous optic neuropathy evaluation project: a standardized internet system for assessing skills in optic disc examination

BACKGROUND

Development of a standardized internet-based system to self-assess skills in optic disc examination for glaucoma risk assessment.

DESIGN

Prospective internet-based observational study.

PARTICIPANTS

Total of 197 participants (glaucoma subspecialists, general ophthalmologists and trainees) from 22 countries.

METHODS

Forty-two optic disc images demonstrating a range of features were selected from 2500 monoscopic disc photographs of normal and glaucomatous eyes. Images were presented to clinicians via website (http://www.gone-project.com). Participants were asked to assess nine topographic features and make a subjective assessment of glaucoma likelihood.

MAIN OUTCOME MEASURES

Inter-observer agreement using kappa (κ) or weighted kappa (κ(w) ).

RESULTS

There was substantial level of inter-observer agreement between glaucoma subspecialists for assessment of glaucoma likelihood (κ(w) = 0.63). Inter-observer agreement was high for haemorrhage (κ= 0.83) and substantial for disc size, disc shape, cup:disc ratio, peripapillary atrophy and cup shape (κ(w) = 0.59-0.68). Subspecialists had stronger inter-observer agreement for glaucoma likelihood and for most disc characteristics than did trainees: the greatest difference being the assessment for retinal nerve fibre layer loss. Analysis of individual disc answers from ophthalmology trainees showed that discs leading to lower agreement of glaucoma likelihood tend to produce lower agreement for the assessment of cup:disc ratio, cup shape, cup depth and retinal nerve fibre layer. Discs with features of moderate to deep cup or cup:disc ratio between 0.6 and 0.8 also lead to lower agreement in glaucoma likelihood.

CONCLUSIONS

This internet-based system is a readily accessible and standardized tool, for clinicians globally, that permits self-assessment and benchmarking of skills in optic disc examination.

Role of imaging in glaucoma diagnosis and follow-up

The purpose of the review is to provide an update on the role of imaging devices in the diagnosis and follow-up of glaucoma with an emphasis on techniques for detecting glaucomatous progression and the newer spectral domain optical coherence tomography instruments. Imaging instruments provide objective quantitative measures of the optic disc and the retinal nerve fiber layer and are increasingly utilized in clinical practice. This review will summarize the recent enhancements in confocal scanning laser ophthalmoscopy, scanning laser polarimetry, and optical coherence tomography with an emphasis on how to utilize these techniques to manage glaucoma patients and highlight the strengths and limitations of each technology. In addition, this review will briefly describe the sophisticated data analysis strategies that are now available to detect glaucomatous change overtime.

Automated segmentation of optic disc region on retinal fundus photographs: Comparison of contour modeling and pixel classification methods

The automatic determination of the optic disc area in retinal fundus images can be useful for calculation of the cup-to-disc (CD) ratio in the glaucoma screening. We compared three different methods that employed active contour model (ACM), fuzzy c-mean (FCM) clustering, and artificial neural network (ANN) for the segmentation of the optic disc regions. The results of these methods were evaluated using new databases that included the images captured by different camera systems. The average measures of overlap between the disc regions determined by an ophthalmologist and by using the ACM (0.88 and 0.87 for two test datasets) and ANN (0.88 and 0.89) methods were slightly higher than that by using FCM (0.86 and 0.86) method. These results on the unknown datasets were comparable with those of the resubstitution test; this indicates the generalizability of these methods. The differences in the vertical diameters, which are often used for CD ratio calculation, determined by the proposed methods and based on the ophthalmologist's outlines were even smaller than those in the case of the measure of overlap. The proposed methods can be useful for automatic determination of CD ratios.

Agreement between the Heidelberg Retina Tomograph (HRT) stereometric parameters estimated using HRT-I and HRT-II

PURPOSE

To assess agreement between Heidelberg Retina Tomograph (HRT)-I and HRT-II stereometric parameters and to determine whether parabolic error correction (PEC) to the topographies improves agreement.

METHODS

University of California San Diego Diagnostic Innovations in Glaucoma Study participants with two HRT-II examinations (n = 380) or one HRT-I and one HRT-II examinations (n = 344) acquired on the same day were included. From the group of 380 eyes, 200 eyes were randomly selected to estimate the repeatability coefficients of HRT-II rim area and volume, cup area and volume, and mean retinal nerve fiber layer (RNFL) thickness parameters (HRT-II control group), and the remaining 180 eyes were used to assess agreement between two HRT-II examinations (HRT-II study group). Agreement between stereometric parameters of HRT-I and HRT-II examinations (HRT-I vs. HRT-II study group) were assessed with (1) no PEC, (2) HRT PEC, and (3) a modified PEC. Bland-Altman plots were used to assess agreement using estimates of bias and clinical limits of agreement (CLA) based on repeatability coefficients.

RESULTS

In the HRT-II study group, agreement between stereometric parameters was good, with no statistically significant biases. For all parameters, differences were within the CLA in 94% of participants. In the HRT-I vs. HRT-II study group, there was a small statistically significant bias between the stereometric parameters, but all differences were within CLA for ≥95% of participants. In both study groups, PEC did not improve agreement.

CONCLUSIONS

Agreement between HRT-I and HRT-II stereometric parameters was good, and PEC did not improve agreement. These results suggest that HRT-I and HRT-II examinations can be used interchangeably to detect changes in stereometric parameters over time.

Agreement between Heidelberg Retina Tomograph-I and -II in detecting glaucomatous changes using topographic change analysis

PURPOSE

To evaluate the suitability of including both Heidelberg Retina Tomograph-I (HRT-I) and HRT-II examinations in the same longitudinal series for HRT topographic change analysis (TCA) and to evaluate parabolic error correction (PEC) to improve the agreement between HRT-I and HRT-II examinations.

METHODS

A total of 66 eyes from the University of California San Diego Diagnostic Innovations in Glaucoma Study with baseline HRT-I and HRT-II examinations obtained on the same day and ≥ 3 HRT-II follow-up examinations were included. Two TCA analyses, HRT-I examination at baseline (HRT-I-mixed series) and HRT-II examination at baseline (HRT-II-only series) were compared. Agreement between the HRT-I-mixed and HRT-II-only series were estimated using Bland-Altman plots. Agreement was assessed: (1) using the current HRT software settings (PEC applied only to HRT-II-only series), and (2) modified HRT settings (PEC also applied to HRT-I-mixed series).

RESULTS

With current HRT software settings, the HRT-I-mixed series significantly overestimated change locations (ie, red pixels) compared with the HRT-II-only series as indicated by statistically significant proportional biases in the Bland-Altman analysis. By applying PEC to HRT-I-mixed series there were no statistically significant biases in the TCA parameter estimates compared with the HRT-II-only series.

CONCLUSION

In some eyes, HRT-I and HRT-II baseline examinations are not interchangeable in TCA analysis without parabolic error correction. HRT-I-mixed series detected more changes characteristic of glaucoma when there were only minimal changes in the HRT-II-only series. Our results suggest that in the majority of cases, with PEC, HRT-I examinations may be included in a longitudinal series containing HRT-II examinations.

Correlation of disc morphology quantified on stereophotographs to results by Heidelberg Retina Tomograph II, GDx variable corneal compensation, and visual field tests

PURPOSE

Quantification of disc morphology by computer-assisted planimetry on stereophotographs and to compare the results with those by other fundus structure imaging devices and visual field tests.

DESIGN

Cross-sectional study.

PARTICIPANTS

Fifteen normal eyes of 15 normal subjects and 69 glaucomatous eyes of 69 open-angle glaucoma patients.

METHODS

Sequential digital stereophotographs of the optic discs were obtained by a nonmydriatic fundus camera. Intraobserver reproducibility for the 3 repeated traces and interobserver reproducibility for the three individual examiners were determined. The obtained results were correlated to those by Heidelberg Retina Tomograph (HRT) II, GDx with variable corneal compensation (VCC) and visual field tests obtained within 2 months of the photography.

MAIN OUTCOME MEASURES

Optic disc parameters such as disc area, cup area, rim/disc area ratio, vertical cup diameter/disc diameter (Cd/Dd) ratio and rim width/disc diameter (Rw/Dd) ratio at radial arbitrary angles. Interclass correlation coefficients (ICC) and coefficients of variance (CV) for the optic disc parameters and their correlation with HRT II, GDx VCC, and visual field test results.

RESULTS

Intra- and interobserver ICC and CV were 0.88 to 0.99, 1.1% to 9.6%, and 0.72 to 0.98, 2.2% to 11.9% for area parameters, 0.81 to 0.92, 2.4% to 4.0% and 0.71 to 0.78, 4.3% to 5.0% for Cd/Dd ratio, and 0.73 to 0.91, 9.2% to 23.0% and 0.65 to 0.82, 16.9% to 27.2% for Rw/Dd ratios, respectively. Significant correlation was seen between disc and cup area and the corresponding parameters by HRT II (r(s) > 0.81; P<0.001), as well as between vertical Cd/Dd ratio and global TSNIT average by GDx VCC and visual field test indices (|r(s)| > 0.51; P<0.001), and between superior and inferior Rw/Dd ratio and their corresponding sectoral HRT II, GDx VCC and visual field test indices (r(s) > 0.31; P<0.006), respectively. Correlation with visual field test indices was comparable among stereophotographs, HRT II, and GDx VCC parameters.

CONCLUSIONS

Optic disc morphologic parameters quantified on stereophotographs by our software showed satisfactory reproducibility and good correlation with those measured by other fundus structure imaging devices and visual field test results, indicating its clinical usefulness, especially in the mass screening of glaucoma.

[Measurement of the disc area by indirect ophthalmoscopy]

PURPOSE

The aim of this study was to assess the accuracy of measuring the optic disc area by indirect ophthalmoscopy.

PATIENTS AND METHODS

In a prospective clinical trial, 57 eyes of 29 subjects (age 57.3+/-12.1 years) were examined. The refractive error was -0.67+/-2.69D (+3.75D to -8D).The vertical and horizontal disc diameters (DD) were measured using a Haag-Streit slit lamp and 60D-, 78D-, 90D-, and Super Field lenses (Volk Optical, Mentor, USA). Afterwards the disc area was calculated by an ellipse formula (horizontal DD x vertical DD x pi/4). The magnification factor given by the manufacturer was taken into account for each lens: 1.15x (60D lens), 0.93x (78D lens), 0.76x (90D lens), and 0.76x (Super Field lens), respectively. As reference for the disc size, the same eyes were examined by HRT II (Heidelberg Engineering, Heidelberg, Germany). Bland-Altman plots were used to assess the agreement between measurements obtained by indirect ophthalmoscopy and HRT.

RESULTS

The results of the disc estimate compared with the HRT measurements were as follows: 0.119+/-0.51 mm(2) (60D lens), 0.224+/-0.57 mm(2) (78D lens), 0.10+/-0.51 mm(2) (90D lens), and -0.07 s+/-0.47 mm(2) (Super Field lens). Differences were not statistically significant for the 60D (p=0.083), 90D (p=0.147), or Super Field lenses (p=0.257). However, the difference between the 78D lens and HRT was statistically significant (Student's t-test; P=0.004).

CONCLUSION

Measuring the disc size by indirect ophthalmoscopy is possible. The 90D lens showed the smallest and the 78D lens the largest deviation.

Variance owing to observer, repeat imaging, and fundus camera type on cup-to-disc ratio estimates by stereo planimetry

OBJECTIVE

To determine and compare variance components in linear cup-to-disc ratio (LCDR) estimates by computer-assisted planimetry by human experts, and automated machine algorithm (digital automated planimetry).

DESIGN

Prospective case series for evaluation of planimetry.

PARTICIPANTS

Forty-four eyes of 44 consecutive patients from the outpatient Glaucoma Service at University of Iowa with diagnosis of glaucoma or glaucoma suspect were studied.

METHODS

Six stereo pairs of optic nerve photographs were taken per eye: 3 repeat stereo pairs using simultaneous fixed-stereo base fundus camera (Nidek 3Dx) and another 3 repeat stereo pairs using sequential variable-stereo base fundus camera (Zeiss). Each optic disc stereo pair was digitized and segmented into cup and rim by 3 glaucoma specialists (computer-assisted planimetry) and using a computer algorithm (digital automated planimetry), and LCDR was calculated for each segmentation (either specialist or algorithm). A linear mixed model was used to estimate mean, SD, and variance components of measurements.

MAIN OUTCOME MEASURES

Average LCDR, interobserver, interrepeat, intercamera coefficients of variation (CV) of LCDR and their 95% tolerance limits.

RESULTS

There was a significant difference in LCDR estimates among the 3 glaucoma specialists. The interobserver CV of 10.65% was larger than interrepeat (6.7%) or intercamera CV (7.6%). For the algorithm, the LCDR estimate was significantly higher for simultaneous stereo fundus images (Nidek, mean: 0.66) than for sequential stereo fundus images (Zeiss, mean: 0.64), whereas interrepeat CV for Nidek (4.4%) was lower than Zeiss (6.36%); the algorithm's interrepeat and intercamera CV were 5.47% and 7.26%, respectively.

CONCLUSIONS

Interobserver variability was the largest source of variation for glaucoma specialists, whereas their interrepeat and intercamera variability is comparable with that of the algorithm. DAP reduces variability on LCDR estimates from simultaneous stereo images, such as the Nidek 3Dx.

[Optic nerve head analysis using the confocal scanning laser ophthalmoscope (CSLO) of big cups versus normal cups]

PURPOSE

To comparatively evaluate the optic nerve head (ONH) using the confocal scanning laser ophthalmoscope (CSLO) in normal eye subjects (NE) versus big cup (BC) patients.

MATERIALS AND METHODS

A total of 83 NEs (43 subjects) and 44 BC eyes (22 patients) were evaluated in this prospective study. The ONH was imaged using the CSLO (HRT-II; Heidelberg Engineering, Heidelberg, Germany). Disc area, cup area, rim area, rim volume (horizontal integrated rim volume), cup/disc ratio, and cup volume were evaluated. Additionally, cup depth and retinal nerve fiber layer thickness (RNFL th) and csa (RNFL csa) of the ONH were also measured. These ONH parameters were compared between normal subjects and eyes with a big cup. We used the t test for statistical analysis.

RESULTS

The mean age was 46 and 44 years in the NE and BC groups, with a sex ratio of 1.26 and 0.69 (p=0.381), respectively. There was a significant difference in disc area (2.07+/-0.38, 2.73+/-0.45, p<0.01), cup area (0.52+/-0.8, 1.26+/-0.3, p<0.01), cup volume (0.14+/-0.14, 0.44+/-0.16, p<0.01), cup/disc ratio (0.24+/-0.1, 0.69+/-0.1, 0.46+/-0.06, p<0.01), Me cup disc (0.24+/-0.1, 0.37+/-0.1, p<0.01), rim volume (0.44+/-0.1, 0.34+/-0.09, p<0.01), rim area (1.56+/-0.27, 1.45+/-0.27, p=0.345), RNFL th (0.28+/-0.05, 0.23+/-0.05, p<0.001) and CSM-0.19+/-0.14,-0.08+/-0.06, p<0.001), RNFL csa (1.43+/-0.25, 1.37+/-0.29, p=0.849) in normal vs big cup eyes, respectively.

DISCUSSION

There is correlation between the cup and disc areas. Parameters evaluating the retinal nerve fiber were comparable in the two groups, except the RNFL thickness, which was lower in the BC group. This may be explained by the larger surface of the optic disc in this group.

CONCLUSION

HRT parameters might be part of the criteria favoring the physiologic character of big cups.

Sensitivity and specificity with the glaucoma probability score in Heidelberg Retina Tomograph II in Japanese eyes

PURPOSE

To evaluate the sensitivity and specificity with glaucoma probability score (GPS), a new glaucoma diagnosis program in Heidelberg Retina Tomograph II ver 3.0 in detecting early glaucoma in Japanese eyes.

PATIENTS AND METHODS

Sensitivity and specificity with GPS were determined in 148 early stage open-angle glaucoma eyes of 148 hospital-based patients (mean deviation >-5 dB) and age-matched and refraction-matched 148 eyes of 148 ophthalmologically normal subjects selected from a population-based sample of the Tajimi Study. The results with GPS were compared with those with Moorfields regression analysis (MRA) and FS Mikelberg discriminant function (FSM). Analysis was also carried out by stratifying the eyes by refraction and disc area.

RESULTS

For all glaucoma eyes, sensitivity with GPS (71.7%) was similar to MRA and FSM (75.2%, P=0.8; 77.2%, P=0.5; Fisher exact test). Specificity with GPS (80.7%) was also statistically equivalent with MRA and FSM (91.1%, P=0.06; 87.4%, P=0.3). Although sensitivity with GPS was lower in higher myopic eyes than in the less myopic eyes, specificity did not differ with refractive error. Specificity with GPS was lower and sensitivity higher (P<0.0001) in larger discs (>2.0 mm) than in smaller discs (< or =2 mm). In larger discs, specificity with GPS was lower (P=0.005) than with MRA and in smaller discs, sensitivity with GPS was lower than with MRA or FSM.

CONCLUSIONS

In Japanese eyes, sensitivity and specificity with GPS to detect early glaucoma were similar with MRA or FSM, although the specificity tended to be lower than results in white eyes. Subgroup analysis revealed a notable dependency in diagnostic ability on disc area and refractive error. The diagnostic ability of GPS is still not sufficient for glaucoma detection in the Japanese population requiring further investigation.

Intracranial pressure in primary open angle glaucoma, normal tension glaucoma, and ocular hypertension: a case-control study

PURPOSE

To compare intracranial pressure (ICP) in subjects with primary open-angle glaucoma (POAG), normal-tension glaucoma (NTG; subset of POAG), and ocular hypertension (OHT) with that in subjects with no glaucoma.

METHODS

The study was a retrospective review of medical records of 62,468 subjects who had lumbar puncture between 1985 and 2007 at the Mayo Clinic. Of these, 57 POAG subjects, 11 NTG subjects (subset of POAG), 27 OHT subjects, and 105 control subjects met the criteria and were analyzed. A masked comparison of the relationship between ICP and other ocular and nonocular variables was performed by using univariate and multivariate analyses.

RESULTS

ICP was significantly lower in POAG compared with age-matched control subjects with no glaucoma (9.1 +/- 0.77 mm Hg vs. 11.8 +/- 0.71 mm Hg; P < 0.0001). Subjects with NTG also had reduced ICP compared with the control subjects (8.7 +/- 1.16 mm Hg vs. 11.8 +/- 0.71 mm Hg; P < 0.01). ICP was higher in OHT than in age-matched control subjects (12.6 +/- 0.85 mm Hg vs. 10.6 +/- 0.81 mm Hg; P < 0.05).

CONCLUSIONS

ICP is lower in POAG and NTG and elevated in OHT. ICP may play an important role in the development of POAG and NTG and in preventing the progression of OHT to POAG. Further prospective and experimental studies are warranted to determine whether ICP has a fundamental role in the pathogenesis of glaucoma.

Heidelberg Retina Tomograph 3 machine learning classifiers for glaucoma detection

AIMS

To assess performance of classifiers trained on Heidelberg Retina Tomograph 3 (HRT3) parameters for discriminating between healthy and glaucomatous eyes.

METHODS

Classifiers were trained using HRT3 parameters from 60 healthy subjects and 140 glaucomatous subjects. The classifiers were trained on all 95 variables and smaller sets created with backward elimination. Seven types of classifiers, including Support Vector Machines with radial basis (SVM-radial), and Recursive Partitioning and Regression Trees (RPART), were trained on the parameters. The area under the ROC curve (AUC) was calculated for classifiers, individual parameters and HRT3 glaucoma probability scores (GPS). Classifier AUCs and leave-one-out accuracy were compared with the highest individual parameter and GPS AUCs and accuracies.

RESULTS

The highest AUC and accuracy for an individual parameter were 0.848 and 0.79, for vertical cup/disc ratio (vC/D). For GPS, global GPS performed best with AUC 0.829 and accuracy 0.78. SVM-radial with all parameters showed significant improvement over global GPS and vC/D with AUC 0.916 and accuracy 0.85. RPART with all parameters provided significant improvement over global GPS with AUC 0.899 and significant improvement over global GPS and vC/D with accuracy 0.875.

CONCLUSIONS

Machine learning classifiers of HRT3 data provide significant enhancement over current methods for detection of glaucoma.

Automated assessment of the optic nerve head on stereo disc photographs

PURPOSE

To develop automated software for optic nerve head (ONH) quantitative assessment from stereoscopic disc photographs and to evaluate its performance in comparison with human expert assessment.

METHODS

A fully automated system, including three-dimensional ONH modeling, disc margin detection, cup margin detection, and calculation of stereometric ONH parameters, was developed and tested. One eye each from 54 subjects (23 healthy, 17 suspected glaucoma, and 14 glaucoma) was enrolled. The majority opinion of three experts defined disc and cup margins on the disc photographs was used for comparison. Seven ONH parameters, disc area, rim area, rim volume, cup area, cup volume, cup-to-disc (C/D) area ratio, and vertical C/D ratio, were computed based on both machine- and expert-defined margins and compared between the methods.

RESULTS

All automated ONH measurements showed good correlation with the expert defined margins (Pearson r = 0.90, disc area; 0.56, rim area; 0.78, rim volume; 0.88, cup area; 0.93, cup volume; 0.69, C/D area ratio; and 0.67, vertical C/D ratio; all P <or= 0.0001). No statistically significant difference was found in the glaucoma-discriminating ability of all seven ONH parameters (P >or= 0.21). The mean or median of automatically defined disc and cup areas was significantly higher than the subjective assessment (disc area P = 0.0001, t-test; cup area P = 0.036, Wilcoxon signed ranks test), although they had high correlation coefficients. The software failed to detect the disc margin for all the disc photographs with peripapillary atrophy.

CONCLUSIONS

The automated ONH analysis method provides an objective and quantitative ONH evaluation using widely available stereo disc photographs.

Comparison of optic disc margin identified by color disc photography and high-speed ultrahigh-resolution optical coherence tomography

OBJECTIVE

To determine the correspondence between optic disc margins evaluated using disc photography (DP) and optical coherence tomography (OCT).

METHODS

From May 1, 2005, through November 10, 2005, 17 healthy volunteers (17 eyes) had raster scans (180 frames, 501 samplings per frame) centered on the optic disc taken with stereo-optic DP and high-speed ultrahigh-resolution OCT (hsUHR-OCT). Two image outputs were derived from the hsUHR-OCT data set: an en face hsUHR-OCT fundus image and a set of 180 frames of cross-sectional images. Three ophthalmologists independently and in a masked, randomized fashion marked the disc margin on the DP, hsUHR-OCT fundus, and cross-sectional images using custom software. Disc size (area and horizontal and vertical diameters) and location of the geometric disc center were compared among the 3 types of images.

RESULTS

The hsUHR-OCT fundus image definition showed a significantly smaller disc size than the DP definition (P <.001, mixed-effects analysis). The hsUHR-OCT cross-sectional image definition showed a significantly larger disc size than the DP definition (P <.001). The geometric disc center location was similar among the 3 types of images except for the y-coordinate, which was significantly smaller in the hsUHR-OCT fundus images than in the DP images.

CONCLUSION

The optic disc margin as defined by hsUHR-OCT was significantly different than the margin defined by DP.

Glaucoma probability score vs Moorfields classification in normal, ocular hypertensive, and glaucomatous eyes

PURPOSE

To evaluate the Heidelberg Retina Tomograph III (HRT III) glaucoma probability score in differentiating normal from pathologic eyes and to compare the glaucoma probability score with Moorfields regression analysis (MRA).

DESIGN

Prospective cross-sectional study.

METHODS

Fifty-nine normal, 40 hypertensive, and 83 glaucomatous eyes were examined with Swedish interactive threshold algorithm standard 24-2 visual fields and HRT III. Sensitivity and specificity were evaluated using global and sectorial glaucoma probability score and MRA compared with damage in visual fields. Areas under receiver operating characteristic (ROC) curves were evaluated. Agreement between MRA and glaucoma probability score was calculated using the kappa coefficient. Glaucoma probability score was considered to be displaced when a symbol was outside and the opposite symbol was inside the optic disk.

RESULTS

MRA sensitivity and specificity were 39.8% and 93.2% (most specific criteria) and 68.7% and 83.1% (least specific criteria), respectively. Glaucoma probability score sensitivity and specificity were 71.1% and 69.5% (most specific criteria) and 85.5% and 54.2% (least specific criteria), respectively. Visual field parameters were related to the global (P = .001) and sectorial (P < .05) glaucoma probability score. A displaced glaucoma probability score was found in 35 eyes, but with unchanged glaucoma probability score sensitivity and specificity. Areas under the ROC curves of glaucoma probability score was 0.77. The kappa coefficient was 0.34.

CONCLUSIONS

Glaucoma probability score analysis tends to be more sensitive but less specific than MRA. Glaucoma probability score did not differentiate normal and hypertensives eyes. When displaced, glaucoma probability score sensitivity and specificity were unchanged. MRA and glaucoma probability score agreement was low. Glaucoma probability score is advantageous over MRA in early-stage glaucoma.

Quantification of neuroretinal rim loss using digital planimetry in long-term follow-up of normals and patients with ocular hypertension

PURPOSE

The purpose of this study was to investigate if digital planimetry is appropriate for quantification of neuroretinal rim loss in patients with ocular hypertension (OHT) and if there is an age-related neuroretinal rim loss in normals.

PATIENTS AND METHODS

Fifty-six patients with OHT without optic disc change, 13 patients with OHT and conversion to early glaucoma during follow-up and 42 age-matched controls were recruited from the Erlangen Glaucoma Registry. Annually, all patients underwent complete ophthalmologic examination including detailed diagnostic testing concerning glaucoma. Gold standard for morphologic evaluation of the optic nerve head was the semiquantitative 2-dimensional-method described by Jonas. Optic disc images from baseline and after 5 or 10 years follow-up were used for digital planimetry. Optic disc area and cup area were measured using commercial software: Soft Imaging System analysis. The investigator was masked for diagnosis and time point of examination.

RESULTS

Mean neuroretinal rim loss was 0.36% per year in controls, 0.54% per year in patients with OHT without progressive disease, and 0.95% per year in OHT and conversion.

CONCLUSIONS

Neuroretinal rim loss was highest in the group of OHT with conversion to early glaucoma during follow-up. In the control group we detected a very low mean neuroretinal rim loss during 10-year follow-up. In ocular hypertensive patients without progressive disease mean neuroretinal rim loss was approximately twice compared with normals.

Comparison of parameters from Heidelberg Retina Tomographs 2 and 3

PURPOSE

To compare stereometric parameters and classification results from the Heidelberg Retina Tomograph version 2 (HRT2); HRT3; and HRT3 Glaucoma Probability Score (GPS), an automated method of obtaining optic nerve head analysis without the need for manual definition of disc margin.

DESIGN

Retrospective cross-sectional study.

PARTICIPANTS

Five hundred four eyes from 281 consecutive subjects (glaucoma, glaucoma suspect, and healthy) evaluated in a glaucoma clinic.

METHODS

All participants had HRT2 scanning of the optic nerve head. Inclusion criteria were scans with good centration and focus, even illumination, an overall quality score by HRT3 of acceptable or better, and standard deviation < 50 mum. A Bland-Altman analysis was used for the comparison of HRT2 and HRT3. From these results, calibration equations were determined to permit conversion of the measurements between devices. The agreement between HRT2 and HRT3 Moorfields regression analysis (MRA) and HRT3 GPS classification methods was measured using kappa statistics.

MAIN OUTCOME MEASURES

Heidelberg Retina Tomograph version 2 and HRT3 stereometric parameters, MRA, and global GPS.

RESULTS

There was a statistically significant difference between HRT2 and HRT3 global disc area, rim area, cup area, rim volume, cup volume, height variation contour, and retinal nerve fiber layer cross-sectional area stereometric parameters. All of those parameters were smaller using HRT3, due to a manufacturer-reported horizontal scaling error of 4% in HRT2 that was corrected in HRT3. kappas for agreement were 0.60 between classifications (within normal limits, borderline, and outside normal limits) of MRA by HRT2 and HRT3 and 0.47 between HRT3 MRA and GPS.

CONCLUSIONS

The HRT3 generally provided smaller stereometric disc measurements than HRT2. There was no clear conversion between HRT3 and GPS parameters, as the 2 methods for measuring the stereometric parameters differ.

Diagnosing glaucoma progression: current practice and promising technologies

PURPOSE OF REVIEW

An update on recent work is provided that has broadened our understanding of the evaluation of visual function and structure, and their use in evaluating glaucoma progression.

RECENT FINDINGS

The challenge of determining visual-field progression and the implications of long-term fluctuation are reviewed and data to support the magnitude of the fluctuation are cited. The use of confirmatory testing can limit the over diagnosis of glaucoma progression. Focusing visual-field testing on the locations of present scotomas or using frequency doubling technology may provide new approaches to assessing visual function. New standardized techniques to interpret visual fields, including neural networks, unsupervised machine learning and pointwise linear regression, may provide more quantitative means for visual-field interpretation. These techniques, along with structural evaluation of the optic nerve and nerve fiber layer, are essential in glaucoma management. Optic-nerve-head photography is still a mainstay in evaluating glaucoma progression, although many technologies including scanning laser tomography, scanning laser polarimetry and optical coherence tomography offer more quantitative means to follow structural change. These modalities, in different ways, show promise in providing additional information regarding the stability of glaucoma.

SUMMARY

Identifying the functional visual component as well as structural changes is essential in evaluating glaucoma progression. New techniques of testing and evaluating visual fields, the optic-nerve head, and the retinal nerve fiber layer offer exciting opportunities to more accurately identify glaucoma progression, and are likely to become more central as imaging devices and software support develop further.

Linear Regression Modeling of Rim Area to Discriminate Between Normal and Glaucomatous Optic Nerve Heads

PURPOSE

To create a diagnostic algorithm by modeling the normal variability of rim area (measured by Heidelberg Retina Tomography) using linear regression in normal elderly subjects.

METHODS

Multiple linear regression was performed between log rim area and disc area, age, and sex in 712 normal elderly subjects. The relationship between log rim area and disc area was not linear and showed significant heteroscedascity (increasing variability of rim area with increasing disc area). These factors violate the assumptions of linear regression as performed by the Moorfields Regression Analysis (MRA), and were overcome by conducting linear regression separately for each disc area quartile. Actual rim area in each disc sector was compared with the 95% lower limit predicted by the analysis in the normal subjects and 58 glaucoma patients.

RESULTS

Specificity and sensitivity of the New Regression Analysis (NRA) was 83% and 81%, respectively (84.4% and 82.7%, respectively for MRA). NRA specificity was unaffected by disc size (81.6% and 86.3% in the smallest and largest disc area quartiles, respectively, P=0.36). MRA specificity was reduced in larger discs (91.1% and 73.1% in smallest and largest disc area quartiles, P<0.001), with reduced sensitivity in smaller discs (58.3% and 85.0% in smallest and largest quartiles, P=0.05).

CONCLUSIONS

Nonlinearity and heteroscedascity in the relationship between log rim area and disc area explain reduced specificity of the MRA in bigger discs as a statistical error. Although overall not offering better diagnostic performance, the NRA performed consistently across different disc sizes, offering better performance than the MRA in large discs.

Agreement between stereophotographic and confocal scanning laser ophthalmoscopy measurements of cup/disc ratio: effect on a predictive model for glaucoma development

OBJECTIVE

To evaluate risk estimates obtained by incorporating confocal scanning laser ophthalmoscopy measurements of cup/disc (C/D) ratio into a previously described and validated predictive model to estimate the risk of glaucoma development in ocular hypertension. These risk estimates were compared with those obtained by the original model in which vertical C/D ratio was estimated from stereophotographs.

DESIGN

Cross-sectional study.

METHODS

The study included 118 eyes of 59 patients with normal optic discs, normal visual fields, and high intraocular pressure. The original predictive model contained information on 6 baseline factors: age, intraocular pressure, central corneal thickness, vertical C/D ratio, visual field pattern standard deviation, and presence of diabetes. Information regarding these baseline factors was collected for each patient. For the original model, vertical C/D ratio was estimated from stereophotographs. The Heidelberg Retina Tomograph (HRT) model used an identical model, except that the parameter linear C/D ratio was used to provide C/D ratio estimates. All patients underwent confocal scanning laser ophthalmoscopy imaging HRT II within 6 months of stereophotographs. The agreement between stereophotograph and HRT II risk estimates was evaluated by Bland-Altman plots.

RESULTS

The difference between HRT II and stereophotograph estimates of C/D ratio was within 0.2 in 95% of the patients. When incorporated into the predictive model, estimates of risk using the HRT II parameter linear C/D ratio were highly correlated to those obtained using stereophotographs (rho=0.954; P<0.001). The 95% limits of agreement were -4.57% to 4.65%.

CONCLUSIONS

HRT II and stereophotograph estimates of C/D ratio can be used interchangeably when incorporated into a predictive model to estimate the risk of conversion from ocular hypertension to glaucoma.

Interobserver variability in confocal optic nerve analysis (HRT)

BACKGROUND/AIMS

To study the observer-related variability of optic nerve head (ONH) measurements using confocal laser scanning tomography (HRT I) in a screening setting.

METHODS

Six experienced glaucoma specialists independently evaluated 50 ONH topographies from 25 adults using HRT software ver. 2.01 in a masked fashion. ONH topographies were obtained from a cohort study of 882 healthy adults and additionally included one patient with glaucomatous eyes. A glaucoma-screening-like setting was intended. The mean interobserver difference was defined as the mean percentual difference between an observer's analysis and the mean of all six observers for all eyes and all observers. The interobserver range was calculated for each eye as the percentual difference between the lowest and highest measurement, with the highest measurement as denominator. Additionally, Kendall's coefficient of rank concordance was assessed for the main HRT parameters.

RESULTS

Mean disc area ranged from 1.83 +/- 0.49 to 2.21 +/- 0.40 mm(2) (mean interobserver difference: 8.3%; interobserver range: 5-50%; rank concordance: 0.86). The lowest mean interobserver differences were found for mean retinal nerve fibre layer thickness (RNFLT; 6.5%), maximum cup depth (2.9%) and cup shape (6.8%). An increased interobserver range was significantly correlated to a low cup to disc area ratio (r=0.64, P<0.0001).

CONCLUSIONS

The observer-dependent diagnostic variability of HRT measurements can lead to divergent diagnostic evaluation of the ONH in a screening setting. Any HRT software relying on a reference database is exposed to relevant observer-related variability of the disc area. For screening purposes, HRT measurements should be completed by other diagnostic methods to compensate for possible diagnostic uncertainty.

Imaging of the optic disc and retinal nerve fiber layer in acute optic neuritis

PURPOSE

To demonstrate whether optical coherence tomography (OCT-3) and scanning laser ophthalmoscopy (HRT-2) can be used to measure changes of the optic disc and peripapillary retinal nerve fiber layer (RNFL) in eyes with acute retrobulbar optic neuritis that have no clinically apparent optic disc swelling. To correlate these findings with presentation magnetic resonance imaging (MRI) of the affected optic nerve.

METHODS

Eight consecutive patients with acute retrobulbar optic neuritis, who had no prior optic neuritis in either eye, were prospectively investigated at presentation and at between 1 and 3 months with clinical examination, OCT-3, HRT-2. At presentation, MRI of the optic nerves were performed in 7/8 patients.

RESULTS

Compared to unaffected eyes, affected eyes without clinically seen optic disc swelling at baseline, there was a non-significant trend to increased thickness in the total RNFL, superior and nasal measurements. Baseline HRT in affected eyes showed smaller mean cup to disc ratio (p=0.003) and a smaller cup area (p=0.002) compared with the unaffected eye. The MRI-demonstrated optic nerve lesion did not correlate with OCT RNFL thickening or HRT decrease of the physiological cup. Follow-up imaging of the affected eyes showed normalization of HRT cup size parameters and OCT RNFL thickness (p<0.04). At follow-up, the temporal RNFL had thinning in 7/8 affected eyes (46.8 mum, p=0.021) compared with fellow unaffected eyes (57.8 mum), which did not change.

CONCLUSION

OCT-3 and HRT demonstrate mild RNFL thickening or optic disc swelling in acute optic neuritis, even when swelling is not seen clinically. OCT-3 appears to reveal measurable RNFL thinning in the temporal quadrant after retrobulbar optic neuritis, even though vision improves. RNFL imaging may be useful in future studies of residual injury after optic neuritis.

Glaucoma detection with the Heidelberg retina tomograph 3

PURPOSE

To compare the ability of the Heidelberg retina tomograph version 3 (HRT 3) and HRT version 2 (HRT 2) to discriminate between healthy and glaucomatous eyes.

DESIGN

Retrospective cross-sectional study.

PARTICIPANTS

Seventy-one eyes of 71 healthy volunteers and 50 eyes of 50 glaucoma patients were studied. The average visual field mean deviation of the glaucoma group was -6.03+/-5.78 dB.

INTERVENTION

All participants had comprehensive ocular examinations, perimetry, and HRT scanning within 6 months. HRT 2 data were analyzed using HRT 3 software without modifying the disc margin.

MAIN OUTCOME MEASURES

Discrimination capabilities between healthy and glaucomatous eyes were determined by areas under the receiver operating characteristics (AROCs) curves. Comparisons between corresponding AROCs obtained by HRT 2 and HRT 3 analyses were performed using the nonparametric DeLong method. Agreement between classifications as defined by the different analysis methods was quantified by kappa analysis.

RESULTS

The individual stereometric parameters with the best discrimination were linear cup/disc ratio (AROC = 0.897; 95% confidence interval [CI], 0.836-0.958) for standard HRT 3 analysis and horizontal retinal nerve fiber layer curvature (0.905) for HRT 3 glaucoma probability score (GPS) analysis. Areas under the receiver operating characteristics for discrimination between glaucomatous and healthy eyes of the overall classification by HRT 2 Moorfields regression analysis (MRA), HRT 3 MRA, and GPS were 0.927 (95% CI, 0.877-0.977), 0.934 (0.888-0.980), and 0.880 (0.812-0.948), respectively. The difference between the 3 AROCs was not significant (P = 0.44). The agreement between HRT 2 and HRT 3 overall MRA classification was good (kappa = 0.70; CI, 0.59-0.80) with HRT 3 tending to report more abnormalities than HRT 2 analysis. The agreement between overall HRT 3 MRA and overall GPS was kappa = 0.58 (CI, 0.45-0.70).

CONCLUSIONS

The glaucoma discriminating ability of the new HRT 3 software is similar to that of the previous generation HRT 2. The GPS analysis showed promising results in differentiating between healthy and glaucomatous eyes without the need for subjective operator input.

Detecting glaucoma with RADAAR: the Bridlington Eye Assessment Project

AIMS

To develop a diagnostic algorithm using rim/disc area asymmetry ratio (RADAAR) to discriminate between normal and glaucomatous eyes using the Heidelberg retina tomograph (HRT).

METHODS

RADAAR was calculated by dividing rim/disc area ratio of the larger disc by that of the smaller disc. RADAAR normal tolerances in each disc sector were generated using 611 population based normal elderly individuals (mean age 72.5 years). Overall status for each individual was equal to the worst disc sector. Sensitivity was assessed in 45 patients with open angle glaucoma.

RESULTS

RADAAR variance was significantly greater in glaucoma compared with normality in all disc sectors (p<0.001). Normal RADAAR limits were not dependent on age or sex. Overall, at the 99th percentile limit, the diagnostic algorithm's specificity was 95.1% with a sensitivity of 55.6%. Males with glaucoma showed greater RADAAR variance than females. Corresponding sensitivity of the diagnostic algorithm was 63.0% and 44.4% in males and females respectively, although this difference was not statistically significant (p = 0.43).

CONCLUSION

The RADAAR diagnostic algorithm detected glaucoma with moderate sensitivity but was limited by significant numbers of glaucoma patients with symmetrical optic disc changes. RADAAR may have greater sensitivity in males than in females although this requires confirmation in further studies.

Keratometry, Optic Disc Dimensions, and Degree and Progression of Glaucomatous Optic Nerve Damage

PURPOSE

To evaluate whether keratometric readings as a measure of corneal shape are associated with optic disc dimensions and with the degree and rate of perimetric progression of chronic open-angle glaucoma or ocular hypertension.

METHODS

The hospital-based observational study included 1826 eyes of 936 patients with ocular hypertension, patients with chronic open-angle glaucoma, or normal individuals. For 733 ocular hypertensive or glaucomatous eyes, follow-up examinations were performed with a mean follow-up time of 58.0+/-34.7 months. Observation procedures were keratometry, morphometric optic disc analysis, tonometry, and perimetry.

RESULTS

In the normal study group, area of the neuroretinal rim, alpha zone and beta zone of parapapillary atrophy, and retinal vessel diameter were not significantly associated with keratometric readings. In the entire study population, the optic disc area was significantly (P<0.001; r=-0.27) correlated with low keratometric readings as expressed in diopters. Keratometric readings were significantly (P<0.001 adjusted for age, intraocular pressure, baseline damage, and corneal asphericity) smaller in the normal-pressure glaucoma group than in the normal study group and in the groups with ocular hypertension or primary and secondary open-angle glaucoma. Rate of perimetric progression was not significantly associated with low keratometric readings, either in simple or in multiple Cox regression analysis, controlling for baseline damage, ocular hypertension, age, corneal asphericity, and intraocular pressure.

CONCLUSIONS

Large optic disc area is statistically significantly, but clinically weakly, correlated with low keratometric readings (diopters). In Caucasian individuals with ocular hypertension and patients with chronic open-angle glaucoma, the rate of development or progression of glaucomatous visual field defects is not significantly associated with keratometric readings.

Optical Coherence Tomography of the Optic Nerve Head

PURPOSE

Optical coherence tomography may be a new technique for quantitative 3-dimensional assessment of the optic nerve head for diagnosis of optic nerve anomalies and diseases such as the glaucomas. The purpose of the present study was to examine its reproducibility.

PATIENTS AND METHODS

The clinical noninterventional study included 10 randomly chosen eyes of 10 healthy individuals who underwent 24 optical coherence tomographic examinations with the high-resolution and fast scan program, interactively corrected or uncorrected. The pupils were not dilated. All examinations were performed by 3 examiners independently of one another. The coefficient of variation was calculated as the ratio of the mean of the standard deviations divided by the mean of the means.

RESULTS

The coefficients of variation for redetermination of optic disc area, ranging between 0.047 and 0.119, were lowest for the manually corrected fast scan mode and highest for the uncorrected fast scan mode. For remeasurements of the neuroretinal rim area, the best mean coefficient of variation was 0.073+/-0.026 (corrected fast scan mode).

CONCLUSIONS

In healthy eyes, the morphometric measurements of the optic nerve head by optical coherence tomography show a relatively high reproducibility with mean coefficients of variation lower than 10% for remeasurements of the optic disc and neuroretinal rim area. With undilated pupils, reproducibility is better with the fast scan mode with interactive correction of the outlining of the optic disc border than it is with the high-resolution mode.

Comparison of optical coherence tomography and fundus photography for measuring the optic disc size

PURPOSE

To assess the agreement and repeatability of optic nerve head (ONH) size measurements by optical coherence tomography (OCT) as compared to conventional planimetry of fundus photographs in normal eyes.

METHODS

For comparison with planimetry the absolute size of the ONH of 25 eyes from 25 normal subjects were measured by both OCT and digital fundus photography (Zeiss FF camera 450). Repeatability of automated Stratus OCT measurements were investigated by repeatedly measuring the optic disc in five normal subjects.

RESULTS

Mean disc size was 1763 +/- 186 vertically and 1632 +/- 160 microm horizontally on planimetry. On OCT, values of 1772 +/- 317 microm vertically (p = 0.82) and a significantly smaller horizontal diameter of 1492 +/- 302 microm (p = 0.04) were obtained. The 95% limits of agreement were (-546 microm; +527 microm) for vertical and (-502 microm; +782 microm) for horizontal planimetric compared to OCT measurements. In some cases large discrepancies existed. Repeatability of automatic measurements of the optic disc by OCT was moderately good with intra-class correlation coefficients (ICC) of 0.78 horizontally and 0.83 vertically. The coefficient of repeatability indicating instrument precision was 80 microm for horizontal and 168 microm for vertical measurements.

CONCLUSIONS

OCT can be used to determine optic disc margins in moderate agreement with planimetry in normal subjects. However, in some cases significant disagreement with photographic assessment may occur making manual inspection advisable. Automatic disc detection by OCT is moderately repeatable.

Specificity of the Heidelberg Retina Tomograph’s Diagnostic Algorithms in a Normal Elderly Population

PURPOSE

To determine the specificity of the Moorfields regression analysis (MRA) and R. Burk (RB) and F. S. Mikelberg (FSM) linear discriminant functions (LDFs) in classifying the optic discs of normal elderly patients imaged with the Heidelberg Retina Tomograph.

DESIGN

Cross-sectional study.

PARTICIPANTS

Optic nerve head analysis of 459 eyes of 459 normal elderly patients was performed by 2 operators. All patients were consecutive in a cohort screened for eye disease. Normals were defined as having a normal visual field (VF) on automated suprathreshold screening, normal intraocular pressure, and normal visual acuity.

MAIN OUTCOME MEASURES

Specificity of the optic nerve head classification by MRA and the RB and FSM LDFs.

RESULTS

Subjects' mean age (262 female, 197 male) was 72.6 years. Males were found to have significantly larger cups than females. Eighty-three percent, 11.3%, and 5.7% of discs were classed as within normal limits, borderline, and outside normal limits by MRA. Percentages of optic discs classified as outside normal limits in the smallest and largest quartiles for disc size were 0.9% and 14.9%. The association of disc size with the classification of outside normal limits was significant in males only. Agreement in MRA classification between 2 investigators separately placing the optic disc contour was moderate (Cohen's kappa = 0.54, P<0.001). The RB and FSM LDFs classified 6.8% and 11.8% of discs as diseased. The RB LDF was significantly more likely to return a diagnosis of disease in larger discs, though only in males. The FSM LDF tended to the same drop in specificity, although not significantly.

CONCLUSIONS

Using the MRA and RB/FSM LDFs, specificity fell dramatically with increasing disc size, particularly in males. This may reflect the finding of significantly larger cups in males than in females in the normal elderly population. This divergence was not predicted by these diagnostic functions, which were developed on samples of younger subjects.