Screening for glaucoma using GDx-VCC in a population with ≥1 risk factors.

OBJECTIVE: To estimate the diagnostic accuracy of the GDx-VCC for glaucoma screening in a population with ≥1 risk factors for glaucoma. DESIGN: Cross-sectional evaluation of a diagnostic test for screening. PARTICIPANTS: Two hundred forty-seven community-based volunteer participants with risk factors for glaucoma. METHODS: The peripapillary retinal nerve fibre layers (RNFL) of participants' eyes were scanned using the GDx-VCC. Based on an ophthalmologic examination and frequency doubling perimetry, eyes were classified into 4 categories: normal, possible glaucoma, probable glaucoma, and definitive glaucoma. The sensitivities, specificities, positive and negative predictive values, and positive and negative likelihood ratios of the RNFL parameters were calculated. RESULTS: The right eyes were retained for analyses. After excluding 5 eyes because of missing data and 23 eyes because of poor scan quality, the data of 219 right eyes were analyzed. Four eyes had definitive glaucoma. The best performing parameter was the nerve fibre indicator using a cutoff of 35 with a sensitivity of 75% (95% CI 19.4-99.4) at a specificity of 95% (95% CI 91.3-97.3), a positive predictive value of 25 (95% CI 4.3-48.1), a negative predictive value of 99 (95% CI 97.5-100.0), a positive likelihood ratio of 16 (95% CI 6.69-32.5), and a negative likelihood ratio of 0.20 (95% CI 0.05-1.44). CONCLUSIONS: The GDx-VCC has inadequate sensitivity for screening of definitive glaucoma.

Screening for glaucoma in high-risk populations using optical coherence tomography.

OBJECTIVE: To estimate the diagnostic accuracy of Stratus optical coherence tomography (OCT) for glaucoma screening in high-risk populations. DESIGN: Cross-sectional evaluation of a diagnostic test for screening. PARTICIPANTS: Three hundred thirty-three community-based volunteer participants with risk factors for glaucoma. METHODS: The optic nerve and peripapillary retinal nerve fiber layer (RNFL) of participants' eyes were scanned using the Stratus OCT. Based on an ophthalmologic examination and frequency doubling perimetry, eyes were classified into 4 categories: normal, possible glaucoma, probable glaucoma, and definitive glaucoma. MAIN OUTCOME MEASURES: The sensitivities, specificities, positive and negative likelihood ratios of the RNFL, optic disc parameters, and their combinations were calculated. RESULTS: The right eyes were retained for analyses. After excluding eyes with missing data or with poor quality scans, the data of 210 right eyes were analyzed. Six eyes had definitive glaucoma. Combining the best performing optic nerve head parameters (cup diameter or cup/disc vertical ratio or cup/disc area ratio) and RNFL parameters (superior average or inferior average or overall average) using AND-logic resulted in a sensitivity of 67% (95% confidence interval [CI], 24%-94%), specificity of 96% (95% CI, 92%-98%), a positive likelihood ratio of 17.08 (95% CI, 7.06-41.4), and a negative likelihood ratio of 0.35 (95% CI, 0.11-1.08). CONCLUSIONS: When adequate quality scans may be obtained, the Stratus has moderate sensitivity and high specificity for definitive glaucoma. Specificity is increased when parameters from both the optic nerve head and RNFL scans are combined.

Validity of screening for glaucomatous optic nerve damage using confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph II) in high-risk populations: a pilot study.

PURPOSE: To evaluate whether confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph II [HRT II], Heidelberg Engineering, Heidelberg, Germany) is a valid tool for the detection of glaucomatous optic nerve damage. DESIGN: Observational, cross-sectional, nonconsecutive study in Montreal, Canada. PARTICIPANTS: Three hundred three nonconsecutive, high-risk persons were enrolled during a 6-month period. METHODS: Participants underwent HRT II testing and a standard ophthalmologic examination, including gonioscopy, intraocular pressure measurement, and optic disc grading. MAIN OUTCOME MEASURES: Positive likelihood ratio (PLR) and negative likelihood ratio (NLR), sensitivities and specificities, positive predictive value (PPV) and negative predictive value (NPV), and kappa coefficients of agreement of Moorfields regression analysis (MRA), cup shape measure (CSM), height variation contour (HVC), and mean retinal nerve fiber layer thickness (MRNFL). RESULTS: Three hundred three participants were enrolled, and 291 were examined clinically; 21 (7.2%) were found to have glaucoma. Heidelberg Retina Tomograph II testing was performed successfully and was of acceptable quality in 531 of 601 eyes (88%). When MRA was compared with the clinically based diagnosis, the weighted kappa coefficient was kappa = 0.567 (95% confidence interval [CI], 0.42-0.71) for the right eye and kappa = 0.516 (95% CI, 0.37-0.66) for the left eye. Best kappa coefficient of agreement was seen when normals were grouped with suspects in both clinical and MRA diagnosis (kappa = 0.604; 95% CI, 0.409-0.799 in the right eye). Depending on the gold standard and test-positive definitions for glaucoma, specificity ranged from 87% to 97%, sensitivity from 25% to 100%, PPV from 28% to 68%, NPV from 84% to 100%, PLR from 5.0 to 19.2, and NLR from 1.3 to 6.2. When CSM, HVC, and MRNFLT were compared with clinical diagnosis, all outcome measures were found to have lower ranges: specificity from 46.9% to 83.7%, sensitivity from 36.5% to 76.9%, PPV from 6% to 36%, NPV from 80% to 99%, PLR from 0.8 to 4.0, NLR from 0.9 to 3.0. CONCLUSIONS: The results of this study suggest that a glaucoma screening program may be effective in detecting glaucoma when targeting high-risk populations. Heidelberg Retina Tomograph II testing may prove to be a useful tool in detecting glaucomatous optic nerve damage and could be used as part of a complete glaucoma screening protocol.

The effect of compression on clinical diagnosis of glaucoma based on non-analyzed confocal scanning laser ophthalmoscopy images.

BACKGROUND AND OBJECTIVE: To evaluate the effect of different image compression formats of non-analyzed Heidelberg Retina Tomography (HRT; Heidelberg Engineering, Heidelberg, Germany) images on the diagnosis of glaucoma by ophthalmologists. MATERIALS AND METHODS: Thirty-three topographic and reflectance images taken with the HRT representing different levels of disease were transformed using nine different compression formats. Three independent ophthalmologists, masked as to contour line and stereometric parameters, classified the original and compressed HRT images as normal, suspected glaucoma, or glaucoma, and Kappa agreement coefficients were calculated. RESULTS: The Tagged Image File Format had the largest file size and the Joint Photographic Experts Group (JPEG) 2000 format had the smallest size. The highest Kappa coefficient value was 1.00 for all ophthalmologists using the Tagged Image File Format. Kappa values for JPEG formats were all in the range of good to excellent agreement. Kappa values were lower for Portable Network Graphic and Graphics Interchange Format compression formats. CONCLUSION: Image compression with JPEG 2000 at a ratio of 20:1 provided sufficient quality for glaucoma analysis in conjunction with a relatively small image size format, and may prove to be attractive for HRT telemedicine applications. Further clinical studies validating the usefulness of interpreting non-analyzed HRT images are required.