advertisement

Editors Selection IGR 9-3

Clinical examination methods: Structural progression

Comment by Paul Healey on:

12377 Use of progressive glaucomatous optic disk change as the reference standard for evaluation of diagnostic tests in glaucoma, Medeiros FA; Zangwill LM; Bowd C et al., American Journal of Ophthalmology, 2005; 139: 1010-1018

Find related abstracts

A major flaw in the evaluation of imaging in glaucoma diagnosis is the use of narrow range of advanced glaucoma cases. While this improves test sensitivity, it has little diagnostic utility as the cases could be easily distinguished from the controls on clinical examination. This problem is addressed by Medeiros et al. (545), who report a clinic-based case-control study of the diagnostic accuracy and utility of the GDx VCC scanning laser polarimeter to detect open-angle glaucoma. The authors tried to identify a wider range of glaucoma states than previous studies by using progression of glaucomatous optic neuropathy (GON) from optic disc photographs to select cases. We do not know whether they excluded nerve fibre layer loss (the feature the GDx is presumed to measure) as a means of case selection to avoid bias. Nor do we know the severity of the GON in the chosen cases. But by assessing visual field tests for repeatable glaucomatous defects, the authors classified disease state from pre-perimetric to advanced. The influence of state on sensitivity of the GDx was tested using the visual field mean deviation. The predictive utility of the GDx was also explored by calculating likelihood ratios for the Nerve Fibre Layer Indicator (NFI) score as well as number of other GDx parameters. The principal findings were that sensitivity of the NFI increased rapidly with decreasing mean visual field defect, reaching 100% (with 95% specificity) at about -15dB. In contrast in the average pre-perimetric case (with an MD of -2 dB) the sensitivity was only about 60%. The predictive utility of the NFI was greatest when it was <15 (negative prediction) or >50 (positive prediction). But even when the NFI lay between 35 and 50, the likelihood of pre-perimetric glaucoma was increased by more than eight-fold compared to controls. This paper is to be commended on several grounds. First, it selected a wide range of disease states including early disease. Second, it studied patients with progressing glaucoma, the group at the highest risk of further damage. Third, it demonstrated that the GDx VCC was capable of providing useful information for clinicians who face the diagnostic challenge of an optic disc morphology suspicious for glaucoma but without a reproducible visual field defect on standard automated perimetry.

• ← Previous
• Next →