Purchase this article with an account.
LM Zangwill, C Bowd, DS Greenfield, RN Weinreb; Correction of corneal polarization axis and magnitude improves discriminating ability of GDx Nerve Fiber Analyzer (GDx) . Invest. Ophthalmol. Vis. Sci. 2002;43(13):930.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: The GDx Nerve Fiber Analyzer measures retinal nerve fiber layer birefringence and compensates for corneal birefringence using a fixed corneal polarization axis (CPA) (15 degrees nasally downward) and fixed corneal polarization magnitude (CPM) (60 nm).To determine whether correction for CPA and CPM improves the discriminating ability of the GDx. Methods: The GDx was modified to enable the measurement of CPA and CPM so that compensation for corneal birefringence is eye specific. 52 normal eyes and 55 eyes with repeatable glaucomatous visual field damage (average MD = -6.5 + 4.9 dB) were examined using both fixed compensation GDx and GDx with variable corneal compensation (GDx VCC). Results: Mean (+ SD) CPA was 24.4 + 17.2 degrees (range -13 to 74 degrees). Mean CPM was 40.1 nm + 15.3 nm (range 7 to 91 nm). After adjusting for age, there was no statistically significant difference in CPA or CPM between normal and glaucoma eyes. There was no significant correlation between CPA and CPM (p ≷ .05). The mean average RNFL thickness measured using the fixed compensation GDx and GDx VCC was 59.0 microns and 41.0 microns in glaucoma patients, and 70.7 microns and 47.7 microns in normal subjects, respectively. The area under the ROC curve for discriminating between normal and glaucoma eyes improved using the GDx VCC for some but not all GDx parameters. For ellipse modulation the ROC area improved from .68 using fixed compensation GDx to .82 using GDx VCC. Other parameters for which the area under the ROC curve improved by ≷ 0.08 from fixed compensation GDx to GDx VCC included total polar integral (.73 versus .83), inferior integral (.75 versus .85), ellipse average (.75 versus .84), inferior average (.77 versus .85), nasal average (.57 versus .67), and several temporal thickness parameters. The area under the ROC curve for "the GDx number", superior maximum, inferior maximum and nasal maximum did not differ between fixed compensation GDx and GDx VCC. Conclusion: Correction for CPA and CPM improves the discriminating ability of many GDx nerve fiber layer parameters.
This PDF is available to Subscribers Only