Purchase this article with an account.
J. Stein, C.A. Girkin, N. Harizman, S. Arthur, A. Chiang, J.D. Ortega, J.M. Liebmann, C. Tello, R. Ritch; A Comparison of False–Positive Test Results Between the Stratus OCT III and the GDx–VCC . Invest. Ophthalmol. Vis. Sci. 2005;46(13):3635.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To compare the proportions of false–positive findings with optical coherence tomography (OCT) and scanning laser polarimetry in an independent cohort of participants with healthy eyes. Methods: Monocular Stratus OCT–3 imaging and GDx scanning with variable cornea compensation (VCC) were performed in 234 prospectively enrolled participants with no evidence of glaucoma based on optic nerve head stereoscopic photography (assessed by masked review), repeated intraocular pressure measurement (<21 mm Hg), and achromatic perimetry. Participants were considered to have false–positive findings on OCT–3 imaging if the mean retinal nerve fiber layer (RNFL) thickness or the RNFL thickness of at least one quadrant was abnormal (based on 99% confidence intervals [CI] for normative values). GDx VCC findings were considered false–positive if the mean temporal, superior, nasal, inferior, temporal (TSNIT) score or the RNFL value in at least one quadrant was abnormal (based on 99% CI). Results: With OCT–3, RNFL thickness values were abnormal in one quadrant only for 21 eyes (9% of participants), in two quadrants only for seven eyes (3.1%), and in three or more quadrants for four eyes (1.7%). GDx VCC scanning indicated an abnormality in one quadrant only for 21 eyes (9.0%), in two quadrants only for six eyes (2.6%), and in three or more quadrants for two eyes (0.85%). Global measurement for OCT–3 (mean RNFL thickness) was abnormal in five eyes (2.1%), and for GDx VCC (mean TSNIT score) in nine eyes (3.8%). The techniques did not differ in the proportion of eyes with false–positive findings for global measurement (p = 0.07) or for one quadrant only (p = 1.0), two quadrants only (p = 0.36), or three or more quadrants (p = 0.69). The agreement between techniques was poor (highest Κ = 0.27, for global RNFL thickness). Conclusions: The OCT–3 and the GDx VCC had similar proportions of false–positive findings but tended to identify different eyes as abnormal. To maintain a false–positive rate of 4% or less, the global indicators can be used. However, achieving a false–positive rate of 1% or less would require demonstration of an abnormality in three or more quadrants on the GDx VCC. Currently, it does not appear possible to achieve a false–positive rate of 1% or less with OCT.
This PDF is available to Subscribers Only