Purchase this article with an account.
A. Zanei, O. Vattovani, G. Di Stefano, G. Ravalico, D. Tognetto, S. Da Pozzo; The Diagnostic Ability of the GDx-VCC Nerve Fiber Indicator (NFI) in Separating Healty From Glaucomatous Eyes. Invest. Ophthalmol. Vis. Sci. 2010;51(13):4898.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
to identify which values of the GDx-VCC parameter NFI are associated wich a significant effect on post-test probability (i.e.both diagnosing or excluding glaucoma).
five-hundred twenty-four healthy subjects (healthy-looking disc and normal automated perimetry), 28 glaucoma suspects (GS, whith normal perimetry and glaucomatous disc changes) and 262 glaucomatous eyes (reproducible rield defects with corresponding glaucomatous disc change) were enrolled.GDx-VCC scan were obtained and NFI recorded. The area under ROC curve (AUROC) was calculated adding GS first to healty subjects and then to glaucomatous eyes.The same procedure was adopted for calculating multilevel likelihood ratios (LR), searching for values <0.1 or >10, considered to be associated with large effect on post-test probability.
mean NFI values were 19.2±5.6, 39.1±13.4, 47.8±21.5 for healty subjects, GS and glaucomatous eyes, respectively (p<0.001 ANOVA).AUROC was 0.920 and 0.937 (p not significant, Hanley-McNeil test) after adding GS to healty and then to glaucomatous eyes, respectively. N.FI value <18 was associated with a large effect on post-test probability for excluding glaucoma with GS grouped both wich healty and glaucomatous eyes. On the contrary, a similar effect on post-test probability was identifield for NFI values >34 and >50 when GS were grouped firstwith glaucomatous eyes and then with healthy subjects, respectively. At values between 24 and 29, the probability to belong to healty or to glaucomatous is nearly similar.
in a large sample of eyes, NFI values associated with significant chances of having healty or glaucomatous eyes, as well as a range of NFI values with little or no effect on post-test probability, were identified. Moreover, even after considering GS and glaucomatous eyes as a simple group, NFI diagnostic ability did not decrease. This model needs further validation on a different population sample.
This PDF is available to Subscribers Only