Purchase this article with an account.
Jean-Remi Fénolland, Jean-Marie Giraud, Marlene Francoz, Hussam El Chehab, Damien Sendon, Franck May, Jean Paul Renard; Peripapillary RNFL thickness measurement : agreement among 4 SD-OCT instruments. Invest. Ophthalmol. Vis. Sci. 2012;53(14):685.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare the peripapillary retinal nerve fiber layer (RNFL) thickness measure with Spectralis (Heidelberg Engineering, Dossenheim, Germany), Topcon 3D OCT-2000 (Topcon, Tokyo, Japan), RTVue-100 OCT (Optovue Inc., Fremont, CA, USA) and Nidek RS-3000 SD-OCT (Nidek, Gamagory, Japan) devices in a population with normal, ocular hypertensive and glaucomatous patients.
Observationnal study. A total of 127 eyes of 69 patients were enrolled after complete clinical examination and a Humphrey SITA standard 24.2 visual field testing. Inclusion criteria were best-corrected visual acuity of ≥20/25, spherical equivalent between -3 and +3 diopters, no history of diabetes, no clinical evidence of maculopathy nor retinopathy (diabetic retinopathy, macular degeneration), no ocular surgery excepted phacoemulsification or glaucoma surgery within 6 months of enrollement.Images were performed by the same operator with the 4 OCTs for each patient the same day. Routinely-used glaucoma protocols were performed with the different instruments : RNFL circle scan for the Spectralis, 3D scan disc for the Topcon, 3D optic nerve head plus ONH map for the RTVue, disc map for the Nidek. Statistical analysis included ANOVA with Bonferroni multiple t-tests to appeciate the differences between the OCTs and Bland Altman plots to evaluate the level of agreement.
A total of 126 eyes of 69 patients were included. In the whole studied population, mean age was 60,5± 2,7 years, spherical equivalent was -0,015±1,4 diopters. The overall RNFL values were : 85,5±18,4 microm for the Spectralis, 96,8±14,0 microm for the Topcon, 94,5±16,5 microm for the RTVue and 94,85±17,8 microm for the Nidek. The values were significantly differents with ANOVA testing for overall, superior, inferior, nasal and temporal RNFL thickness between the 4 instruments (p<0,01). With Bonferroni adjusted multiple t-tests, for overall values, Spectralis values were differents compare to the 3 other OCTs (p<0,01) wheras RTVue, Nidek and Topcon were not differents (p=1). Bland Altman plots were constructed and showed a moderate agreement between the instruments.
It is the first study, in our knowledge, to compare four SD-OCT in RNFL thickness measurements. These 4 SD-OCT are not interchangeable because of moderate agreement. Hardware and software industrial choices may explain a part of the OCTs measurements differences but it is today impossible to affirm which instrument is the more accurate. It would be useful to test these machines on animal models to compare the values obtained to histological thckness.
This PDF is available to Subscribers Only