Purchase this article with an account.
N. M. Buchser, H. Ishikawa, G. Wollstein, R. A. Bilonick, Y. Ling, L. S. Folio, L. Kagemann, R. J. Noecker, E. Albeiruti, J. S. Schuman; Retinal Nerve Fiber Layer (RNFL) Thickness Measurement Reproducibility Comparison Across Three Spectral Domain Optical Coherence Tomography (SD-OCT) Devices With Systematic Bias Correction. Invest. Ophthalmol. Vis. Sci. 2010;51(13):4394.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare the RNFL measurement reproducibility across three SD-OCT devices using their native default scan protocol and analysis while correcting for systematic measurement bias.
Forty eyes of 20 healthy volunteers underwent peripapillary RNFL thickness measurements using three SD-OCT devices: Cirrus HD-OCT optic nerve head (ONH) cube 200x200 protocol (Carl Zeiss Meditec, Dublin, CA); RTVue ONH protocol (16 radial lines and 8 concentric circles of various diameters, each centered on the ONH; Optovue, Fremont, CA); 3D OCT-1000 3D Scan 256x256 protocol (Topcon, Paramus, NJ) on the same day. All images were qualified for the manufacturer recommended image quality. RNFL measurement reproducibility was assessed by calculating bias and imprecision using a structural equation model. Estimated imprecision after correcting for bias represented measurement reproducibility that was directly comparable across the different devices.
Global mean RNFL thickness measurement showed estimated imprecision (scale bias corrected) of 2.76 (Cirrus), 7.76 (RTVue), and 4.36 (3D OCT-1000). In other words, RTVue was 2.8x more imprecise and 3D OCT-1000 was 1.6x more imprecise than Cirrus in global RNFL measurement.
Cirrus RNFL measurement had the best measurement reproducibility among the three tested SD-OCT devices. Three dimensional cube scanning with post-hoc data sampling may be a factor reducing imprecision.
This PDF is available to Subscribers Only