Abstract
Purpose: :
to assess the intraobserver and interobserver variability of macular thickness measurements of three Fourier-domain (Spectral OCT/SLO Combo System, Ophthalmic Technologies Inc., Toronto, Canada; 3D OCT-1000,Topcon, GB Ltd, Newbury,Berkshire; RTVue-100, Optovue, Fremont, California) and one Time-domain (Stratus OCT, Carl Zeiss Meditec, Inc, Dublin, California, USA) devices.
Methods: :
macular thikness was assessed in 18 randomly chosen consecutive eyes of 18 healthy volunteers by two masked operators: A (L.P), B (E.M) using four OCT devices. Both observers repeated two consecutive measurements on one eye of each patient during the same day at the same time. All patients underwent OCT imaging using the four devices over a period of four weeks. Intraobserver reproducibility was evaluated by intraclass coefficient correlation (ICC) in a random model. Interobserver variability was evaluated by Bland-Altman analysis. Instrument-to-instrument reproducibility was determined by ANOVA for repeated measures with two within factors (instrument-operator).
Results: :
the mean macular thichkness was for observer A: 213,52 micron (Spectral OCT/SLO), 202,88 micron (Stratus OCT), 219,13 micron (3D OCT-1000), 237.5 (RTVue-100); for observer B: 215,02 micron (Spectral OCT/SLO), 206,27 micron (Stratus OCT), 226,69 micron (3D OCT-1000), 236.92 micron (RTVue-100).The ICC (95% CI) was for observer A: 0.94 (Spectral OCT/SLO), 0.88 (Stratus), 0.63 (3D OCT-1000), 0,75 (RTVue-100); for observer B: 0.92 (Spectral OCT/SLO), 0.92 (Stratus OCT), 0.91 (3D OCT-1000), 0.76 (RTVue-100). In Bland Altman analysis the mean difference inter-operators was: -1,5 (Limits of Agreement +/-15,42) for Spectral OCT/SLO; -3,38(LoA +/- 19,32) for OCT Stratus; -7,55 (LoA +/- 26,5) for 3D OCT-1000; 0,5 (LoA +/- 12,12) for RTVue-100. Inter-instruments reproducibility analysis was statistically different (p< 0.001); the interaction instrument * operator was not statistically significant (p=0,19).
Conclusions: :
macular thickness measurements showed low intraobserver variability for Spectral OCT/SLO and Stratus OCT and higher variability for 3D OCT-1000 and RTVue-100. 3D OCT-1000 is more operator dependent. Interobserver variability was high for each instrument. High instrument-to-instrument inter-observer variability showed that OCT does not provide reproducible measurements when different examiners are using different instruments.