Comparison of Topcon Optical Coherence Tomography and Ultrasound Pachymetry

Purpose. To establish within-rater repeatability and minimum measurements required for reliable assessment of central corneal thickness (CCT) using Topcon three-dimensional OCT-2000 (Topcon Medical Systems, Oakland, NJ) Fourier-domain optical coherence tomography (FD-OCT) and to test agreement against ultrasound pachymetry (USP). Methods. Twenty participants underwent five scans using FD-OCT followed by five CCT measurements using USP. Each FD-OCT scan produced 12 meridional optical sections, giving 60 sections in total. The FD-OCT CCT was calculated from each section using Topcon FastMap software. Within-rater repeatability was assessed using intraclass correlation coefficient (ICC), within-subject variance (S-w), andwithin-subject coefficients of repeatability (COR) and variation (COV). Agreement was assessed using analysis of variance, simple linear regression, Bland-Altman analysis, and ICC. The minimum number of FD-OCTscans required to achieve a mean CCT within the 99% confidence intervals of all 60 measurements was calculated. Results. Mean CCTs were 528 T 27 mu m (FD-OCT) and 544 T 29 mu m (USP). The FD-OCT within-rater repeatability values assessed using ICC, S-w, COR, and COV were 0.99, 5.07 mu m, 14.07 mu m, and 0.01%, respectively. The USP within-rater repeatability values assessed using ICC, S-w, COR, and COV were 0.96, 5.49 mu m, 15.22 mu m, and 0.01%, respectively. Both instruments were similarly correlated (p < 0.05), with FD-OCT underestimating CCT by 16.08 mu m relative to USP. Four FD-OCT scans were required to achieve a mean within the 99% confidence interval of all measurements for each participant. Conclusions. Both instruments displayed high repeatability, with strong agreement between devices, although FD-OCT significantly underestimated CCT compared with USP. The FD-OCT agreement with USP remained strong when only four scans were analyzed. (Optom Vis Sci 2012; 89: 1708-1714)