Purpose: To measure the interobserver reproducibility and intra-observer reproducibility of subfoveal choroidal thickness measurements performed by enhanced depth imaging of spectral-domain optical coherence tomography (EDI-OCT) in a population-based setting.
Methods: The Beijing Eye Study 2011 was a population-based study performed in rural and urban regions of Greater Beijing. The study included 3468 individuals with a mean age of 64.6 ± 9.8 years (range, 50-93 years). The participants underwent EDI-OCT and the subfoveal choroidal thickness (SFCT) was measured. To examine the interobserver variability, all images were assessed by two examiners independently of each other within 2 months. To examine the intra-observer reproducibility, a smaller study sample consisting of 21 eyes of 21 healthy subjects from the Tongren Eye Center was included in the study. These latter subjects were scanned 10 times with 1 minute breaks between each examination. The SFCT was measured by the same observer within 2 weeks. The intrasession within subject SD, the coefficient of variation, and the intraclass correlation coefficient (ICC) were calculated.
Results: EDI-OCTs were performed for 3233 subjects. Mean SFCT measured by grader one and grader two were 254.6 ± 107.3 μm and 253.8 ± 107.4 μm, respectively, with a mean difference of 3.14 ± 13.1 μm (95% confidence interval, 0.0, 24.0). Bland-Altman plot showed 1.9% (61/3233) points outside the 95% limits of agreement. For the assessment of the intra-observer reproducibility, the ICC was 1.00 (P < 0.001, and the mean coefficient of variation was 0.85% ± 1.48%).
Conclusions: Under routine examination conditions, SFCT measurements by EDI-OCT showed a high intra-observer reproducibility and interobserver reproducibility.