Intraobserver repeatability of automated versus adjusted optical coherence tomography measurements in patients with neovascular age-related macular degeneration

Abstract

Aim: To assess the intraobserver repeatability of automated versusadjusted optical coherence tomography (OCT) measurements in patients with neovascular age-related macular degeneration (NVAMD). Methods: Ten eyes with NVAMD from 10 consecutive patients underwent two OCT measurements within 5 days by a single operator. Automated and adjusted central 1-mm foveal thickness and automated and adjusted total macular volume were measured in each study eye. The term 'adjusted' refers to manually corrected values, in which the interface landmarks for measurements are selected by the operator using Stratus® scan profiling and custom software. Bland-Altman method and bootstrap comparison of intraclass correlations (ICCs) were used for repeatability analysis. Results: Bland-Altman comparison did not reveal any statistically significant difference in any parameter, when results at first and second examination were compared (p > 0.05), indicating that the repeated measurements are similar. Further analysis was conducted using the bootstrap comparison of ICCs. The difference between adjusted and automated foveal thickness ICCs (r = 0.945 and 0.635, respectively) was significant (p = 0.031), indicating higher repeatability for adjusted foveal thickness. The ICCs for adjusted and automated total macular volume (r = 0.873 and 0.863, respectively) showed no statistically significant difference (p = 0.881). Conclusion: The repeatability of adjusted retinal thickness measurements, in which the errors of retinal boundary detection by OCT analysis software is corrected by the operator using scan profiling, is found to be higher than that of automated ones in this small group of NVAMD patients when performed by a single experienced operator. Copyright © 2006 S. Karger AG.