Comparison of Manually-Corrected Retinal Thickness Measurements from Multiple Spectral Domain Optical Coherence Tomography Instruments

摘要

Purpose: : To compare retinal thickness measurements from three different spectral domain optical coherence instruments (SD-OCT) when manual segmentation is employed to standardize retinal boundary locations. Methods: : 40 eyes of 21 healthy subjects were scanned consecutively on the Cirrus HD-OCT, Topcon 3D OCT-2000, and the Heidelberg Spectralis OCT using dense macular cube scanning protocols (512 x 128 for Cirrus and Topcon, and 1024 x 37 for Heidelberg). Raw OCT data was imported into the previously validated custom OCT grading software (3D-OCTOR) of the Doheny Image Reading Center (DIRC) which allows the user to draw the inner and outer retinal boundaries at a consistent location on every B-scan or a subset of scans. For this study, manual segmentation was performed on every Spectralis scan, and every 4th scan from the denser Cirrus and 3D OCT-2000 cube scans, such that manually segmented B-scans spaced no more than 188 microns apart were available to compute retinal thickness and volume values for each subject and each instrument. The outer retinal boundary was designated according to reading center convention to be at the band believed to correspond to the interdigitation between the photoreceptor tips and the retinal pigment epithelium. Retinal thickness values in the foveal central subfield were computed for every case. Results: : 37 eyes of 20 subjects were gradable on every machine. By standardizing the location of the outer retinal boundary, the average retinal thicknesses for these 37 eyes were 236.7 ?μm (SD 20.1), 235.7 ?μm (SD 20.4), and 236.5 ?μm (SD 18.0) for the Cirrus, the 3D OCT-2000, and the Spectralis, respectively. When comparing manual retinal thickness measurements between any two machines among individual eyes, the maximum difference was 18.2 ?μm. The mean absolute differences per eye between two machines were: = 4.9 ?μm for Cirrus vs 3D OCT-2000, 3.7 ?μm for Cirrus vs Spectralis, and 4.4?μm for 3D OCT-2000 vs. Spectralis. Conclusions: : When a uniform position is used for the location of the outer retinal boundary, retinal thickness measurements derived from three different SD-OCT devices are virtually identical. Manual correction to a standardized outer retinal boundary may provide a viable method to allow OCT-derived thickness measurements to be compared between devices in clinical trials and clinical research.