PURPOSE: To determine the effect of multiple B-scan averaging on retinal nerve fiber layer (RNFL) thickness measurement taken with a spectral-domain optical coherence tomographer (OCT). METHODS: Twenty-five normal individuals and 29 patients with glaucoma were examined. One eye was selected randomly of each participant for circumpapillary RNFL scanning by the Spectralis OCT (Heidelberg Engineering, Heidelberg, GmbH, Dossenheim, Germany). Three sets of OCT image series comprising 2, 8, and 16 consecutive B-scans were obtained in the same visit. After multiple B-scan averaging, the respective global and sectoral (superonasal, nasal, inferonasal, inferotemporal, temporal, and superotemporal) RNFL thicknesses were compared with linear mixed modeling and analyzed for RNFL measurement variability. RESULTS: There was no RNFL segmentation failure in all the scans. No significant differences were found in global or sectoral RNFL thicknesses among the image series averaged with different number of B-scan except for the inferonasal sector in the glaucoma group (P=0.036). The intraclass correlation coefficients were all >0.971 and the coefficients of variation were all below 4.33% for both sectoral and global RNFL thicknesses. No significant differences in the image quality score were detected among the image series (P≥0.141). CONCLUSIONS: Although multiple B-scan averaging is commonly applied to enhance visualization of the retinal layers in OCT images, it may not have a significant impact on global and sectoral RNFL thicknesses measured by the Spectralis OCT.
PURPOSE: To determine the effect of multiple B-scan averaging on retinal nerve fiber layer (RNFL) thickness measurement taken with a spectral-domain optical coherence tomographer (OCT). METHODS: Twenty-five normal individuals and 29 patients with glaucoma were examined. One eye was selected randomly of each participant for circumpapillary RNFL scanning by the Spectralis OCT (Heidelberg Engineering, Heidelberg, GmbH, Dossenheim, Germany). Three sets of OCT image series comprising 2, 8, and 16 consecutive B-scans were obtained in the same visit. After multiple B-scan averaging, the respective global and sectoral (superonasal, nasal, inferonasal, inferotemporal, temporal, and superotemporal) RNFL thicknesses were compared with linear mixed modeling and analyzed for RNFL measurement variability. RESULTS: There was no RNFL segmentation failure in all the scans. No significant differences were found in global or sectoral RNFL thicknesses among the image series averaged with different number of B-scan except for the inferonasal sector in the glaucoma group (P=0.036). The intraclass correlation coefficients were all >0.971 and the coefficients of variation were all below 4.33% for both sectoral and global RNFL thicknesses. No significant differences in the image quality score were detected among the image series (P≥0.141). CONCLUSIONS: Although multiple B-scan averaging is commonly applied to enhance visualization of the retinal layers in OCT images, it may not have a significant impact on global and sectoral RNFL thicknesses measured by the Spectralis OCT.