Purpose:
To study the reproducibility of SD-OCT macular and optic nerve head (ONH) parameters obtained using our own custom-developed segmentation algorithms and using the proprietary algorithms provided in the commercial Zeiss Cirrus software.
Methods:
Repeat macular and ONH-centered scans (200×200×1024 voxels, 6×6×2 mm3) were obtained within 90-day period with the interval of 75 ± 20 days from both eyes of 58 glaucoma and glaucoma suspects (62.9 ± 14.8 years, 35% male) using a CirrusTM HD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA). The ONH scans with obvious motion artifacts through the disc, or clipped A-scans through the 3.46mm peripapillary circle were excluded from analysis. Macular images with >20% of the clipped A-scans were also excluded. Remaining 198 scans from 99 eyes of 56 subjects were analyzed. The central subfield thickness, cube volume, cube average thickness from the macular scans, and the average retinal nerve fiber layer (RNFL) thickness, rim area, disc area, average (linear) cup-to-disc ratio from the ONH scans were calculated using our custom automated segmentation methods (Lee K et al, IEEE Trans Med Imaging 29:159, 2010). The reproducibility was analyzed using absolute difference, Pearson’s correlation coefficient, linear regression, and paired t-test. The reproducibility of the same parameters provided by Cirrus software (v.5.1.0.96) was also analyzed.
Results:
The absolute difference, correlation coefficient (r), linear regression, and p-value of the macular and ONH parameters obtained from our custom algorithm (Iowa) and Cirrus software are provided in Table.
Conclusions:
The macular and ONH analysis parameters obtained from SD-OCT using our custom automated algorithms showed excellent reproducibility in our subjects . The reproducibility was similar to those provided by commercial Cirrus software.
Keywords: imaging/image analysis: clinical • optic nerve • macula/fovea