Purchase this article with an account.
Gijs Thepass, Johannes F de Boer, Hans G Lemij, Koenraad Arndt Vermeer; RNFL attenuation coefficients for normal and glaucomatous eyes from standard peripapillary SD-OCT circle scans. Invest. Ophthalmol. Vis. Sci. 2014;55(13):4742.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Regular peripapillary SD-OCT circle scans can be used to obtain the attenuation coefficient (AC) of the retinal nerve fiber layer (RNFL). To obtain information based on the strength of the OCT signal, we determined the AC, an optical property derived from scattering and absorption in tissue. (Vermeer, ARVO E-abstract 395, 2013) Our aim was to use a commercially available OCT system and the system’s proprietary fully automated RNFL segmentation to determine the RNFL ACs in a next step. Resulting ACs were evaluated for use of this marker as a new glaucoma diagnostic.
Peripapillary circle scans of the Spectralis OCT system (Heidelberg Engineering) were made in 30 normal eyes and in 30 mild (MD 0 to -6 dB), 30 moderate (MD -6 to -12 dB) and 30 advanced (MD <-12 dB), age matched, glaucomatous eyes. The OCT images were converted into depth-resolved AC images in an automated fashion by using our own algorithm. Next, the Spectralis segmentation was used to determine RNFL AC. The RNFL ACs were then compared between the four groups.
Figure 1 shows a conventional OCT circle scan image and three AC images in a normal and 2 glaucomatous eyes. Figure 2 shows the ACs for all eyes and median values for each group. The connecting line shows the highly significant trend (Jonckheere-Terpstra, p<0.001). There was a statistically significant relationship (p<0.001) between MD and RNFL AC (Spearman ρ=0.529). Between group comparisons showed statistically significant differences (p<0.05) for all groups for RNFL AC, except between normal and mild glaucoma (p=0.209) and between moderate and advanced glaucoma (p=0.117) (Fig.2).
We used a commercially available OCT system to obtain RNFL ACs. We found a highly statistically significant relationship between RNFL AC and glaucoma severity. These results are consistent with previous work based on manual segmentation (Van der Schoot, IOVS, 2012). By using automated segmentation in a regular OCT system, we have shown the possibility of implementation of our algorithm in a currently used clinical device. Future work will focus on improving the segmentation, excluding blood vessels and adding spatial analysis to improve the performance of our method.
This PDF is available to Subscribers Only