Repeatability of Automated Neural Canal Opening Segmentation in SD-OCT Images of Optic Nerve in Glaucoma Patients

Abstract

 

Purpose:

 

Having more reproducible structural parameters for the assessment of glaucoma is desirable. Use of the neural canal opening (NCO) from spectral-domain optical coherency tomography (SD-OCT) volumes may enable more reproducible parameters to be obtained. The purpose of this study is to investigate the repeatability of an automated NCO segmentation algorithm in SD-OCT volumes of optic nerve.

 

Methods:

 

72 SD-OCT (Zeiss Cirrus) volumes centered at the optic nerve head (ONH) from 9 subjects with glaucoma (2 eyes x 2 repeats x 2 visits for each subject) were obtained at the University of Iowa. Each subject had OCT scans done on 2 separate visits within a 3 month period. Multiple images were taken during each visit. The NCO was automatically segmented on each volume using our 3-D graph-theoretic approach (Hu et al., IOVS 2010). The repeatability was evaluated by the intra-visit and inter-visit coefficient of repeatability (CR) and Pearson's correlation coefficient (r) with respect to the NCO area.

 

Results:

 

Both the intra- and inter- visit repeatability of the NCO segmentation showed the r and CR for the NCO area of 0.97 and 87 respectively (Table 1).

 

Conclusions:

 

Area measurements of the NCO from SD-OCT are robust across repeated acquisitions and visits and may serve as a stable reference landmark for describing structural changes that occur in glaucoma.  

 

Keywords: image processing • optic disc • optic nerve