Author

Jennifer Lee

Date of Award

8-24-2009

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)

First Advisor

M. Bruce Shields

Abstract

We hypothesized that nasal or temporal peak contour shifts from the normative database observed on the Fast retinal nerve fiber layer (RNFL) scan is a common occurrence because there is a range of distribution of peak contours due to anatomical variation, not necessarily related to glaucoma, and this displacement may lead to misclassification as glaucoma under current commercial optical coherence tomography (OCT) criteria. The purpose of this study was to evaluate the prevalence these shifts with Stratus OCT and possible associations with demographic or glaucoma-related variables. This was a retrospective case series of glaucoma patients and glaucoma suspects (one eye per patient) who underwent a Fast RNFL thickness study with the Stratus OCT. A study was considered to have a clinically significant horizontal deviation if there was greater than 20° shift from the normative database in both superior and inferior peaks. A second cut-off value of 12° was also used to examine smaller deviations. A linear regression model was used to assess correlations of demographic and glaucoma-related variables between eyes with and without significant deviations. Of 400 subjects screened, 273 met the inclusion criteria. Thirty-nine eyes (14.3%) had clinically significant horizontal deviation using the 20° cut-off (95%CI 10%-19%), while 122 (44.7%) met the definition with the 12° cut-off (95%CI 38%-51%). Additionally, 121 eyes (44.3%) had a >20° horizontal shift of either the superior or inferior peak (95% CI 38%-51%). There was no correlation between the demographic or glaucoma-related variables and the horizontal deviation of peak contours. This study suggests that significant horizontal deviation of peak RNFL contours with the Stratus OCT Fast RNFL is common and emphasizes the need for caution when interpreting the influence of such deviations on clock hour segment thinning. It is not possible with this technology to distinguish between scan circle misalignment (horizontal or vertical) and anatomical variation as the explanation for the finding, and further study with newer technology is needed.

Comments

This is an Open Access Thesis.

Share

COinS