A method for mapping retinal nerve fiber layer birefringence

Abstract

 

Purpose

 

We have shown that RNFL retardance, measured by scanning laser polarimetry (SLP), declines prior to RNFL thickness measured by spectral domain optical coherence tomography (SDOCT) using global average peripapillary scan locations for non-human primates (NHP) with experimental glaucoma (EG). These results are consistent with the hypothesis that axonal cytoskeletal disruption precedes axon bundle thinning within the RNFL during the course of glaucomatous neurodegeneration. An important step forward in testing this hypothesis is to develop a method for mapping RNFL birefringence beyond the conventional peripapillary scan locations.

 

Methods

 

SLP scans (GDxVCC, Carl Zeiss Meditec, Inc) and SDOCT cube scans consisting of 290 horizontal raster lines (Spectralis, Heidelberg Engineering, GmbH) were acquired longitudinally in NHPs with EG. Custom software was developed to rapidly co-localize pairs of reflectance images and derive birefringence maps.

 

Results

 

Fig1 demonstrates each step of the method and an example of glaucomatous change in RNFL birefringence.

 

Conclusions

 

We developed software to rapidly and effectively co-localize the reflectance image from SLP scans to the reflectance image from SDOCT scans and to derive RNFL birefringence maps from the retardance and thickness values, respectively. Registration of scans over time and across instruments reduces noise in longitudinal series of sectoral data.

 

 

Fig1: Focal RNFL birefringence changes can be followed longitudinally (by sector and eccentricity). (A) Infrared reflectance image acquired during baseline SDOCT cube scan; red lines indicate position of 290 raster scans. (B) Examples of individual B-scans with RNFL segmentations (red and green lines) from positions indicated by green arrows. (C) RNFL thickness map for baseline derived from data in panels A&B. (D) RNFL thickness map for final time point; note mild decrease in thickness. (E) Infrared reflectance image obtained from the baseline SLP scan and (E’) corresponding RNFL retardance map. (F) Colocalization step to register SLP scan to SDOCT scan: the SLP reflectance image is pseudo-colored, scaled, translated and rotated to optimize match. This step enables calculation of birefringence (nm retardance per µm thickness) and derivation of maps for (G) baseline and (H) final time point demonstrating birefringence changes in this eye.

 

Keywords: 610 nerve fiber layer • 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 550 imaging/image analysis: clinical