Retinal Nerve Fiber Layer Thickness Map Determined With SD–OCT and Performance Assessment of the Automatic Algorithm

Abstract

 

Purpose:

 

To determine large area retinal nerve fiber layer (RNFL) thickness maps from Spectral–Domain Optical Coherence Tomography (SD–OCT) images and to assess the performance and the reproducibility of the algorithm.

 

Methods:

 

A clinical (SD–OCT) system was used for high–speed (29fps) and ultra–high axial resolution (better than 3 µm) cross–sectional scanning of the retina of a normal volunteer. An automated algorithm based on anisotropic noise suppression and deformable splines was implemented for determining the vitreous–RNFL and RNFL–ganglion cell/inner plexiform layer boundary, respectively.

 

Results:

 

The data is presented in a correlated manner that includes the RNFL thickness map and an integrated reflectance image together with an ultra–high axial resolution OCT movie (Fig. 1). Very good agreement was found between the results of the automatic algorithm and expert evaluation of the RNFL boundaries by two ophthalmologists. Excellent reproducibility was also found for the results of the algorithm applied to the same scan area measured on different sessions.

 

Conclusions:

 

The RNFL thickness maps can potentially be used for a thorough evaluation of the RNFL thickness in longitudinal studies of glaucoma progression. These studies may then use large area RNFL thickness maps, which may allow for more accurate correlations of RNFL thinning with visual field defects, as opposed to individual circular scans. The representation shown in Fig.1 provides a comprehensive picture to clinicians, and we anticipate that it will be a valuable tool for interpreting the OCT data for diagnostic purposes. Fig 1: Integrated reflectance map (top left), RNFL thickness map (bottom left), and retinal cross–sectional image (right). The color scheme for the RNFL thickness map is scaled in microns, the darkest blue meaning no thickness, and the darkest red being a maximum of 177 µm. Each map has a size of 8.85x5.73 mm², while the vertical size of the cross–sectional image is 1.2 mm. The processing of the RNFL thickness map took about 3 hours.  

 

Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • nerve fiber layer • image processing