Developing Clinically Implementable Approaches for Directional Optical Coherence Tomography

Directional optical coherence tomography (OCT) provides a true outer nuclear layer (ONL) measure not masqueraded by the Henle fiber layer. This study aimed to test different image acquisition and analysis approaches to develop a clinically implementable tool.

OCT images from 12 participants (12 eyes) were taken using the Spectralis OCT2 (Heidelberg Engineering, Heidelberg, Germany). Directional OCT was obtained by the operator displacing the OCT beam or the participants fixating on three parts of an internal fixation cross. ONL thickness was segmented (1) using the built-in automated algorithm and (2) refining the segmentation border ("corrected segmentation"), compared to manual segmentation. Test-retest comparison between sessions was analyzed.

Automated segmentation returned thicker ONL (nasal, P = 0.05; temporal, P < 0.01), whereas corrected segmentation gave similar ONL thickness (nasal, P = 0.43; temporal, P = 0.06) compared to manual segmentation. Similar ONL thicknesses were obtained with the different approaches (nasal, P = 0.09; temporal, P = 0.26), which were repeatable between sessions.

Directional OCT can be obtained with either the operator altering the angle of the OCT device, which is less reliant on the patient's subjective involvement, or the patient adjusting their fixation. Correcting the ONL segmentation returns similar retinal thickness profiles to manual segmentation, with the benefit of being more time and resource efficient. These methods also demonstrated good test-retest reliability.

This study provides an approach for directional OCT image acquisition and analysis that can be readily implemented into clinical environments.