Quantitative shadow compensated optical coherence tomography of choroidal vasculature

Abstract

Purpose : Conventionally rendered optical coherence tomography (OCT) images of the posterior segment contain shadows which mask the visualization of deep structures such as the choroid. The choroidal vascularity index (CVI) is an OCT-derived estimated ratio of luminal to total choroidal volume (hyporeflective to total voxels). The purpose of this paper is to determine whether OCT shadow compensation alters the appearance of the choroid and the apparent CVI.

Methods : Twenty-seven subjects without any ocular disease underwent swept-source OCT imaging of the right eye, without averaging, and the resulting raw voxel maps were exported for further processing. The inner and outer choroidal boundaries were established using a structural similarity and tensor voting based algorithm. All scans were shadow compensated using a previously published algorithm, then binarized, using a novel algorithm recently described by our group, so that a CVI could be calculated from choroidal voxels. Images at each stage were inspected subjectively and the effect of shadow compensation on CVI was quantified objectively by Bland-Altman analysis.

Results : In shadow compensated OCT scans, the choroid was visualized with higher reflectivity than the neurosensory retina and the masking of deep tissues by retinal blood vessels was greatly reduced (figure). Mean CVI was 0.51 when calculated from non-shadow compensated scans, and 0.55 with shadow compensation. Bland-Altman analysis revealed a statistically significant CVI difference of 0.04 per eye (95% CI -0.05, -0.02).

Conclusions : Conventionally acquired OCT underestimates both choroidal reflectivity and calculated CVI. Shadow compensation corrects for masking which may improve accuracy of quantitative analysis. In healthy eyes, the darkest shadows are cast by retinal blood vessels. In eyes harboring pathology, OCT shadows cast by lesions such as hemorrhage or pigment epithelial detachment may predominate, particularly in regions of interest. Although the magnitude of the difference in CVI calculated from shadow compensated vs raw scans may be small in healthy eyes, this difference may become more pronounced in eyes with disease. Since shadow compensated OCT better represents actual tissue reflectivity, we believe it may improve accuracy for choroidal analysis in pathologic states. Further study is needed to test this hypothesis.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.

 

Choroid vasucularity: Raw OCT vs shadow compensated OCT.