Optical coherence tomography angiography of optic disc perfusion in glaucoma

Abstract

Purpose: To compare optic disc perfusion between normal subjects and subjects with glaucoma using optical coherence tomography (OCT) angiography and to detect optic disc perfusion changes in glaucoma.

Design: Observational, cross-sectional study.

Participants: Twenty-four normal subjects and 11 patients with glaucoma were included.

Methods: One eye of each subject was scanned by a high-speed 1050-nm-wavelength swept-source OCT instrument. The split-spectrum amplitude-decorrelation angiography (SSADA) algorithm was used to compute 3-dimensional optic disc angiography. A disc flow index was computed from 4 registered scans. Confocal scanning laser ophthalmoscopy (cSLO) was used to measure disc rim area, and stereo photography was used to evaluate cup/disc (C/D) ratios. Wide-field OCT scans over the discs were used to measure retinal nerve fiber layer (NFL) thickness.

Main outcome measures: Variability was assessed by coefficient of variation (CV). Diagnostic accuracy was assessed by sensitivity and specificity. Comparisons between glaucoma and normal groups were analyzed by Wilcoxon rank-sum test. Correlations among disc flow index, structural assessments, and visual field (VF) parameters were assessed by linear regression.

Results: In normal discs, a dense microvascular network was visible on OCT angiography. This network was visibly attenuated in subjects with glaucoma. The intra-visit repeatability, inter-visit reproducibility, and normal population variability of the optic disc flow index were 1.2%, 4.2%, and 5.0% CV, respectively. The disc flow index was reduced by 25% in the glaucoma group (P = 0.003). Sensitivity and specificity were both 100% using an optimized cutoff. The flow index was highly correlated with VF pattern standard deviation (R(2) = 0.752, P = 0.001). These correlations were significant even after accounting for age, C/D area ratio, NFL, and rim area.

Conclusions: Optical coherence tomography angiography, generated by the new SSADA, repeatably measures optic disc perfusion and may be useful in the evaluation of glaucoma and glaucoma progression.

Figure 1

Disc photographs (A1, A2), optical coherence tomography (OCT) reflectance (B1, B2), whole depth OCT angiograms (C1, C2, en face maximum projection), cross-sectional angiograms (D1, D2, overlaying on OCT reflectance in gray scale) in the right eye of a normal subject (A1-H1) and, the left eye of a perimetric glaucoma subject (A2-H2). Disc margins are marked by the black elliptical outlines (B1, B2, C1, C2). The position of the cross-section is shown by dotted blue lines (C1, C2). A dense microvascular network was visible on the OCT angiography of the normal disc (C1). This network was greatly attenuated in the glaucomatous disc (C2). In order to appreciate the ability of OCT angiography to detect blood flow within the various vascular beds, the 3D angiograms were separately projected into en face maximum projection in 3 layers, i.e. retinal angiograms (E1, E2), choroidal angiograms (F1, F2) and scleral/lamina cribrosa angiograms (G1, G2). The boundaries used for segmentation are indicated by dotted green lines on cross-sectional OCT reflectance (H1, H2).

Figure 2

Box plot showing the disc flow index in normal and glaucoma groups. The median (line inside the box), interquartile range (box) and the whole range of values (whiskers) are shown. This plot shows that the two groups are completely separated. The minimum of normal group is 0.1516, and the maximum of glaucoma group is 0.1513, indicating sensitivity and specificity are both 100% using a cutoff value of 0.1515.

Figure 3

Plots of disc flow index versus age (A), pattern standard deviation (PSD) (B), rim area (C), cup/disc (C/D) area ratio (D) and nerve fiber layer (NFL) thickness (E) in both normal group and glaucoma group. The linear regression trend lines are gray in the normal group and black in the glaucoma group.