Quantitative Analysis of Choriocapillaris Using Swept-Source Optical Coherence Tomography Angiography in Eyes with Angioid Streaks

Abstract

Purpose: to quantitatively analyze choriocapillaris perfusion using swept-source optical coherence tomography angiography (SS-OCTA) in eyes presenting with angioid streaks in comparison with control eyes. Methods: Macular 6 × 6 mm SS-OCTA scans were retrospectively analyzed in eyes with angioid streaks and in control eyes. En face choriocapillaris flow images were compensated with en face choriocapillaris structure images, followed by the Phansalkar local thresholding method (with a window radius of four and eight pixels). Quantitative analysis was performed in the four peripheral 1 × 1 mm corners of the 6 × 6 mm SS-OCTA image to include equidistant and comparable regions. The percentage of flow deficits (FD%), the number and size of the flow deficits (FDs) and the total area of FDs were then calculated. Results: 54 eyes of 31 patients were included in the study: 27 eyes diagnosed with angioid streaks and 27 controls. Analysis of the four 1 × 1 mm peripheral corners of the 6 × 6 mm SS-OCTA image showed that eyes with angioid streaks had a higher FD% compared to the control group (47.62 ± 8.06 versus 38.90 ± 6.38 using a radius of four pixels (p < 0.001); 48.37 ± 7.65 versus 39.66 ± 6.51 using a radius of eight pixels (p < 0.001). The average size of FDs as well as the total area size of the FDs were significantly higher in eyes with angioid streaks compared to control eyes (p < 0.001). Eyes with angioid streaks present reduced choriocapillaris flow compared to control eyes. Decreased choriocapillaris perfusion may contribute, among other factors, to the development of neovascularization and atrophy in patients with angioid streaks.

Keywords: angioid streaks; choriocapillaris; optical coherence tomography-angiography; retina.

Figure 1

Image processing methodology used for quantification of the choriocapillaris flow. En face choriocapillaris structure image (A) was transformed using the Fiji “Invert” function (B). Gaussian blur filter was then used for smoothing (C). The en face choriocapillaris flow image (D) was multiplied with the resulting en face choriocapillaris structure image (C) with the “Image Calculator” function. A compensated en face choriocapillaris image was created (E). The compensated en face choriocapillaris flow image was then binarized, to quantitatively measure the flow deficits as reported in the current literature (F). Analysis was performed within the four corners of the OCTA 6 × 6 mm images measuring each 1 × 1 mm. This analysis was performed in order to include comparable and equidistant regions outside of the area of choroidal neovascularization, when present.

Figure 2

Quantitative analysis of the choriocapillaris and multimodal imaging of a 55 year-old patient presenting with angioid streaks. (A): Ultra-wide field color fundus photography revealing the presence of grey peripapillary jagged break lines. (B): Ultra-wide field autofluorescence fundus photography in the same patient showing hypo- and hyperautofluorescent break lines radiating from the optic disc. (C): Spectral domain optical coherence tomography in the same patient showing peripapillary atrophy. (D): Choriocapillaris en face flow image of the patient in panels A–C, as well as the resulting image after image processing using the Phansalkar local thresholding method with a window radius of four pixels (E) and eight pixels (F). Of note, the right panel shows the images of a 57-year-old healthy control choriocapillaris en face (G) flow image as well as the resulting image after image processing using the Phansalkar local thresholding method with a window radius of four pixels (H) and eight pixels (I).