Quantitative optical coherence tomography angiography of vascular abnormalities in the living human eye

Abstract

Retinal vascular diseases are important causes of vision loss. A detailed evaluation of the vascular abnormalities facilitates diagnosis and treatment in these diseases. Optical coherence tomography (OCT) angiography using the highly efficient split-spectrum amplitude decorrelation angiography algorithm offers an alternative to conventional dye-based retinal angiography. OCT angiography has several advantages, including 3D visualization of retinal and choroidal circulations (including the choriocapillaris) and avoidance of dye injection-related complications. Results from six illustrative cases are reported. In diabetic retinopathy, OCT angiography can detect neovascularization and quantify ischemia. In age-related macular degeneration, choroidal neovascularization can be observed without the obscuration of details caused by dye leakage in conventional angiography. Choriocapillaris dysfunction can be detected in the nonneovascular form of the disease, furthering our understanding of pathogenesis. In choroideremia, OCT's ability to show choroidal and retinal vascular dysfunction separately may be valuable in predicting progression and assessing treatment response. OCT angiography shows promise as a noninvasive alternative to dye-based angiography for highly detailed, in vivo, 3D, quantitative evaluation of retinal vascular abnormalities.

Keywords: ocular circulation; ophthalmic imaging; optical coherence tomography angiography.

Conflict of interest statement

Conflict of interest statement: Oregon Health & Science University (OHSU), Y.J., J.G.F., and D.H. have a significant financial interest in Optovue, Inc., a company that may have a commercial interest in the results of this research and technology. These potential conflicts of interest have been reviewed and managed by OHSU. J.G.F. and D.H. receive royalties on an optical coherence tomography patent licensed by the Massachusetts Institute of Technology (MIT) to Carl Zeiss Meditec. J.G.F. and J.H. receive royalties from intellectual property owned by MIT and licensed to Optovue, Inc. Other authors do not have financial interest in the subject of this article.

Figures

Fig. 1.

Optical coherence tomography (OCT) angiography (3 × 3 mm) of a healthy human eye, acquired using a 70-kHz spectral OCT system with an 840-nm center wavelength. (A) Cross-sectional composite OCT angiogram. Depth layer segmentation lines are shown in green. BM, Bruch’s membrane; ILM, internal limiting membrane; OPL, outer plexiform layer. Flow signals are color coded by depth: purple, anterior to the OPL; red, posterior to BM. (B) En face OCT angiogram above the ILM shows the normal, avascular vitreous. (C) En face OCT angiogram between the ILM and OPL shows the normal retinal vasculature. (D) En face OCT angiogram between the OPL and BM shows the normal, avascular outer retina. (E) En face OCT angiogram of the inner 10 µm of the choroid shows dense, relatively even flow throughout the central macula (3 × 3 mm). (F) En face OCT structural image with an inverse gray scale shows the deeper choroid with medium- and large-sized vessels.

Fig. 2.

Quantification of inner retinal blood flows in normal control (A1–D1) and nonproliferative diabetic retinopathy (NPDR) with macular edema (A2–D2). OCT angiography was acquired using a 70-kHz spectral OCT system with a center wavelength of 840 nm. White dashed circle indicates normal foveal avascular zone (FAZ, 0.6 mm diameter white dashed circle). Area between white and blue dashed circles indicates parafoveal zone. Area between blue and green dashed circles indicates perifoveal zone. First column (A1 and A2) indicates fundus photo (A1) and fluorescein angiography (A2). Second column (B1 and B2) indicates en face 3 × 3 mm OCT angiograms. Enlargement of the FAZ is present in the parafoveal region for NPDR (B2). Parafoveal and perifoveal retinal flow indexes (vessel densities) are shown on en face 6 × 6 mm OCT angiograms (C1 and C2). Nonperfusion areas (blue) are shown of 6 × 6 mm OCT angiograms (D1 and D2). The normal subject had a FAZ of 0.30 mm2, whereas the NPDR case showed an enlarged FAZ and scattered areas of macular nonperfusion totaling 7.07 mm2. Decorrelation (shown by the scale bar) as calculated by the SSADA algorithm measures fluctuation in the backscattered OCT signal amplitude (intensity). It is linear to blood flow velocity up to a saturation limit after which it reaches a maximum value. The linear range is approximately within the limits of capillary flow velocity.

Fig. 3.

Proliferative diabetic retinopathy (PDR) case imaged using a 100-kHz swept-source OCT system with a center wavelength of 1,050 nm. (A) Confocal scanning laser ophthalmoscope (cSLO) shows retinal neovascularization at the optic disk (NVD) and attenuated retinal vessels. (B) FA showing NVD and peripapillary capillary dropout. The green squares in A and B outline the 3 × 3 mm area shown on the OCT angiogram below. (C) En face OCT angiography showing NVD and areas of capillary dropout that correspond to FA (NV is shown in light red gold; normal retinal vessels are in purple). The area of NVD was 0.47 mm2. The vitreous flow index was 0.022. (D) Cross-sectional composite OCT angiogram showing NVD above the inner limiting membrane (red/gold).

Fig. 4.

Type I CNV case imaged by a 100-kHz swept-source OCT system with a center wavelength of 1,050 nm. The CNV is identified by OCT angiography (3 × 3 mm), but it is ill defined by fluorescein angiography (FA). (A) Fundus photograph. The black square outlines the areas shown on the angiograms. (B) Late stage fluorescein angiograph showing an occult CNV. (C) Composite en face color-coded OCT angiogram with CNV flow highlighted in yellow. The CNV area was 0.96 mm2 and the outer retina flow index was 0.012. The yellow dashed line indicates the position of the OCT cross-section. (D) Cross-sectional color OCT angiogram. Both composite en face (C) and cross-sectional color OCT angiograms (D) show inner retinal flow in purple, outer retinal flow (CNV) in yellow, and choroidal flow in red. The CNV is predominantly under the RPE.

Fig. 5.

Geographic atrophy case imaged by a 100-kHz swept-source OCT system with a center wavelength of 1,050 nm. (A) Fundus photograph showing the area of geographic atrophy (GA) adjacent to the foveal center. (B) An autofluorescence image sharply outlines the area of absent RPE and is surrounded by a halo of hyperautofluorescence. The green squares in A and B outline the area shown in C–F. (C) A drusen–RPE complex thickness map shows the area of RPE thinning (the dark area nasally). (D) An en face OCT structural image on an inverse gray scale of the deeper choroid reveals medium- and large-sized vessels. (E) An en face OCT angiogram (3 × 3 mm) of the choriocapillaris shows dramatically decreased, but not absent, choriocapillaris flow in the area of GA. (F) The light blue color represents the choriocapillaris nonperfusion area (2.75 mm2). (G and H) Cross-sectional composite OCT angiograms show the absence of choriocapillaris flow in most of the area of GA, but the flow at the edge of the atrophy is spared (shown by the green arrows in the magnified views).

Fig. 6.

Choroideremia case imaged by a 100-kHz swept-source OCT system with a center wavelength of 1,050 nm. The large-field en face OCT angiograms (∼3 × 8.5 mm) were produced by stitching together three 3 × 3 mm scans. (A) OCT angiography of inner retinal blood flow. (B) OCT angiography with quantification of inner retinal blood flow demonstrating patchy areas of nonperfusion (blue) in the extrafoveal macula. The total nonperfusion area of the inner retina was 7.65 mm2. (C) OCT angiography of the choroidal blood flow, including choriocapillaris and deep choroid. It should be noted that OCT angiography is able to image deeper choroidal vessels in this case due to the relative absence of the overlying choriocapillaris and RPE. (D) OCT angiography of the choroidal blood flow with quantification of the choriocapillaris nonperfusion area (purple), which was 12.11 mm2 (47.5% of image area). (E) An autofluorescence image outlines the area of existing RPE (hyperautofluorescent area).