11

Spectral Domain Optical Coherence Tomography Angiography Using NIDEK RS-3000 Advance

Mayss Al-Sheikh, MD and SriniVas R. Sadda, MD

Optical coherence tomography angiography (OCTA) produces 3-dimensional volumetric flow information, which can be segmented into different layers to visualize and evaluate the microvascular structure of the retina and choroid.¹ The technology relies on motion contrast to separate moving structures (eg, blood cells) from stationary structures (eg, surrounded extravascular retinal structures). To visualize the motion contrast image on a macula, 2 or more cross-sectional images (B-scans) at the same position but separated in time are acquired, the speckle difference calculated, and additionally averaged inside the scanning region in a square from 3 mm to 9 mm set in advance. Afterward the difference of fluctuation of the OCTA signal (phase and amplitude) is obtained from the following formulas (1) and (2).

• N: B-scan numbers of optical coherence tomography (OCT) images at the same scan line position
  
• ΔA(x,z): Additionally averaged absolute complex difference on the coordinates (x,z)
  
• ΔA_n(x,z): Absolute complex difference of (N)th B-scan on the coordinates (x,z)
  
• A_n+1(x,z): Complex OCT signal of (N+1)th B-scan on the coordinates (x,z)
  
• A_n(x,z): Complex OCT signal of (N)th B-scan on the coordinates (x,z)

Different retinal reference surfaces are used to create retinal and choroidal slabs to allow en face visualization of the corresponding vascular supply for a specific layer.

PRINCIPLES OF OPERATION

The NIDEK OCTA (NIDEK) has a central wavelength of 880 nm and an acquisition speed of 53,000 A-scans/second. Each B-scan consists of 256 A-scans and repeats 4 times at the same location to compute OCTA. The optical axial and transverse resolution is 7 μm and 20 μm in tissue.

The allowable field of view using the NAVIS-EX software (NIDEK) is 3 × 3 mm to 9 × 9 mm, but since the number of A-scans is fixed for each B-scan, smaller fields of acquisition are associated with a higher resolution. However, a panorama image acquisition of 6 × 6 mm, 9 × 9 mm, and 12 × 9 mm is available, allowing montage of individual high-resolution OCTA over very large regions (Figure 11-1).

Cross-sectional OCT angiograms provide flow information of the different layers with corresponding segmentation lines on B-scan (Figure 11-2A). The segmentation can be created automatically or manually. An automated segmentation algorithm for retinal and choroidal layers is provided with the OCTA NAVIS-EX software. Because automated segmentation will not always create an appropriate segmentation when there is retinal disruption due to various pathologies, manual correction may be considered to adjust the positions of the relevant boundaries. The thickness of every C-scan may also be modified, which allows the analysis of different structures included in a slab of tissue (Figure 11-2B).

En face OCT angiograms are generated by summarizing the blood flow information within a specified depth of relevant anatomic layers or slabs. It takes the maximum or average flow values depending on the protocol. In this way, the 3-dimensional angiogram can be compressed into two-dimensional images.

Different retinal boundaries are defined as reference for the segmentation of OCTA volume. The reference boundaries include the internal limiting membrane (ILM), outer boundary of the inner plexiform layer (IPL), outer boundary of the outer plexiform layer (OPL), and the retinal pigment epithelium/Bruch’s membrane layer (RPE/BM). Using those references, NIDEK OCTA software generates en face images from slabs at 4 different retinal and choroidal layers (see Figure 11-2B): the superficial retinal layer (SRL), the deep retinal layer (DRL), the outer retina (avascular), and the choriocapillaris. The automated segmentation defines the en face slab for the SRL to extend from 8 μm below the ILM to the inner boundary of the inner nuclear layer (INL). The en face slab for the DRL extends from the inner boundary of the INL to 84 μm below.

In a healthy eye, the SRL shows the vessels that are located in the nerve fiber layer (NFL) and ganglion cell layer (GCL). It includes large retinal blood vessels as well as the fine capillaries of the superficial plexus (Figure 11-2C). The deep retinal plexus shows the fine capillary network, which consists of 2 separate layers, that “bracket” along the inner and outer borders of the INL (Figure 11-2D). The outer retina is avascular, but may show projections of the vessels of the inner retina due to the tails of the flow signal (Figure 11-2E). The projection is most apparent on a highly reflective surface such as the RPE. The choriocapillaris in the macula has very small inter-sinusoidal spaces that are generally below the resolution-limit of OCTA, as a result OCTA images at the level of the choriocapillaris show a “confluent” flow (Figure 11-2F).

QUANTIFICATION

Automated Vessel Density Measurement

The NIDEK OCTA system provides a software tool for automated vessel density measurement using a threshold method. After creating the en face images of the different tissue slabs (Figures 11-3A and 11-3B representative for SRL and DRL), the software algorithm identifies all bright pixels corresponding to the vasculature. The software automatically calculates the vessel density as the percentage of area occupied by vessels (Figures 11-3C and 11-3D).

Semiautomated Foveal Avascular Zone Area Measurement

The area of the foveal avascular zone (FAZ) is defined as the area inside the inner border of the capillary network. It can be manually outlined in each retinal layer, and the area can be computed automatically (Figures 11-3E and 11-3F).

Stay updated, free articles. Join our Telegram channel

Join