Optical Coherence Tomography Angiography (OCTA) of the eye: A review on basic principles,advantages,disadvantages and device specifications

Abstract

Optical Coherence Tomography Angiography (OCTA) is a relatively new imaging technique in ophthalmology for the visualization of the retinal microcirculation and other tissues of the human eye. This review paper aims to describe the basic definitions and principles of OCT and OCTA in the most straightforward possible language without complex mathematical and engineering analysis. This is done to help health professionals of various disciplines improve their understanding of OCTA and design further clinical research more efficiently. First, the basic technical principles of OCT and OCTA and related terminology are described. Then, a list of OCTA advantages and disadvantages, with a special reference to blood flow quantification limitations. Finally, an updated list of the basic hardware and software specifications of some of the commercially available OCTA devices is presented.

Keywords

OCTA eye basic principles advantages disadvantages pros and cons microvessels blood flow device specifications

1 Introduction

Up to 1980, the commonly available instruments for visualizing the ocular fundus were the direct and indirect ophthalmoscope, the slit lamp microscope, and the fundus camera. Imaging of the retinal tissue is one of the most common diagnostic procedures in ophthalmology since the clinical doctor has access to both the microcirculation and the neural tissue and their related diseases.

In 1981 an original paper appeared [1], introducing a new device that combined optics with electronics under the name Scanning Laser Ophthalmoscope (SLO). Significant advantages of SLOs over conventional fundus cameras were more than 100 times less luminous irradiance, insensitivity to the quality of the collection optics, and rejection of stray light (in their confocal versions [2]). Due to these advantages, fluorescein angiography could be performed with one-tenth of the dye, in both eyes [3].

In 1991 the pioneer paper on Optical Coherence Tomography (OCT) appeared [4], demonstrating that in vitro measurements at the human retina and the interior of the coronary arteries are possible. The basic components of the experimental setup were a low-coherence light source (instead of a laser in SLOs) and a Michelson Interferometer. Since then, the applications moved from in vitro to in vivo and into the clinic and many wonderful things happened in the field of OCT. OCT was the fastest-adopted imaging technique in the history of ophthalmology [5] and in 2010 the number of ophthalmic OCT scans was comparable to the number of MRI, SPECT, PET, and CT examinations [5]. One of the relatively novel OCT imaging applications is OCT angiography (OCTA) which was commercially introduced in 2014 [6]. OCTA utilizes the blood movement within the vasculature in order to construct an angiogram of the retina and the choroid and to provide volumetric 3D maps of the fundus microvessels. However, there are important limitations and especially on absolute blood flow quantification.

In this review paper, we are going to describe the basic principles of OCT and OCTA, advantages and disadvantages, and basic hardware and software specifications of some of the commercially available OCTA devices. This is done in the simplest possible language for helping health professionals without deep knowledge of optoelectronics. In more detail, the paper consists of the following sections: section 1 is the introduction, section 2 includes a basic description of the basic OCT principles and terms, section 3 focuses on the OCT methods for visualizing the microvessels of the retina described with the collective term OCTA (OCT Angiography), section 4 includes the advantages of OCTA, section 5 includes the disadvantages of OCTA with a special reference to blood flow quantification, section 6 includes up to date basic hardware and software specifications of some of the commercially available OCTA devices and section 7 is a brief conclusion.

2 OCT

Huang et al. in 1991 [4], were the first to extend low-coherence reflectometry to tomographic imaging in biological systems called Optical Coherence Tomography (OCT). The central component of their apparatus was a fiber optic Michelson interferometer (Fig. 1) illuminated by a low-coherence superluminescent diode (at 830 nm). The Michelson interferometer has two arms, a movable mirror arm (or reference arm) and a fixed mirror arm, but in biological applications, the fixed mirror is replaced by the tissue sample of interest. Huang et al. demonstrated the feasibility of OCT on the human retina and coronary artery in vitro.

Fig. 1

A schematic diagram of the basic components of a Michelson Interferometer. There are two arms: the movable mirror arm (or reference arm) and the fixed mirror arm. In biological applications, the fixed mirror is replaced by the tissue sample. There is also a half-silvered mirror (beam splitter).

The basic mechanism of OCT is very similar to that of ultrasound but instead of sound waves, OCT uses the reflection of low-coherence light. However, even though the basic mechanism is the same, the resolution is far superior due to a much lower wavelength and in contrast to the ultrasound devices, no contact is needed between the probe and the tissue. Huang et al. [4] reported a lateral and axial resolution of the order of 10 μm and a sensitivity of 10 femtowatt (fW) demonstrating better performance than confocal SLO.

2.1 A-scan

The amplitude scans (or A-scans) are one-dimensional and give the reflection intensity at various depths in the axial direction of light transmission. Under the condition that the path length difference between the two light paths in the two arms is less than the coherence length of the light source when the reflected beams recombine, interference occurs. The interference profile can give intensity information, in the form of a reflectivity profile in depth. Changing the location of the reference scanning mirror allows backscattered tissue intensity levels to be detected from different depths in the tissue sample for the A-scan (Fig. 2a). This approach of obtaining an A-scan is called Time-Domain OCT (TD-OCT) because time-encoded signals are obtained using a movable (scanning) reference mirror. The basic block diagram of a TD-OCT instrument is similar to that shown in Fig. 1 where in the place of the observer a photodetector exists (Fig. 3a). The first experimental instruments required manual adjustments of contrast (and other parameters) until these procedures were automated in the commercial products. The first TD-OCT instruments became available by Zeiss in 1996 and after 6 years, a third-generation device was available (Stratus OCT3, Zeiss) with an improved axial scan speed of 400 A-scans per second [7].

Fig. 2

(a) An OCT A-scan gives the signal intensity in the axial direction of the eye. At fundus depths of about 240 and 330 pixels, there are steep increases in the received light intensity (b) The white column in the image gives the position of the A-scan that corresponds to the intensity diagram below the black arrow. Many A-scans combined along a line in the horizontal direction, form a B-scan cross-sectional image. The B-scan here demonstrates a cross-section of the human retina depicting the macula. The B-scan image was taken with Moptim Mocean 4000.

Fig. 3

(a) A schematic diagram of the basic components of a time-domain OCT (TD-OCT). The TD-OCT includes a movable reference scanning mirror. Changing the location of the reference scanning mirror allows backscattered tissue intensity levels to be detected from different depths in the tissue sample for the A-scan. Light is then detected by a photodetector. (b) A schematic diagram of the basic components of a spectral-domain OCT (SD-OCT). In contrast to the TD-OCT, it comprises a fixed reference mirror from which the light gets reflected to a diffraction grating and the interference pattern obtained is divided into its frequency components which are all detected at the same time by a 1D light detector. (c) A schematic diagram of the basic components of a swept-source OCT (SS-OCT). Like the SD-OCT, it uses a fixed reference mirror that reflects light. But here, instead of a spectrometer and a linear detector, a much simpler photo-detector is used because the light source is a tunable laser with a central wavelength λ o and a “tuning range” (shown here by the range of a bell-shaped curve). An instantaneous line with a small width sweeps across the “tuning range” producing narrow bands of wavelengths.

2.2 B-scan

Many A-scans at a series of lateral locations along a line can be combined to create a two-dimensional image (B-scan) which gives a cross-sectional view of the tissue sample in grayscale (Fig. 2b). With the aid of the B-scans, cross-sectional slices of the human retina became available for the first time, along any line drawn by the operator on the fundus area.

The next step would be the ability to acquire an image of a fundus area but, in order to achieve this, many B-scans would have to be combined. This means there was a great need for speed and as it was clear from the nineties, the moving parts of the TD-OCT were a great disadvantage. The basic idea to overcome this disadvantage was to acquire the sample reflectance as a function of all depths together through Fourier transform, without a moving mirror, thus facilitating rapid A-scan collection [8]. The acquired signal is the integrated spectrum of the light source, superimposed by fringes whose frequency encodes the path length imbalance of the interferometer. Using the Fourier transform, the sample reflectance is represented as a function of depth, which permits rapid A-scan collection.

In practice, the Fourier transform idea can be implemented in two ways: first, by utilizing a broadband light source, a spectrometer, and a one-dimensional (1D) detector as a receiver, and second, by utilizing a wavelength-swept light source and a standard single-point photodiode receiver. Even though the terminology is not yet fully standardized, the first configuration comprising a spectrometer is referred to as spectral radar or spectral-domain OCT (SR-OCT or SD-OCT), and the second configuration comprising a wavelength-swept laser is referred to as swept-source OCT (SS-OCT) [9].

2.3 SD-OCT

After the presentation of the initial experimental works on SD-OCT [8, 10] the first commercial device was launched by Optovue in 2006 [7]. In SD-OCT the reference mirror becomes fixed in place and the light gets reflected back to a diffraction grating where it is analyzed to its frequency components (Fig. 3b). The frequency components obtained from the diffraction grating (Fourier Transform) are all detected at the same time by a line (1D) scan charge-coupled device (CCD). This device has 1D light detectors sensitive to different frequencies which correspond to the different tissue depths within the tissue. It is of note, that all points of an SD-OCT A-scan are received simultaneously, generating an increased scanning speed of around 20,000–40,000 A-scans/sec [11]. This increased scanning speed reduces the possibility of motion artifact, increases the resolution, and diminishes the possibility of not locating lesions. The top A-scan rate is limited by the CCD device reading rate. Current commercial SD instruments operate at 70,000 to 125,000 axial scans/sec. In research systems, speeds of 147 and 300 kHz have been presented [12, 13].

Spectral-domain systems increase the signal-noise ratio by image-averaging multiple B-scans at the same location. Spectral-domain systems originally operated at 800–870 nm wavelengths with a difficulty in penetrating deeper into fundus tissue due to high reflection and absorption in the retinal pigment epithelium and in the choroid. This problem was solved in two ways: 1) a kind of image averaging called “Enhanced Depth Imaging (EDI) [14] and 2) by increasing the wavelength of the superluminescent diode (SLD) to the area of 1050 – 1060 nm. The systems with longer wavelengths (1050–1060 nm) offer deeper penetration and better posterior eye imaging with better sensitivity and specificity of diagnosis [13]. The water absorption window at ∼ 1050 nm, in the near-infrared part of the electromagnetic spectrum [15], has the advantage of increased penetration to the choroid due to reduced scattering by retinal pigment epithelium. In addition, it has the advantage of increased penetration to the optic nerve and through cataracts and various ocular opacities (eg. drusen). In this way, the shadowing artifact (see section 5.1) where existent blood flow is not displayed, is reduced.

In addition to deeper penetration and increased speed, new SD-OCT machines approach a 3-micron axial resolution (see Table 1 in section 6) in comparison to 10–15 microns of TD-OCT. While most TD-OCT instruments image about 6 radial slices, SD-OCT continuously images a 6 mm area. This decreases the chance of inadvertently missing pathology.

Table 1
OCTA device hardware specifications

OCTA device HARDWARE SPECIFICATIONS

FDT Speed λ (nm) AR TR MFOV (mm×mm)

1 SD 70,000 855 1.6/3 –/20 10×10

2 SD 125,000 * 880 3.9/7 5/14 9×9

3 SD 85,000 880 –/7 –/20 9×9

4 SD 120,000 940 –/5 –/15 12×12

5 SS 100,000 1050 2.6/8 –/20 12×12

6 SS 200,000 ** 1040–60 1.9/6.3 –/20 12×12

OCTA device	HARDWARE SPECIFICATIONS
1	SD	70,000	855	1.6/3	–/20	10×10
2	SD	125,000 *	880	3.9/7	5/14	9×9
3	SD	85,000	880	–/7	–/20	9×9
4	SD	120,000	940	–/5	–/15	12×12
5	SS	100,000	1050	2.6/8	–/20	12×12
6	SS	200,000 **	1040–60	1.9/6.3	–/20	12×12

OCTA devices are in alphabetical order. 1: CANON XEPHILIO OCT-A1, 2: HEIDELBERG ENGINEERING SPECTRALIS-SHIFT, 3: NIDEK RS-3000 ADVANCE-2 ANGIOSCAN, 4: OPTOVUE SOLIX ANGIOVUE, 5: TOPCON DRI OCT TRITON, 6: ZEISS PLEX ELITE-9000 ANGIOPLEX. FDT: Fourier Transform Technology (Spectral-Domain (SD) or Swept-Source (SS)), Speed: A-scan speed (A-scans/s), λ: wavelength (nm), AR: Axial Resolution (Digital/Optical); Digital resolution is measured in μm/pixel and Optical resolution is measured in μm, TR: Transverse or Lateral Resolution (Digital/Optical), MFOV: Maximum Field Of View (mm) on the retina. *A-scan speed can take values of 20 kHz/42,5 kHz/85 kHz/125 kHz (SHIFT technology). **A-scan speed can change at values of 100 kHz and 200 kHz (Dual-speed).

2.4 SS-OCT

The basic concept of Swept Source OCT (SS-OCT) was initially described in experimental studies in 1997 [16]. Its name refers to the sweeping light used in its tunable laser light source (Fig. 3c). This tunable laser light source has a central wavelength but is also able to sweep across numerous, discrete, narrow bands of wavelengths thus having a “tuning range” (for example from ∼ 1260 nm to ∼ 1360 nm in the work of Choma et al. [17] and a tuning range of 72 nm, with a center wavelength of 1060 nm, in the work of Yasuno et al. [18]). Similarly to an SD-OCT, the SS-OCT occupies a fixed reference arm but does not use a spectrometer and a linear detector with their associated losses. Instead, the light waves returning from the sample are detected by a point photodetector which is simpler and faster than the linear CCD detector giving increased scanning speeds (Fig. 3c). The resolution of the interferogram fringes in the SS-OCT systems is dictated by the instantaneous laser light line and not by the spectrometer-linear CCD detector combination as in SD-OCT. Ideally, the laser light emitted would be of a single wavelength at a moment but in reality, it is constructed of more than one wavelength spanning across a range that is named the “instantaneous line width”.

SS-OCT has many advantages over SD-OCT such as better sensitivity and imaging depth, longer imaging range, higher detection efficiency, and the ability of dual-balanced detection [19]. These characteristics permit the SS-OCT to have significant use in the choroid, vitreous and retinal visualization and research [20]. Early studies by Yun et al. [21] utilizing an SS-OCT demonstrated speeds of 15,700 A-scans/sec, with a central wavelength of 1310 nm, an axial resolution of 13.5 μm and a high sensitivity of –110 dB on the ventral portion of a human finger. Later, increased acquisition speeds of 100,000 to 400,000 A-scans/sec were reported depending on the configuration [19 , 23] and with Fourier Domain Mode Locking (FDML) the incredible speed of 20,8 million A-scans/sec was achieved on the human finger [24]. Klein et al. [25] managed to take 6,7 million A-scans/sec on the human retina with a central wavelength of 1050 nm, since water absorption does not permit retinal imaging at 1310 nm [19].

The axial resolution of SS-OCT systems, is inversely proportional to the full width half maximum (FWHM) bandwidth of the light source, the optical amplifier and the tunable filter. Lee et al. [26], managed to build an ultrahigh resolution (UHR) SS-OCT system with an axial resolution of 2.9 μm in tissue. The transverse resolution of OCT systems depends on the OCT beam size which is typically much larger than a red blood cell and so, individual blood cells cannot be discriminated [27]. SS-OCT was introduced in clinical practice in Japan in 2012 [20] but it was FDA approved for posterior imaging in 2016.

2.5 C-scan and other scans

The T-scan, an abbreviation for transverse scan, is performed in the transverse axis of the eye (Fig. 4) along a lateral direction at a given depth. During one T-scan the axial position is fixed and multiple T-scans are necessary for the generation of a C-scan [28]. The C-scan is also known as en face OCT image or en face OCT section and is a topographic frontal view of the structure at a given depth (Fig. 4). A red-free image of the fundus has essentially the same appearance as an en face scan of the retina and both demonstrate the layout of the retinal surface. In some cases, T-scans can also be performed radially (at a given angle) or circularly (at a given polar radius) of the eye [28].

Fig. 4

The relative orientation of an axial scan (A-scan), a transverse scan (T-scan), a longitudinal section (B-scan) and, an en face section (C-scan) on the human eye. Modified from Rosen et al. [28].

There are 2 ways to generate an en face image of a visualized structure: the flying spot scanning way and the full-field way. In the first way (flying spot scanning), imaging systems use a spot of light as a scanning source and are able to scan the sample in 2 orthogonal directions. Flying spot imaging systems are heavily utilized in ophthalmic scanning systems like the modern OCT devices. The older OCT models were using the A-scan at adjacent transverse positions to generate the A-based B-scan (Fig. 2b). Newer OCT models utilize multiple T scans (Fig. 4) to the generation of a T-based B-scan [28].

The second way (full-field) is similar to a single snapshot of a structure analogous to a bone x-ray and can be captured at an instant. In contrast to conventional flying spot scanning OCT devices which require lateral scans to generate the image, in full-field imaging it is possible to get a 2D en face image in one capture by using a Linnik-based interferometer configuration occupying an array detector (CCD or CMOS camera) and two identical high numerical aperture microscope objectives. The microscopy application of full-field OCT imaging is called Full-Field Optical Coherence Tomography (FF-OCT) or Full-Field Optical Coherence Microscopy (FF-OCM) and was developed in ESPCI Paris [29]. The use of FF-OCT in vivo is limited by the imaging speed of the CCD and CMOS cameras [30].

The M-scan is basically a series of repeated A-scans at a specific location represented as a function of time, thus generating time-lapse imaging of a single spot.

3D cube raster scans are volume scans equivalent to CT or MRI scans that obtain volumetric cubes of data through a fast series of parallel and adjacent B-scans (Fig. 5). 3D cube raster scans are used today to scan the optic nerve and the macula in OCTA image generation [28]. The data inside the volumetric window can be processed and compared to subsequent scans. The distance between the B-scans and the pixels in both x and y directions can be manipulated generating a rectangle or square scanning region. According to Sinha et al. [31] the Cirrus HD-OCT is equipped with 2 macular cube protocols, the 512×128 protocol that analyses 128 horizontal scans at high scan density (512 A-scans per B-scan) and the 200×200 protocol that analyses 200 horizontal scans at lower scan density (200 A-scans per B-scan) but with higher imaging speed as it takes only 1.6 sec. A 3D-volume rendering software recreates a 3D image of the complete retinal cube.

Fig. 5

An en face image from a 3D cube raster scan of the inner plexiform layer (IPL) of the fundus of the normal right eye with the Foveal Avascular Zone (FAZ) of the macula close to the center. This image was taken with Optovue SOLIX.

3 Optical Coherence Tomography Angiography (OCTA)

Following the invention and the clinical use of the OCT, various theoretical and experimental studies were performed to find further potential applications of this technology in visualizing important anatomical structures. Concerning the field of ophthalmology, the visualization of retinal and choroidal vessels in high detail has always been considered an important step for a better diagnosis, follow-up, and management of pathologies. OCT angiography (OCTA) detects changes in backscattered OCT signal in order to effectively differentiate locations of blood flow from locations of static tissue. The basic concept of the imaging technique is depicting as black the locations without variance (static tissue) and as white the locations of high variance due to the detected blood flow. In this way the microvessels appear bright in a dark background. The identification of the pixels with blood flow is accomplished with a motion contrast algorithm.

3.1 Motion contrast algorithms

There are many algorithmic techniques for creating motion contrast such as phase-based, amplitude-based and both phase and amplitude based but the most widely used techniques today are amplitude-based [27].

Phase-based flow detection techniques [32 –36] detect phase variations from microvascular flow and work better with SD-OCT which provides high phase stability. Makita et al. [32] were the first to describe the idea of an angiography using SD-OCT. Phase-based Doppler OCT is a combination of optical coherence tomography (OCT) with Doppler velocimetry and permits imaging of blood flow locations (microvessels). Doppler shift of the OCT signal was used to contrast blood vessels and angiographic images were produced by integrating the power of Doppler shift images [32]. In bi-directional flow Doppler OCT the average phase change between adjacent A-scans is used to calculate the Doppler frequency shift [32 , 38]. Then the location of the moving particles can be determined and furthermore the average particle velocity, provided that the angle between the velocity and the optical beam is known. In variance Doppler OCT, the standard deviation of the Doppler frequency spectrum of the A-scans is used and provides a better mapping of microvessel location with higher contrast [32 , 39].

Yasuno et al. [18] produced choroidal angiographs using SS-OCT en face images instead of B-scans. Their scattering technique was a software-based 3D segmentation method for choroidal vessels using intensity-threshold binarization (ITB) that does not require phase stability as Doppler angiography does.

In amplitude-based flow detection techniques, changes in the backscattered OCT signal amplitude (intensity) between sequential OCT B-scans of exactly the same cross-section are computed and a microvessel blood flow map is created. Amplitude-based flow detection techniques are better suited to SS-OCT [40] and are presented in more detail below.

3.2 Amplitude-based motion contrast algorithms

In amplitude-based OCTA techniques, multiple B-scans (B₁, B₂,...B_N) are performed at the same tissue plane and the structural images are compared pixel by pixel in order to detect intensity (amplitude) changes due to flowing erythrocytes. From these changes, a motion contrast visualization of the vasculature is created. Acquiring each B-scan image takes a time T_A (acquisition time) and then the OCT beam must be rapidly scanned back to the initial position during the flyback time T_F which is around 10 – 20% of the acquisition time (Fig. 6). The B-scans are repeated after a time interval T_I (interscan time) which is equal to the sum of the acquisition time and the flyback time (T_I = T_A + T_F). Therefore, the interscan time T_I depends on 1) the A-scan speed, 2) the number of A-scans per B-scan and 3) the flyback time T_F.

Fig. 6

Visualization of 3 basic time intervals in optical coherence tomography angiography (OCTA): acquisition time (T_A), flyback time (T_F), and interscan time (T_I). B1 and B2 are two successive B-scans. Each A-scan of a B-scan is also repeated at the interscan time.

Three representatives of amplitude-based OCTA algorithms are the “speckle variance” [41], the “correlation mapping” [42], and the “split-spectrum decorrelation” which is a subcategory of decorrelation algorithms [40]. Here, we are going to focus on decorrelation motion contrast algorithms.

After acquiring a series of B-scans (assuming a number of N scans with the exact value of N depending on the instrument) the basic steps of amplitude-based decorrelation OCTA image processing [27] are 1) registration of B-scans, 2) implementation of a decorrelation motion contrast algorithm for producing an unthresholded OCTA image and 3) a thresholding operation for producing the final thresholded OCTA image. More information on these steps is given below.

3.2.1 Registration

With the first step of image registration, the target is to cancel signal artifacts caused by eye motion. There are many different eye movements that can cause artifacts such as microsaccades, bulk eye motion, and expansions from increases in intraocular pressure. With registration, bulk tissue displacements in the axial direction are removed, ensuring that the detected changes result from the movement of red blood cells. However, registration of successive B-scans can only correct motions within the B-scan image plane.

3.2.2 Implementation of decorrelation motion contrast algorithms

In the decorrelation category of the amplitude-based techniques, each pair of B-scans produces a decorrelation frame D and therefore a total of N-1 decorrelation frames are produced which are then averaged to a single decorrelation image (Dav). Irrespective of the particular decorrelation method, the signal in the final decorrelation image Dav should have a value between zero (no flow) and one (maximum flow). Many averaged decorrelation images from different parallel locations in the tissue are combined to form 3D blood perfusion rectangles of the posterior eye [40].

A special kind of decorrelation algorithm is the “split-spectrum amplitude-decorrelation angiography” (SSADA) which was proved particularly optimal for imaging of retinal and choroidal flow compared to other decorrelation algorithms [40]. With the SSADA algorithm, the spectrum of each pair of B-scans is split by applying M equally spaced bandpass filters, and hence, M decorrelation frames are produced from each B-scan pair. In this way, the signal-to-noise ratio and microvessel connectivity are improved in the resulting angiography showing many orders of microvascular branching in the optic nerve head and the macula. In addition, it was demonstrated for the first time that the superficial vascular network supplying the optical disk ends at the disk boundary [40].

As it was mentioned above, irrespective of the particular decorrelation method, the signal in the final decorrelation image Dav should have a normalized value [43] between zero (no flow) and one (maximum flow). The exact relationship between blood flow and OCT signal is usually sigmoidal [27] as it is shown in Fig. 7. The red blood cell (RBC) motion inside a microvessel (which is at a particular pixel group position) affects pixel intensity at repeated A-scans with the repetition interval equal to the interscan time T_I. However, when the blood flow is very slow, the RBC translocation is very small in comparison to the OCT beam size and therefore not enough to cause a change in the OCT signal intensity. This case defines a sensitivity threshold (Fig. 7) which corresponds to the slowest detectable blood flow (SF). In contraast, when the RBCs move very fast, there is a limit above which the OCT signal does not change. This case defines a saturation threshold (Fig. 7) which corresponds to the fastest detectable blood flow (FF).

Fig. 7

The sigmoidal curve describing the relationship between blood flow velocity (on the horizontal axis) and OCTA signal (on the vertical axis) (modified from Spaide et al. [27]). The OCTA sensitivity threshold corresponds to the Slowest detectable blood Flow (SF) and the OCTA saturation threshold corresponds to the Fastest detectable blood Flow (FF). NF stands for “No Flow” representing the zero blood velocity case. Longer interscan times improve the detection of slower RBC motion and consequently, SF and FF move to the left. Shorter interscan times improve the detection of faster RBC motion and consequently, SF and FF move to the right.

The RBC motion detection depends not only on the OCT beam size but also on the interscan time T_I. Therefore, by changing the T_I it is possible to affect the detected values of blood flow. Longer interscan times improve the sensitivity and detection of slower RBC motion at the expense of lowering the saturation threshold and consequently, SF and FF decrease and move to the left in the sigmoidal diagram (Fig. 7). The slowest detectable flow (SF) depends also on the background noise [43]. Shorter interscan times improve the detection of faster RBC motion at the expense of increasing the sensitivity threshold and consequently, SF and FF increase and move to the right in the sigmoidal diagram (Fig. 7).

As it is evident from the above, the interscan time T_I is an important parameter for OCTA sensitivity and saturation and therefore for OCTA image contrast. The range of detectable blood flow speeds can be increased by using different interscan times with a technique known as Variable Interscan Time Analysis (VISTA) and the slower flows in the smaller vessels can be discriminated from the faster flows in the larger vessels by using a false color scale [44, 45].

3.2.3 Thresholding

In many ocular regions without blood flow, such as the vitreous and parts of the choroid, the OCT signal is low and consists mainly of noise. In these cases, OCTA images are produced with false flow depictions. A solution to this problem is to determine a threshold and apply it on a pixel-by-pixel basis to the decorrelation image: OCTA pixels with values below the threshold are assumed invalid and are converted to black. A clever way to determine the value of the threshold for each image pixel is to make a “threshold mask” from a series of registered B-scans [27].

3.3 En face OCTA

En face OCTA is based on structural OCT tissue layer segmentation and on the subsequent projection of 3D flow layer data into a 2D en face image (a single-pixel layer). After selecting and segmenting the desirable retinal layers, an assembly of voxel columns is created from the segmented volume, called 3D flow data. For the construction of the en face OCTA image, the 3D flow data must be converted to 2D flow data and this conversion is done using various techniques such as intensity summation [46], maximal intensity projection, average intensity projection, and histogram distribution projection [27]. With maximal intensity projection, the maximum voxel value from the voxel column is selected for the final 2D en face image. Regardless of the conversion method and the corresponding advantages and disadvantages, in the final en face 2D image, only one voxel value will be represented and depth information is lost. So, if there are more than one vessel in the same voxel column, they will appear as one.

4 Advantages of OCTA

There are important advantages of the OCTA imaging technique which are presented in the following subsections.

4.1 High contrast

OCTA images have much higher contrast than structural OCT images [27] and contrast is one of the most important parameters for diagnosing pathology from clinical images. The difference in contrast between a structural OCT and OCTA en face image can be seen by comparing the macular images from Figs. 5 and 8, respectively.

Fig. 8

High contrast en face OCTA image of the fundus of the normal eye with the Foveal Avascular Zone (FAZ) of the macula in the center. The reader can compare the contrast of this image to the contrast of the image without angiography technology in Fig. 5. This image was taken with the same instrument (Optovue SOLIX) but using angiography technology (ANGIOVUE) in addition.

4.2 Co-registration to structural OCT

OCTA data are co-registered with 3D structural OCT data and this fact gives to the user the versatility of portraying vascular OCTA information superimposed on structural data obtained from the OCT [47]. This technique greatly improves visualization and localization of vascular lesions in the retina or the choroid, allowing the user to make structural correlations.

4.3 Non-invasive

OCTA on a patient can be completed non-invasively, in a few minutes, without any special preparation in the clinical environment. In comparison, Fluorescein Angiography (FA) and IndoCyanine-Green Angiography (ICGA) take about an hour per patient since they require administration of drugs for inducing mydriasis and the injection of a fluorescent dye. On surplus, ophthalmic fluorescent dyes have been associated with systemic adverse effects such as nausea, vomiting [45] and even allergic and anaphylactic reactions [43]. The probability of those reactions increases with the frequency of use hence, OCTA is preferable for the assessment of patients requiring frequent follow-up examinations. In addition, the intravenous administration of ICG dye is contraindicated in pregnancy and kidney disease [47].

4.4 Complementarity to FA/ICGA

In the cases that FA and ICGA must be performed, an asset of the OCTA is that it can be complementary to the FA/ICGA providing 3D imaging information of the fundus. FA and ICGA give only 2-D en-face images with poor depth resolution [18]. Segmentation of different layers is not easily possible with FA/ICGA whereas, localization of the depth of the lesion/pathology is possible with OCTA which provides flow information at a fixed point in time [47].

4.5 Online estimation of digital image parameters

The fifth advantage of OCTA is the ability to calculate online important microvessel digital image parameters-biomarkers. One group of imaging biomarkers is related to microvessel density. First, Vessel Area Density (VAD) is defined (Optovue definition [6]) as the total perfused vessel area per unit area in the region of measurement and is measured in percent (%). It is a biomarker that is more influenced by larger vessels and we agree with Sampson et al. [48] that VAD should replace the terms “vessel density’ and “perfusion density”. Second, Vessel Length Density (VLD) is defined (Zeiss definition [6]) as the total length of perfused vasculature per unit area in the region of measurement and is measured in mm^-1. Length-based measurements of vessel density are more influenced by small capillaries. Third, Vessel Perimeter Index (VPI) is defined as the total perfused vessel perimeter per unit area in the measurement region and is measured in mm^-1. Fourth, the Distribution of Vessel Diameter (DVD) is estimated from the diameters of all identified microvessel segments with each segment measured at least 3 times. Average and median diameter values (in μm) can give valuable information on microvessel dilation [48].

Another group of OCTA imaging biomarkers is relative to the Foveal Avascular Zone (FAZ) cited in the center of the macula within the fovea. In general, the avascular areas are very important in all fundus regions and not just in the fovea. FAZ area (mm²) is calculated from the number of black pixels inside the FAZ perimeter and it is a very important parameter that correlates to various ocular pathologies and to the retinal non-perfusion [45]. Other biomarkers are FAZ diameter and/or perimeter ( μm) and FAZ circularity which is defined as an index of how circular the FAZ perimeter is. FAZ circularity values closer to 1 indicate a FAZ perimeter shape close to a circle whereas higher circularity values indicate a more irregular FAZ shape (less circular) [49].

Other OCTA imaging biomarkers are Fractal Dimension (FD) describing the complexity of vessel branching in the region of measurement and Vessel Tortuosity (VT) describing the abnormal curvature of the vessels (the higher the value the higher tortuosity). FD and VT are both dimensionless.

All the above biomarkers may help in diagnosing and monitoring disease progression. However, they have not been standardized across manufacturers [45 , 50] and this makes it hard to compare results from different instruments, and build large databases. The members of the Diabetic Retinopathy expert committee of the European Vision Clinical Research network have reported [50] that there is an urgent need for a consensus OCTA nomenclature to standardize a clinical and scientific language. In addition, not all manufacturers provide the above biomarkers online and in many situations, image processing calculations have to be performed offline.

4.6 A new tool for microvascular anatomy and physiology

The OCTA’s ability to visualize different retinal vascular layers is a new tool for normal vascular anatomy and physiology investigations (Fig. 9). Four vascular plexuses have been identified at the inner retina [6 , 51]: 1) the Radial Peripapillary Capillary Plexus (RPCP) surrounds the papilla (head of the optic nerve) radially and supplies the superficial retinal nerve fiber layer (RNFL) at the same level, 2) the Superficial Vascular Plexus (SVP) is located in the ganglion cell layer and is the only plexus with larger microvessels including arterioles and venules which lie more superficially close to the retinal nerve fibers, 3) the Intermediate Capillary Plexus (ICP) is located between the inner plexiform layer and the inner nuclear layer; the ICP collocates with the amacrine cells, and 4) the Deep Capillary Plexus (DCP) which is located between the inner nuclear layer and the outer plexiform layer; the DCP collocates with the horizontal cells. The ICP and DCP are near the high-oxygen demand synapses of the inner and outer plexiform layers, respectively [43].

Fig. 9

High contrast en face OCTA image of the fundus of the normal right eye with the Foveal Avascular Zone (FAZ) of the macula in the center (image size of 6×6 mm). The image shows the Superficial Vascular Plexus (SVP) comprising microvessels in the segmentation volume defined between the Internal Limiting Membrane (ILM) and the Inner Plexiform Layer (IPL). Also, part of the Radial Peripapillary Capillary Plexus (RPCP) is shown around the optic nerve head. This image was taken with Canon XEPHILIO-OCT-A1.

In the choroidal area, even though there are no specific anatomic boundaries, there are three vascular layers [27]: 1) the “choriocapillaris”, a capillary anastomotic layer 10–30 μm below the Bruch’s membrane [43], 2) the Sattler’s layer with medium-sized vessels and 3) the Haller’s layer with larger vessels.

4.7 A new tool for eye pathology and other diseases

Even though it is only 8 years since the OCTA was commercially introduced in 2014 [6], there are already numerous studies demonstrating the usefulness of OCTA in various eye pathologies [51, 52] such as choroidal neovascularization (CNV) [53], non-neovascular age-related macular degeneration (AMD), diabetic retinopathy (DR) [50, 54], retinal vascular occlusions, retinal vasculitis [55], sickle cell retinopathy (SCR) [56], myopic degeneration, macular telangiectasia, inherited retinal dystrophies, inflammatory diseases, radiation maculopathy, glaucoma [57, 58], pseudoexfoliative glaucoma (PXG) [59], ocular burns [60], choroidal melanoma [61] and ocular oncology [62]. Wang et al. [53] reported that OCTA is a high diagnostic value method for active CNV with diagnostic accuracy independent of device type and algorithm. On many occasions, OCTA exhibited higher specificity than fluorescence angiography [49].

Cardiovascular disease patients including those with systemic hypertension [63], diabetes mellitus, kidney disease, preeclampsia, coronary artery disease, and carotid artery stenosis exhibit, in general, lower retinal and choroidal vessel density and increased FAZ area and perimeter [64].

After the onset of the COVID-19 pandemic, the OCTA proved useful in supporting that COVID-19 has a microvascular component, demonstrating significant COVID-19-related retinal micro-vasculopathy [65 –67].

Furthermore, OCTA use has been expanded to the study of multiple sclerosis [68] and many other neurological disorders with promising results. However, no neurological disease-specific pattern of OCTA has been demonstrated that would give a diagnosis solely based on OCTA [6].

5 Disadvantages and Limitations of OCTA

Although OCTA shows great promise for a bright future, interpretation of OCTA imaging must be done carefully, taking into consideration the limitations of the technology and the appearance of image artifacts. Important limitations and disadvantages of the OCTA were grouped in the following subsections.

5.1 Vascular leakage and shadowing artifacts

Unlike the other invasive retinal imaging modalities of FA and ICGA, the OCTA is not able to assess slow flow, microaneurysms [45], and vascular leakage which gives valuable information related to the integrity of the vasculature. In addition, a recent meta-analysis [69] did not find a superiority of OCTA over FA in the detection of neovasculature in age-related macular degeneration. In general, the OCTA signal-to-noise ratio (S/N) can be extremely low causing the disappearance of smaller microvessels that produce a weak signal. Apart from slow flow, low S/N can be caused by defocusing mechanisms and local opacities from vitreous floaters, cell aggregates, hemorrhages, tumors, swellings, and drusen. These opacities attenuate the OCT beam below them, reducing the S/N and producing “shadowing artifacts” [43] or “false negative signals” (shadows) at deeper tissue layers. Shadowing artifacts lead to wrong estimations of vascular density and avascular areas.

5.2 Inability of arteriolar from venular side discrimination

OCTA cannot discriminate the arteriolar from the venular side of the ocular microcirculation. These sides are not only anatomically and structurally different but also very different from the hemodynamic point of view. On the arteriolar side of the microcirculation there is considerable blood flow pulsation, as has been demonstrated in animal tissues [70, 71] and in the ocular tissues of humans [72 –76].

5.3 Projection artifacts

An important class of OCTA artifacts is “projection artifacts” caused by the projection of superficial microvessels on deeper tissue layers. Superficial microvessels are closer to the OCT beam source and deeper tissue microvessels are those farther from the beam source. So, blood flow inside the superficial vessels causes fluctuating projections to appear at deeper tissue layers as “false positive signals” [51]. For example, microvessels of the retinal layer can appear as false positive signals (projections) on the choriocapillaris deeper layer. This kind of artifacts can be reduced by simple digital image processing techniques such as subtraction of the superficial image from the deeper layer image or by more complex techniques such as “projection-resolved OCTA” (PR-OCTA, [77]) which are in most cases proprietary [27]. Projection artifacts lead to wrong estimations of vascular density and avascular areas.

5.4 Segmentation artifacts

As we have seen in section 3.2, en face OCTA is based on structural OCT tissue layer segmentation and on the subsequent projection of 3D flow layer data into a 2D en face image. Segmentation errors are the source of the most important OCTA artifacts [27] and are caused by the fact that segmentation is based on the retinal architecture of healthy eyes. However, retinal architecture changes drastically in various eye pathologies [78] such as high myopia and retinal edema. In these cases, the processes of segmentation and projection can cause serious en face OCTA errors, for example, different microvessels can appear to be the same or to be anastomosed. To avoid OCTA errors it is recommended [27] to 1) confirm the correct position of segmentation contours from OCT B-scan images, 2) visualize en face OCTA together with OCT B-scans, 3) use software (if provided) with optional manual correction of segmentation contours and 4) use software (if provided) with volume rendering (orthoplane visualization) which does not depend on segmentation and presents microvascular flow data in 3 axes of rotation (3D visualization).

5.5 Restricted field of view

An additional limitation of OCTA is the restricted field of view. Currently, 3×3 mm to 12×12 mm is available. Increasing the field of view comes with decreasing scanning density. As a result, 3×3 mm provides the highest quality image. Thus, the scanning speed of the device limits the field of view and lowers resolution in images with larger fields of view [45]. The development of automatic montage software will make feasible the generation of wider OCT angiograms which are very useful in clinical practice.

5.6 Motion artifacts

OCTA is based on the hypothesis that if the eye is static the only movement is blood flow and thus any observed decorrelation comes from this blood flow. However, many sources of movement can coexist while performing the scan such as bulk movement of the head or eye, saccades, poor fixation (slow drift), and movement within the eye from intraocular pressure variations. All these movements not related to blood flow produce false signals and distortions in OCTA which are referred to as “motion artifacts”. Makita et al., [32] emphasized the importance of motion compensation in OCTA and presented two different algorithms for reducing axial motion artifacts. Motion artifacts appear as white lines on en face images, representing decorrelation over the entire B scans in those line locations [51]. Image quality can be improved by registration, motion correction software and reduced interscan time but still, motion artifacts may act as confounding factors to the motion signal and rectangular gaps may appear in en face OCTA images [27]. Therefore, motion correction software combined with active eye tracking is recommended for further reducing motion artifacts [79].

5.7 Blood flow quantification

OCTA cannot quantify blood flow in an absolute way, since it normalizes the signal between 0 and 1 (Fig. 7) and provides an image showing either the presence or absence of flow at a particular region. This situation can be improved by VISTA providing information on relative velocities in a color map [45] however, not in a fully quantitative way of absolute speed (mm/s), direction, and volume flow. In addition, with a best transverse resolution of the order of 10–20 μm (see Table 1 in section 6) it is not possible for OCTA to measure in the smallest microvessels with diameters in the range of 4 to 20 μm.

FA and ICGA are capable of investigating retinal blood flow dynamics, up to a degree, with the arteriovenous passage time and the dye-dilution method [80]. Laser Doppler velocimetry (LDV) measures maximum velocity in a single vessel at a time and may be used to measure maximum blood velocity in microvessels with calibers greater than about 40 μm [72 , 82].

Regarding the ability to quantify microvessel diameter and absolute blood flow, all the above techniques have some limitations. First, for the estimation of absolute distance (AD) in μm, on the eye fundus, a correction of transverse magnification error is needed. This correction is performed commonly with the Littmann-Bennet method using the axial length of the eye [48]. However, it is not accustomed to measuring the actual axial length and only about 8% of the OCTA studies have corrected fundus images for magnification error [48]. Second, other factors such as refractive errors due to pathology or optical interfaces have to be taken into account [83]. Third, it is obvious that the correct estimation of AD is a prerequisite for the measurement of vessel diameter and blood velocity. Fourth, from a fluid dynamics point of view, there is a difference between axial microvessel velocity (Vax) and cross-sectional velocity (Vs) which is necessary for the correct estimation of blood volume flow (Q). There is no method for the direct estimation of the Vs. However, there is a mathematical type for the conversion of the Vax to Vs [84, 85].

Conjunctival Video Capillaroscopy (CVC) gives an attractive alternative for the in vivo microvascular study in humans since it can provide fully quantitative hemodynamic information (AD, Vax, Vs) from the smallest diameter microvessels [86] without mydriasis. Conjunctival tissue is flat and there are no tissues between the microvessels of interest and the optics of the instrument. Moreover, the field of measurement is translucent approaching the quality of in vitro and intravital [89] setups. There are already numerous CVC studies providing quantitative hemodynamic information both in normal [86 , 90–93] and abnormal-pathological cases [94 –97]. CVC is a type of slit-lamp biomicroscopy (SLB) that is routinely used in ophthalmic examinations and recently, Xiao et al. [98] presented a novel instrument combining SLB with OCT. However, CVC has also standardization issues to be solved [99].

6 Current OCTA instrumentation specifications

There are several commercially available OCTA systems in the global market. The basic hardware and software specifications of six of them are shown in Tables 1 and 2, respectively. The number of A-scans per mm in the X direction (Fig. 4) and the number of B-scans per mm in the Y direction (Fig. 4) define the scan density (SCD) in the corresponding directions. Scan density and transverse resolution define the sampling density (SAD) which should be higher than 2 according to the Nyquist criterion however, this is not always satisfied [48]. Some of the special and unique hardware and software characteristics of each instrument are described below (alphabetically).

Table 2
OCTA device software specifications

SOFTWARE SPECIFICATIONS OCTA device

1 2 3 4 5 6

Algorithm FSADA FSP CIPD SSADA OCTARA OMAG

ET &MC YES YES YES YES YES YES

Panorama YES N/A YES YES YES YES

PAR YES YES YES YES N/A N/A

MSC YES YES N/A YES N/A N/A

3D Visualization YES NO N/A YES N/A YES

VISTA NO NO * N/A NO N/A N/A

Vessel Density YES NO YES YES N/A N/A

Avascular Area YES NO YES YES N/A N/A

SOFTWARE SPECIFICATIONS	OCTA device
Algorithm	FSADA	FSP	CIPD	SSADA	OCTARA	OMAG
ET &MC	YES	YES	YES	YES	YES	YES
Panorama	YES	N/A	YES	YES	YES	YES
PAR	YES	YES	YES	YES	N/A	N/A
MSC	YES	YES	N/A	YES	N/A	N/A
3D Visualization	YES	NO	N/A	YES	N/A	YES
VISTA	NO	NO *	N/A	NO	N/A	N/A
Vessel Density	YES	NO	YES	YES	N/A	N/A
Avascular Area	YES	NO	YES	YES	N/A	N/A

OCTA devices are in alphabetical order. 1: CANON XEPHILIO OCT-A1, 2: HEIDELBERG ENGINEERING SPECTRALIS-SHIFT, 3: NIDEK RS-3000 ADVANCE-2 ANGIOSCAN, 4: OPTOVUE SOLIX ANGIOVUE, 5: TOPCON DRI OCT TRITON, 6: ZEISS PLEX ELITE-9000 ANGIOPLEX. ET & MC: Eye Tracking & Motion Correction. Panorama: automatic montaging of different fundus areas to a single wide field image. PAR: Projection Artifact Reduction. MSC: Manual Segmentation Correction; this is an option for overriding automatic segmentation technology in order to reduce segmentation errors. VISTA: Variable Interscan Time Analysis. FSADA: Full Spectrum Amplitude Decorrelation Angiography. FSP: Full-Spectrum Probabilistic. CIPD: Combined Intensity and Phase Decorrelation (Hagag et al., 2017). SSADA: Split-Spectrum Amplitude-Decorrelation Angiography (Jia et al., 2012). OCTARA: Optical Coherence Tomography Angiography Ratio Analysis (Kalra et al., 2022). OMAG: Optical MicroAngioGraphy (Kalra et al., 2022). *: the SHIFT technology of Heidelberg Engineering SPECTRALIS is a kind of VISTA, since by changing the scan speed is like changing the interscan time T_I (see Section 3.2). N/A: Not Available information.

Canon XEPHILIO-OCT-A1 is accompanied by the “Denoise” function, a deep learning technology to effectively remove noise and enhance details in a single scan (Fig. 8) and, with “Flow Fusion” an SLO-enabled averaging technology for improved image quality and reduced noise. It is the only instrument in Table 1 with an optical axial resolution as low as 3 μm.

Heidelberg Engineering SPECTRALIS OCTA is using a special Full-Spectrum Probabilistic (FSB) algorithm taking advantage of the difference in the OCT signal probabilistic distribution between static and perfused tissue. From a short-time series of B-scans, it seems to be possible to calculate the probability that the signal from a specific pixel fits one of the two distributions (distribution of flow rather than the distribution of static tissue). In addition, the spectrum is not split (no resolution sacrificing) and there is no implementation of spatial and additional filtering (no artifact introduction). The SPECTRALIS Scan Planning Tool permits the acquisition of OCTA scans, based on previous OCT, FA, or ICGA images, giving the option of follow-up studies including different imaging modalities. SPECTRALIS SHIFT has the unique technological capability of changing A-scan speed among 4 different values: 20, 42.5, 85, and 125 kHz.

Nidek RS-3000 ADVANCE 2 ANGIOSCAN is accompanied by NAVIS-EX image filing software, which includes optionally “long axial length normative database” for assisting the diagnosis of eye diseases in patients with long axial lengths. In addition, there is deep learning “B-scan denoising software” to automatically display a denoised image once the B-Scan acquisition is complete.

Optovue SOLIX ANGIOVUE is accompanied by OCT Wellness software that generates a single, comprehensive report including en face images of four different layers (superficial vascular plexus, deep capillary plexus, outer retina layer, and choriocapillaris) and metrics on retinal and ganglion cell layer (GCL) thickness. In addition, it gives automatically metrics about these layers such as superficial and deep vessel density (%) and FAZ area (mm²), and FAZ perimeter (mm).

Topcon DRI OCT TRITON has Swept-Source (SS) technology with a long wavelength laser of 1050 nm and a tuning range of 100 nm [100]. The long wavelength makes OCTA imaging possible for patients with media opacities. An area of interest can be examined with en face OCTA, B-scans, and fundus photography on a single screen using IMAGEnet software. Dynamic Focus and Enhanced Vitreous Visualization (EVV) software help the operators to enhance the contrast in the vitreous and in the vitreoretinal interface. PixelSmart is an image processing algorithm for reducing speckle noise thus improving image quality.

Zeiss PLEX ELITE-9000 ANGIOPLEX has SS technology with a long wavelength laser (1040–1070 nm) exhibiting better opacity penetration (as in the case of Topcon TRITON). There is an option of an ultrawide 15×9 mm view captured in a single OCTA scan and an ultrawide OCTA en face montage corresponding to a 70° field of view. PLEX ELITE-9000 is the first commercial dual-speed SS-OCT instrument with dual A-scan speed values of 100 and 200 kHz.

All the above information about OCTA instrumentation and software was acquired from the English manuals (brochures) and official webpages, and from the companies that answered our request, in the period of September-October 2022.

7 Conclusion

OCTA is an exciting new, fast and noninvasive imaging technique for the examination of the ocular fundus in ophthalmology and the same instrument can be used for the anterior segment of the eye. However, excessive image processing is employed and clinicians must be familiar with the basic principles and the associated artifacts to make correct decisions and perform diagnoses.

Footnotes

Acknowledgments

The authors would like to thank Arnoud Loot from Canon Medical Systems Europe, Emilia Kostadinova and Anna Kleinert from Heidelberg Engineering Ltd, Nick Morogiannis from Ommalite (on behalf of Heidelberg Engineering) and Patricia Peley from Visionix (on behalf of Optovue) for providing useful data, images and specifications about the OCTA instruments.

Declaration of competing interest

The authors have declared no conflicts of interest.

Funding

No specific funding was received from any funding bodies in the public, commercial or not for-profit sectors to carry out the work described in this manuscript

References

Webb

, Hughes

. Scanning laser ophthalmoscope. IEEE Trans Biomed Eng. 1981;28(7):488–92. doi: 10.1109/TBME.1981.324734.

Webb

. Scanning laser ophthalmoscope. In: Masters

Barry R.

, editor. Noninvasive Diagnostic Techniques in Ophthalmology. Springer-Verlag; 1990. pp. 438–450.

Webb

. Manipulating laser light for ophthalmology. IEEE Eng Med Biol Mag. 1985;4(4):12–6. doi: 10.1109/MEMB.1985.5006222.

Huang

, Swanson

, Lin

, Schuman

, Stinson

, Chang

, Hee

, Flotte

, Gregory

, Puliafito

, et al. Optical coherence tomography. Science. 1991;254(5035):1178–81. doi: 10.1126/science.1957169.

Drexler

, Fujimoto

. Optical Coherence Tomography. 2^nd Edition, Springer International; 2015.

Augustin

, Atorf

. The Value of Optical Coherence Tomography Angiography (OCT-A) in Neurological Diseases. Diagnostics (Basel). 2022;12(2):468. doi: 10.3390/diagnostics12020468.

Fujimoto

, Swanson

. The Development, Commercialization, and Impact of Optical Coherence Tomography. Invest Ophthalmol Vis Sci. 2016;57(9):1–13. doi: 10.1167/iovs.16-19963.

Fercher

, Hitzenberger

, Kamp

, Elzaiat

. Measurement of intraocular distances by backscattering spectral interferometry. Opt Comm. 1995;117:43–8. doi:10.1016/0030-4018(95)00119-S.

Bouma

, Yun

, Vakoc

, Suter

, Tearney

. Fourier-domain optical coherence tomography: recent advances toward clinical utility. Curr Opin Biotechnol. 2009;20(1):111–8. doi: 10.1016/j.copbio.2009.02.007.

10.

Ha Usler

, Lindner

. “Coherence radar” and “spectral radar"-new tools for dermatological diagnosis. J Biomed Opt. 1998;3(1):21–31. doi: 10.1117/1.429899.

11.

Wojtkowski

, Srinivasan

, Ko

, Fujimoto

, Kowalczyk

, Duker

. Ultrahigh-resolution, high-speed, Fourier domain optical coherence tomography and methods for dispersion compensation. Opt Express. 2004;12(11):2404–22. doi: 10.1364/opex.12.002404.

12.

Potsaid

, Gorczynska

, Srinivasan

, Chen

, Jiang

, Cable

, Fujimoto

. Ultrahigh speed spectral/Fourier domain OCT ophthalmic imaging at 70,000 to 312,500 axial scans per second. Opt Express. 2008;16(19):15149–69. doi: 10.1364/oe.16.015149.

13.

Zhang

, Zhang

, Huang

, Zhong

, Wang

. Multifunctional nm Spectral Domain OCT System at 147kHz for Posterior Eye Imaging. Sovrem Tekhnologii Med. 2015;7(1):7–12. doi: 10.17691/stm2015.7.1.01.

14.

Spaide

, Koizumi

, Pozzoni

. Enhanced depth imaging spectral-domain optical coherence tomography. Am J Ophthalmol. 2008;146(4):496–500. doi: 10.1016/j.ajo.2008.05.032.

15.

Koutsiaris

. Deep tissue near infrared second derivative spectrophotometry for the assessment of claudication in peripheral arterial disease. Clin Hemorheol Microcirc. 2017;65(3):275–84. doi: 10.3233/CH-16181.

16.

Chinn

, Swanson

, Fujimoto

. Optical coherence tomography using a frequency-tunable optical source. Opt Lett. 1997;22(5):340–2. doi: 10.1364/ol.22.000340.

17.

Choma

, Sarunic

, Yang

, Izatt

. Sensitivity advantage of swept source and Fourier domain optical coherence tomography. Opt Express. 2003;11(18):2183–9. doi: 10.1364/oe.11.002183.

18.

Yasuno

, Youngjoo

, Shuichi

, Masahiro

, Toyohiko

. In vivo high-contrast imaging of deep posterior eye by 1- μm swept source optical coherence tomography and scattering optical coherence angiography. Opt. Express 15. 2007;15(10):6121–39. doi: 10.1364/oe.15.006121.

19.

Potsaid

, Baumann

, Huang

, Barry

, Cable

, Schuman

, Duker

, Fujimoto

. Ultrahigh speed 1050nm swept source/Fourier domain OCT retinal and anterior segment imaging at 100,000 to 400,000 axial scans per second. Optics express. 2002;18(19):9–48. doi: 10.1364/OE.18.020029.

20.

Kishi

. Impact of swept source optical coherence tomography on ophthalmology. Taiwan Journal of Ophthalmology. 2016;6(2):58–68. doi: 10.1016/j.tjo.2015.09.002.

21.

Yun

, Tearney

, de Boer

, Iftimia

, Bouma

. High-speed optical frequency-domain imaging. Opt Express. 2003;11(22):2953–63. doi: 10.1364/oe.11.002953.

22.

Grulkowski

, Liu

, Potsaid

, Jayaraman

, Lu

, Jiang

, Cable

, Duker

, Fujimoto

. Retinal, anterior segment and full eye imaging using ultrahigh speed swept source OCT with vertical-cavity surface emitting lasers. Biomed Opt Express. 2012;3(11):2733–51. doi: 10.1364/BOE.3.002733.

23.

Chen

, Liu

, Minneman

, Ginner

, Hoover

, Sattmann

, Bonesi

, Drexler

, Leitgeb

. Phase-stable swept source OCT angiography in human skin using an akinetic source. Biomed Opt Express. 2016;7(8):3032–48. doi: 10.1364/BOE.7.003032.

24.

Wieser

, Biedermann

, Klein

, Eigenwillig

, Huber

. Multi-Megahertz OCT: High quality 3D imaging at 20 million A-scans and 45 GVoxels per second. Optics Express. 2010;18(14):14685. doi: 10.1364/oe.18.014685.

25.

Klein

, Wieser

, Reznicek

, Neubauer

, Kampik

, Huber

. Multi-MHz retinal OCT. Biomedical Optics Express. 2013;4(10):1890. doi: 10.1364/boe.4.001890.

26.

Lee

, Song

, Jung

, Kim

. Wide tuning range wavelength-swept laser with a single SOA at 1020nm for ultrahigh resolution Fourier-domain optical coherence tomography. Opt Express. 2011;19(22):21227–37. doi: 10.1364/OE.19.021227.

27.

Spaide

, Fujimoto

, Waheed

, Sadda

, Staurenghi

. Optical coherence tomography angiography. Prog Retin Eye Res. 2018;64:1–55. doi:10.1016/j.preteyeres.2017.11.003.

28.

Rosen

, Garcia

, Podoleanu

, Cucu

, Dobre

, Trifanov

, van Velthoven

MEJ

, de Smet

, Rogers

, Hathaway

, Pedro

, Weitz

. En-face Flying Spot OCT/Ophthalmoscope. In: Drexler

, Fujimoto

, editors. Optical coherence tomography. 2nd ed. Springer, Cham; 2015. pp. 1797–1832.

29.

Beaurepaire

, Boccara

, Lebec

, Blanchot

, Saint-Jalmes

. Full-field optical coherence microscopy. Opt Lett. 1998;23(4):244–6. doi: 10.1364/ol.23.000244.

30.

Subhash

. Full-Field and Single-Shot Full-Field Optical Coherence Tomography: A Novel Technique for Biomedical Imaging Applications. Advances in Optical Technologies. 2012:435408. doi: 10.1155/2012/435408.

31.

Sinha

, Phadikar

, Saxena

. Optical coherence tomography: a primer. In: Meyer

, Saxena

, Sadda

, editors. Spectral domain optical coherence tomography in macular diseases. Springer, New Delhi; 2017. pp. 3–13.

32.

Makita

, Hong

, Yamanari

, Yatagai

, Yasuno

. Optical coherence angiography. Opt Express. 2006;14(17):7821–40. doi: 10.1364/oe.14.007821.

33.

, Wang

. In vivo volumetric imaging of vascular perfusion within human retina and choroids with optical micro-angiography. Opt Express. 2008;16(15):11438–52. doi: 10.1364/oe.16.011438.

34.

Fingler

, Zawadzki

, Werner

, Schwartz

, Fraser

. Volumetric microvascular imaging of human retina using optical coherence tomography with a novel motion contrast technique. Opt Express. 2009;17(24):22190–200. doi: 10.1364/OE.17.022190.

35.

, Chen

. Doppler variance imaging for three-dimensional retina and choroid angiography. J Biomed Opt. 2010;15(1):016029. doi: 10.1117/1.3302806.

36.

Kim

, Fingler

, Werner

, Schwartz

, Fraser

, Zawadzki

. In vivo volumetric imaging of human retinal circulation with phase-variance optical coherence tomography. Biomed Opt Express. 2011;2(6):1504–13. doi: 10.1364/BOE.2.001504.

37.

Zhao

, Chen

, Saxer

, Xiang

, de Boer

, Nelson

. Phase-resolved optical coherence tomography and optical Doppler tomography for imaging blood flow in human skin with fast scanning speed and high velocity sensitivity. Opt Lett. 2000;25(2):114–6. doi: 10.1364/ol.25.000114.

38.

Park

, Pierce

, Cense

, Yun

, Mujat

, Tearney

, Bouma

, de Boer

. Real-time fiber-based multi-functional spectral-domain optical coherence tomography at 1. 3 microm. Opt Express. 2005;13(11):3931–44. doi: 10.1364/opex.13.003931.

39.

Zhao

, Chen

, Saxer

, Shen

, Xiang

, de Boer

, Nelson

. Doppler standard deviation imaging for clinical monitoring of in vivo human skin blood flow. Opt Lett. 2000;25(18):1358–60. doi: 10.1364/ol.25.001358.

40.

Jia

, Tan

, Tokayer

, Potsaid

, Wang

, Liu

, Kraus

, Subhash

, Fujimoto

, Hornegger

, Huang

. Split-spectrum amplitude-decorrelation angiography with optical coherence tomography. Opt Express. 2012;20(4):4710–25. doi: 10.1364/OE.20.004710.

41.

Mariampillai

, Standish

, Moriyama

, Khurana

, Munce

, Leung

, Jiang

, Cable

, Wilson

, Vitkin

, Yang

. Speckle variance detection of microvasculature using swept-source optical coherence tomography. Opt Lett. 2008;33(13):1530–2. doi: 10.1364/ol.33.001530.

42.

Jonathan

, Enfield

, Leahy

. Correlation mapping method for generating microcirculation morphology from optical coherence tomography (OCT) intensity images. J Biophotonics. 2011;4(9):583–7. doi: 10.1002/jbio.201000103.

43.

Borrelli

, Sadda

, Uji

, Querques

. Pearls and Pitfalls of Optical Coherence Tomography Angiography Imaging: A Review. Ophthalmol Ther. 2019;8(2):215–26. doi: 10.1007/s40123-019-0178-6.

44.

Ploner

, Moult

, Choi

, Waheed

, Lee

, Novais

, Cole

, Potsaid

, Husvogt

, Schottenhamml

, Maier

, Rosenfeld

, Duker

, Hornegger

, Fujimoto

. Toward quantitative optical coherence tomography angiography: Visualizing Blood Flow Speeds in Ocular Pathology Using Variable Interscan Time Analysis. Retina. 2016;36(Suppl 1):S118–S126. doi: 10.1097/IAE.0000000000001328.

45.

Arya

, Rashad

, Sorour

, Moult

, Fujimoto

, Waheed

. Optical coherence tomography angiography (OCTA) flow speed mapping technology for retinal diseases. Expert Rev Med Devices. 2018;15(12):875–82. doi: 10.1080/17434440.2018.1548932.

46.

Jiao

, Knighton

, Huang

, Gregori

, Puliafito

. Simultaneous acquisition of sectional and fundus ophthalmic images with spectral-domain optical coherence tomography. Opt Express. 2005;13(2):444–52. doi: 10.1364/opex.13.000444.

47.

de Carlo

, Romano

, Waheed

, Duker

. A review of optical coherence tomography angiography (OCTA). Int J Retina Vitreous. 2015;1:5. doi:10.1186/s40942-015-0005-8.

48.

Sampson

, Dubis

, Chen

, Zawadzki

, Sampson

. Towards standardizing retinal optical coherence tomography angiography: a review. Light Sci Appl. 2022;11(1):63. doi: 10.1038/s41377-022-00740-9.

49.

Kalra

, Zarranz-Ventura

, Chahal

, Bernal-Morales

, Lupidi

, Chhablani

. Optical coherence tomography (OCT) angiolytics: a review of OCT angiography quantitative biomarkers. Surv Ophthalmol. 2022;67(4):1118–34. doi: 10.1016/j.survophthal.2021.11.002.

50.

Vujosevic

, Cunha-Vaz

, Figueira

, Löwenstein

, Midena

, Parravano

, Scanlon

, Simó

, Hernández

, Madeira

, Marques

, C-V Martinho

, Santos

, Simó-Servat

, Salongcay

, Zur

, Peto

. Standardization of Optical Coherence Tomography Angiography Imaging Biomarkers in Diabetic Retinal Disease. Ophthalmic Res. 2021;64(6):871–87. doi: 10.1159/000518620.

51.

Hagag

, Gao

, Jia

, Huang

. Optical coherence tomography angiography: Technical principles and clinical applications in ophthalmology. Taiwan J Ophthalmol. 2017;7(3):115–29. doi: 10.4103/tjo.tjo_31_17.

52.

Laíns

, Wang

, Cui

, Katz

, Vingopoulos

, Staurenghi

, Vavvas

, Miller

. Retinal applications of swept source optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA). Prog Retin Eye Res. 2021;84:100951. doi:10.1016/j.preteyeres.2021.100951.

53.

Wang

, Gao

, Zhang

. Sensitivity and specificity of optical coherence tomography angiography in the diagnosis of active choroidal neovascularization: a systematic review and meta-analysis. Graefes Arch Clin Exp Ophthalmol. 2021;259(12):3529–36. doi: 10.1007/s00417-021-05239-4.

54.

Russell

, Han

. Toward a New Staging System for Diabetic Retinopathy Using Wide Field Swept-Source Optical Coherence Tomography Angiography. Curr Diab Rep. 2021;21(9):28. doi: 10.1007/s11892-021-01401-8.

55.

Agarwal

, Rübsam

, Zur Bonsen

, Pichi

, Neri

, Pleyer

. A Comprehensive Update on Retinal Vasculitis: Etiologies, Manifestations and Treatments. J Clin Med. 2022;11(9):2525. doi: 10.3390/jcm11092525.

56.

Cano

, Farzad

, Khansari

, Tan

, Huang

, Lim

, Shahidi

. Relating retinal blood flow and vessel morphology in sickle cell retinopathy. Eye (Lond). 2020;34(5):886–91. doi: 10.1038/s41433-019-0604-y.

57.

Fan

, Ying

, Zhai

, Sheng

, Sun

, Xu

, Kong

. The characteristics of fundus microvascular alterations in the course of glaucoma: a narrative review. Ann Transl Med. 2022;10(9):527. doi: 10.21037/atm-21-5695.

58.

Shin

, Wolf

, Harris

, Verticchio Vercellin

, Siesky

, Rowe

, Packles

, Oddone

. Vascular biomarkers from optical coherence tomography angiography and glaucoma: where do we stand in 2021? Acta Ophthalmol. 2022;100(2):e377–e385. doi: 10.1111/aos.14982.

59.

Chatziralli

, Milionis

, Christodoulou

, Theodossiadis

, Kitsos

. The Role of Vessel Density as Measured by Optical Coherence Tomography Angiography in the Evaluation of Pseudoexfoliative Glaucoma: A Review of the Literature. Ophthalmol Ther. 2022;11(2):533–45. doi: 10.1007/s40123-022-00483-1.

60.

Kate

, Basu

. Role of Anterior Segment-Optical Coherence Tomography Angiography in Acute Ocular Burns. Diagnostics (Basel). 2022;12(3):607. doi: 10.3390/diagnostics12030607.

61.

Obuchowska

, Konopińska

. Importance of Optical Coherence Tomography and Optical Coherence Tomography Angiography in the Imaging and Differentiation of Choroidal Melanoma: A Review. Cancers (Basel). 2022;14(14):3354. doi: 10.3390/cancers14143354.

62.

Pellegrini

, Staurenghi

, Preziosa

. Clinical Applications of Optical Coherence Tomography Angiography in Ocular Oncology: Pearls and Pitfalls. Ocul Oncol Pathol. 2022;8(2):79–87. doi: 10.1159/000520951.

63.

Anjos

, Ferreira

, Barkoudah

, Claggett

, Abegão Pinto

, Miguel

. Application of Optical Coherence Tomography Angiography Macular Analysis for Systemic Hypertension. A Systematic Review and Meta-analysis. Am J Hypertens. 2022;35(4):356–64. doi: 10.1093/ajh/hpab172.

64.

Monteiro-Henriques

, Rocha-Sousa

, Barbosa-Breda

. Optical coherence tomography angiography changes in cardiovascular systemic diseases and risk factors: A Review. Acta Ophthalmol. 2022;100(1):e1–e15. doi: 10.1111/aos.14851.

65.

McGrath

, Aslam

. Use of Imaging Technology to Assess the Effect of COVID-19 on Retinal Tissues: A Systematic Review. Ophthalmol Ther. 2022;11(3):1017–30. doi: 10.1007/s40123-022-00509-8.

66.

Wang

, Wang

, Hu

, Wang

. Retinal microvascular impairment in COVID-19 patients: A meta-analysis. Immun Inflamm Dis. 2022;10(6):e619. doi: 10.1002/iid3.619.

67.

Teo

, Invernizzi

, Staurenghi

, Cheung

CMG

. COVID-19-Related Retinal Micro-vasculopathy - A Review of Current Evidence. Am J Ophthalmol. 2022;235:98–110. doi:10.1016/j.ajo.2021.09.019.

68.

Chalkias

, Bakirtzis

, Pirounides

, Boziki

, Grigoriadis

. Optical Coherence Tomography and Optical Coherence Tomography with Angiography in Multiple Sclerosis. Healthcare (Basel). 2022;10(8):1386. doi: 10.3390/healthcare10081386.

69.

Maesa

, Baños-Álvarez

, Rosario-Lozano

, Blasco-Amaro

. Diagnostic accuracy of optical coherence tomography angiography in the detection of neovasculature in age-related macular degeneration: a meta-analysis. Acta Ophthalmol. 2022;100(2):e368–e376. doi: 10.1111/aos.14979.

70.

Tangelder

, Slaaf

, Muijtjens

, Arts

, oude Egbrink

, Reneman

. Velocity profiles of blood platelets and red blood cells flowing in arterioles of the rabbit mesentery. Circ Res. 1986;59(5):505–14. doi: 10.1161/01.res.59.5.505.

71.

Koutsiaris

, Pogiatzi

. Velocity pulse measurements in the mesenteric arterioles of rabbits. Physiol Meas. 2004;25(1):15–25. doi: 10.1088/0967-3334/25/1/002.

72.

Nagaoka

, Yoshida

. Noninvasive evaluation of wall shear stress on retinal microcirculation in humans. Invest Ophthalmol Vis Sci. 2006;47(3):1113–9. doi: 10.1167/iovs.05-0218.

73.

Koutsiaris

, Tachmitzi

, Papavasileiou

, Batis

, Kotoula

, Giannoukas

, Tsironi

. Blood velocity pulse quantification in the human conjunctival pre-capillary arterioles. Microvasc Res. 2010;80(2):202–8. doi: 10.1016/j.mvr.2010.05.001.

74.

Koutsiaris

. The resistive index as a function of vessel diameter in the human carotid arterial tree. Microvasc Res. 2013;89:169–71. doi:10.1016/j.mvr.2013.06.005.

75.

Koutsiaris

. Pulsatility Index quantification in the human precapillary arterioles of the eye. Microvasc Res. 2016;106:36–8. doi:10.1016/j.mvr.2016.03.008.

76.

Koutsiaris

. Hemodynamics in the Microcirculation. Ann Biomed Eng. 2016;44(4):1321–2. doi: 10.1007/s10439-016-1568-8.

77.

Zhang

, Hwang

, Campbell

, Bailey

, Wilson

, Huang

, Jia

. Projection-resolved optical coherence tomographic angiography. Biomed Opt Express. 2016,9;7(3):816–28. doi: 10.1364/BOE.7.000816.

78.

Borrelli

, Sadda

, Uji

, Querques

. OCT angiography: guidelines for analysis and interpretation. In: Grzybowski

, Barboni

, editors. OCT and imaging in central nervous system diseases. Springer; 2020. pp. 41–54.

79.

Camino

, Zhang

, Gao

, Hwang

, Sharma

, Wilson

, Huang

, Jia

. Evaluation of artifact reduction in optical coherence tomography angiography with real-time tracking and motion correction technology. Biomed Opt Express. 2016;7(10):3905–15. doi: 10.1364/BOE.7.003905.

80.

Wolf

, Jung

, Kiesewetter

, Körber

, Reim

. Video fluorescein angiography: method and clinical application. Graefes Arch Clin Exp Ophthalmol. 1989;227(2):145–51. doi: 10.1007/BF02169788.

81.

Riva

, Feke

, Eberli

, Benary

. Bidirectional LDV system for absolute measurement of blood speed in retinal vessels. Appl Opt. 1979;18(13):2301–6. doi: 10.1364/AO.18.002301.

82.

Riva

, Petrig

. Retinal blood flow: Laser Doppler Velocimetry and blue field simulation technique. In: Masters

Barry R.

, editor. Noninvasive Diagnostic Techniques in Ophthalmology. Springer-Verlag; 1990. pp. 390–409.

83.

Lal

, Alonso-Caneiro

, Read

, Carkeet

. Induced Refractive Error Changes the Optical Coherence Tomography Angiography Transverse Magnification and Vascular Indices. Am J Ophthalmol. 2021;229:230–41. doi:10.1016/j.ajo.2021.04.012.

84.

Koutsiaris

. Volume flow estimation in the precapillary mesenteric microvasculature in vivo and the principle of constant pressure gradient. Biorheology. 2005;42(6):479–91.

85.

Sun

, Li

, Liu

, Ma

, Zhou

, Duan

, Bian

, Li

, Qi

. Progress of Bulbar Conjunctival Microcirculation Alterations in the Diagnosis of Ocular Diseases. Dis Markers. 2022:4046809. doi: 10.1155/2022/4046809.

86.

Koutsiaris

. Wall shear stress in the human eye microcirculation in vivo, segmental heterogeneity and performance of in vitro cerebrovascular models. Clin Hemorheol Microcirc. 2015;63(1):15–33. doi: 10.3233/CH-151976.

87.

Koutsiaris

. Digital micro PIV ( μPIV) and velocity profiles in vitro and in vivo. In: The particle image velocimetry-characteristics, limits and possible applications. Cavazzini

, editor. InTech, Rijeka Croatia; 2012. pp. 283–318.

88.

Koutsiaris

. Loss of field of view due to optical mismatch at the inner diametric plane of cylindrical artificial microvessels. Results in Optics. 2022;8:100241. doi:10.1016/j.rio.2022.100241.

89.

Janssen

BGH

, Najiminaini

, Zhang

, Omidi

, Carson

JJL

. Multispectral intravital microscopy for simultaneous bright-field and fluorescence imaging of the microvasculature. Appl Microsc. 2021;51(1):12. doi: 10.1186/s42649-021-00059-6.

90.

Jung

, Körber

, Kiesewetter

, Prünte

, Wolf

, Reim

. Measuring the microcirculation in the human conjunctiva bulbi under normal and hyperperfusion conditions. Graefes Arch Clin Exp Ophthalmol. 1983;220(6):294–7. doi: 10.1007/BF00231359.

91.

Shahidi

, Wanek

, Gaynes

, Wu

. Quantitative assessment of conjunctival microvascular circulation of the human eye. Microvasc. Res. 2010;79(2):109–13. doi: 10.1016/j.mvr.2009.12.003.

92.

Brennan

, McNeil

, Jing

, Awuah

, Finlay

, Blighe

, McLaughlin

JAD

, Wang

, Moore

, Nesbit

, Trucco

, Spence

, Moore

TCB

. Quantitative assessment of the conjunctival microcirculation using a smartphone and slit-lamp biomicroscope. Microvasc. Res. 2019;126:103907. doi:10.1016/j.mvr.2019.103907.

93.

Moka

, Koutsiaris

, Garas

, Messinis

, Tachmitzi

, Giannoukas

, Tsironi

. Blood flow velocity comparison in the eye capillaries and postcapillary venules between normal pregnant and non-pregnant women. Microvasc Res. 2020;127:103926. doi:10.1016/j.mvr.2019.103926.

94.

Kord Valeshabad

, Wanek

, Saraf

, Gaynes

, Gordeuk

, Molokie

, Shahidi

. Changes in Conjunctival Hemodynamics Predict Albuminuria in Sickle Cell Nephropathy. Am J Nephrol. 2015;41(6):487–93. doi: 10.1159/000438678.

95.

Khansari

, Wanek

, Tan

, et al. Assessment of Conjunctival Microvascular Hemodynamics in Stages of Diabetic Microvasculopathy. Sci. Rep. 2017;7:45916. doi.org/10.1038/srep45916.

96.

Brennan

, McNeil

, Jing

, Awuah

, Moore

, Mailey

, Finlay

, Blighe

, McLaughlin

JAD

, Nesbit

, Trucco

, Moore

TCB

, Spence

. Assessment of the conjunctival microcirculation for patients presenting with acute myocardial infarction compared to healthy controls. Sci Rep. 2021;11(1):7660. doi: 10.1038/s41598-021-87315-7.

97.

Koutsiaris

, Riri

, Boutlas

, Panagiotou

, Kotoula

, Daniil

, Tsironi

. COVID-19 hemodynamic and thrombotic effect on the eye microcirculation after hospitalization: A quantitative case-control study. Clin Hemorheol Microcirc. In press. 2022. doi: 10.3233/CH-221554.

98.

Xiao

, Duan

, Wang

, Deng

, Wang

, Zhang

, Liang

, Yuan

. Multi-modal Anterior Eye Imager Combining Ultra-High Resolution OCT and Microvascular Imaging for Structural and Functional Evaluation of the Human Eye. Applied Sciences. 2020;10(7):2545. doi: 10.3390/app10072545.

99.

Koutsiaris

. Meta-analysis of conjunctival microvascular hemorheology metrics. Microvasc Res. 2022;142:104369. doi:10.1016/j.mvr.2022.104369.

100.

Ang

, Cai

, Tan

. Swept Source Optical Coherence Tomography Angiography for Contact Lens-Related Corneal Vascularization. J Ophthalmol. 2016;9685297. doi: 10.1155/2016/9685297.

OCTA device	HARDWARE SPECIFICATIONS
	FDT	Speed	λ (nm)	AR	TR	MFOV (mm×mm)
1	SD	70,000	855	1.6/3	–/20	10×10
2	SD	125,000 *	880	3.9/7	5/14	9×9
3	SD	85,000	880	–/7	–/20	9×9
4	SD	120,000	940	–/5	–/15	12×12
5	SS	100,000	1050	2.6/8	–/20	12×12
6	SS	200,000 **	1040–60	1.9/6.3	–/20	12×12