We introduce a new type of scanning protocols, called segmented protocols, which enable extracting multi-range flow velocity information from a single Spectral OCT data set. The protocols are evaluated using a well defined flow in a glass capillary. As an example of in vivo studies, we demonstrate two- and three-dimensional imaging of the retinal vascular system in the eyes of healthy volunteers. The flow velocity detection is performed using a method of Joint Spectral and Time domain OCT. Velocity ranging is demonstrated in imaging of retinal vasculature in the macular region and in the optic disk area characterized by different flow velocity values. Additionally, an enhanced visualization of retinal capillary network is presented in the close proximity to macula.
© 2009 OSA
Optical coherence tomography (OCT) is a noninvasive optical imaging modality which uses low coherence interferometry to obtain two-dimensional (2-D) cross-sectional images with micrometer resolution. Three-dimensional (3-D) virtual representation of measured semi-transparent object can be reconstructed from a set of 2-D OCT images.
Since the introduction of OCT technique in early 1990’s [1, 2] there has been a dramatic progress in imaging resolution and speed as well as in development of image processing methods. Today the performance of OCT systems allows for eye imaging with resolutions as high as ~2 μm [3, 4] and speeds exceeding 300,000 axial lines/s . Image processing methods have been developed for semi-automatic and automatic segmentation of morphological structures [6, 7]. Time domain and Spectral/Fourier domain OCT instruments have been commercialized and are widely used in ophthalmology clinics.
Although OCT is mostly used for imaging of tissue morphology it is also capable of visualizing functional processes in biological tissues. In particular, there has been a considerable interest in ocular blood flow assessment. Changes of retinal blood circulation are considered to be a condition associated with a number of eye diseases such as age related macular degeneration, glaucoma and diabetic retinopathy. Therefore, blood flow detection and measurement techniques may play an important role in ophthalmic diagnostics.
High-speed OCT instruments allow for in vivo blood flow imaging, monitoring of blood pulsation [8, 9] and 3-D visualization of retinal vasculature [10–14]. There are various flow detection techniques referred to as “Doppler OCT methods”. Most of them rely on determination of the phase difference of the OCT signals between consecutive axial scans (A-scans) [8, 15–18]. Another approach, called resonant Doppler OCT imaging, uses a moving reference mirror to introduce reference velocity for phase matching of OCT signals. This provides flow-based contrast changes in OCT images [19, 20]. There are also reports which present speckle [21–23] or light attenuation  analysis in flow detection. Recent methods such as microangiography  use spatial/temporal filtering to determine volumetrically positions of moving scattering particles within tissue based on the initial idea of B-M mode . An extension of this technique employs multiple spatial/temporal filtering for quantitative flow velocity extraction [14, 26].
Another variant of aforementioned Doppler OCT technique is a method proposed by our group, called Joint Spectral and Time domain OCT (STdOCT) [27, 28]. This method performs the Fourier analysis of the spectral OCT signal in wavenumber and time domain revealing simultaneously structural and flow information. The results are quantitative 3-D blood flow velocity maps of the retinal vasculature . The STdOCT method can be used for efficient suppression of the complex conjugate images in both structural and flow maps, as well .
Most of the Doppler OCT methods measure only the velocity vector component parallel to the light propagation direction. This causes anatomical restrictions to the detectable velocity bandwidth in retinal imaging: arteries and veins in the optic nerve head are oriented with larger angles with regard to the scanning beam. Together with the fact that such vessels exhibit faster flow they give rise to large Doppler frequency shifts. Vessels in the macular region – on the other hand – are nearly perpendicular to the direction of light propagation and exhibit slower flow which altogether results in very small Doppler frequency shifts. As a consequence, the axial blood flow velocity component ranges from a fraction to several tens of millimeters per second. Measurement of such a wide range of velocities is challenging for Doppler OCT techniques.32]. If we assume constant ΔΦerr, we immediately observe the dilemma of the above mentioned bandwidth issue: increasing the line rate 1/Δt causes the minimal velocity to exceed the slow flow velocity values. On the other hand, decreasing the line rate will result in slower acquisition rates, which is disadvantageous for in vivo imaging. Additionally, the decrease of the OCT line rate will introduce stronger effects of motion artifacts such as fringe washout or image blurring which will reduce the capability of the system to reveal reliable Doppler signatures and vascular structure.
The challenge that we address in the present work is to achieve a flexible and wide velocity bandwidth in Doppler optical coherence tomography without sacrificing acquisition speed and image quality.
This goal can be achieved by using specialized lateral scan protocols for ultrahigh speed OCT imaging. This possibility has not been fully exploited in Doppler OCT yet. Most systems employ a standard raster pattern where Doppler analysis is performed along the fast scanning direction (A-scan based). However, there are also reports on double-circular scanning patterns that have been introduced to determine vessels orientation for simplified blood flow assessment [9, 33]. Recently, Vakoc et al. has used multiplicated rapid scanning for extraction the information about the variance of signal phase, which made it possible to visualize blood vascular networks surrounding tumor tissue .
The central idea of our approach is to perform the Doppler analysis at the basis of consecutive B-scans or so called segments rather than subsequent A-scans. In this case the effective period Δt will be adjusted by altering the width (number of A-scans) of the B-scans or segments used for the analysis, while the total time needed for collecting 3-D structural data will remain unchanged.
Based on the new protocols we demonstrate how to achieve a high dynamic range of flow measurements with preserved high imaging speed. The performance is evaluated with well defined flow of Intralipid in a glass capillary. The applicability of the new scanning protocols for in vivo examination of the eye is presented in 2-D and 3-D imaging of retinal vasculature at different locations of the fundus in healthy volunteers. Finally, we illustrate the capabilities of the method for visualization of the retinal capillary network.
2. Materials and methods
2.1 Experimental setup and materials
The Spectral OCT (SOCT) system used for the study was based on a standard fiber-optics Michelson interferometer configuration (Fig. 1 ). The light emitted by a superluminescent diode (SLD; λ0 = 820 nm, Δλ = 71 nm; Superlum, Ireland) passes through an optical isolator (AC Photonics, Inc., USA) and is split via 70:30 fiber coupler (AC Photonics, Inc., USA) into reference and object arms. The object arm is equipped with an optical system for retinal imaging: optical scanners (Cambridge Technology, Inc., USA), scan lens and an ocular lens. In experiments with a glass capillary we used an additional lens placed in the front of the ocular lens to focus light on the object (inset in Fig. 1). The OCT signal is detected by a custom designed spectrometer containing: a collimating lens, a volume holographic diffraction grating (1200 lines/mm; Wasatch Photonics, USA), a telecentric lens, and a line-scan CMOS camera (Sprint, Basler AG, Germany).
The axial resolution of the presented instrument is 4 μm in tissue or Intralipid. The diameter of the beam incident at the cornea is 3 mm. In an aberration free optical system of the eye with the focal length of ~22.3 mm (Gullstrand’s model of the eye ) the transverse resolution would be ~8 μm. However, we assume the axial resolution of ~15 μm due to possible aberrations in the eye. In case of glass capillary imaging the transverse resolution is 30 μm. The experimentally determined axial imaging range is 1.5 mm in tissue. The sensitivity of the system is 89 dB measured with 800 μW power of light incident at the object and an exposure time of 3.4 μs. The measured signal roll-off is 20.5 dB over the entire imaging depth (1.5 mm). The imaging speed depends on the camera settings: the number of active pixels and exposure time . We used 1024 from 4096 camera pixels. The shortest possible exposure time for this number of active pixels is 3.4 μs resulting in 4.7 μs line period (or repetition time t rep). Thus, the maximal achievable imaging speed is ~213 000 A-scans/s (image lines/s). The standard deviation of the phase noise measured for 1000 spectral fringe signals at exposure time of 3.4 μs is 0.083 rad for the regular interferometer set-up. To compare this value with fundamentally limited phase noise we have also measured standard deviation of phase for the common path configuration with glass cover slip to be 0.0034 rad for the same exposure time.
Experiments were performed in a test object and in human eyes in vivo. As a test object we used a glass capillary with a solution of Intralipid (0.5% v/v) in water. The flow was initialized and maintained by a high-pressure liquid chromatography (HPLC) syringe pump (Selko Industries, Ltd., Poland). The power of light incident on the object in phantom experiment is set to 1.5 mW.
In vivo OCT imaging was performed in human retinas of two healthy young volunteers. The power of light incident at the eye was 800 µW which is below the safety limits for the wavelength range of the light source. The study was conducted in accordance with the tenets of the Declaration of Helsinki.
2.2. Joint Spectral and Time domain OCT
The flow velocity was retrieved using Joint Spectral and Time-domain OCT (STdOCT) method . In the STdOCT technique, the OCT signal is acquired while the object is scanned laterally with sufficient oversampling for Doppler signal analysis (oversampling is defined as a ratio between the beam diameter and the scanning step size). This way the spectral interference fringes are acquired over time. Two-dimensional Fourier transformation is then applied to the spectral OCT signal. Transformation along the wavenumber axis generates structural images similar to conventional Spectral OCT tomograms. Transformation along the time axis provides information about Doppler beating frequencies corresponding to flow velocities in the object.
One of the most critical sources of errors in studies on slow flow in retinal vessels is associated with involuntary movements of the object and/or instability of the scanning and interferometric systems. In both cases the phase shift or the beating Doppler frequency caused by these motions can be larger than the slow axial component of blood flow. We will refer to both effects as “bulk motion”. One of the most straightforward methods for bulk motion correction is offered by the STdOCT method. In STdOCT bulk motion is detected as an average local Doppler frequency shift introducing an offset to the measured flow velocities. This offset can be easily removed to achieve bulk motion free flow images. A detailed description of the STdOCT bulk motion correction algorithm can be found in .
2.3. Data visualization
The STdOCT method provides structural and flow information simultaneously. The structural information is displayed in cross-sectional gray scale images or, in the case of three dimensional imaging, in projection OCT fundus images . The projection OCT method performs axial summation of the OCT data in selected depth ranges. This highlights weakly scattering features in selected retinal layers.
The flow information is displayed in color-coded velocity maps. Red and blue indicate flow in opposite axial directions. The value of the axial velocity is displayed as the color saturation [27, 28]. Additionally, single axial scans are displayed as velocity profiles – velocity values are plotted versus the depth location in scatter graphs. Retinal capillaries are visualized in en face views of rendered 3-D OCT data sets. The rendering was performed using commercial software Amira (Visage Imaging, Inc., USA).
2.4. Scanning protocols for smart velocity ranging
Scanning protocols most commonly used for Doppler OCT are raster scans. For standard 3-D acquisition the voltage signals for the X and Y scanners are shown in Fig. 2(a) . One of the scanners is moving fast to acquire the cross-sectional OCT images (B-scans). This axis is usually referred to as the fast scan axis. The second scanner is moving slowly in the direction perpendicular to the fast scanner’s oscillations. Subsequent B-scans are acquired along the slow scan axis providing 3-D OCT data sets. The driving signal for the fast scanner can be either sawtooth or triangle waveform allowing for unidirectional or bidirectional acquisition of B-scans, respectively. The slow scanner is driven with a step-like voltage signal but may also be moved continuously.
Usually, the Doppler OCT analysis is performed for consecutive image lines (A-scans) along the fast scan axis (i.e. within a B-scan). The time period and in effect the detectable axial speed is set by the time between A-scans according to Eq. (1). Hence, the only parameters for tuning the velocity range are the camera speed or the lateral oversampling factor with all the consequences discussed in the introduction.
To overcome this limitation we propose to exploit alternative scan protocols that enable flexible velocity ranging with maintained high imaging speeds. As mentioned before the new scan protocols should allow for Doppler analysis not only along the fast scan axis (A-scan based) but also between consecutive B-scans or segments. In this case the velocity range is tunable through the time between the B-scan or segment lengths. The corresponding scanning voltage patterns are plotted in Figs. 2(b) and 2(c). Figure 2(b) illustrates the case where several B-scans are recorded at the same horizontal position. In Fig. 2(c) each B-scan is divided into smaller parts called segments, and each segment is repeated several times (Fig. 2(d)) at a given horizontal position. The segmented B-scans can be acquired for a number of vertical positions in the same way as it is done in standard raster scanning. The additional degree of freedom for velocity ranging is now the time between the segments which is determined by the number of A-scans within each segment. Clearly, the scan patterns in Fig. 2 could also be performed at a constant vertical position giving rise to two-dimensional versions of scanning patterns.
2.5. Selection of A-scans for STdOCT data analysis
To extract flow information the STdOCT method performs two-dimensional Fourier transformation in the data collected over time. The STdOCT data set can be constructed in several ways depending on the implemented scanning protocol. Figure 3 illustrates schematically which A-scans are chosen for further STdOCT processing in the case of scanning protocols suitable for a flexible velocity ranging. M axial scans can be acquired either within one B-scan (green rectangles in Figs. 3(a) and 3(b)) or sequentially for consecutive B-scans or segments (red rectangle in Fig. 3(a)). The same scheme can be applied for bidirectional scanning (triangular driving waveform). In this case one needs to distinguish between odd or even B-scans or segments (red and blue circles in Fig. 3(b)).
If the STdOCT data set encloses subsequent axial scans (green frame in Fig. 3), the time interval Δt between data taken for STdOCT analysis is equal to t rep. Thus, the maximal measurable axial flow velocity value is given by the formula (Eq. (1)):
The FFT window can be also selected to enclose axial scans from subsequent B-scans or segments. In the case of the unidirectional scanning (Fig. 3(a)) the time interval between data for the STdOCT analysis is Δt = M∙trep and the maximal axial flow velocity can be calculated using the formula:Fig. 3(b)) data points for the STdOCT data set must be selected from odd or even B-scans / segments to maintain a constant value of Δt. Two sets of images are thus obtained for the two scanning directions. They can be used for averaging of the flow maps. The sampling time interval is Δt = 2M∙trep and the maximal axial velocity value is then given by the formula:
Depending on the data processing method different flow regimes (large flow values and small values below the A-scan frequency limit) can be assessed using the same data sets. The design of scan patterns (i.e. the number of A-scans per B-scan / segment, number of segments in B-scan, number of B-scans, etc.) depends on the specific application and the expected flow velocity values in the object. However, scan protocols must be designed to provide data with oversampling sufficient for Doppler OCT analysis. During the STdOCT analysis the spectra in the STdOCT data set must be selected to include only data containing signal from the same location in the object. For short time intervals between axial scans the STdOCT data set are limited by the degree of oversampling. For long time intervals the motion of the object may be the main factor restraining the number of spectra selected for the Fourier transformation.
In the following discussion we refer to the A-scan based Doppler analysis as fast-flow detection and to the B-scan or segment based Doppler analysis as slow-flow detection method.
3. Results and discussion
3.1 Flow of Intralipid solution through a glass capillary
The initial experiments were performed in the test object to evaluate the proposed scan protocols for enhanced velocity ranging. To demonstrate high flow-sensitive imaging based on the new scan protocols as compared to the standard Doppler analysis a glass capillary of radius R0 = (450.0 ± 0.5) μm was placed in the object arm of the OCT instrument at an angle α = (85.7 ± 0.5) deg relative to the imaging beam. The information about the angle was extracted from 3-D OCT data set. Stable laminar flow of Intralipid solution at ΔV/Δt = 100 µl/min was maintained by the HPLC syringe pump. Volume flow is defined as volume ΔV of liquid flowing through the cross-section of the capillary in time Δt. The peak velocity of the parabolic flow profile can be calculated according to the formula:
Three scanning protocols were used to evaluate the extended flow detection capabilities with the STdOCT method: standard 2-D scan protocol with sufficiently slow A-scan rate to allow for comparison, 2-D segmented scan and raster with repetitive 2-D scans. The details of the protocols are given in Table 1 . The camera repetition times and scanning parameters were chosen to ensure the same maximal flow velocity values vmax in all scan protocols, according to Eqs. (2)-(4).
The velocity maps of the capillary flow are presented in Fig. 4 . Figures 4(a) and 4(b) show B-scans obtained with 2-D scan protocols. Figure 4(c) is a cross-section along the axis of the capillary reconstructed from a 3-D data set: A-scans located at the center of the capillary were extracted from consecutive B-scans and displayed as YZ cross-sectional image. Flow velocity profiles are shown below the B-scans. The parabolic distribution of the velocity is characteristic for laminar flow. A broadening of the profiles can be noticed. This effect was reported in the literature on Time-domain OCT systems [37–40]. The broadening increases in the vicinity of the capillary axis and is often associated with factors like beam geometry, Brownian motion, optical inhomogenity of measured medium etc.
To compare flow velocities obtained in STdOCT imaging with known velocity value provided by the HPLC pump we determined peak velocities for each scan protocol (Table 2 ) from parabolic functions fitted to the flow profiles. In the case of 2-D scan protocols, they were obtained from single profiles (indicated with green lines in Figs. 4(a) and 4(b)). For the 3-D scan pattern we calculated an average peak velocity from all 300 lines shown in Fig. 4(c). The novel scan protocols for enhanced flow detection correctly retrieve information about the peak velocity in the capillary even though the actual A-scan rate is two orders of magnitudes higher. As it can be noticed in Tables 1 and 2, the new segmented scan protocol can acquire more scans within less time without loss of velocity reconstruction performance.
The full potential and flexibility of the velocity ranging to cover different velocity regimes is demonstrated using a 2-D segmented scan protocol with unidirectional data acquisition (compare Fig. 2(d), Table 3 ). The glass capillary is mounted at the angle α = (79.8 ± 0.5) deg relative to the imaging beam. The flow of Intralipid was set to ΔV/Δt = 100 µl/min and to ΔV/Δt = 2000 µl/min in two separate STdOCT measurements. Based on Eq. (5) the calculated peak axial flow velocities were v0zI = (1.03 ± 0.02) mm/s and v0zII = (20.68 ± 0.18) mm/s, respectively.
Figure 5 shows flow velocity maps obtained with two methods of selection A-scans for the STdOCT data set: enclosing consecutive A-scans within a segment (fast flow detection) and including A-scans from consecutive segments (slow flow detection) as illustrated in Fig. 3(a). We effectively have two flow speed regimes for the same A-scan rate of 213 kHz. Velocity profiles at the center of the capillary are demonstrated below the flow maps.
The fast flow detection method correctly retrieves flow profiles for large values of the axial velocity component (Fig. 5(c)). However, if the velocity is too small, it falls below the sensitivity limit and reliable velocity measurement is lost (Fig. 5(a)). Nevertheless, in this case correct velocity values can be extracted using the complementary method for slow flow detection offered by the new scan protocols (Fig. 5(b)). Not surprisingly, the slow flow method cannot retrieve large flow velocities (Fig. 5(d)). Despite the fact that the Intralipid was flowing in one direction, the map reveals both positive and negative Doppler frequency shifts visualized as alternating red and blue concentric rings. This “velocity wrapping” is caused by aliasing of Doppler signals with frequencies exceeding the Nyquist limit.
3.2 STdOCT imaging of retinal vessels
As outlined in the introduction, retinal vessels exhibit a large range of axial velocities depending on the measurement location. Hence, it is an ideal sample for in vivo testing of our scan protocols enabling velocity ranging. We performed STdOCT imaging in the human retina using three scanning protocols: segmented B-scan, raster with segmented B-scans and the standard raster scan (Table 4 ). These scan protocols were designed to assess different axial flow velocities associated with the following retinal vessel locations: the vessels in the proximity of the optic disk, vasculature in the macular area, and the capillary network.
3.2.1. 2-D imaging of vessels in the optic disc area
The 2-D segmented protocol was implemented for STdOCT imaging in the vicinity of the optic disc where large vessels with fast blood flows are located but also smaller vessels with lower flow velocities are present. In the fast flow detection method consecutive A-scans from one segment were selected to generate both the structural and the velocity image (Figs. 6(a) and 6(c)). In the slow flow assessment method A-scans from consecutive segments were included in the STdOCT data set (Figs. 6 (b) and 6(d)). The structural cross-sections reveal no difference in imaging of retinal morphology with the two methods. The flow images however contain different but complementary information. The fast flow detection method allows for the assessment of axial flow velocity in large vessels (Fig. 6(c)). Quantitative flow profiles are shown in Figs. 6(e) and 6(f). Vessels characterized by slow axial blood flow velocities are not visualized. They are revealed with the method for slow flow detection (Fig. 6(d)). In this case however, measurement of velocity distribution in the large vessels is not possible for the reason explained earlier in the paper (section 3.1).
These results again demonstrate the advantage of segmented scan protocols which offer complementary velocity ranging for retinal perfusion imaging in a single data set.
3.2.2. 3-D imaging of the vasculature in the macular region
The segmented B-scan protocol was also extended to 3-D imaging, Fig. 2(c). The details of the protocol are provided in Table 4. The most important information to notice here is the oversampling value in the slow axis (vertical) direction of 1. This means that the scanning step between segmented B-scans was equal to the beam spot size. However, in this case there is no need for oversampling between segmented B-scans in a 3-D data set. The sampling density in slow axis direction just impacts the quality of reconstruction On the other hand, the oversampling must be assured in the fast axis direction within each segment in order to be able to perform the fast axial velocity detection. In this experiment the horizontal oversampling was 20 (Table 4).
3-D imaging was performed in the macular region of the retina. The results are presented in Fig. 7 . The structural information is presented as a projection OCT fundus image (Fig. 7(a)) with the 3-D OCT data set axially summed over the entire imaging depth.
The 3-D data set was processed twice using both fast and slow flow detection method. The fundus view in Fig. 7(a) reveals the vasculature within the imaged retinal region. Figure 7(b) shows the velocity map obtained with the fast flow method. Despite the high contrast of the blood vessel in the fundus view the fast flow detection method fails to visualize any blood flow. In the slow flow detection method, we set the STdOCT data set to include 5 A-scans from consecutive segments. The en face view reconstructed from velocity maps reveals information about blood flow (Fig. 7(c)). In this case, the velocity range has been narrowed by a factor of 20 as compared to the velocity range of fast flow detection.
3.3 Imaging of retinal capillaries
One of the most challenging tasks in Doppler OCT retinal imaging is mapping the flow velocities in the capillary network located in the inner retinal layers of the macular area. Due to small sizes of the vessels and their almost perpendicular orientation relative to the imaging beam, the axial velocity components can have magnitudes smaller than fractions of mm/s. On the other hand, detection of such small structures requires high imaging speeds in order to avoid image blurring and loss of structural details [5, 32]. High-speed acquisition, in turn, means that the velocity window is set to high velocity flows. In this section we demonstrate how this discrepancy can be avoided by using smart velocity ranging. Data were acquired in raster scan with high oversampling in the slow scan axis direction as opposed to usually implemented oversampling in the fast axis direction. The details of the scanning protocol are given in Table 4. The oversampling defined as a ratio between the beam spot size (~15 μm) and the scanning step (1.2 μm) was 12.5 suggesting that ~12 A-scans could be included in the STdOCT analysis. However, only five axial scans selected from consecutive B-scans were included in the STdOCT data set, and this value was determined experimentally. This discrepancy may be partially attributed to the underestimation of the transverse resolution of our instrument.
The results of imaging of the capillary network are shown in Figs. 8 -10 . The OCT fundus image in Fig. 8 is provided to indicate the location of the area examined with the STdOCT method. The structural data are visualized in projection OCT fundus images. The fundus projections are generated by axial summation of the OCT signal within depth ranges indicated by yellow dashed lines in Fig. 9(g) . The vessels within the ganglion cell layer are clearly visible in Fig. 9(a). The location in the innermost layers of the retina results in a large amount of light reaching these vessels. In addition, due to their relatively large diameters (~50 μm) the light scattering is very efficient. As a consequence, the OCT signal is strong and vessels are clearly visualized in structural images. The vasculature within the depth region of the inner plexiform layer is rather poorly visualized in intensity images (Fig. 9(b)). The inner plexiform layer is a highly scattering tissue as can be seen in Fig. 9(g) (the structure between lines II and III). Light scattering at small vessels located in this layer is not sufficient to provide enough contrast against the highly scattering surrounding. The depth range of the inner nuclear layer (Fig. 9(c)) provides a very clear image of the fine retinal capillary system. In the low scattering tissue (the depth range between lines III and IV in Fig. 9(g)) the light reflection at small vessels provides contrast sufficient for good visualization of the vasculature in the projection OCT images.
Figures 9(d)-(f) show the flow maps obtained in the same depth ranges as demonstrated in the projection OCT images. As compared to structural imaging, better visualization of retinal vasculature is provided especially in the ganglion cell layer and in the inner plexiform layer. Figure 9(d) reveals a network of small vessels in the ganglion cell layer, which is not visualized in Fig. 9(a). Figure 9(e) shows better contrast in visualization of the capillaries in the inner plexiform layer than the projection OCT fundus image in Fig. 9(b). Figure 9(f) clearly visualizes the capillary network in the inner nuclear layer, which can be compared to the structural image in Fig. 9(c).
The Doppler OCT provides images of retinal capillaries with increased contrast. However, we observe randomly varying colors in the flow maps. There may be few hypotheses that could explain the observed effects. One possible explanation is that blood in such small vessels cannot be treated as optically homogeneous medium anymore since the vessel diameter is already similar to that of individual red blood cells. Different blood constituents may flow with different, varying velocities which results in a random Doppler signal. In this context, due to the capillary diameter and the nature of blood flow we should not expect to observe well defined flow profiles across the vessel. Another possibility is that flow in these vessels exceeds the maximal measurable value prohibiting unambiguous velocity assessment.
The cross-sectional flow image shows capillaries (Fig. 9(h)). However, the velocity profiles cannot be reliably extracted due to small vessels’ diameters. In the vessel pointed by the green arrow, the colors change from blue to red (blue and red indicate flows in opposite directions) and therefore do not provide reliable information about flow velocity value. Nevertheless, the Doppler signal can provide a contrast for enhanced visualization of the capillary network as shown in Fig. 10. In these images the flow direction was ignored, and the quantitative information is lost. The axial flow information was displayed in en face view of the rendered OCT data. Figure 10(a) shows the image of all vessels in the inner retina (depth range between lines I-IV in Fig. 9(g)). Figure 10(b) displays capillaries in the inner plexiform and inner nuclear layers (depth range between the boundaries II-IV), and Fig. 10(c) visualizes vessels in the inner nuclear layer (region between lines III-IV).
New, smart scanning protocols are demonstrated which enable 2-D and 3-D OCT imaging of flow at two velocity ranges. We have demonstrated that these protocols increase the flexibility in velocity ranging, which is especially important in the case of retinal imaging. The obtained velocity maps reveal different details of retinal vasculature, from large vessels to small capillaries. We have shown that appropriate scanning protocols followed by processing leads to enhanced visualization of small capillaries.
The main advantage of the proposed approach is that relatively high and low velocity values can be visualized in the imaged region depending on the data processing method. Thus, the protocols provide information about flow velocity within both ranges. Additionally, this information is extracted from the same data set and from the same area of interest. As a consequence, there is no need to collect data twice with different settings of imaging parameters. The imaging speed is kept at maximal value. Thus, the total imaging time is decreased in comparison to standard protocols, where two measurements at two speeds must be performed to ensure detection of two different velocity ranges.
The protocols presented in this manuscript constitute only a part of the whole family of scanning protocols that can be designed in this way. The number of segments within a B-scan, the number of repeated segments or the number of A-scans per segment can be varied. It is possible to choose between unidirectional and bidirectional acquisition. Clearly, the particular version of the protocol must be adjusted to the expected velocity values in the imaged region of interest.
This work was supported by EuroHORCs-European Science Foundation EURYI Award EURYI-01/2008-PL. Maciej Szkulmowski and Anna Szkulmowska acknowledge additional support of Foundation for Polish Science (scholarship START 2009). The authors would like to express their gratitude to Prof. Bogusław Buszewski (Department of Environmental Chemistry and Ecoanalytics; Faculty of Chemistry; Nicolaus Copernicus University, Torun, Poland) for providing the HPLC pump.
References and links
1. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical Coherence Tomography,” Science 254(5035), 1178–1181 (1991). [CrossRef] [PubMed]
2. A. F. Fercher, C. K. Hitzenberger, G. Kamp, and S. Y. Elzaiat, “Measurement of intraocular distances by backscattering spectral interferometry,” Opt. Commun. 117(1-2), 43–48 (1995). [CrossRef]
3. M. Wojtkowski, V. J. Srinivasan, T. H. Ko, J. G. Fujimoto, A. Kowalczyk, and J. S. Duker, “Ultrahigh-resolution, high-speed, Fourier domain optical coherence tomography and methods for dispersion compensation,” Opt. Express 12(11), 2404–2422 (2004). [CrossRef] [PubMed]
4. K. Bizheva, A. Unterhuber, B. Hermann, B. Povazay, H. Sattmann, W. Drexler, A. Stingl, T. Le, M. Mei, R. Holzwarth, H. A. Reitsamer, J. E. Morgan, and A. Cowey, “Imaging ex vivo and in vitro brain morphology in animal models with ultrahigh resolution optical coherence tomography,” J. Biomed. Opt. 9(4), 719–724 (2004). [CrossRef] [PubMed]
5. B. Potsaid, I. Gorczynska, V. J. Srinivasan, Y. L. Chen, J. Jiang, A. Cable, and J. G. Fujimoto, “Ultrahigh speed spectral / Fourier domain OCT ophthalmic imaging at 70,000 to 312,500 axial scans per second,” Opt. Express 16(19), 15149–15169 (2008). [CrossRef] [PubMed]
6. M. Wojtkowski, B. L. Sikorski, I. Gorczynska, M. Gora, M. Szkulmowski, D. Bukowska, J. Kaluzny, J. G. Fujimoto, and A. Kowalczyk, “Comparison of reflectivity maps and outer retinal topography in retinal disease by 3-D Fourier domain optical coherence tomography,” Opt. Express 17(5), 4189–4207 (2009). [CrossRef] [PubMed]
7. M. Szkulmowski, M. Wojtkowski, B. Sikorski, T. Bajraszewski, V. J. Srinivasan, A. Szkulmowska, J. J. Kaluzny, J. G. Fujimoto, and A. Kowalczyk, “Analysis of posterior retinal layers in spectral optical coherence tomography images of the normal retina and retinal pathologies,” J. Biomed. Opt. 12(4), 11 (2007). [CrossRef]
8. B. R. White, M. C. Pierce, N. Nassif, B. Cense, B. H. Park, G. J. Tearney, B. E. Bouma, T. C. Chen, and J. F. de Boer, “In vivo dynamic human retinal blood flow imaging using ultra-high-speed spectral domain optical coherence tomography,” Opt. Express 11(25), 3490–3497 (2003). [CrossRef] [PubMed]
9. H. Wehbe, M. Ruggeri, S. Jiao, G. Gregori, C. A. Puliafito, and W. Zhao, “Automatic retinal blood flow calculation using spectral domain optical coherence tomography,” Opt. Express 15(23), 15193–15206 (2007). [CrossRef] [PubMed]
12. L. An and R. K. K. Wang, “In vivo volumetric imaging of vascular perfusion within human retina and choroids with optical micro-angiography,” Opt. Express 16(15), 11438–11452 (2008). [CrossRef] [PubMed]
13. Y. K. Tao, A. M. Davis, and J. A. Izatt, “Single-pass volumetric bidirectional blood flow imaging spectral domain optical coherence tomography using a modified Hilbert transform,” Opt. Express 16(16), 12350–12361 (2008). [CrossRef] [PubMed]
14. Y. K. Tao, K. M. Kennedy, and J. A. Izatt, “Velocity-resolved 3D retinal microvessel imaging using single-pass flow imaging spectral domain optical coherence tomography,” Opt. Express 17(5), 4177–4188 (2009). [CrossRef] [PubMed]
15. Y. H. Zhao, Z. P. Chen, C. Saxer, S. H. Xiang, J. F. de Boer, and J. S. 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. 25(2), 114–116 (2000). [CrossRef] [PubMed]
17. R. A. Leitgeb, L. Schmetterer, W. Drexler, A. F. Fercher, R. J. Zawadzki, and T. Bajraszewski, “Real-time assessment of retinal blood flow with ultrafast acquisition by color Doppler Fourier domain optical coherence tomography,” Opt. Express 11(23), 3116–3121 (2003). [CrossRef] [PubMed]
18. R. Leitgeb, L. Schmetterer, W. Drexler, F. Berisha, C. Hitzenberger, M. Wojtkowski, T. Bajraszewski, and A. F. Fercher, “Real-time measurement of in-vitro and in-vivo blood flow with Fourier domain optical coherence tomography,” in Conference on Coherence Domain Optical Methods and Optical Coherence Tomography in Biomedicine VIII, V. V. I. J. A. F. J. G. Tuchin, ed. (San Jose, CA, 2004), pp. 141–146.
19. A. H. Bachmann, M. L. Villiger, C. Blatter, T. Lasser, and R. A. Leitgeb, “Resonant Doppler flow imaging and optical vivisection of retinal blood vessels,” Opt. Express 15(2), 408–422 (2007). [CrossRef] [PubMed]
20. R. Michaely, A. H. Bachmann, M. L. Villiger, C. Blatter, T. Lasser, and R. A. Leitgeb, “Vectorial reconstruction of retinal blood flow in three dimensions measured with high resolution resonant Doppler Fourier domain optical coherence tomography,” J. Biomed. Opt. 12(4), 041213 (2007). [CrossRef] [PubMed]
21. Z. Luo, Z. Wang, Z. Yuan, C. Du, and Y. Pan, “Optical coherence Doppler tomography quantifies laser speckle contrast imaging for blood flow imaging in the rat cerebral cortex,” Opt. Lett. 33(10), 1156–1158 (2008). [CrossRef] [PubMed]
22. A. Mariampillai, B. A. Standish, E. H. Moriyama, M. Khurana, N. R. Munce, M. K. Leung, J. Jiang, A. Cable, B. C. Wilson, I. A. Vitkin, and V. X. Yang, “Speckle variance detection of microvasculature using swept-source optical coherence tomography,” Opt. Lett. 33(13), 1530–1532 (2008). [CrossRef] [PubMed]
23. M. C. Skala, A. Fontanella, H. Hendargo, M. W. Dewhirst, and J. A. Izatt, “Combined hyperspectral and spectral domain optical coherence tomography microscope for noninvasive hemodynamic imaging,” Opt. Lett. 34(3), 289–291 (2009). [CrossRef] [PubMed]
24. Y. Hong, S. Makita, M. Yamanari, M. Miura, S. Kim, T. Yatagai, and Y. Yasuno, “Three-dimensional visualization of choroidal vessels by using standard and ultra-high resolution scattering optical coherence angiography,” Opt. Express 15(12), 7538–7550 (2007). [CrossRef] [PubMed]
25. Y. Yasuno, S. Makita, T. Endo, G. Aoki, M. Itoh, and T. Yatagai, “Simultaneous B-M-mode scanning method for real-time full-range Fourier domain optical coherence tomography,” Appl. Opt. 45(8), 1861–1865 (2006). [CrossRef] [PubMed]
26. S. Makita, T. Fabritius, and Y. Yasuno, “Quantitative retinal-blood flow measurement with three-dimensional vessel geometry determination using ultrahigh-resolution Doppler optical coherence angiography,” Opt. Lett. 33(8), 836–838 (2008). [CrossRef] [PubMed]
27. M. Szkulmowski, A. Szkulmowska, T. Bajraszewski, A. Kowalczyk, and M. Wojtkowski, “Flow velocity estimation using joint Spectral and Time domain Optical Coherence Tomography,” Opt. Express 16(9), 6008–6025 (2008). [CrossRef] [PubMed]
28. A. Szkulmowska, M. Szkulmowski, A. Kowalczyk, and M. Wojtkowski, “Phase-resolved Doppler optical coherence tomography--limitations and improvements,” Opt. Lett. 33(13), 1425–1427 (2008). [CrossRef] [PubMed]
29. A. Szkulmowska, M. Szkulmowski, D. Szlag, A. Kowalczyk, and M. Wojtkowski, “Three-dimensional quantitative imaging of retinal and choroidal blood flow velocity using joint Spectral and Time domain Optical Coherence Tomography,” Opt. Express 17(13), 10584–10598 (2009). [CrossRef] [PubMed]
30. M. Szkulmowski, I. Grulkowski, D. Szlag, A. Szkulmowska, A. Kowalczyk, and M. Wojtkowski, “Flow velocity estimation by complex ambiguity free joint Spectral and Time domain Optical Coherence Tomography,” Opt. Express 17(16), 14281–14297 (2009). [CrossRef] [PubMed]
31. B. H. Park, M. C. Pierce, B. Cense, S. H. Yun, M. Mujat, G. J. Tearney, B. E. Bouma, and J. F. de Boer, “Real-time fiber-based multi-functional spectral-domain optical coherence tomography at 1.3 µm,” Opt. Express 13(11), 3931–3944 (2005). [CrossRef] [PubMed]
33. Y. Wang, B. A. Bower, J. A. Izatt, O. Tan, and D. Huang, “Retinal blood flow measurement by circumpapillary Fourier domain Doppler optical coherence tomography,” J. Biomed. Opt. 13(6), 064003 (2008). [CrossRef] [PubMed]
34. B. J. Vakoc, R. M. Lanning, J. A. Tyrrell, T. P. Padera, L. A. Bartlett, T. Stylianopoulos, L. L. Munn, G. J. Tearney, D. Fukumura, R. K. Jain, and B. E. Bouma, “Three-dimensional microscopy of the tumor microenvironment in vivo using optical frequency domain imaging,” Nat. Med. 15(10), 1219–1223 (2009). [CrossRef] [PubMed]
35. R. Rabbetts, Bennett & Rabbetts' Clinical Visual Optics (Elsevier, 2007).
36. I. Gorczynska, V. J. Srinivasan, L. N. Vuong, R. W. S. Chen, J. J. Liu, E. Reichel, M. Wojtkowski, J. S. Schuman, J. S. Duker, and J. G. Fujimoto, “Projection OCT fundus imaging for visualising outer retinal pathology in non-exudative age-related macular degeneration,” Br. J. Ophthalmol. 93(5), 603–609 (2009). [CrossRef] [PubMed]
37. H. W. Ren, K. M. Brecke, Z. H. Ding, Y. H. Zhao, J. S. Nelson, and Z. P. Chen, “Imaging and quantifying transverse flow velocity with the Doppler bandwidth in a phase-resolved functional optical coherence tomography,” Opt. Lett. 27(6), 409–411 (2002). [CrossRef] [PubMed]
38. S. G. Proskurin, Y. H. He, and R. K. K. Wang, “Determination of flow velocity vector based on Doppler shift and spectrum broadening with optical coherence tomography,” Opt. Lett. 28(14), 1227–1229 (2003). [CrossRef] [PubMed]
39. D. Q. Piao, L. L. Otis, and Q. Zhu, “Doppler angle and flow velocity mapping by combined Doppler shift and Doppler bandwidth measurements in optical Doppler tomography,” Opt. Lett. 28(13), 1120–1122 (2003). [CrossRef] [PubMed]
40. D. Q. Piao and Q. Zhu, “Quantifying Doppler angle and mapping flow velocity by a combination of Doppler-shift and Doppler-bandwidth measurements in optical Doppler tomography,” Appl. Opt. 42(25), 5158–5166 (2003). [CrossRef] [PubMed]