High resolution imaging of microvasculature

High-resolution imaging of microvasculature in
human skin in-vivo with optical coherence
tomography
Gangjun Liu,1,2,*
Wangcun Jia,1
Victor Sun,1
Bernard Choi,1,2
and Zhongping Chen1,2,3
1
Beckman Laser Institute, University of California, Irvine, Irvine, California 92612, USA
2
Department of Biomedical Engineering, University of California, Irvine, Irvine, California 92697, USA
3
z2chen@uci.edu
*
gangjun@gmail.com
Abstract: In this paper, the features of the intensity-based Doppler variance
(IBDV) method were analyzed systemically with a flow phantom. The
effects of beam scanning density, flow rate and the time interval between
neighboring A-lines on the performance of this method were investigated.
The IBDV method can be used to quantify the flow rate and its sensitivity
can be improved by increasing the time interval between the neighboring A-
lines. A higher sensitivity IBDV method that applies the algorithm along
the slower scan direction was proposed. In comparison to laser speckle
imaging maps of blood flow, we demonstrated the ability of the method to
identify vessels with altered blood flow. In clinical measurements, we
demonstrated the ability of the method to image vascular networks with
exquisite spatial resolution and at depths up to 1.2 mm in human skin.
These results collectively demonstrated the potential of the method to
monitor the microvasculature during disease progression and in response to
therapeutic intervention.
© 2012 Optical Society of America
OCIS codes: (170.4500) Optical coherence tomography; (170.3890) Medical optics
instrumentation.
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).
2. Z. Chen, T. E. Milner, D. Dave, and J. S. Nelson, “Optical Doppler tomographic imaging of fluid flow velocity
in highly scattering media,” Opt. Lett. 22(1), 64–66 (1997).
3. Z. Chen, T. E. Milner, S. Srinivas, X. Wang, A. Malekafzali, M. J. C. van Gemert, and J. S. Nelson,
“Noninvasive imaging of in vivo blood flow velocity using optical Doppler tomography,” Opt. Lett. 22(14),
1119–1121 (1997).
4. J. A. Izatt, M. D. Kulkarni, S. Yazdanfar, J. K. Barton, and A. J. Welch, “In vivo bidirectional color Doppler flow
imaging of picoliter blood volumes using optical coherence tomography,” Opt. Lett. 22(18), 1439–1441 (1997).
5. Y. Zhao, Z. Chen, C. Saxer, S. 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).
6. Y. Zhao, Z. Chen, C. Saxer, Q. Shen, S. Xiang, J. F. de Boer, and J. S. Nelson, “Doppler standard deviation
imaging for clinical monitoring of in vivo human skin blood flow,” Opt. Lett. 25(18), 1358–1360 (2000).
7. H. Ren, K. M. Brecke, Z. Ding, Y. Zhao, J. S. Nelson, and Z. 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).
8. G. Liu, W. Qi, L. Yu, and Z. Chen, “Real-time bulk-motion-correction free Doppler variance optical coherence
tomography for choroidal capillary vasculature imaging,” Opt. Express 19(4), 3657–3666 (2011).
9. L. Yu and Z. Chen, “Doppler variance imaging for three-dimensional retina and choroid angiography,” J.
Biomed. Opt. 15(1), 016029 (2010).
10. J. Barton and S. Stromski, “Flow measurement without phase information in optical coherence tomography
images,” Opt. Express 13(14), 5234–5239 (2005).
#161014 - $15.00 USD Received 5 Jan 2012; revised 4 Mar 2012; accepted 4 Mar 2012; published 20 Mar 2012
(C) 2012 OSA 26 March 2012 / Vol. 20, No. 7 / OPTICS EXPRESS 7694

11. A. Mariampillai, B. A. Standish, E. H. Moriyama, M. Khurana, N. R. Munce, M. K. K. Leung, J. Jiang, A.
Cable, B. C. Wilson, I. A. Vitkin, and V. X. D. Yang, “Speckle variance detection of microvasculature using
swept-source optical coherence tomography,” Opt. Lett. 33(13), 1530–1532 (2008).
12. A. Mariampillai, M. K. K. Leung, M. Jarvi, B. A. Standish, K. Lee, B. C. Wilson, A. Vitkin, and V. X. D. Yang,
“Optimized speckle variance OCT imaging of microvasculature,” Opt. Lett. 35(8), 1257–1259 (2010).
13. Y. Yasuno, Y. J. Hong, S. Makita, M. Yamanari, M. Akiba, M. Miura, and T. Yatagai, “In vivo high-contrast
imaging of deep posterior eye by 1-microm swept source optical coherence tomography and scattering optical
coherence angiography,” Opt. Express 15(10), 6121–6139 (2007).
14. E. Jonathan, J. Enfield, and M. J. Leahy, “Correlation mapping method for generating microcirculation
morphology from optical coherence tomography (OCT) intensity images,” J Biophotonics 4(9), 583–587 (2011),
http://onlinelibrary.wiley.com/doi/10.1002/jbio.201000103/pdf.
15. G. Liu, L. Chou, W. Jia, W. Qi, B. Choi, and Z. Chen, “Intensity-based modified Doppler variance algorithm:
application to phase instable and phase stable optical coherence tomography systems,” Opt. Express 19(12),
11429–11440 (2011).
16. J. Fingler, R. J. Zawadzki, J. S. Werner, D. Schwartz, and S. E. Fraser, “Volumetric microvascular imaging of
human retina using optical coherence tomography with a novel motion contrast technique,” Opt. Express 17(24),
22190–22200 (2009).
17. L. An, J. Qin, and R. K. Wang, “Ultrahigh sensitive optical microangiography for in vivo imaging of
microcirculations within human skin tissue beds,” Opt. Express 18(8), 8220–8228 (2010).
18. S. Zotter, M. Pircher, T. Torzicky, M. Bonesi, E. Götzinger, R. A. Leitgeb, and C. K. Hitzenberger,
“Visualization of microvasculature by dual-beam phase-resolved Doppler optical coherence tomography,” Opt.
Express 19(2), 1217–1227 (2011).
19. S. Makita, F. Jaillon, M. Yamanari, M. Miura, and Y. Yasuno, “Comprehensive in vivo micro-vascular imaging
of the human eye by dual-beam-scan Doppler optical coherence angiography,” Opt. Express 19(2), 1271–1283
(2011).
20. 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).
21. W. Jia, V. Sun, N. Tran, B. Choi, S. W. Liu, M. C. Mihm, Jr., T. L. Phung, and J. S. Nelson, “Long-term blood
vessel removal with combined laser and topical rapamycin antiangiogenic therapy: implications for effective port
wine stain treatment,” Lasers Surg. Med. 42(2), 105–112 (2010).
22. O. Yang, D. Cuccia, and B. Choi, “Real-time blood flow visualization using the graphics processing unit,” J.
Biomed. Opt. 16(1), 016009 (2011).
23. A. K. Bui, K. M. Teves, E. Indrawan, W. Jia, and B. Choi, “Longitudinal, multimodal functional imaging of
microvascular response to photothermal therapy,” Opt. Lett. 35(19), 3216–3218 (2010).
24. J. Enfield, E. Jonathan, and M. Leahy, “In vivo imaging of the microcirculation of the volar forearm using
correlation mapping optical coherence tomography (cmOCT),” Biomed. Opt. Express 2(5), 1184–1193 (2011).
1. Introduction
Optical coherence tomography (OCT) is a non-invasive, high resolution imaging modality
which is able to capture three-dimensional images from within highly scattering biological
tissue [1]. The functional extension of the OCT technique to image biological tissue has
numerous applications in biomedical imaging. Doppler optical coherence tomography
(DOCT) or optical Doppler tomography (ODT) is one functional extension of OCT which
combines the Doppler principle with OCT [2–5]. Phase-resolved ODT is one of the preferred
ODT methods to extract the Doppler frequency because of its high velocity sensitivity [5].
The phase-resolved ODT uses the phase difference among the adjacent A-lines to extract the
Doppler frequency shift and quantitative information such as blood flow speed and blood flow
direction can be obtained [4–9].
There are considerable interests in imaging blood vessels using intensity or amplitude
information instead of phase information [10–14]. Barton et al. proposed a method based on
the speckle of conventional amplitude optical coherence tomography images [10].
Mariampillai et al. used the speckle variance in a small 3D volume to image blood vessels
[11, 12]. Yasuno et al. used the intensity threshold binarization-based method, scattering
optical coherence angiography (SOCA), for retinal and choroidal blood vessel imaging [13].
Jonathan et al. used a two dimensional correlation map based on the OCT intensity images for
blood vessel extraction [14]. Recently, we proposed an intensity-based method that uses an
algorithm derived from a modified Doppler variance algorithm [15]. In our previous paper,
the algorithm is used between adjacent A-lines in a single B-scan frame. Although the

performance of the algorithm is comparable to the phase-resolved methods, the sensitivity of
this algorithm is not high enough for imaging small vessel where blood flow is very slow.
For phase-resolved ODT method, the velocity sensitivity can be improved by increasing
the time interval between adjacent A-lines used in the algorithm. Although increasing Doppler
sensitivity by using phase-resolved algorithms between adjacent frames or along the slow
scanning direction without sacrificing the system speed was proposed in 2000 [5], recent
development of high speed Fourier domain OCT makes it possible to be implemented in
Doppler OCT systems [16–19]. In this paper, the way to increase the velocity sensitivity of
the intensity-based Doppler variance (IBDV) method will be investigated. The features of
IBDV method were analyzed systemically with a flow phantom. The effects of beam scanning
density, flow rate and the time interval on the performance of this method were investigated.
It is found that the IBDV method can be used to quantify flow rate and the velocity sensitivity
of the method can be improved by increasing the time interval between the neighboring A-
lines. An improved sensitivity IBDV method that applies the algorithm along the slower scan
direction was proposed. To demonstrate the capability of this improved method, we imaged
the subdermal side of hamster skin in a dorsal skin fold window chamber. The performance of
the proposed IBDV method was evaluated by comparing to the blood flow maps obtained
with laser speckle imaging (LSI). In-vivo imaging of microcirculation in human skin was
conducted and capillary loops in the dermal papillae were detected.
2. Method
IBDV method is derived from the phase resolved Doppler variance method based on an
autocorrelation algorithm [6, 7]. By replacing the complex value term in the autocorrelation
algorithm with an amplitude term, the IBDV can be written as follows [15]:
, 1,
1 12
2 2
, 1,
1 1
1
( )
2
J N
j z j z
j z
J N
j z j z
j z
A A
A A


 

 
 



(1)
where ,j z
A is the complex value OCT data of the jth A-line at a depth of z, ,j z
A is the
amplitude (absolute value) of ,j z
A , J is the number of averaged A-lines and N is the number
of averaged depth pixels. Performance of the IBDV method is comparable with that of
traditional phase-resolved Doppler variance and color Doppler methods for a phase-stable
situation. IBDV demonstrates better performance without instability induced artifacts in a
phase-unstable situation [15]. Previously, we reported the IBDV algorithm that was used
along the adjacent A-lines in one single B-scan and it is called inter-A-line IBDV(IA-IBDV)
here [15]. In this manuscript, the features of the IBDV method are analyzed systemically with
a flow phantom. The effects of beam scanning density, flow rate and the time interval of
neighboring A-lines on the performance of this method are investigated. It is found that the
IBDV method can be used to quantify the flow rate and its sensitivity can be improved by
increase the time interval between the neighboring A-lines. A high sensitivity IBDV method
that applies the algorithm along the slower scan direction without sacrificing the system speed
was proposed and was called inter-frame IBDV (IF-IBDV). In the following sections of this
manuscript, if not specified otherwise, the values for J and N in Eq. (1) are both set at 4 to
balance between signal to noise ratio, resolution and computation time [15].
3. Doppler OCT system
A swept source OCT system was used in this study. The schematic of the system setup is
shown in Fig. 1. The system used a compact swept source laser with a central wavelength of

1310 nm, an A-line rate of 50 kHz and a total average power of 16 mW (SSOCT-1310, Axsun
Technologies Inc, Billerica, MA). For the OCT module, A Mach-Zehnder type interferometer
was used and 90% of the laser light power was sent to the sample arm and 10% of the light to
the reference arm. A dual-balanced detection scheme was used to acquire the signal. The
system utilized a K-trigger mode so that no linear wavenumber re-calibration was needed. In
the sample arm, a fiber collimator, a two-axis galvo mirror scanning system and an
achromatic doublet with a focal length of 30mm was used. The collimator, galvo mirror
scanner and the achromatic doublet were packaged in a handheld probe, achieving a beam
diameter of 2 mm and a lateral resolution of 14.6 µm. For in-vivo awake human imaging, the
probe worked in the contact mode so that the motion induced artifact was minimized. The
FWHM bandwidth of the laser source was approximately 80 nm, resulting in a depth
resolution of 9.3 µm in air. A data acquisition software written in C + + , running on a 64 bit
Windows 7 operating system, was used to control the galvo scanner and acquire the data. The
software was able to display the OCT images in real time and save the acquired fringe data
into the computer memory then onto the hard disk.
Fig. 1. Schematic of the swept source Doppler OCT system.
4. Data processing flow
The saved data was processed afterwards to obtain the IBDV images and en-face images with
a custom software written in C++ . Figure 2 shows the data processing flow to generate the
IBDV image. The acquired fringes in one frame were firstly averaged to obtain the
background fringe. After subtracting the background fringe, the fringes were shaped with a
Gaussian function and the complex analytical depth encoded signals were converted from the
shaped fringes using fast Fourier transform (FFT). The amplitude values of the complex
analytical depth encoded signals were used to calculate the OCT images after logarithm
calculation and IBDV values using Eq. (1). For the final IBDV image, an intensity threshold
that is 1dB above the OCT intensity noise floor was used to eliminate the low scattering but
high noise region. For in-vivo human patient imaging especially when using the IF-IBDV
method, there may be strong axial movement induced by involuntarily motion of the imaging
subject and axial direction movement may degrade the performance of the IBDV method.
Axial movement induced displacement can be compensated by cross-correlating the
amplitude values of neighboring A-lines used in the algorithm [16]. Both OCT and IBDV
images were then obtained from the compensated A-line amplitudes. The en-face IBDV
images at different depth were extracted from the three dimensional IBDV tomographic
images and projection view en-face images were processed from the en-face IBDV images.

Fig. 2. The data processing Flowchart for generating the IBDV image.
5. Features of IBDV
To further understand the features of the IBDV method, the effects of several important
parameters, such as beam scan density (sampling density), flow rate and the time interval on
the performance of the IBDV were investigated. A flow phantom was used for this study. A
capillary tube with an inner diameter of 500 µm was embedded in a tissue phantom made with
silicon and Titanium dioxide. A 0.1% Intralipid solution composed of particles of 0.356 µm
diameter was pumped into the tube with a syringe pump.
5.1 Effect of sampling density level
Fig. 3. The effect of sampling density on the performance of IBDV. (a) OCT image of a flow
phantom. (b)-(m) IBDV images at different sampling density levels. The Sds in (b)-(m) mean
the sampling density level. (n) The relationship between background average IBDV versus the
sampling density level. The scale bars in (a) represent 500 µm.
The IBDV method is based on the intensity (or amplitude) autocorrelation between
adjacent beam scanning locations. Autocorrelation requires certain correlation among the
adjacent scanning locations (or A-lines) and the lateral shift between adjacent scanning
locations will decrease autocorrelation level and increase the decorrelation noise. The
decrease in autocorrelation will degrade the performance of the method. Therefore, the beam
scanning density (or sampling density) is important to the performance of the method. In
order to investigate the effect of beam scanning density on the performance of the IBDV, we
analyzed the average background IBDV values at different beam scanning density levels.
Figure 3(a) shows the OCT image of the flow phantom. Figures 3(b)-3(m) show the IBDV
images at different sampling density levels (Sd in these images means sampling density
level). Sampling density level was defined as the ratio of laser beam spot size (14.6 µm in this
case) to the lateral scanning step size. It can be clearly seen that t noise level decreases as the

sampling density level increases. It order to quantity the noise level, the average background
IBDV values (scaled to numbers between 0 and 255 for displaying purpose) in the region
indicated by the yellow rectangle in Fig. 3(b) were calculated for all the IBDV images. The
relationship between the background IBDV value and the sampling density level is plotted in
Fig. 3(n). A sampling density level of more than 3 is required in order to have a noise level of
less than 4.
5.2 Effect of flow rate
Fig. 4. The dependence of IBDV value on the flow rate. (a) OCT image of a flow phantom.
(b)-(g) IBDV images of the flow phantom at different flow rates. The flow rates are shown in
the figures. (h) The relationship between average IBDV value in the tube and the flow rate.
The red line shows the exponential decay fit of the data. (i) The relationship between the
average IBDV value in the tube versus the flow rate at lower speed. The red line shows the
linear fit of the data. The scale bars in (a) represent 500 µm.
One of the attractive features of phase resolved Doppler OCT is that quantitative information
such as blood flow speed, flow velocity profile, blood flow direction can be obtained. In our
previous publication of IBDV, quantifying the flow parameters was not demonstrated. Here,
we investigated the dependence of average IBDV value in the tube on flow rate. The results
are shown in Fig. 4. The images each contained 2048 A-lines covering a scanning range of 1
mm, and the corresponding sampling density level was 29.9. Figure 4(a) shows the OCT
image of the sample. Figures 4(b)-4(g) show the IBDV images at the flow rates of,
respectively, 8.4 mm/s, 16.8 mm/s, 42 mm/s, 58.8 mm/s, 84 mm/s, and 588 mm/s. The
improvement of sensitivity of the IBDV with the increase of flow rate could be seen from
these images. In order to quantify this improvement, we calculated the average IBDV value in
the tube when the flow rate was changed. The relationship between the average IBDV value
in the tube and the flow rate is plotted in Fig. 4(h). Figure 4(h) clearly shows that the average
IBDV value increases with the flow rate until it becomes saturated when the flow rate exceeds
certain level (around 168 mm/s in this case). An exponential decay fit of the data shows the
coefficient of determination to be 0.99397 and the saturated average IBDV to be 51.5. This

result demonstrates that the IBDV method could be used to quantify flow rate. Figure 4(i)
shows relationship between average IBDV value in the tube and the flow rate at a relative low
flow rate before the averaged IBDV value becomes saturated. A linear fit of the data
demonstrates a correlation coefficient of 0.99745 between the two parameters.
5.3 Effect of time interval
Fig. 5. The dependence of IBDV value on neighboring A-line time interval. (a) OCT image of
a flow phantom. (b)-(g) IBDV images of the flow phantom at different neighboring A-line time
interval. The neighboring A-line time intervals are shown in the figures. (h) The relationship
between average IBDV value in the tube and neighboring A-line time interval. The red line
shows the exponential decay fit of the data. (i) The relationship between average IBDV value
in the tube and neighboring A-line time interval. The red line shows the exponential decay fit
of the data. The scale bars in (a) represent 500 µm.
Recently, high sensitivity imaging of microvasculature has been demonstrated with OCT by
several groups [16–19]. For phase resolved DOCT, the blood flow sensitivity could be
improved by increasing the time intervals between neighboring A-lines used in the algorithm.
This can be realized by reducing the A-line rate of the system, by using the algorithm along
the slowing scanning direction or by using dual beam setup [16–19]. Here, the effect of time
interval between neighboring A-lines used in the IBDV algorithm on the sensitivity of IBDV
was investigated and the results are shown in Fig. 5. For all the images shown in Fig. 5(a)-
5(g), the flow rate was set at 12.6 mm/s. The image in Fig. 5(a) contains 9216 A-lines
covering a scanning range of 1 mm and the corresponding sampling density level was 134.6.
Figure 5(b) shows the IBDV image processed directly from the original data. Because the A-
line rate of the current swept source laser was fixed at 50 KHz with a neighboring A-line time
interval of 20 µs, we cannot directly test the effect of time interval on the sensitivity of IBDV.
Since the data was acquired at a very high sampling density level (134.6), we could change
the time interval by using only partial data. For example, to test the time interval of 40 µs, we
took data points from all the even (or odd) number A-lines in the original data set to form a
new data set with a total of 4608 A-lines. . The algorithm as Eq. (1) was then applied on this

new data set to obtain an IBDV image with a neighboring A-line time interval of 40 µs.
Accordingly, to test the algorithm with a neighboring A-line time interval of 60 µs, every
third A-line was extracted from the original data set to form a new data set with a total of
3072 A-lines and the IBDV algorithm was applied to the new data set. With the same
downsampling method, we obtained the IDBV images with different neighboring A-line time
intervals from 20 µs to 400 µs. Figures 5(c)-5(g) show the IBDV images corresponding to
neighboring A-line time intervals of, respectively, 20 µs, 60 µs, 100 µs, 140 µs, 180 µs, and
400 µs. It should be pointed out that in Figs. 5(c)-5(g), each image contains different total
number of A-lines and Fig. 5(g) has the lowest sampling density level. Figure 5 (g) contains
460 A-lines and the corresponding sampling density level is 6.71. Figures 5(b)-5(g)
demonstrate that the sensitivity of IBDV improves with the increase of the neighboring A-line
time interval. It could be also seen that the shadowing effect below the tube also increases
with the increase of the neighboring A-line time interval. The average IBDV value in the tube
was used to quantify the sensitivity improvement and its relationship with the neighboring A-
line time interval is plotted in Fig. 5(h). Figure 5(h) clearly shows that the average IBDV
value increases with the time interval and it becomes saturated when the time interval exceeds
certain level. An exponential decay fit of the data shows a coefficient of determination value
of 0.99785 and a saturated average IBDV value of 49. Figure 5(i) shows that at relatively
small time interval region, there is a liner relationship between average IBDV value and the
time interval.
These findings suggested that the ways to improve the blood vessel detection sensitivity
for the phase resolved methods could also be used to improve the sensitivity of the IBDV
method. This means the use of the algorithm along the slowing scanning direction (inter-
frame algorithm) and the using of dual beam setup can also been used to improve the
sensitivity of the IBDV method. Because the dual beam setup requires hardware modification
of the system, we only tested the inter-frame algorithm and Fig. 6 shows the result. Figure
6(a) shows the OCT image of the flow phantom. The image contains 1024 A-lines covering a
scanning range of 1 mm. The 1mm range beam scanning was repeated for 4 times and the
data were processed using IF-IBDV algorithm. Figure 6(b) shows the IBDV image processed
using the IF-IBDV when the syringe pump was stopped and the corresponding flow rate was
zero. This demonstrates that this method is sensitive enough to detect the Brownian motion of
the Intralipid uid in the tube. Figure 6(c) shows the histogram of the IBDV value distribution
inside the tube. An average IBDV value of 53.28 was found and this value was very close to
the saturation average IBDV value predicted by the exponential decay model in Fig. 4(h) and
Fig. 5(h). Figure 6(d) shows the IBDV image processed using the IF-IBDV when the flow
rate was set at 92.4 mm/s. Figure 6(e) shows the histogram of the IBDV value distribution
inside the tube for the image in Fig. 6(d) and the average IBDV value was 53.86. Because the
IBDV values were scaled to 0-255 in a 8-bit grey level image and the maximum of the
average IBDV value was around 50, in the following content of this manuscript, we will
multiply the final value obtained from Eq. (1) by 3 to enhance the contrast for displaying
purpose.

Fig. 6. The IF-IBDV images of a flow phantom. (a) OCT image of a flow phantom. (b) IF-
IBDV images of the flow phantom with the syringe pump stopped. (c) The histogram of the
IBDV value distribution inside the tube in (b). (d) IF-IBDV images of the flow phantom with
flow rate of 92.4 mm/s. (e) The histogram of the IBDV value distribution inside the tube in (d)
. The scale bars in (a) represent 500 µm.
6. In-vivo imaging of vasculature network in hamster skin
The subdermal side of a dorsal skinfold window chamber (DSWC) on an anesthetized
hamster [20] was imaged with the OCT system. Figure 7(a) shows the OCT images obtained
with our system. The frame has 1024 A-lines covering a scanning range of 4 mm. Figure 7(b)
shows the IA-IBDV results generated with the algorithm that used adjacent A-lines (inter-A-
lines) in the single B-scan (fast scan) frame. Blood vessels were not detected with this
algorithm. In Fig. 7(c), the IF-IBDV algorithm was used along the C-scan (slow scan)
direction and a few vessels were imaged as indicated by the red arrows. Figure 7(c) was
processed from 4 B-scan frames with 1024 A-lines per frame. The distance between adjacent
B-scan frames is 4 µm and the corresponding sampling density level is 3.65. The improved
sensitivity to blood flow is clear from these pictures. However, the obtained blood vessel
image suffers from shadowing effect, especially for large blood vessels. Shadowing effect
was seen from phase resolved DOCT and may be corrected with sophisticated numerical
method [20].
Fig. 7. OCT images of hamster skin. (a) OCT B-scan image; (b) IA-IBDV tomography image
and (c) IF-IBDV tomography image.
Laser irradiation was then performed on the subdermal side of the DSWC. Blood vessels
were irradiated with a sequence of three pulses from a frequency-doubled Nd:YAG laser
(Dualis VP+
, Fotona Laser, Ljublijana, Slovenia) with pulse width of 1 millisecond (ms),
energy density of 5 J/cm2
, repetition rate of 20 Hz, and beam sport size of 2 mm. To verify the
OCT IF-IBDV results, we also took transillumination images with a green color filter [Fig.
8(a)] and LSI blood flow map [Fig. 8(b)]. The red arrows in Fig. 8(a) indicate the laser

irradiation locations. After laser irradiation, constriction of the targeted blood vessels was
observed in the transillumination image, but blood-flow changes cannot be determined from
this image. Therefore, LSI was used to visualize blood flow dynamics in the window [21, 22].
During LSI, the window was imaged directly by a camera (Nuance, CaliperLife Sciences,
Hopkinton, MA) equipped with a 5 × zoom lens. A HeNe laser (633 nm, 30 mW) was used to
irradiate the epidermal side of the window chamber and the transmitted speckle pattern
imaged with the camera. An image exposure time of 10 milliseconds was used and a custom
software written in LabVIEW (Version 8.6, National Instruments, Austin, TX) and MATLAB
(The Mathworks,Natick, MA) was used to acquire and process image. A sequence of 10 raw
speckle images was collected. Each image was converted to a speckle contrast image using a
7 × 7 sliding window. As shown in Fig. 8(b), blood flow was absent in the irradiated vessels.
In comparison, the corresponding maximum intensity projection (MIP) OCT IF-IBDV image
[Fig. 8(c)] depicts a similar functional blood flow map to that shown in the LSI image. The
blood vessels without flow do not show up in both the IBDV image and LSI flow map. The
IBDV image shows better result in the capillary vessels region as indicated by circles. The
reason for this may be due to the exposure time (10 ms) used for the LSI image, which is too
short to allow LSI to have any appreciable sensitivity to capillary-level flow. With longer
exposure times (~1000 ms), capillary visibility for LSI increases considerably [23]. Note that
the OCT IF-IBDV image in Fig. 8(c) was extracted from a 3D OCT data volume containing
4096 frames with 1024 A-lines per frame.
Fig. 8. Comparison of LSI and IF-IBDV images of microvascular network in DSWC model,
after laser irradiation of select vessels. (a) With green filtering of broadband light used to
transilluminate the DSWC, the blood-vessel architecture is clearly visible. (b) With LSI,
vessels containing flowing blood are displayed, with an absence of flow in the irradiated
vessels. (c) With IF-IBDV method, a similar functional map of the vasculature is obtained,
with enhanced visualization of smaller arteriolar, venules, and capillaries. The scale bars in (a)
represent 1 mm.
7. In vivo imaging of microvascular network in human skin
Imaging of the microcirculation in human skin is important for diagnosing and monitoring
diseases of microvascular origin. There have been demonstrations of human skin imaging
with optical based techniques, such as laser Doppler imaging, LSI and photoacoustic imaging.
Although these techniques have high sensitivity to specific features of skin microvasculature,
they are limited by either spatial temporal resolution or lack of depth discrimination [17].
Recently, high resolution imaging of human skin with techniques based on OCT has been
demonstrated [17, 24].
To demonstrate the capability of the IF-IDBV method, measurements were taken from
human skin. Imaging was performed on a healthy adult male volunteer. The volunteer sat on a
chair and held the probe himself. The imaging region was on the thigh of the volunteer [Fig.

9(a)] and the size of this region is 4 × 4 mm2
. The 3D OCT data volume consisted of 4096
frames with 1024 A-lines per frame and took approximately 80 seconds to capture the entire
3D OCT data volume [Fig. 9(b)-9(g)]. Figures 9(b)~9(e) show the MIP views of the IF-IBDV
images at different depth. Figure 9(b) shows the MIP view at depths from 120 μm to 360 μm.
A dense capillary network was identified. With increasing depth, larger but less dense blood
vessels can be seen [Figs. 9(c), 9(d) and 9(e)]. Figures 9(c), 9(d) and 9(e) show, respectively,
the MIP view at the depths from 360 μm to 600 μm, 600 μm to 840 μm, and 840 μm to 1300
μm. The red arrows indicate the larger blood vessels detected in the corresponding images.
Clear blood vessels branches can also be found at deeper region as shown in Fig. 9(e). The red
circles in Fig. 9(b)-9(d) show the branch locations. It is difficult to differentiate the depth of
the blood vessels from the grey scale MIP view shown in Fig. 9(b)-9(e). A color encoded
depth (CED) MIP image can provide the depth information of the vessels with a single image
and it can also minimize the shadowing effect. The CED MIP images for depth from 120 μm
to 360 μm and depth from 360 μm to 1300 μm are shown in Fig. 9(f) and Fig. 9(g).
Fig. 9. In-vivo imaging of human thigh skin. (a) The photograph shows the imaging location
(white rectangle area) on the thigh of the volunteer. (b)-(e) MIP view IF-IBDV images of
microcirculation network at different depths of human skin with depth of: (b) 120 μm-360 μm;
(c) 360 μm-600 μm; (d) 600 μm-840 μm; (e) 840 μm-1.3 mm. The arrows indicate new blood
vessels detected in each image and the circles indicate new branches detected in the image. (f)
The CED MIP view IBDV image for the depth of 120 μm-360 μm. (g) The CED MIP view
IBDV image for the depth of 360 μm-1300 μm . The scale bars in (a) represent 1 mm.
To further demonstrate the performance of this method, we imaged the skin on a finger of
a healthy volunteer [region enclosed by white line in Fig. 10(a)]. During imaging, the OCT
probe was placed on a table and the forearm of the volunteer was also rested on the table.
Figure 10(b) shows an OCT B-scan image of the region. Figures 10(c)-10(e) show the MIP
view of the IF-IBDV images at different depths. Figure 10(b) shows one OCT image to
indicate the depth region for Fig. 10(c), 10(d) and 10(e). Figure 10(c) shows the MIP view at
the depths from 270 μm to 450 μm. At this depth, the blood vessels manifest as dots or short
lines instead of a connected network. Based on Fig. 10(b), we identify the imaged region in
Fig. 10(c) to be in the papillary dermis, near the dermal-epidermal junction. Hence, the dots or
short lines in Fig. 10(c) are the capillary loops. Figure 10(d) shows the projection view image
at the depths from 450 μm to 810 μm. At this depth, a dense capillary network is visible,
corresponding to the superficial horizontal plexus. In Fig. 10(e), several larger blood vessels

(indicated by the red arrows) are visible. The deepest blood vessels detected are at a depth of
approximately 1.2 mm. The CED MIP image for the depth from 270 μm to 1300 μm is shown
in Fig. 10(f).
Fig. 10. In-vivo imaging of human finger skin. (a) The photograph shows the imaging location
(area enclosed by white line) on the middle finger of the volunteer. (b) An OCT B-scan Image
with the blue bars indicating the corresponding depth region for Fig. 5(c)~5(e). (c), (d) and (e)
are the MIP view IF-IBDV images of microcirculation network at different depths with depths
of: (c) 270 μm-450 μm (d) 450 μm-810 μm (e) 810 μm-1300 μm. (f) CED MIP view IBDV
image for the depth of 270 μm-1300 μm. All scale bars in (b) and (c) represent 1 mm.
8. Conclusions
In summary, the features of the IBDV method are analyzed systemically with a flow phantom.
The effects of beam scanning density, flow rate and the time interval on the performance of
this method are investigated. It is found that the IBDV method can be used to quantify the
flow rate and its sensitivity can be improved by increase the time interval between the
neighboring A-lines. We have demonstrated high resolution blood vessel imaging with the IF-
IBDV method. By applying the algorithm along the slow scanning direction, the sensitivity of
this method to blood flow is greatly enhanced. With comparison to LSI maps of blood flow,
we demonstrate the ability of the IF-IBDV method to identify vessels with altered blood flow.
Furthermore, with clinical measurements, we demonstrate the ability of the IF-IBDV method
to image vascular networks with exquisite spatial resolution and at depths up to 1.2 mm in
skin. These results collectively demonstrate the potential of the IF-IBDV method to monitor
the microvasculature during disease progression and in response to therapeutic intervention.
Acknowledgments
The authors thank Ms. Wenjuan Qi for preparing the tissue phantom used in the experiment
and Mr. Danny Chou for proofreading the manuscript. This work was supported by the
National Institutes of Health (EB-10090, HL-105215, HL-103764, and RR-01192), Air Force
Office of Scientific Research (FA9550-04-0101), and the Arnold and Mabel Beckman
Foundation. Dr. Chen has a financial interest in OCT Medical Imaging Inc., which, however,
did not support this work.

High resolution imaging of microvasculature

More Related Content

What's hot (18)

Viewers also liked (20)

Similar to High resolution imaging of microvasculature (20)

More from Lucia Regina Cavalcanti (6)

Recently uploaded (20)

High resolution imaging of microvasculature