DEVELOPMENT OF A COMPACT X-RAY PIV SYSTEM AND

NEW X-RAY FLOW TRACERS

Visualization of Opaque Biofluid Flows using X-ray PIV System

Sang Joon Lee and Sung Yong Jung

Center for Biofluid and Biomimic Research, Department of Mechanical Engineering, Pohang University of Science and

Technology (POSTECH), San 31, Hyojadong, Pohang 790-784, South Korea

Keywords: Medical X-ray, PIV (Particle Image velocimetry), Microparticles, X-ray contrast.

Abstract: A compact X-ray particle image velocimetry (PIV) system employing a medical X-ray tube as a light source

was developed to obtain quantitative velocity field information of opaque flows. The X-ray PIV system

consists of a medical X-ray tube, an X-ray CCD camera, a programmable shutter for generating a pulse-type

X-ray beam, and a synchronization device. Through performance tests, the feasibility of the developed X-

ray PIV system as a flow measuring device was verified. In applying the developed system to biofluid

flows, the most important prerequisite is to develop suitable flow tracers which should be detected clearly

by the X-ray imaging system. Iopamidol was encapsulated into the poly(vinyl alcohol) (PVA)

microparticles to fabricate such flow tracers. The characteristics of the fabricated microparticles were

checked. With increasing the amount of crosslinker, the degree of crosslinking and the efficiency of the

Iopamidol encapsulation were increased. This compact X-ray PIV system is a unique and useful for

investigating various biofluid flows in laboratory experiments.

1 INTRODUCTION

Recently, abnormal blood flows such as formation

of recirculation flow or low wall shear stress have

been known to play a key role in the pathology of

atherosclerosis (Malek et al., 1999). For more

detailed and meaningful elucidation about these

hemodynamic phenomena related with vascular

diseases, it is definitely needed to get quantitative

hemodynamic information of blood flows related

with the vascular diseases with high spatial

resolution of tens micrometer and temporal

resolution in the order of millisecond.

Currently, various medical instruments such as

Doppler ultrasonography, MRI (Magnetic

Resonance Imaging) and X-ray angiography have

been used to measure the blood flow. However, each

method has intrinsic problems in diagnosing the

hemodynamic aspects of vascular diseases in more

detail. Doppler ultrasonography is one dimensional

point measurement tool and has relatively high

errors caused by its angle dependency. For the case

of MRI method, real time measuring is very limited

due to low temporal resolution and its spatial

resolution is not so good enough. X-ray angiography

provides morphological information of blood vessels

and blood velocity can be evaluated using

concentration difference of iodine contrast media

from consecutive digital images (Rhode et al., 2005).

Unfortunately, this method is also not suitable for

acquiring blood flow information inside blood

vessels due to low spatial resolution and it is also

limited to apply to flow in a short vessel. These

limitations of conventional clinical diagnosis

methods bring about a strong demand on a more

effective and advanced analysis tool which can

provide more detailed, quantitative hemodynamic

information of blood flows.

On the other hand, The particle image

velocimetry (PIV) method, which extracts velocity

vectors of tracer particles seeded in a flow by

obtaining their displacement, has been accepted as

one of the reliable velocity measurement techniques

with high spatial and temporal resolutions (Adrian,

1991). However, it is difficult for the PIV method to

visualize the blood flows inside blood vessels of

vascular diseases under in vivo condition, because

the conventional PIV techniques have used a visible

light source which cannot transmit an opaque blood.

To overcome these limitations on conventional PIV

Lee S. and Jung S. (2010).

DEVELOPMENT OF A COMPACT X-RAY PIV SYSTEM AND NEW X-RAY FLOW TRACERS - Visualization of Opaque Bioﬂuid Flows using X-ray PIV

System.

In Proceedings of the Third International Conference on Biomedical Electronics and Devices, pages 32-37

DOI: 10.5220/0002738100320037

 SciTePress

systems, a transmission-type light source such as X-

ray or ultrasonic wave should be used instead of

visible light.

An X-ray PIV technique using a synchrotron X-

ray source was developed (Lee and Kim, 2003).

Thereafter, several application studies have been

performed, such as blood flows (Kim and Lee,

2006), flow of micro bubbles (Lee and Kim, 2005),

glycerin flow (Im et al. 2007), sap flow in a bamboo

leaf (Lee and Kim, 2008), and 3D X-ray PIV

application (Fouras, 2008). The synchrotron X-ray

PIV system, however, is not easy to use for general-

purpose research works because synchrotron facility

is very limited to getting beam times and

synchrotron X-rays are usually dangerous for

biological samples due to its high flux. Therefore, in

the consideration of X-ray PIV applications, the

substitution of the synchrotron X-ray source with a

medical X-ray tube is a straightforward and natural

consequence.

As far as we surveyed, there has been no trial

conducted to combine the medical X-ray imaging

method and the PIV velocity field measurement

technique with seeding fine tracer particles in the

flow. In this study, we established a new compact X-

ray PIV system employing a conventional medical

X-ray tube, an X-ray CCD camera equipped with the

frame-straddling feature, and a newly developed

mechanical shutter for generating double pulses of

X-rays. This compact X-ray PIV system has a spatial

resolution of tens of micrometers and a temporal

resolution of several milliseconds.

In applying the developed system to blood flows,

the most important prerequisite is to develop suitable

flow tracers which should be detected clearly by the

X-ray imaging system. The tracers should be able to

follow properly the fluid flow to be measured. In

addition, they should be bio-compatible and can be

removed safely without any harm to the living

organism. In earlier X-ray PIV experiments, alumina

particles (Lee and Kim, 2003) and silver-coated

hollow glass beads (Fouras et al., 2007; Im et al.,

2007) were used to track the fluid flows. However

none of them is suitable for visualizing blood flows,

because their specific weight is relatively high and

they are not bio-compatible nor bio-degradable. The

speckle pattern of red blood cells was also used to

get their velocity information (Kim and Lee, 2006).

However, this method is not conducive to in-vivo

applications due to image blurring and noises

generated by the surrounding materials such as

tissues. Therefore, the in-vivo velocity field

measurements of opaque biofluid flows have not

been fully satisfactory so far.

Poly(vinyl alcohol) (PVA) is a carbon-

backboned polymer that is biodegradable under both

aerobic (Nord, 1936) and anaerobic (Matsumura et

al., 1993) conditions. In this point of view, the

grafting functional oligomers were incoporated into

PVA chain. The PVA chain has been tried to be

modified widely where the biodegradable vinyl

alcohol block plays an essential role in the

functional polymer composites (Matsumura et al.,

1988). Iopamidol (also known as Solutrast®) is a

nonionic, low-osmolar iodinated contrast agent

developed by Bracco. It has been commonly used in

clinical angiography to enhance the absorption

contrast. In this study, Iopamidol was encapsulated

into bio-compatible and bio-degradable PVA

microparticles. The Iopamidol encapsulated PVA

was templated by W/O (water-in-oil) emulsion. It

was specially designed as flow tracers for

quantitative X-ray imaging of biofluid flows. The

compact X-ray PIV system and new X-ray flow

tracers are expected to expand the application areas

of the X-ray PIV technique even to the in vivo

velocity field measurement of various blood flows.

2 COMPACT MEDICAL X-RAY

PIV SYSTEM

Figure 1 shows the schematic diagram of the

compact X-ray PIV system developed in this study.

In the compact X-ray PIV system, a conventional

medical X-ray tube was used as a light source. The

most important issue in PIV measurements is

acquiring two consecutive flow images within a

short time interval. For this, double pulses of X-rays

should be generated from an X-ray tube and then a

detector needs to record two consecutive X-ray

Figure 1: Schematic diagram of the established compact

X-ray PIV system.

DEVELOPMENT OF A COMPACT X-RAY PIV SYSTEM AND NEW X-RAY FLOW TRACERS - Visualization of

Opaque Biofluid Flows using X-ray PIV System

images with high spatial resolution and good

sensitivity.

The pixel size of most clinical X-ray detectors

used in medical radiography is larger than 50 μm. In

addition, the conventional X-ray detectors cannot

provide the frame-straddling feature indispensable

for PIV measurements. In this study, we purchased a

customized X-ray CCD camera of 4000

× 2672

pixels from Hamamatsu Co. The architecture of this

CCD camera is identical with that of a cooled CCD

camera, the exception is that a CsI scintillation

crystal of 100μm in thickness is adhered to the top of

the CCD sensor array. Due to high sensitivity of the

CsI scintillator, this X-ray CCD is well fitted to the

compact X-ray PIV system for acquiring X-ray

images suitable for PIV velocity field measurement.

Related to the medical X-ray source, because an

X-ray tube with a smaller focal spot size can provide

clearer particle images (Jenneson et al., 2003), we

employed an X-ray tube (Varian A272) with a focal

spot size of 0.3/0.6mm at 100/300 mA for 40~150

kVp. The maximum exposure time of the X-ray tube

(Varian A272) is 5sec at 200mA and 60kVp. The

total pulse length of X-ray beam fired from the X-

ray tube is an input parameter on the X-ray tube

controller (SMS-525). In this experiment, the total

pulse length was fixed at 100 ms. This long pulse

was chopped into two short X-ray pulses by using a

mechanical shutter newly developed in this study.

The actual pulse length of each short pulse was 5 ms

and the time interval between two X-ray pulses was

20 ms. This X-ray tube was synchronized through an

external triggering switch by using a photocoupler

chip and a transistor transistor logic (TTL) signal

from a delay generator.

The X-ray shutter system connected with a 4-

channel synchronizer plays a key role in

synchronizing the X-ray tube, the X-ray detector,

and the rotating disc of the shutter. The delay

generator takes the pulse signal of 1 pulse/rev from

the encoder of the DC motor and then the X-ray tube

takes two pulse signals of “X-ray ready” and “X-ray

shot” from the delay generator. The X-ray CCD

camera is synchronized by using the delay generator.

The timing sequence for synchronization of the

compact X-ray PIV system is mainly determined

from the disk geometry and motor rotation speed.

Finally, the X-ray PIV system can provide two

chopped X-ray pulses for recording two consecutive

X-ray images.

3 PERFORMANCE EVALUATION

To check the feasibility of the developed compact

X-ray PIV system, we measured velocity fields of a

flow of glycerin (

= 1.260 g/cm

) inside an opaque

tube of 4 mm in diameter. The width of an X-ray

pulse fired from the X-ray tube was 100 ms. This

pulse was chopped into two short X-ray pulses of 5

ms width with a time interval of 20 ms using the

developed X-ray shutter. The X-ray CCD camera

then recorded two X-ray images consecutively.

Tungsten oxide particles were used as tracer

particles. The particle has a relatively high density

(ρ

) of 7.2 g/cm

and a high attenuation coefficient

for X-rays. Since the size of the tracer particles is in

the range of 10~50 μm in diameter, it is appropriate

for recording images with the X-ray CCD camera.

Glycerin solution having a viscosity coefficient (μ

)

of 1.50 N·s/m

was used as a working fluid. The

free-falling speed of the tracer particles (U

) in the

glycerin fluid is in the range of 0.216~5.395 μm/s, a

negligibly small under the given experimental

condition. We tested a tube flow at two different

mean velocities of 1mm/s and 2mm/s. As the two

velocity fields obtained were substantially the same,

the experimental result measured at the mean

velocity of 2mm/s is only represented in this paper.

The size of interrogation area was 16

(H)

× 128

(V)

Figure 2: Mean velocity field of glycerin flow in a tube

measured by the compact X-ray PIV system.

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

Figure 3: Comparison of velocity profiles of the test flow

in a tube.

pixels in the consideration of dominant longitudinal

flow. The number of particles in each interrogation

window was about 4-8 on average. We obtained 25

pairs of X-ray images to get velocity field

information.

The images of tracer particles seeded into the

opaque tube are apparent. In addition, there is no

optical distortion in the region near the tube wall.

From the two consecutive X-ray images, quantitative

velocity field information was obtained by applying

a cross-correlation PIV algorithm. Figure 2

represents a typical mean velocity field obtained by

ensemble averaging 25 instantaneous velocity fields.

The glycerin flow in the circular tube seems to have

a parabolic velocity profile with a maximum value at

the tube center. Actually, it is nearly impossible for

other flow measurement techniques to obtain this

kind of quantitative velocity field information of a

flow inside an opaque conduit.

Figure 3 shows the velocity profile across the

tube diameter, extracted from the mean velocity

field. The solid dots are the experimental data

obtained and the parabolic line represents the

theoretical velocity profile. The experimental results

measured with the developed compact X-ray PIV

system agree well with the theoretical one. The

spatial resolution between two adjacent velocity

vectors is about 91 μm, which is much better than

that of the MRI method (Pahernik et al. 2001).

The small discrepancy between the two velocity

profiles seems to be attributed to the density

difference of the tracer particles and the working

fluid. In addition, the cohesion feature of tungsten

particles in glycerin also has some influence on the

velocity difference.

4 FABRICATION OF NEW X-RAY

FLOW TRACERS

In applying the compact X-ray PIV system to blood

flows, new microparticles were developed as

suitable flow tracers. The detailed flow chart of

particle synthesis procedure is schematically shown

in Fig. 4. 3 ml of oil soluble Span 80 (HLB=4.3) was

dissolved in 100 ml n-hexane (A). 41 wt%

Iopamidol stock solution and 16 wt% of PVA

aqueous solution were prepared overnight at the

room temperature and 70°C, respectively. 3 ml of

PVA aqueous solution (4.8 x10

-2

g/mL) and 7 ml of

Iopamidol stock solutions (4.8 x10

-1

g/mL) were

mixed (B). The oil phase (A) was added in drop

forms to the water phase (B) by stirring at 4400 rpm.

After stirring for one hour at the room temperature,

emulsion picture was taken using an optical

microscopy to check the size of the microreactor.

Controlled amount of glutaraldehyde (GA) was

introduced as a crosslinker with proper amount of

hydrochloric acid (0.04 ~ 0.16g) for activation. The

homogenizing process was performed for additional

6 hours at the room temperature with keeping the

same stirring rate. The emulsion formation (4400

rpm) was followed by the overnight stabilization. To

remove unreacted residues, the filtering and washing

process were repeated several times with n-hexane

and de-ionized (DI) water and then centrifuged to

get the settle-downed particles. The filtered particles

were de-moisturized in the drying oven at 60°C and

then placed in the desiccator until there is no further

variation in their mass.

The microdroplets of the fabricated water-in-oil

(W/O) emulsion in this study were several

micrometers in diameter and their size distribution

was relatively narrow. After the formation of stable

emulsion, the designed amount of crosslinker was

Figure 4: Schematic diagram of the microparticle

synthesis.

DEVELOPMENT OF A COMPACT X-RAY PIV SYSTEM AND NEW X-RAY FLOW TRACERS - Visualization of

Opaque Biofluid Flows using X-ray PIV System

Table 1: Specifications of the particles fabricated in this study.

Particle

Crosslinker

ratio (a)

Average

diameter [μm]

Degree of

swelling

[%]

Degree of

crosslinking

[%]

Amount of Iopamidol

EDS

[atomic %]

H NMR

[mole]

2.5 6.14 250 77 1.52 2.5

5 6.92 183 93 4.83 49.0

10 8.64 125 93 4.04 34.5

Empty

capsule

1 9.17 133 74 0 0

incorporated. Four hydroxyl groups of the PVA are

supposed to be connected by one glutaraldehyde

molecule. Considering the amount of hydroxyl

groups in the PVA, three different types of particle

sensors were fabricated by changing the amount of

glutaraldehyde (GA). For the case of particle #1,

0.00625 mole of GA was incorporated, which

corresponds to 2.5 times of the possible crosslinker

units in PVA. Likewise, 5 times (0.0125 mole) of

GA were added into the particle #2. For the particle

#3, 10 times of those (0.025 mole) were added,

corresponding to 0.1 mole of hydroxyl groups in the

PVA. Table 1 summarizes the specifications of each

particle fabricated in this study.

5 THE CHARACTERISTICS OF

TRACER PARTICLES

The particle size and their distributions were counted

and averaged from several images produced through

scanning electron microscopy (SEM). The average

diameter of the particles ranged from 6 to 9 μm, and

this value increased with the increase of the

crosslinker (GA) ratio from particles #1 to #3.

Therefore, the added crosslinker used in this study

contributed to the increased size rather than led to

the network shrinkage. The degree of crosslinking

was determined by the

H NMR. The chemical shift

of particles #1, #2, and #3 were compared with pure

Iopamidol, empty capsule, and uncrosslinked PVA.

The methylene group newly formed by crosslinking

increased along with an increase in crosslinker ratio

from particles #1 to #3, while the peaks in the

methylene groups in uncrosslinked PVA decreased

by that order. As summarized in Table 1, the

quantitative degree of crosslinking exhibited a value

of 77% in particle #1 and then reached the optimum

of 93% in particle #2; no further increase in particle

#3 was observed.

Meanwhile, the degree of swelling was measured

by dissolving the particles in deionized water until

there was no further increase in their mass according

to the relation, W = (W

− W

)/W

x100 [%], where

is the mass of the fully swollen particle, and W

that of the dried particle. As the added crosslinker

increased, the degree of swelling in DI water was

observed to show a decreasing trend. The degree of

maximum swelling reflects the equilibrium between

the dilution of the polymer chain into the solvent and

the retractive force produced by the crosslinked

junction points that serve to restrict further swelling.

Given that the chemical compositions of all the

particles are supposed to be similar (similar Flory-

Huggins interaction parameter, χ), the difference in

the degree of swelling seemed to be mainly caused

by the network structure. The particles became less

elastic from particles #1 to #3, reflecting more

densely crosslinked network structure by that order.

The amount of the encapsulated Iopamidol was

obtained by measuring the energy dispersive X-ray

spectroscopy (EDS) connected with SEM and then

averaging the value from the three randomly

Figure 5: Degree of crosslinking vs. swelling and

Iopamidol encapsulation according to the crosslinker ratio.

selected particles. Iopamidol encapsulation was also

determined by

H NMR where the chemical shift of

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

particles #1, #2, and #3 were compared with those of

the characteristic methyl group peaks in the pure

Iopamidol. In both methods, the efficiency of the

Iopamidol encapsulation increased dramatically in

particle #2 compared with particle #1 brought about

by the increase in the added crosslinker. The amount

of Iopamidol reached the maximum in particle #2

and decreased slightly in particle #3. Although the

degree of crosslinking detected by the

H NMR was

almost same in particles #2 and #3, the degree of

swelling in particle #2 was slightly higher. From this,

we can assume that particle #2 has a more flexible 3-

dimensional structure compared to particle #3,

which could lead to higher encapsulation of

Iopamidol in the former. The relation between the

degrees of crosslinking and swelling, as well as the

amount of Iopamidol encapsulations are graphically

summarized in Figure 5. Due to the presence of the

elastic force equilibrium between the retractive and

extended forces, the degrees of crosslinking and the

swelling exhibited opposite tendencies, while

Iopamidol reached the maximum amount at the

optimized encapsulation in particle #2.

6 CONCLUSIONS

The compact X-ray PIV system combining the

conventional X-ray radiography technique and PIV

velocity field measurement method was developed.

Through preliminary tests, the spatial and temporal

resolution of this system was found to be higher than

any conventional clinical instruments. In addition,

new X-ray flow tracers were fabricated by

encapsulating Iopamidol into bio-compatible

polymer PVA. The functional characteristics of the

fabricated microparticles were checked. With

increasing the amount of crosslinker, the degree of

crosslinking and the efficiency of the Iopamidol

encapsulation were increased. In near future, the

developed system would be employed usefully for

measuring in vivo velocity field information of blood

flows.

ACKNOWLEDGEMENTS

This work was supported by the Creative Research

Initiatives (Diagnosis of Biofluid Flow Phenomena

and Biomimic Research) of MEST/KOSEF.

REFERENCES

Adrian, R. J. (1991) “Particle-imaging techniques for

experimental fluid mechanics”, Annual Review of

Fluid Mechanics, Vol. 23, pp. 261-304.

Fouras, A., Dusting, J., Lewis, R. and Hourigan, K. (2007)

“Three-dimensional synchrotron X-ray particle image

velocimetry”, Journal of Applied Physics, Vol. 102,

064916.

Im, K.S., Fezzaa, K., Wang, Y. J., Liu, X., Wang, J. and

Lai, M. C. (2007) “Particle tracking velocimetry using

fast X-ray phase-contrast imaging”, Applied Physics

Letters, Vol. 90, 091919.

Jenneson, P. M., Gilboy, W. B., Morton, E. J. and Gregory,

P. J. (2003) “An-ray micro-tomography system

optimised for the low-dose study of living organisms”,

Applied radiation and isotopes, Vol. 58, pp. 177-181.

Kim, G. B. and Lee, S. J. (2006) “X-ray PIV

measurements of blood flows without tracer particles”,

Experiments in Fluids, Vol. 41, pp. 195-200.

Lee, S. J. and Kim, G. B. (2003) “X-ray Particle Image

Velocimetry for measuring quantitative flow

information inside opaque objects”, Journal of Applied

Physics, Vo. 94, pp. 3620-3623.

Lee, S. J. and Kim, S. (2005) “Simultaneous measurement

of size and velocity of microbubbles moving in an

opaque tube using an X-ray particle tracking

velocimetry technique”, Experiments in Fluids, Vol.

39, pp. 492-497.

Lee, S. J. and Kim, Y. M. (2008) “In vivo Visualization of

the Water-refilling Process in Xylem Vessels Using X-

ray Micro-imaging”, Annals of Botany, Vol. 101, pp.

595-602.

Malek, A., Alper, S. and Izumo, S. (1999) “Hemodynamic

shear stress and its role in atherosclerosis”, JAMA, Vol.

282, pp. 2035-2042.

Matsumura, S., Kurita, H. and Shimokobe, H. (1993)

“Anaerobic biodegradability of polyvinyl alcohol”,

Biotechnology Letters, Vol. 15, pp. 749-754.

Matsumura, S., Takahashi, J., Maeda, S. and Yoshikawa,

S. (1988) “Molecular design of biodegradable

functional polymers, 1. Poly(sodium vinyloxyacetate)”,

Die Makromolekulare Chemie, Rapid

Communications, Vol. 9, pp. 1-5.

Nord, F. F. (1936) “Dehydrogenation activity of Fusarium

lini B”, Naturwissenschaften, Vol. 24, 763.

Pahernik, S., Griebel, J., Botzlar, A., Gneiting, T., Brandl,

M., Dellian, M. and Goetz, A.E. (2001) “Quantitative

imaging of tumour blood flow by contrast-enhanced

magnetic resonance imaging”, British Journal of

Cancer, Vol. 85, pp. 1655-1663.

Rhode K.S., Lambrou T., Hawkes D.J. and Seifalian A.M.

(2005) “Novel approaches to the measurement of

arterial blood flow from dynamic digital x-ray images”,

IEEE TRANSACTIONS ON MEDICAL IMAGING,

Vol. 24, pp. 500-513.

DEVELOPMENT OF A COMPACT X-RAY PIV SYSTEM AND NEW X-RAY FLOW TRACERS - Visualization of

Opaque Biofluid Flows using X-ray PIV System