Coupling intravascular shear stress with endoluminal impedance to detect pre-atherosclerosis regions

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

COUPLING INTRAVASCULAR SHEAR STRESS WITH ENDOLUMINAL
IMPEDANCE TO DETECT PRE-ATHEROSCLEROSIS REGIONS

by

Juhyun Lee

A Thesis Presented to the
FACULTY OF THE USC VITERBI SCHOOL OF ENGINEERING
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(BIOMEDICAL ENGINEERING)

August 2012

Copyright 2012 Juhyun Lee

ii
Table of Contents

List of Figures iii
Abstract iv
1. Introduction 1
2. Methods 4
2.1. MEMS thermal sensor fabrication and packaging 4
2.2. MEMS sensor working theory and circuit design 6
2.3 Rabbit experiment and data acquisition 8
2.4 Electrochemical impedance spectroscopy measurement 9
2.5 Immunohistochemistry 11
2.6 Computational fluid dynamics simulation and analysis 11
2.7 Statistical analysis 12
3. Results 13
3.1 Properties and calibration of sensors 13
3.2 Intravascular shear stress in normal diet vs. high-fat diet 14
3.3 Simulation results 18
3.4 Intravascular ultrasound and histochemistry 20
3.5 EIS measurements between normal and atherosclerotic plaque 21
4. Discussions 23
References 28

iii
List of Figures

Figure 1 Fabrication process of the heat transfer based shear stress MEMS
sensor

5
Figure 2 Sensor diagram and picture 6
Figure 3 Angiography image 9
Figure 4 High-fat diet rabbit aorta sample 10
Figure 5 Properties and calibration of flexible polymer sensor 13
Figure 6 ND rabbit shear stress data comparison between baseline and after
8 week

15
Figure 7 HD rabbit shear stress data comparison between baseline and after
8 week

17
Figure 8 MEMS sensor captured distinct pulsatile shear stress profiles
along the aorta

19
Figure 9 IVUS and histology image to determine the plaque in endolumen 20
Figure 10 EIS measurements were performed at the different region 22

iv
Abstract

Coronary artery disease (CAD) is the leading cause of morbidity and mortality in the
United States. Hemodynamic parameters such as spatial and temporal variations of wall
shear stress are intimately linked with the development and progression of CAD.
However, the real-time detection of pre-atherosclerotic regions with minimally invasive
technology remains an unmet clinical challenge. Here, we sought to compare the changes
in hemodynamic parameters in New Zealand White (NZW) rabbits fed on a normal diet
(ND) and NZW rabbits fed a high-fat diet (HD). A convective heat transfer-based micro-
electromechanical system (MEMS) sensor was deployed for measurement of
intravascular hemodynamic parameters. Baseline measurements were taken from the
descending aortic arch, thoracic, abdominal, and proximal to the renal arteries in NZW
rabbits (n=4). Following 8 weeks of a high-fat (HD, n=2) and a normal diet (ND, n=2),
intravascular measurements were repeated from the similar regions where baseline
measurements were acquired. Experimental measurements were validated by the
computational fluid dynamics (CFD) simulation. To further assess endoluminal vascular
oxidative stress in terms of electrochemical impedance spectroscopy (EIS), we used
concentric bipolar microelectrodes to interrogate explanted aortas from the high-fat fed
rabbits. The hemodynamic assessment revealed an increase in time-averaged shear stress
by 23.53% and peak values by 27.67% in the descending aorta, respectively (n=10, p <
0.05). There were no significant changes in hemodynamic parameters when compared to
baseline measurements (n=10, p>0.05) in the normal dietfed rabbits. Also, frequency-
v
dependent EIS ranging from 10 kHz to 100 kHz was elevated in the shoulder and center
regions (n=3, p<0.05). The pre-atherosclerotic lesions in rabbits on high-fat diets were
further validated by intravascular ultrasound (IVUS) and immunohistochemistry for
oxidized LDL and foam cells. Our findings revealed that high-fat fed rabbits developed
an increase in intravascular shear stress (ISS) accompanied by endoluminal
atherosclerotic lesions and increased serum viscosity. Thus, our convective heat transfer
strategy using the flexible MEMS thermal sensors offered an entry point to assess ISS
with high temporal and spatial resolution; thereby, providing a new approach to predict
changes in ISS in the pre-atherosclerotic regions.

1
1. Introduction

Despite advances in diagnosis and therapy, atherosclerotic cardiovascular disease remains
the leading cause of morbidity and mortality in industrialized nations. US Health
expenditure is reaching greater than 17% of GNP. As the increasing numbers of baby
boomers join the rank of the rising geriatric population, the annual cost to treat heart
disease—including high blood pressure, coronary heart disease, heart failure, and
stroke—will triple by 2030, from $273 billion to $818 billion [16]. Thus, identifying
active plaques will identify high-risk patients for selective treatment; thereby, reducing
incidence of heart attacks and stroke.

The development of atherosclerotic plaques is a key factor to the clinical manifestation of
acute coronary syndromes and stroke [24]. Atherosclerosis poses a major public health
problem that kills 7.2 million people yearly and remains the number one killer worldwide
[5]. It is a chronic inflammatory disease that is characterized by the deposition of
oxidized Low Density Lipoprotein (Ox-LDL) into arterial walls, leading to thickening,
hardening, and eventual occlusion of the coronary arteries [8]. The arteries are exposed to
wall shear stress (WSS) caused by blood flow patterns, which is intimately linked with
the development of atherosclerosis [15]. Shear stress on the endothelium is closely
involved in the regulation of oxidative stress in the vascular system [1]. Increased
oxidative stress initiates and promotes an inflammatory response that promotes
atherogenesis [5, 10, 17, 22].

2
Atherosclerotic lesion development is influenced by a combination of chemical and
physical factors. Sun et al. reported an LDL transport model in which increased LDL
concentration increases the probability of initiating atherosclerosis [21]. Stone et al. has
further shown that restenosis occurs at the regions of low endothelial shear stress, and
endothelial shear stress is significantly related to atherosclerosis and outward vascular
remodeling [19]. However, others proposed that increasing endothelial shear stress plays
a key role in decreasing the probability of initiating atherosclerosis [19]. Thus, the
characteristics of shear stress modulate the initiation of atherosclerosis.

LDL particles transmigrate into the subendothelial layers where they undergo oxidation
to form ox-LDL and participate in vascular oxidative stress and inflammatory responses
[23]. Recruitment of additional circulating monocytes to the inflammatory regions
further promotes monocytes transmigration into subendothelial layers where they
transformed to macrophages and engulfed ox-LDL to become the foam cells. Further
deposition of ox-LDL and recruitment of monocytes promote active oxidative stress and
inflammatory responses, leading to atherosclerotic plaques prone to rupture [15].

In this context, the goal of our investigation was to predict pre-atherosclerotic regions by
coupling intravascular shear stress (ISS) with endoluminal impedance. Here, we
introduced a micro-electromechanical system (MEMS) thermal sensor to measure local
changes in ISS in hyperlipidemic NZW rabbits that have developed pre-atherosclerotic
lesions. The MEMS sensor operates based on convective heat transfer. Using a constant

3
temperature driving circuit [14, 22, 25], the sensing element of the thermal sensor was
maintained at constant temperature in response to pulsatile blood flow. The heat
dissipation from the sensor was calibrated to blood flow velocity, from which the
changes in voltage output was inferred as intravascular shear stress (ISS) [13]. Following
ISS measurements, the rabbits’ aortas were excised. Ex vivo study was performed to
measure endoluminal electrochemical impedance in rabbits fed on a high-fat diet versus
rabbits fed on a normal diet. The electrochemical impedance spectroscopy (EIS) signals
provided bulk electro-chemical compositional changes as measured by impedance
changes in the presence of oxidative stress (oxLDL, foam cells, etc.). In parallel, ISS
measurements were validated by CFD analytical solutions.

Overall, we observed increased time-averaging shear stresses, peak shear stresses, and
flow pattern variance in hyperlipidermic rabbits. Moreover, EIS measurements revealed
increased impedance around pre-atherosclerotic tissues and at plaque centers. Taken
together, our methodology offer a means to detect changes in local ISS and tissue
impedance associated with the pre-atherosclerotic lesions.
.

4
2. Method

2.1. MEMS Thermal Sensor Fabrication and Packaging

MEMS sensors were fabricated with Parylene C, Titanium (Ti), and Platinum (Pt) to
maintain biocompatibility. First, the 0.3μm

layer was deposited on the 1μm
sacrificial silicon layer through electron beam evaporation. The sensing element, made of
Ti/Pt layers with 0.06μm/0.015μm thickness, was also deposited on the silicon layer by
electron beam evaporation. Using a vacuum coating system (PDS2010, Specialty Coating
Systems, Inc., IN), the entire surface was coated with 9μm-thick Parylene C, then a 2μm
thin Gold/Chromium metal layer was deposited onto the electrode leads through electron
beam evaporation. Parylene C was deposited to completely insulate the sensing element.
The Parylene C layer was removed from the sensor with space on the sides to the end of
the thin Gold/Chromium metal wire. Finally the

was etched away with

dry
etching (Fig 1). The resulting sensor dimension was 320μm x 21μm and 4cm long [25].

5

Figure. 1. Fabrication process of the heat transfer based shear stress MEMS sensor. (a) 1 μm sacrificial layer
was deposited on SiO
2
substrate. (b) Sensing elements TI/Pt were deposited 0.06 μm and 0.015 μm respectively.
(c) 9 μm parylene C was deposited. (d) Cr/Au metal layers were deposited and patterned as electrode leads. (e)
To foam a MEMS sensor device, parylene C was deposited and patterned 12 μm. (f) Etching the underneath Si
sacrificial layer leading to the final device [25].

To package the sensor onto the coaxial wire, conductive epoxy (EPO-TEK H20E-175:
Epoxy Technology, MA) was used on the exposed tip of the thin Gold/Chromium metal
wire to communicate between the sensing element and the breadboard. The entire sensor
body was then anchored securely to a coaxial wire with biocompatible epoxy (EPO-TEK
301: Epoxy Technology, MA). Final sensor appearance after packaging is shown in Fig
2.

6

Figure 2. Sensor diagram and picture (a) MEMS sensor was packaged in the coaxial wire with conductive epoxy
and covered with biocompatible epoxy to prevent current leakage. Sensing element was placed 4cm down from
the sensor tip to be exposed in fully developed flow. (b) Titanium/Platinum (Ti/Pt) sensing element has a
dimension of 2 μm by 280 μm. The other ends of the electrode leads were the binding pads for connecting to the
coaxial wire. (d) Photograph of the sensor attached into the coaxial wire before completely being covered by
biocompatible epoxy.

2.2 MEMS Sensor Working Theory and Circuit Design

The thermal sensor operates by the heat transfer principle. The thermal sensing element is
exposed to the flow velocity boundary layer and the rate of heat loss from the thermal
element is dependent on the velocity profile. Thus, the change in resistivity of the sensor
is dependent on convective heat transfer, in which a circuit feedback loop maintains
constant temperature.

7
The circuit was designed to operate under a constant temperature (CT) mode. Based on
previous research, the CT circuit provides the fastest time response [14].The temperature
overheat ratio (α
T
) is defined as the temperature variations of the sensor over the average
temperature at a reference condition (T
0
) [1, 14, 101]:

(

)

, (1)
where T denotes the temperature of the sensor. The relation between resistance and
temperature overheat ratios is expressed as [1, 14, 101]:

(

)

(

), (2)
where α is the Temperature Coefficient of Resistance (TCR), R
0
is the average resistance
at a reference condition (T
0
), and R is the resistance of the sensor. The sensor was
calibrated in vitro by steady flow in a platinum-cared silicon tube. A 3% overhear ratio
was applied to stabilize the sensing element temperature.
Theoretically, shear stress values were calculated by:

, (3)
where τ
w
is the WSS, is the blood viscosity, and r is the dimensions of the flow
channel [2]. If the blood was treated as a non-Newtonian flow, the viscosity of the blood
as a function of shear rate was measured using a viscometer.

8
2.3 Rabbit Experiment and Data Acquisition

Under proof of the USC Institutional Animal Care, New Zealand White (NZW) rabbits
(n=4, ten weeks) were acquired from a local breeder (Irish Farms, Norco, CA) and
housed in the Heart Institute of Good Samaritan Hospital. After the quarantine period, the
baseline experiment was performed. The rabbits were anesthetized with 100U/kg of
ketamine (JHP Pharmaceuticals, LLC.), via the ear vein to prevent thrombosis when the
sensor was introduced into aorta. Ultrasound (Philips SONOS 5500 at 12 MHz) was used
to measure the flow velocity in the four aortic regions (aortic arch, thoracic, abdominal,
and renal bifurcation artery) to have an inlet boundary condition in the rabbit aorta
modeled by CFD. A Heparin-rinsed coaxial-wired sensor was then introduced into the
femoral artery and was passed through until it reached the aortic arch. In each region, we
obtained a voltage signal from the sensor and pulled it back to the lower level with
measuring shear stress in the thoracic aorta and abdominal aorta until the renal artery
bifurcation (Fig. 3).

Angiography (Philips BV-22HQ C-arm) allowed for fluoroscopic visualization sensor in
each aortic region with injection of contrast dye through the carotid arteries, (Fig. 4) and
blood viscosity was measured with a cone and plate viscometer (Brookfield Engineering,
Newhall, CA).

9

Figure 3. Angiography image (a) Rabbit aorta visualized by fluoroscopic angiography with contrast for
measurements of the aortic diameter. (b) Fluoroscopy further guided deployment of flexible intravascular
sensor to the ascending aorta and the pacing lead to the right atrium.

The constant temperature circuit was used for real-time voltage signal acquisition in the
aorta. The voltage across the sensing element was monitored by a LabVIEW-based data
acquisition system, including a data acquisition board (USB-6216 DAQ device, National
Instruments, Austin, TX) and a laptop computer (ThinkPad T61, Lenovo) equipped with
LabVIEW. The acquisition was sampled at 2,000 Hz. Wavelet decomposition and low-
pass filters were applied to remove the noise background, resulting in a signal-to-noise
ratio of 4.8.

2.4 Electrochemical Impedance Spectroscopy Eeasurement

Following ISS measurements in the rabbits, the animals were sacrificed and their aortas
excised. Each section (descending aorta, thoracic aorta, abdominal aorta, and renal artery)
was resected about 2cm in length. Specimens were open in a longitudinal direction
individually with their inner lumen faced up, then gross histology of atherosclerosis
region in HF were identified (Fig. 4)

10

Figure 4. High-fat diet rabbit aorta sample (a) After 8 weeks, HD rabbit aortic arch had atherosclerotic regions.
(b) Segment was opened longitudinally. Small atherosclerotic legions were present [24].

Specimens were maintained in phosphate buffer saline (PBS). Endoluminal EIS
measurements were performed at the healthy tissue region, atherosclerotic plaque
shoulder, and plaque center. The concentric bipolar microelectrodes (FHC Co., ME, USA)
consisted of recording, and reference electrodes were used to measure the impedance.
Electric impedance was measured as a function of frequency in response to AC currents
in the biological tissue. Electrical impedance (Z=R+iX
c
) consists of resistance (R) and
reactance (X
c
) [4]. Blood vessels have their own characteristic resistance, but show a
complex number of impedance as a function of frequency. Therefore, frequency
dependence of electric and dielectric behavior by recording impedance in different
frequency gave determination of tissue properties [11]. EIS measurements were
performed with putting the reference electrode immersed in the PBS solution and
recorded by using a Gamry Series G 300 potentiostat (Gemry Instruments, PA, USA)
installed in a Dell desktop computer. A micro-manipulator (World precision Instrument

11
Inc., FL, USA) held the concentric bipolar microelectrodes to avoid interference of
chance with depth of PBS solution and orientation of specimen contact. 10eV peak-to-
peak AC voltage was inputted and varied the frequency range from 10kHz to 100kHz.
The magnitudes and phases of the impedance were acquired at 20 data points per
frequency decade.

2.5 Immunohistochemistry

The aorta rings were cut from each of its aorta section, and immersed into 4%
papraformaldehyde for 24 hours. Samples were fixed in paraffin blocks and serial 5μm
sections of rabbit atheromas were cut (Sakura Finetek, Torrence, CA, USA). Standard
immunostaining was performed using biotinylated secondary antibodies and peroxide
straining. Diaminobenzidine (DAB) was used as a chromogen and each segment were
counterstained with hematoloxylin for visualization. Atherosclerotic regions were
identified (Leica, DM LB2 microscopy, Germany) and images were captured with a CCD
camera (Spot RT-KE, Diagnostic Instruments, MI, USA).

2.6 Computational Fluid Dynamics Simulation and Analysis

The CFD was developed to compare the analytical and experimental data. Three-
dimensional modeling of the healthy rabbit aortic geometries (aortic arch, thoracic,
abdominal, renal aorta) were reconstructed with Solidworks (Concord, Massachusetts,

12
USA) based on the fluoroscopic aorta diameter measurements. The aorta was assumed to
be a rigid tube to simplify simulation results. The inlet pulsatile blood velocity profiles
were obtained from the pulsed-wave Doppler velocity measurements and filtered with a
low-pass filter using MATLAB (Natick, MA, USA). The inlet length was elongated to
allow for fully-developed parabolic flow as the flow streamed down to the sensor. The
outlet boundary condition was determined from the mean arterial pressure of 80 mmHg.
Computational domain was solved with Solidworks Flow Simulation (Massachusetts,
USA). The simulation was solved with a 2 second finish condition and a time step size of
0.005s. Solidworks Flow Simulation was automatically set up to meet the convergence
condition with changing iterating numbers. Solidworks Flow Simulation reckoned the
convergence when the dispersion values from iteration to iteration were smaller than the
specified number. We set it up as automatic. The governing equations were solved by
assuming laminar, incompressible, and unsteady flow under the non-slip condition [5].

2.7 Statistical Analyses

Data was expressed as mean ± standard deviation. Statistical significance was performed
using student t-test, and p values < 0.05 were considered statistically significant.

13
3. Results

3.1 Properties and Calibration of Sensors

The resistance of the sensor at 37.8°C was about 1.189 kΩ with an overheat ratio (α
T
) of
0.046 and a TCR (α) of 0.84 x 10
-3
/°C under in vivo–like operating conditions (Fig 5A).
The sensor was most sensitive at 5 kHz with a gain at 8.72. (Fig 5B). In vitro calibration
of the sensor shows a linear relationship between flow rate and change in output voltage.
Using the in vitro calibration curve, the output voltage was converted to shear stress (Fig.
5C).

Figure 5. Properties and calibration of flexible polymer sensor. (a) Plot of sensing element resistance versus
temperature. A linear relationship was established over a temperature range from 18°C to 45°C with an
operating resistance of 1.189 kΩ at 37.8°C, a temperature overheat ratio (α
T
) of 0.046, and a thermal coefficient
resistance (α) of 0.84 x 10
-3
/°C. (b) The sensor harbors a maximum frequency response at 5 kHz with a gain of
8.72. (c) In vitro calibration of sensor demonstrates a linear relationship between flow rate and change in output
voltage.

14
3.2 Intravascular Shear Stress in Normal Diet vs. High-fat Diet

Initial baseline ISS measurements were taken in the descending aorta, thoracic aorta,
abdominal aorta, and near the renal arteries. Following 8 weeks of a normal diet (ND),
shear stress measurements were repeated and revealed no significant changes in shear
stress patterns per cardiac cycle across all regions (Fig. 7a-h). Furthermore, peak shear
stress and time-averaged ISS values remained unchanged in the individual regions of
aortas (n=10, p < 0.05). (Fig. 7i-j)

15

Figure 6. Intravascular shear stress (ISS) profile in ND fed rabbits. (a) ISS from baseline measurement of
descending aortic arch. (b) Shear stress after 8 weeks of descending aortic arch. (c) ISS from baseline
measurement of thoracic aorta. (d) ISS after 8 weeks of thoracic aorta. (e) ISS from baseline measurement of
abdominal aorta. (f) ISS after 8 weeks of abdominal aorta. (g) ISS from baseline measurement of renal artery. (h)
ISS after 8 weeks of renal artery. Blood flow provides consistent pattern. Center thick lines represent mean
shear stress from 10 samples. Upper envelope and lower envelope represent one standard deviation from mean
shear stress. (i) time-averaged ISS comparison between baseline measurement and after 8 weeks of ND. (j) Peak
ISS comparison between baseline measurement and after 8 weeks of ND.

16
In contrast, shear stress patterns were different from the baseline (Fig7a-f) after 8 weeks
of a high-fat diet (HF). Standard deviations of individual measurements were greater than
those of baseline measurements. ISS patterns were variable across cardiac cycles for all
measured regions. Peak shear stress and time-averaged ISS values were significantly
elevated compared to the baseline values. Peak ISS values increased 27.67%, 15.91%,
32.05%, 26.05% in the descending aortic arch, in thoracic aorta, in abdominal aorta, and
near renal artery, respectively. Also, time-averaged ISS values also increased 23.52%,
20.07%, 30.41%, and 38.22%, respectively (Fig7g-h).

17

Figure 7. Intravascular shear stress (ISS) profile in HD fed rabbits. (a) ISS from baseline measurement of
descending aortic arch. (b) Shear stress after 8 weeks of descending aortic arch. (c) ISS from baseline
measurement of thoracic aorta. (d) Shear stress after 8 weeks of thoracic aorta. (e) ISS from baseline
measurement of abdominal aorta. (f) ISS after 8 weeks of abdominal aorta. (g) Shear stress from baseline
measurement of renal artery. (h) ISS after 8 weeks of renal artery. Blood flow provides consistent pattern.
Center thick lines represent mean shear stress from 10 samples. Upper envelope and lower envelope represent
one standard deviation from mean ISS. (i) Time-averaged ISS comparison between baseline measurement and
after 8 weeks of HD. (j) Peak ISS comparison between baseline measurement and after 8 weeks of HD. * mark
represents values are significantly different from each other

18
3.3 Simulation Results

Experimental real-time ISS values were compared to the CFD simulation. Boundary
conditions were acquired from Doppler ultrasound and angiogram. The aorta was
reconstructed using fluoroscopic aorta diameter measurements and assumed the aorta to
be a rigid pipe. The WSS from simulation closely matched the experimental ISS data
from the MEMS sensor (Fig. 8). The computed WSS values differed from the time-
averaged shear stress and peak ISS by approximately 23% and 14%, respectively. Despite
the rigid tube assumption, the output signals from the MEMS sensors were in close
agreement with the experimental ISS.

19

Figure 8. The MEMS sensor captured distinct pulsatile ISS profiles along the aorta. The sensor was introduced
via the femoral cut-down approach into the abdominal, thoracic aorta, and aortic arch where measurements
were performed. The IVV shear stress is closely superimposed to the WSS from the CFD.

20
3.4 Intravascular Ultrasound and Histochemistry

The echogenicity analysis of individual aortic sections by the high-frequency IVUS
revealed pre-atherosclerotic lesions. These measurements were supported by the presence
of atherosclerotic lesions by the hematoxylin and eosin (H&E) staining (Fig. 9).
Atherosclerotic lesions were absent in the endoluminal surface of the normal diet fed
group, as reflected by the constant echogenicity and negative immunostaining for lipids
throughout the entire segments of aortas in the HD group.

Figure 9. IVUS and histology images to determine endoluminal atherosclerotic lesions. After 8 weeks with ND,
plaque was not formed inside of the lumen (a, b), however, atherosclerotic plaques were found in the HD group
after 8 weeks (c, d)

21
3.5 EIS Eeasurements Between Normal and Atherosclerotic Plaque

Frequency-dependent-EIS measurements between 10 kHz and 100 kHz were assessed
using the ex vivo aortas that harbored atherosclerotic lesions. At 94.9 kHz, the lowest
impedance was measured. Impedance values were inversely proportional to frequency as
affected by internal capacitance of EIS. Central regions of Atherosclerotic lesions
harbored the highest impedance that was nearly 2 fold higher than normal healthy regions
(Fig 9a). Moreover, the highest phase difference at 94.9 kHz occurred in the central
regions of the atherosclerotic lesions. The normal healthy region represented most near
ohmic resistance (Fig 9b). Measurements were performed multiple times in the
representative normal healthy regions, atherosclerotic shoulders, and atherosclerotic
centers to further corroborate statistical T-test (Fig 9c). In the three different frequency
ranges, P-values among the three difference regions were smaller than 0.05. Hence, the
EIS strategy offered to assess endoluminal impedance are significant.

22

Figure 10. EIS measurements were performed at different regions. (a) Impedance is the highest at the
atherosclerotic center, and shoulder has second highest impedance (b) Impedance phase angle. (c) Compared
different frequency range among atherosclerotic center, shoulder, and normal lesion. Shows highest gap at 94.88
kHz. All three frequency ranges have significantly different values. *: all three groups of values are significantly
different from each other.

23
4. Discussion

In this study we sought to measure local changes in ISS and tissue impedance associated
with pre-atherosclerotic plaques in a hyperlipidemic rabbit model. Using our MEMS
thermal sensors, we demonstrated an increase in time-averaged shear stress and peak
shear stress in multiple areas in the aorta. Experimental changes in time-averaged and
peak shear stress were validated by CFD analysis. To exclude contributions from lumen
elasticity, the aorta was assumed as a rigid tube [1, 26]. Using an inlet boundary
condition derived from a Doppler ultrasound and the reconstructed aorta geometries from
the fluoroscopic measurements, we showed that computed WSS profile closely
overlapped with the experimental measurements with a difference of 23% and 14.7% in
time-averaged and peak WSS respectively. The difference between experimental and
analytical data was within the range of acceptable experimental errors [26]. Using our
concentric bipolar microelectrodes, we demonstrated frequency-dependent changes in
EIS signals in the atherosclerotic lesions visualized by IVUS. There is a strong
correlation between an elevated ISS and an increase in EIS signals. Thus, the novel
aspects of our study are twofold: 1) ISS increased in the atherosclerotic regions in high-
fat diet fed NZE rabbits, and 2) increased ISS values were supported by an increase in
endoluminal EIS signals. These two observations provide a new insight into identifying
regions of vascular shear stress with relevance to unstable plaque.

24
While baseline ISS measurements were similar between normal and high-fat diet fed
animals, ISS was significantly increased after eight weeks of fat-fed diet. Numerous
factors could contribute to the changes in wall shear stress after 8 weeks of high-fat diet;
namely, changes in the elastic properties of vasculature, luminal remodeling, increased
blood viscosity, and presence of atherosclerotic lesions. The inward expansion of pre-
atherosclerotic lesions into the lumen influences convective heat transfer at the throat,
upstream and downstream of the stenosis [3]. Experimentally measured shear stress is
also directly dependent on blood viscosity as evidenced in Eqn. 3. Blood viscosity in the
hyperlipidemic rabbits was 30.5% higher than that of ND blood (4.23cP vs. 2.94cP). This
increase in blood viscosity is largely caused by a 138.67-fold increase in serum LDL
concentration of after 8 weeks of high-fat diet (26.2mg/dL to 3659.4mg/dL). In contrast,
LDL concentrations in the ND rabbits decreased by 67% after 8 weeks (from 52.mg/dL
to 15.8mg/dL). In addition to increased blood viscosity in HD rabbits, we observed a
significant increase in ISS in the thoracic aorta and near the renal arteries of HD rabbits.
Shear stress values are also dependent on lumen diameters. As the sensor was introduced
downstream through the rabbit aortas, experimentally measured ISS values increased.
Angiogram images revealed narrowing of the aorta as the sensor was advanced
downstream around the abdominal aorta and near the renal arteries. However, the
velocity profiles obtained by a Doppler ultrasound were similar for all four measured
regions. Shear stresses increased when measured further downstream from the
descending aorta due to narrowing diameters in the presence of unchanged velocity

25
profiles [23]. This analysis suggests that the presence of pre-atherosclerotic lesions
contributed to increased measured WSS [3, 22].

HD rabbits had increased variability in measured shear stress patterns when compared to
ND rabbits. The protruding geometry of pre-atherosclerotic plaques can lead to disturbed
(unstable) blood flow. Plaques can continue to grow if high LDL concentration is
maintained in the blood, leading to further narrowing of the lumen. This could induce
further turbulent flow and generate vortices to occur behind of the plaque [22, 27]. The
plaque obstruction also generates increased blood velocities over the head of the plaque,
increasing the chance of ruptures [2, 10]. Emerging research has suggested that disturbed
flow associated with tachycardia can contribute to atherosclerosis [9, 12, 18]. Therefore,
high variability in shear stress patterns in the aorta lead to greater blood flow disturbance
which can contribute to an increase in plaque as well.

EIS analysis revealed significantly higher impedance in pre-atherosclerotic regions where
ox-LDL and foam cells were present. Previously, Yu et al. determined that EIS
measurements were independent of tissue probing depth as supported by the notion that
current flow via the shortest distance from the central electrode and through the tissue to
the outer shell of concentric bipolar microelectrode [23, 24]. Therefore, EIS data
measurements were independent of lumen diameters and blood viscosity; the flow rate
provided that the microelectrodes were in contact with endoluminal surface. Increased

26
EIS measurements suggest the presence of increased electro-active chemicals such as ox-
LDL which were present in the pre-athersclerotic lesions.
The MEMS thermal sensor operates optimally in the center of the lumen and the
operational efficiently is limited by its lumen position [2]. However, the precise location
of the MEMS sensor in the artery cannot be determined. The signal strength is dependent
on the radial orientation of the sensor and signal strength is reduced when the sensor is
directed proximally to the lumen. The introduction of the sensor itself contributes to flow
disturbances around the sensor. Thus, the sensing element of the sensor was designed to
be sufficiently located behind the catheter tip to allow for fully developed parabolic flow
patterns. Previous research determined that the sensing element located 4cm further down
gives reasonable data measurements [1, 2, 23, 25]. The deployment of MEMS sensor in
humans presents different challenges. While the flow disturbance caused by the sensor is
smaller due to larger lumen to sensor diameter ratio, there is a greater likelihood that the
sensor is not located near the center of the lumen. Furthermore, the plaque geometry in
humans differs from the rabbit model. Thus, human plaques can induce flow disturbances
that alter expected shear stress patterns. Although we can use CFD to predict these
changes, we have yet experimentally test our sensor in humans.

In conclusion, our MEMS sensor provides reasonable shear stress measurements to
determine the approximation of pre-atherosclerotic plaques in an in vivo the NZW rabbit
model. MEMS sensor detected local ISS around pre-atherosclerotic plaques with
increased WSS and shear stress pattern variability. The EIS analysis identified changes in

27
tissue around pre-atherosclerotic plaque shoulders and center. Combining shear stress and
EIS analysis, these methods can identify active atherosclerotic plaques in a real-time
minimally invasive approach.

28
References
1. Ai L, Yu H, Dai W, Hale SL, Kloner RA, Hsiai TK, Real-time intravascular shear
stress in the rabbit abdominal aorta. IEEE Transaction on Biomedical Engineering
56(6):1755-1764, 2009.

2. Ai L, Yu H, Takabe W, Paraboschi A, Yu F, Kim ES, Li R, Hsiai TK, Optimization
of intravascular shear stress assessment in vivo. Journal of Biomechanics
42(10):1429-1437, 2009.

3. Ai L, Zhang L, Dai W, Hu C, Shung KK, Hsiai TK, Real-time assessment of flow
reversal in an eccentric arterial stenosis model. Journal of Biomechanics 43(14):2678-
83, 2010

4. Aroom, KR, Harting MT, Cox CS, Radharkrishnan RS, Smith C, Gill BS,
Bioimpedance analysis: a guide to simple design and implementation. The Journal of
Surgical Research 153(1):23-30, 2009 May 1.

5. Atherosclerosis, [Online], Available:
http://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0001224/

6. Berk, BC, Atheroprotective signaling mechanisms activated by steady laminar flow in
endothelial cells. Circulation 117(8):1082-1089, 2008.

7. Blackwelder, RF, Hot-wire and hot-film anemometers. Academic Press, 1981, pp.
259-314.

8. Cardiovascular Disease, [Online], Available:
http://www.who.int/mediacentre/factsheets/fs317/en/index.html

9. Craven TE, Ryu JE, Espeland MA, Kahl FR, McKinney WM, Toole JF, McMahan
MR, Thompson CJ, Heiss G, Crouse JR, Evolution of the association between carotid
artery atherosclerosis and coronary artery stenosis a case-control study. Circulation.
1990;82:1230-1242.

10. Davies PF, Remuzzi A, Gordon EJ, Dewey CF, Gimbrone MA, Turbulent fluid shear
stress induces vascular endothelial cell turnover in vitro. PNAS 83(7):2114-2117,
1986.

11. Foster KR, Schwan HP, Dielectric properties of tissues. In: Handbook of biological
effects of electromagnetic fields, pp. 25-102, 1996.

12. Guazzi M, Arena R, Endothelial dysfunction and pathophysiological correlates in
atrial fibrillation. Heart 95(2):102-6, 2009 Feb.

29

13. Hsiai, TK, Cho SK, Honda HM, et al. Endothelial cell dynamics under pulsating
flows: significance of high versus low shear stress slew rates (d τ/dt). Annals of
Biomedical Engineering 30(5):646-656, 2002.

14. Huang JB, Liu C, Jiang FK, Tung S, Tai YC, Ho CM, Fluidic shear-stress
measurement using surface-micromachined sensors, Microelectronics and VLSI, 16-
19, 1995.

15. Madamanchi, NR, Vendrov A, and Runge MS, Oxidative stress and vascular disease.
Arteriosclerosis, Thrombosis, and Vascular Biology 25(1):29-38, 2005.

16. National Vital Statistics Reports Volume 59 No. 4 March 16, 2011

17. Nerem, RM, Alexander RW, Chappell DC, Medford RM, Varner SE, Tylor WR, The
study of the influence of flow on vascular endothelial biology. American Journal of
the Medical Scienes 316(3):169-175, 1998.

18. Sakellarios AI, Siogkas PK, Tsakanikas VD, Stefanou KA, Michalis LK, Fotiadis
DI, Simulation of the effect of tachycardia on atherosclerotic plaque development
based on the LDL transport in coronary arteries. Computing in Cardiology, 329-332,
2010.

19. Stone PH, Coskun AU, Kinlay S, Clark ME, Sonka M, Wahle A, Ilegbusi
OJ, Yeghiazarians Y, Popma JJ, Orav J, Kuntz RE,Feldman CL, Effect of endothelial
shear stress on the progression of coronary artery disease, vascular remodeling, and
in-stent restenosis in humans: in vivo 6-month follow-up study. Circulation
108(4):438-444, 2003.

20. Sun N, Wood N, Hughes A, Thom S, Yun XX, Influence of Pulsatile Flow on LDL
Transport in the Arterial Wall, Annals of Biomedical Engineering 35:1782-1790,
2007.

21. Topper JN, Cai J, D. Falb, Gimbrone MA, Identification of vascular endothelial genes
differentially responsive to fluid mechanical stimuli: cyclooxygenase-2, manganese
superoxide dismutase, and endothelial cell nitric oxide synthase are selectively up-
regulated by steady laminar shear stress. PNAS 93(19):10417-10422, 1996.

22. Varghese SS, Frankel SH, Numerical modeling of pulsatile turbulent flow in stenotic
vessels, ASME Journal of Biomedical Engineering 125(4), 445-460, 2003

30
23. Yu F, Ai L, Dai W, Rozengurt N, Yu H, Hsiai TK, MEMS thermal sensors to detect
changes in heat transfer in the pre-atherosclerotic regions of fat-fed New Zealand
white rabbits, Annals of Biomedical Engineering 39(6), 1736-1744, 2011.

24. Yu F, Li R, Ai L, Edington C, Yu H, Barr M, Kim ES, Hsiai. TK, Electrochemical
Impedance Spectroscopy to Assess Vascular Oxidative Stress, Annuls of Biomedical
Engineering 39(1):287-96, 2011.

25. Yu F, Dai X, Beebe T, Hsiai, TK, Electrochemical impedance spectroscopy to
characterize inflammatory atherosclerotic plaques, Biosensors and Bioelectronics
15;30(1):165-73. 2011.

26. Yu H, Ai L, Rouhanizadeh M, Patel D, Kim ES, Hsiai TK, Flexible polymer sensors
for in vivo intravascular shear stress analysis. IEEE/ASME Journal of
Microelectromechanical Systems, 17(5) 1178-1186, 2008.

27. Zhang BX, Zhang Y, Numerical simulation of pulsatile turbulent flow in tapering
stenosed atteries, International journal of numerical methods for heat & fluid flow
19(5) 561-573, 2009.

Abstract (if available)

Abstract Coronary artery disease (CAD) is the leading cause of morbidity and mortality in the United States. Hemodynamic parameters such as spatial and temporal variations of wall shear stress are intimately linked with the development and progression of CAD. However, the real-time detection of pre-atherosclerotic regions with minimally invasive technology remains an unmet clinical challenge. Here, we sought to compare the changes in hemodynamic parameters in New Zealand White (NZW) rabbits fed on a normal diet (ND) and NZW rabbits fed a high-fat diet (HD). A convective heat transfer-based micro-electromechanical system (MEMS) sensor was deployed for measurement of intravascular hemodynamic parameters. Baseline measurements were taken from the descending aortic arch, thoracic, abdominal, and proximal to the renal arteries in NZW rabbits (n=4). Following 8 weeks of a high-fat (HD, n=2) and a normal diet (ND, n=2), intravascular measurements were repeated from the similar regions where baseline measurements were acquired. Experimental measurements were validated by the computational fluid dynamics (CFD) simulation. To further assess endoluminal vascular oxidative stress in terms of electrochemical impedance spectroscopy (EIS), we used concentric bipolar microelectrodes to interrogate explanted aortas from the high-fat fed rabbits. The hemodynamic assessment revealed an increase in time-averaged shear stress by 23.53% and peak values by 27.67% in the descending aorta, respectively (n=10, p < 0.05). There were no significant changes in hemodynamic parameters when compared to baseline measurements (n=10, p>0.05) in the normal dietfed rabbits. Also, frequency-dependent EIS ranging from 10 kHz to 100 kHz was elevated in the shoulder and center regions (n=3, p<0.05). The pre-atherosclerotic lesions in rabbits on high-fat diets were further validated by intravascular ultrasound (IVUS) and immunohistochemistry for oxidized LDL and foam cells. Our findings revealed that high-fat fed rabbits developed an increase in intravascular shear stress (ISS) accompanied by endoluminal atherosclerotic lesions and increased serum viscosity. Thus, our convective heat transfer strategy using the flexible MEMS thermal sensors offered an entry point to assess ISS with high temporal and spatial resolution

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Electrical signals to assess cardiovascular phenomenon

PDF

Interfacing fluid shear stress with vascular oxidative stress: application of nano and micro sensors

Asset Metadata

Creator Lee, Juhyun (author)

Core Title Coupling intravascular shear stress with endoluminal impedance to detect pre-atherosclerosis regions

School Viterbi School of Engineering

Degree Master of Science

Degree Program Biomedical Engineering

Publication Date 05/29/2012

Defense Date 04/10/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag atherosclerosis,computational fluid dynamics,electrochemical impedance spectroscopy,MEMS sensor,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hsiai, Tzung K. (committee chair), Shung, Kirk Koping (committee member), Yamashiro, Stanley M. (committee member)

Creator Email juhyun925lee@gmail.com,juhyunle@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-42057

Unique identifier UC11290051

Identifier usctheses-c3-42057 (legacy record id)

Legacy Identifier etd-LeeJuhyun-864.pdf

Dmrecord 42057

Document Type Thesis

Rights Lee, Juhyun

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA