journal of the mechanical behavior of biomedical materials 34 (2014) 217–230

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

On the necessity of modelling fluid–structure

interaction for stented coronary arteries

Claudio Chiastraa, Francesco Migliavaccaa,

Miguel Ángel Martı́nezb,c, Mauro Malvèb,c,d,n
a
Laboratory of Biological Structure Mechanics (LaBS), Department of Chemistry, Materials and Chemical Engineering
“Giulio Natta”, Politecnico di Milano, Italy
b
Aragón Institute of Engineering Research (I3A), Universidad de Zaragoza, C/María de Luna s/n, E-50018 Zaragoza,
Spain
c
Centro de Investigación Biomédica en Red-Bioingeniería Biomateriales y Nanomedicina (CIBER-BBN), C/Poeta Mariano
Esquillor s/n, E-50018 Zaragoza, Spain
d
Universidad Pública de Navarra, Departamento de Ingeniería Mecánica, Energética y de Materiales, Campus Arrosadía,
E-31006 Pamplona, Spain

art i cle i nfo ab st rac t

Article history: Although stenting is the most commonly performed procedure for the treatment of
Received 30 September 2013 coronary atherosclerotic lesions, in-stent restenosis (ISR) remains one of the most serious
Received in revised form clinical complications. An important stimulus to ISR is the altered hemodynamics with
23 January 2014 abnormal shear stresses on endothelial cells generated by the stent presence.
Accepted 5 February 2014 Computational fluid dynamics is a valid tool for studying the local hemodynamics of stented
Available online 12 February 2014 vessels, allowing the calculation of the wall shear stress (WSS), which is otherwise not directly

Keywords: possible to be measured in vivo. However, in these numerical simulations the arterial wall and

Fluid–structure interaction the stent are considered rigid and fixed, an assumption that may influence the WSS and flow

Computational fluid dynamics patterns. Therefore, the aim of this work is to perform fluid–structure interaction (FSI) analyses

Stent of a stented coronary artery in order to understand the effects of the wall compliance on the

Coronary artery hemodynamic quantities. Two different materials are considered for the stent: cobalt–

Wall shear stress chromium (CoCr) and poly-L-lactide (PLLA). The results of the FSI and the corresponding
rigid-wall models are compared, focusing in particular on the analysis of the WSS distribution.
Results showed similar trends in terms of instantaneous and time-averaged WSS between
compliant and rigid-wall cases. In particular, the difference of percentage area exposed to
TAWSS lower than 0.4 Pa between the CoCr FSI and the rigid-wall cases was about 1.5% while
between the PLLA cases 1.0%. The results indicate that, for idealized models of a stented
coronary artery, the rigid-wall assumption for fluid dynamic simulations appears adequate
when the aim of the study is the analysis of near-wall quantities like WSS.
& 2014 Elsevier Ltd. All rights reserved.

n
Corresponding author at: Public University of Navarra, Department of Mechanical Engineering, Energetics and Materials, Campus
Arrosadía, 31006 Pamplona, Spain. Tel.: þ34 948 16 9294; fax: þ34 948 16 9099.
E-mail address: [email protected] (M. Malvè).

http://dx.doi.org/10.1016/j.jmbbm.2014.02.009
1751-6161 & 2014 Elsevier Ltd. All rights reserved.
218 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

and cyclic bending of the coronary plaque on flow and plaque

1. Introduction stress/strain conditions. A similar study was recently con-
ducted by Asanuma et al. (2013) who investigated the dis-
Among the percutaneous interventions, stenting is the most tributions of shear stress and tissue stresses in the proximal
commonly performed procedure for the treatment of coron- segment of a stenotic human left anterior descending cor-
ary atherosclerotic lesions; however, serious clinical compli- onary artery (LCA). Torii et al. (2009) performed a FSI analysis
cations remain such as in-stent restenosis (ISR), which is the of a human stenotic RCA using physiological velocity and
reduction in lumen size following the stent implantation pressure waveforms to investigate the effects of wall com-
(Park et al., 2012). The primary process leading to ISR is pliance on hemodynamics. A comparison between a FSI and a
neointimal hyperplasia (NH) that consists in an excessive rigid-wall model was carried out showing noticeable differ-
growth of tissue in and around the implanted stent, resulting ences in instantaneous WSS profiles. Finally, Malvè et al.
in a decreased blood flow through the vessel. One of the (2012) made a comparison between the WSS distribution of a
stimuli to NH is the altered hemodynamics due to the stent compliant and a rigid-wall model of a human left main
presence that causes abnormal shear stresses on the coronary artery with its main branches. WSS distributions
endothelial cells (Wentzel et al., 2008). were substantially affected by the arterial wall compliance, in
To date, numerical models have been established as a particular considering the minimum and maximum values
valid tool for studying the local hemodynamics of stented of WSS.
vessels, allowing the calculation of the wall shear stress The aim of the present work is to perform FSI analyses of a
(WSS), which is otherwise difficult to be measured in vivo. stented coronary artery in order to understand the effects of
Many computational fluid dynamic (CFD) studies have been the wall compliance on hemodynamic quantities. Both a
proposed in the literature, considering idealized (Gundert bare-metal (cobalt–chromium – CoCr) and a polymeric (poly-
et al., 2013, 2012, Murphy and Boyle, 2010a, Pant et al., 2010) L-lactide – PLLA) stent are considered. The choice of the
or more complex image-based stented coronary models polymeric stent, with the same design of the metallic one,
(Chiastra et al., 2013, Ellwein et al., 2011, Gundert et al., is done because of the different stiffness of the device, which
2011). Despite providing a great deal of information, in these could produce bigger deformations with a great influence on
fluid dynamic models the arterial wall and the stent are the fluid dynamics. The results of the FSI and the correspond-
considered rigid and fixed, an assumption that may influence ing rigid-wall models are then compared, focusing in parti-
the WSS and flow pattern results. In the recent past the fluid– cular on the analysis of the WSS distribution.
structure interaction (FSI) approach has been applied to a
wide range of arterial problems including the aorta (Figueroa
et al., 2006), abdominal aortic aneurysms (Leung et al., 2006;
Scotti et al., 2008; Scotti and Finol, 2007; Tezduyar et al., 2007; 2. Materials and methods
Wolters et al., 2005), cerebral aneurysms (Gerbeau et al., 2005;
Takizawa et al., 2012, 2011; Tezduyar et al., 2007), carotid 2.1. Geometry
artery bifurcations (Filipovic et al., 2013; Gao et al., 2009;
Gerbeau et al., 2005; Lee et al., 2012; Perktold and Rappitsch, A geometrical model of a straight coronary artery and a
1995; Tezduyar et al., 2007), and anastomoses of bypass grafts typical open-cell stent were created using the CAD software
(Hofer et al., 1996; Leuprecht et al., 2002). The extensive study RHINOCEROS v.4.0 Evaluation (McNeel & Associates, Indiana-
of Tang and coworkers on human carotid atherosclerotic polis, IN, USA). The geometry of the artery was created with a
plaque (Tang et al., 2008, 2004, 2001, 1999, Teng et al., 2010, length of 20 mm, an internal diameter of 2.7 mm and an
Yang et al., 2011, 2010) deserves special mentioning. Indeed, arterial wall thickness of 0.9 mm. The stent model is char-
they provided structural and hemodynamics markers for the acterized by eight rings with a total length of 8.5 mm, an
initiation and development of atherosclerosis in carotid external diameter of 1.55 mm (uncrimped configuration) and
arteries using the FSI approach. In a further study Belzacq strut thickness of 90 μm.
et al. (2012) investigated the effect of the length, the stiffness, In order to obtain the geometrical model of a stented
and the severity of an asymmetric carotid atherosclerotic artery which is not based only on geometrical assumptions
plaque on the mechanical action of the blood flow by but also takes into account the deformation of the vessel
developing a parametric FSI model. caused by the stent deployment, the device was expanded
FSI simulations of coronary artery models have been inside the vessel through a structural analysis reaching the
performed only in a limited number of studies, without final diameter of 3 mm. The simulation was carried out by
considering the presence of stents. Koshiba et al. (2007) means of ABAQUS/Explicit (Dassault Systemes, Simulia Corp.,
simulated blood flow, arterial wall deformation and filtration RI, USA) following the method proposed in Gastaldi et al.
flow in the wall of a coronary artery with multiple bends. The (2010).
results of the FSI simulation were used to analyze the low The final geometrical configuration, after the elastic recoil,
density lipoproteins transport in the arterial lumen and wall was exported as a triangulated surface and used to create the
by solving the advection–diffusion–reaction equation. Yang fluid and solid domains for the subsequent FSI and CFD
et al. (2009) developed a FSI model of the middle segment of analyses (Morlacchi et al., 2011). An extension with a length
the human right coronary artery (RCA) with atherosclerotic of four diameters (10.8 mm) was added at both extremities of
plaques, based on intravascular ultrasound (IVUS) images, in the arterial model (Fig. 1a) in order to obtain developed flow
order to quantify the effects of anisotropic vessel properties near the region of the stent and avoid border effects due to
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 219

Fig. 1 – (A) Geometrical model of the straight coronary artery with a typical deployed open-cell stent. The geometry was
obtained through a structural simulation of stent expansion inside the vessel. Extensions with a length of four diameters
were added at the extremities of the model. The fluid domain is displayed in light gray while the solid domain in dark gray.
Dimensions are in millimeters. (B) Flow rate and pressure waveforms of a human LAD (Davies et al., 2006) applied at the inlet
and at the outlet cross-section of the model, respectively. Peak pressure is at t¼ 0.37 s while peak flow is at t ¼0.59 s. (C) Detail
of the tetrahedral grid obtained at the fluid-structure interface in the vicinity of the stent struts. The mesh is characterized by
smaller elements close to the struts. (D) Particular of the mesh of the solid (left) and the fluid (right) domain at one extremity of
the extensions (inlet cross-section). The solid domain is discretized by prismatic elements with triangular base in the inner
part and by a layer of prismatic elements with rectangular base in the external part. The dots aligned with the z-axis indicate
the constrains which were added in the extended regions to avoid the motion of the vessel outside its axis during the cardiac
cycle. For the set of nodes in x direction (orange dots), movement in y direction was not allowed while for the set of nodes in y
direction (white dots), movement in x direction was not allowed. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

the constrains in the inlet and outlet cross-sections, thus elements. This choice allowed us to create elements oriented
reducing their influence on the results. in the radial direction of the vessel, with an optimal quality
(determinant criteria close to 1). The total number of ele-
2.2. Numerical grids ments was also reduced with respect to a pure tetrahedral
grid. In particular, the mesh contained 899,710 elements
The geometry of the artery is complex in the vicinity of the (189,521 nodes). The final solid mesh is finer than that
stent struts making the compatibility between the fluid and obtained after an independence study performed on the
solid domains critical during a FSI simulation. To minimize artery without stent. In fact, the device adds complexity to
this problem, the two domains were discretized simulta- the geometry and needs to be correctly refined.
neously obtaining coincident nodes at the fluid–structure The Octree method was chosen for the creation of the
interface. tetrahedral meshes for both the domains. This method
The meshing software ANSYS ICEM CFD v.14.0 (ANSYS ensures refinement of the mesh where necessary, but main-
Inc., Canonsburg, PA, USA) was used. An unstructured tetra- tains larger elements where possible. For the solid domain a
hedral mesh was generated in the fluid domain using smaller full tetrahedral mesh was initially created, prisms were then
elements close to the struts (Fig. 1c) as previously done in created by extrusion of the surface mesh, and the resulting
Chiastra et al. (2012). The fluid grid, which was chosen after prisms were made conformal with the existing tetrahedral
an appropriate mesh independence study, had 1,121,130 volume mesh.
elements (222,476 nodes).
In the solid domain an unstructured hybrid tetrahedral
2.3. Material properties
and prismatic grid was created (Fig. 1d). In particular, the
stent and the inner part of the arterial wall were meshed with
The arterial wall was modeled as a hyperelastic incompres-
tetrahedral elements because of the complexity of the geo-
sible isotropic and homogeneous material. The following
metry and the necessity of obtaining coincident nodes
strain energy function was considered:
between the fluid and solid domains. The external part
of the arterial wall was meshed with a layer of prismatic W ¼ A½expðBðI1  3ÞÞ 1 þ UðJÞ ð2:1Þ
220 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

where I1 is the first invariant of the deviatoric right Cauchy- the nodes set in y direction (Fig. 1d – white dots), the
Green tensor C ¼ J  2=3 FT F, J ¼ det ðFÞ is the Jacobian, F is the movement in x direction was not allowed. These conditions
standard deformation gradient, U is the volumetric energy are widely used for FSI biomechanical studies of coronary
function and A and B are the material constants, which were arteries (Malvè et al., 2012; Torii et al., 2009), abdominal
set to 3.71 Pa and 140.2, respectively. This strain energy aneurisms (Scotti et al., 2008; Takizawa et al., 2011, Torii
function was first proposed by Demiray (1972), and later used et al., 2010, 2006) and carotid arteries (Malvè et al., 2014).
by Delfino et al. (1997) for a carotid artery bifurcation model The artery was initially pressurized by applying at the
and Rodriguez et al. (2008) for an abdominal aortic aneurysm inlet cross-section a ramp of velocity from 0 m/s to 0.138 m/s
model. The function parameters were fitted using the experi- and, at the outlet, a ramp of pressure from 0 mmHg to
mental tests of human coronary arteries obtained by 80 mmHg. Then, the velocity and pressure waveforms mea-
Carmines et al. (1991). This material modeling was success- sured by Davies et al. (2006) in a human left coronary artery
fully applied to a coronary artery in a previous study (Malvè were imposed (Fig. 1b) at the inlet and the outlet, respec-
et al., 2012). This model takes into account the effect of tively. The conditions were assumed to be uniform over the
the axial pre-stretch since the used experimental data also cross-section. The average flow rate was 45 mL/min and the
included it. The viscoelasticity, the active behavior of muscle duration of the cardiac cycle was 0.9 s.
fibers of the artery, and the intrinsic anisotropy, due to the The no-slip boundary condition was applied to the fluid-
preferential directions of collagen and muscle fibers, were structure interface (surface representing the endothelial wall
neglected. This assumption might alter the results obtained and the stent struts).
for the structural part in terms of stresses and strains, but
does not alter the overall compliance of the artery in 2.5. Numerical simulations
presence of the stent due to the fact that the material model
used for the artery fitted experimental data and provide the The following four analyses were performed: FSI and rigid-
same stiffness. Since the aim of the work is the analysis of wall simulation of the arterial model with the CoCr stent, and
the compliant artery and not the evaluation of the stresses FSI and rigid-wall simulation of the model with the PLLA
inside it, in first approximation the results have to be stent. All the simulations were carried out using the com-
considered acceptable. mercial finite element package ADINA v.8.7.3 (ADINA R&D,
The stent material was considered linear elastic, isotropic Inc., Watertown, MA, USA).
and homogeneous. Two different cases were analyzed main-
taining the same geometry: CoCr (Young's modulus¼ 233 GPa 2.5.1. FSI simulations
and Poisson's ratio¼ 0.35) and PLLA stent (Young's mod- Fully coupled FSI simulations were performed. The numerical
ulus¼900 MPa and Poisson's ratio¼0.30). approach uses the arbitrary Lagrangian–Eulerian (ALE) formu-
Relative displacements were not allowed between the lation (Bathe and Zhang, 2004; Bathe et al., 1999) for the fluid
stent and the arterial wall. In fact, the volume was meshed domain and a typical Lagrangian formulation of the solid
with coincident nodes at the interface between the stent domain (Bathe and Zhang, 2004; Donea et al., 1982).
and the arterial wall. Although non-physiological, this condi- Considering the moving reference velocity, the Navier–
tion allowed a significant reduction in the complexity of Stokes equation can be written as:
the model.
∂vF B
The blood was modeled as an incompressible Newtonian ρF þ ρF ððvF wÞU ∇ÞvF ¼  ∇p þ μ∇2 vF þ f F ð2:2Þ
∂t
fluid, with a density of 1060 kg/m3 and a dynamic viscosity of
where the term w is the moving mesh velocity vector, vF is
0.0035 Pa s (Rikhtegar et al., 2013). Since the Reynolds number B
the velocity vector, p is the pressure, f F is the body force per
based on the inlet diameter was 210 at peak of flow rate, an
unit volume, ρF is the fluid density, and μ is the dynamic
order of magnitude smaller than the Reynolds number for
viscosity. In the ALE formulation, ðvF  wÞ is the relative
transition to turbulence (2300) (Spurk and Aksel, 2008), the
velocity of the fluid with respect to the moving coordinate
flow was assumed to be laminar under unsteady conditions.
velocity.
The Womersley number was 1.95.
The governing equation of the solid domain is the follow-
ing momentum conservation equation:
2.4. Boundary and flow conditions
B
€S
∇ UrS þ f F ¼ ρS u ð2:3Þ
The extremities of the extensions of the solid model were B
where ρS is the solid density, rS is the solid stress tensor, f F is
constrained by preventing rigid-body axial and transaxial
the body force per unit volume and u € S is the local accelera-
motion. The constrains were applied far from the stented
tion of the solid. The domains described by Eqs. 2.2 and 2.3
region, at a distance longer than four diameters (Fig. 1a).
are then coupled through displacement compatibility and
Therefore, the influence of these conditions on the stented
traction equilibrium (Bathe and Zhang, 2004) with the follow-
region is very limited. In order to avoid the motion of the
ing equations:
vessel outside of its axis (z-axis) during the cardiac cycle,
constrains were added in x and y direction to sets of nodes uS ¼ uF ðx; y; zÞ AΓ Fwall \ Γ Swall ð2:4Þ
along the length of the extensions (Fig. 1d). In particular,
rS UnS þ rF UnF ¼ 0 ðx; y; zÞ AΓ Fwall \ Γ Swall ð2:5Þ
along the extensions, starting from the top and bottom
surfaces, for the nodes set in x direction (Fig. 1d – orange where Γ Fwall and Γ Swall are the boundaries of the fluid and solid
dots), the movement in y direction was not allowed while, for domains, respectively. Eq. 2.5 is an equilibrium condition
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 221

between the stresses acting in normal direction on both In Fig. 2 the comparison between the initial undeformed
domain boundaries Γ Fwall and Γ Swall . configuration and the diastolic configuration of the CoCr and
The governing equations were solved with the finite element PLLA models is presented.
method, which discretizes the computational domain into finite
elements that are interconnected by element nodal points. For 2.6. Results quantification
the structural model, a sparse matrix solver was used to solve
the system. The full Newton–Raphson method (Bathe, 2006a) For the FSI cases, the displacements of the fluid–structure
with a maximum of 500 iterations in each time step was chosen interface during the cardiac cycle were analyzed. Moreover,
as iteration scheme. four different cross-sections (S1, S2, S3, and S4) were chosen
The fluid domain employs special flow-condition-based- to evaluate the variation of area and the corresponding
interpolation (FCBI-C) tetrahedral elements (Bathe, 2006b). All diameter (Fig. 3). The area variation (App–Alp) expressed in
solution variables are defined in the center of the element percentage was calculated as:
and the coupling between the velocity and the pressure is
App  Alp
handled iteratively. FCBI-C elements require the segregated Area variation ½% ¼ 100 ð2:6Þ
Alp
method to solve the nonlinear equations. The sparse linear
equation solver based on Gaussian elimination was used. To where App is the area of the cross-section at peak pressure
solve the coupling between the fluid and the structural and Alp is the area of the cross-section at the minimum
models, the iterative method was chosen (Bathe, 2006b). In pressure during the cardiac cycle.
this computing method, the fluid and solid solution variables The diameter variation (dpp  dlp) expressed in percentage
are fully coupled. The fluid equations and the solid equations was calculated as:
are solved individually in succession, always using the latest dpp  dlp
information provided from another part of the coupled Diameter variation ½% ¼ 100 ð2:7Þ
dlp
system. This iteration is continued until convergence in the
solution of the coupled equations is reached. The maximum where dpp is the area of the cross-section at peak pressure
number of fluid-structure iterations for each time step was and dlp is the area of the cross-section at the minimum
set to 1000. The time step size was set to 0.001 s (900 time pressure during the cardiac cycle. Since cross-sections S1 and
steps were necessary for one cardiac cycle). S2 are not circular (Fig. 3) because of the presence of the stent
Simulations were performed in parallel on one node of a struts, the hydraulic diameter was calculated for these loca-
cluster with an Intel processor, 8 CPUs, with a CPU speed of tions as:
2268 MHz and a total memory of 24 Gb. One cardiac cycle 4A
dh ¼ ð2:8Þ
was modeled. In order to verify that one cardiac cycle was p
enough to guarantee correct results, the CoCr FSI model and where A is the area of the cross-section and p is its wetted
the corresponding rigid-wall model were investigated by perimeter.
simulating three cardiac cycles. No significant differences To compare the FSI and rigid-wall models, the wall shear
were found in the WSS results between the first and the stress (WSS) and the time-averaged WSS (TAWSS) were
third cardiac cycle. It has to be noted that the calculation considered. TAWSS is defined as:
time of the FSI simulation was around 500 h for one Z T
1
cardiac cycle. TAWSS ¼ j!
τ w jdt ð2:9Þ
T 0

2.5.2. Rigid-wall simulations where ! τ w is the WSS vector and T is the duration of one
The same settings chosen for the fluid domain of the cardiac cycle. Low values of WSS have been recognized as
FSI models were used. In order to obtain a more realistic critical for NH. Endothelial cells subjected to WSS lower
comparison between FSI and rigid-wall simulations, the than 0.4 Pa or oscillatory WSS are circular in shape
geometries of the CoCr and PLLA cases, pressurized at without any preferred flow alignment pattern (Malek
80 mmHg, were considered. More in detail, the structural et al., 1999). These cells, coupled with the blood stagna-
model of the stented artery, for both cases, PLLA and CoCr, tion usually observed in regions of low WSS, lead to
is initially not pressurized. For this reason, before a cardiac increased uptake of blood-borne particles to the arterial
cycle can be applied, the diastolic pressure has to be reached. wall, which is prevalent in atherosclerosis, as a result of
In order to start the heart cycle with identical geometrical increased residence time and increased permeability of
configurations and properly perform a comparison of hemo- the endothelial layer (Murphy and Boyle, 2012). An
dynamics variables between rigid-wall and FSI simulations, inverse relationship between WSS magnitude and the
the reference diastolic configuration of the artery was con- extent of ISR was found both by in vivo animal studies
sidered for the rigid-wall analyses. By means of the commer- (Carlier et al., 2003; LaDisa et al., 2005) and human
cial software MATLAB (MathWorks Inc. – Natick, MA, USA), studies (Gijsen et al., 2003; Papafaklis et al., 2009, 2007;
the deformed mesh obtained from the FSI simulation at the Sanmartín et al., 2006; Thury et al., 2002; Wentzel et al.,
beginning of the cardiac cycle was used to generate an input 2001).
grid for the CFD computations. This mesh allows that the WSS and TAWSS were analyzed in the region of the
performed simulations, rigid-wall and FSI analyses, could be fluid-structure interface that contains the stent (in the
compared taking into account only the compliance of the following, we refer to it as the ‘region of interest’ (ROI);
arterial wall due to the cardiac cycle. Fig. 3). In particular, the percentage area exposed to low
222 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

Fig. 2 – (A) Initial undeformed configuration of the FSI simulations, before the pressurization step. This configuration is
identical for both CoCr and PLLA FSI cases. (B) Diastolic CoCr configuration. (C) Diastolic PLLA configuration. In the
magnification boxes on the right, the tissue between two rings of the stent is shown. As indicated by the black arrows, in the
undeformed configuration the tissue is prolapsed between the struts toward the interior of the vessel while in the diastolic
configurations it is prolapsed toward the exterior for the effect of the blood pressure.

Fig. 3 – Location of the four cross-sections that were chosen to evaluate the variation of area and the corresponding diameter
during the cardiac cycle. The region of interest (ROI) for the analysis of the WSS (region of the fluid-structure interface that
contains the stent) is representated in light gray.

WSS was calculated in the ROI as the ratio between the visualized using histograms by displaying the amount
area exposed to WSS lower than 0.4 Pa and the total area of area of the ROI contained between specific intervals of
of the ROI. The area distribution of WSS was also the variable value (Murphy and Boyle, 2010b).
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 223

dilate. On the contrary, PLLA struts, with a Young's modulus

3. Results and discussion of two orders of magnitude lower than the CoCr, significantly
move; the displacement value is between 0.09 mm and
The displacements of the fluid–structure interface that occur 0.15 mm in all the stented region. The same behavior is
during the cardiac cycle for the two FSI models were first evident looking at the displacements on a line in the axial
analyzed. The maximum displacements take place at peak direction of the vessel for the two FSI cases (Fig. 4b).
pressure (t¼ 0.37 s, p¼ 139.6 mmHg); they are localized in the The variation of area and the corresponding diameter of
arterial wall portions immediately before and after the stent the four cross-sections chosen for the FSI models are reported
(Fig. 4a) and they are equal to 0.215 mm and 0.240 mm, in Table 1. These data confirms the rigid behavior of the CoCr
respectively for case FSI CoCr and FSI PLLA. A qualitative stent. The cross-sections characterized by the presence of the
comparison of the contour maps of displacement magnitude stent struts (S1 and S2) show lower area and diameter
shown in Fig. 4a highlights the different influence of the CoCr variations if compared to the cross-sections outside the
and PLLA stents on the arterial wall deformation: CoCr struts stented region (S3 and S4). In particular, S2, which is char-
do not move as a consequence of the blood pressure varia- acterized by the higher amount of struts, has the lowest area
tion; only the arterial wall portions within the stent cell and diameter variation, 0.47% and 0.11%, respectively. Cross-
sections S3 and S4 have values of area and diameter varia-
tions comparable to the FSI PLLA model, about 4.3% and 2.1%,
respectively. The FSI PLLA case shows a more uniform
behavior than the FSI CoCr case, with similar values of area
and diameter variations for all the analyzed cross-sections.
The values of diameter variation outside the stented
region (S3 and S4) can be compared to literature data. The
values are similar to those found by Malvè et al. (2012) and
Torii et al. (2009) with their FSI models of coronary arteries.
Moreover, the FSI simulations of the present study show
comparable values of diameter variation to the experimental
findings by Schaar et al. (2005) based on IVUS palpography
measurements. The values of area variation outside the
stented region are in accordance with those reported for
patients with coronary artery disease by Kelle et al. (2011)
(mean area change equal to 2.9 %, calculated using angio-
graphy, and to 5.3 %, using magnetic resonance imaging) and
by Peters et al. (1996) and Weissman et al. (1995) (range from
2.0 % to 5.9 %, calculated using IVUS).
As previously shown by experimental and computational
studies (Haluska et al., 2007; Tada and Tarbell, 2005), the temporal
variation of vessel diameters is in phase with the pressure.
In order to study the effects of the wall compliance on the
hemodynamics of a coronary stented artery model, the WSS
patterns were analyzed. In Fig. 5 the contour maps of TAWSS
along the fluid–structure interface for the FSI and rigid-wall cases
are depicted. Low TAWSS are located next to the stent struts. For
each repeating stent cell, the values of TAWSS increase from the
zones near the struts towards the center of each cell. These
findings are in agreement with several previous computational
studies (Balossino et al., 2008; Duraiswamy et al., 2008; Gundert
et al., 2013). A qualitative comparison of the contour maps of
both the CoCr and PLLA cases does not point out any significant
difference between the FSI and rigid-wall models. In Fig. 6 the
TAWSS on a line in the axial direction of the vessel is displayed
Fig. 4 – (A) Contour maps of displacement magnitude along for the CoCr and PLLA cases. The values of TAWSS are slightly
the fluid–structure interface of case FSI CoCr (top) and FSI higher in the FSI model in the portions of the fluid–structure
PLLA (bottom) at peak pressure (t ¼0.37 s, p¼ 139.6 mmHg), interface outside the stent, while they are similar in the vicinity
in the region of interest (ROI). (B) Displacements on a line in of the struts. This means that the stent makes the vessel wall
the axial direction of the vessel for the cases FSI CoCr and stiffer, thus reducing the WSS differences.
PLLA at peak pressure. The distance is normalized in the The percentage area exposed to TAWSS lower than 0.4 Pa
ROI. The reference line is shown on the top left image of the in the ROI (Table 2) is similar between the FSI and rigid-wall
vessel model. (For interpretation of the references to color in cases. In particular, the percentage difference between the
this figure legend, the reader is referred to the web version CoCr FSI and rigid-wall cases is equal to 1.57% while between
of this article.) the PLLA cases is 1.05%.
224 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

Table 1 – Variations of area and diameter of the four analyzed locations of FSI CoCr and FSI PLLA models.

Location Area variation [mm2] Area variation [%] Diameter variation [mm] Diameter variation [%]

FSI CoCr FSI PLLA FSI CoCr FSI PLLA FSI CoCr FSI PLLA FSI CoCr FSI PLLA

S1 0.802 3.948 0.86 4.07 0.013 0.071 0.41 2.12

S2 0.418 3.975 0.47 4.23 0.003 0.071 0.11 2.32
S3 3.051 3.370 4.32 4.42 0.064 0.068 2.14 2.19
S4 2.653 2.587 4.31 4.20 0.060 0.058 2.13 2.08

Fig. 5 – Contour maps of TAWSS along the fluid-structure interface: (left) CoCr and PLLA FSI models; (right) CoCr and PLLA
rigid-wall models. Low TAWSS are indicated in red. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)

Fig. 6 – TAWSS on a line in the axial direction of the vessel: (A) FSI and rigid-wall CoCr models; (B) FSI and rigid-wall PLLA
models. The distance is normalized in the region of interest (ROI). The reference line is shown on the top left image. In the
magnification area, an example of WSS peaks due to the presence of stent struts is displayed.

In Fig. 7 the area distribution of TAWSS in the ROI are wall models are indicated by symbols. Differences between
presented for the four analyzed cases. To better visualize the the bars are small: the maximum absolute difference for the
differences between the histograms, the bars with an abso- CoCr cases is 0.022 (interval 1.4–1.5 Pa) while for the PLLA
lute difference greater than 0.005 between the FSI and rigid- cases is 0.008 (interval 1.4–1.5 Pa).
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 225

In addition to the TAWSS, which is an average quantity, the wall approach, some necessary assumptions were adopted.
instantaneous WSS were studied. The time instants correspond- The velocity and pressure waveforms imposed as boundary
ing to peak pressure and peak flow rate (t¼0.59 s, Q¼ 91.1 mL/ conditions of the models were taken from in-vivo data (Davies
min) were chosen. In Figs. 8 and 9 the distributions of WSS in the et al., 2006). Therefore, flow conditions are quite reliable.
ROI are presented, respectively at peak pressure and flow rate. However, the heart motion, typical of the diastolic coronary
No significant differences between FSI and rigid-wall models can flow, was neglected in this work, isolating the effects of the
be detected, also considering the instantaneous quantities. pure coronary artery intravascular flow on the wall compliance.
The motion of coronary arteries can be described by overall
3.1. Limitations vessel translation, stretching, bending and twisting and to
a minor degree by radial expansion and axial torsion
Although this study proposes a novel aspect of the compar- (Ramaswamy et al., 2004). The effects on hemodynamics in
ison between a stented artery analyzed under FSI and rigid- presence of the heart motion are still a subject of study. Several
works have been proposed in the literature with conflicting
Table 2 – Percentage of area exposed to WSS lower than results. In particular, Zeng et al. (2003) studied the hemody-
0.4 Pa in the region of interest (ROI). namics of a RCA model under physiologically realistic cardiac-
induced motion. These authors concluded that the motion
Percentage area with WSSo0.4 Pa [%]
effects were small compared to flow pulsation effects.
CoCr PLLA Ramaswamy et al. (2004) found that arterial motion substan-
tially affects the hemodynamics in the LAD. Prosi et al. (2004)
Time instant FSI Rigid-wall FSI Rigid-wall considered a realistic curved model of the LAD with its first
t ¼ 0.37 s 67.68 66.41 70.96 69.85
diagonal branch by attaching it to the surface of a sphere with
t ¼ 0.59 s 12.58 12.79 16.10 15.68 time-varying radius based on experimental dynamic curvature
Time averaged 40.22 38.65 44.91 43.86 data. Their results showed that the effect of curvature dynamics
on the flow field were negligible. Theodorakakos et al. (2008)

Fig. 7 – TAWSS distributions: (left) CoCr and PLLA FSI models; (right) CoCr and PLLA rigid-wall models. Each bar of the
histograms represents the amount of normalized area with a defined range of TAWSS. Bar widths are 0.1 Pa. The bars with an
absolute difference greater than 0.005 between the FSI and rigid-wall models are indicated by the symbols “xx” and “þþ”,
respectively for the CoCr and PLLA cases.
226 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

Fig. 8 – WSS distributions at peak pressure (t¼ 0.37 s, p¼139.6 mmHg): (left) CoCr and PLLA FSI models; (right) CoCr and PLLA
rigid-wall models. Each bar of the histograms represents the amount of normalized area with a defined range of TAWSS. Bar
widths are 0.1 Pa. The bars with an absolute difference greater than 0.01 between the FSI and rigid-wall models are indicated
by the symbols “xx” and “þþ”, respectively for the CoCr and PLLA cases.

studied the effect of myocardial motion on the flow field and study. In fact, in the current version of the used software, the
WSS distribution of an imaged-based human LAD and its main influence of the aforementioned stresses cannot be taken
branches in the presence of a stenosis. Results indicated that into account. This aspect may over-estimate the computation
fluid dynamics was considerably affected by the pulsatile nature of the displacements in the arterial wall.
of the flow and myocardial motion had only a minor effect on Lastly, the initial structural simulation of stent deploy-
flow patterns within the arterial tree. Although the absolute ment was used exclusively to obtain the geometry for the
values of WSS were different, the WSS spatial distribution was subsequent FSI and rigid-wall fluid dynamic simulations.
very similar between the stationary and the moving coronary The stress state of the artery and the stent due to the
tree. Hasan et al. (2013) investigated the effects of cyclic motion expansion of the device were not considered in the FSI
(i.e. bending and stretching) on blood flow in a 3D model of a analyses. Therefore, the stress in the artery and in the
segment of the LAD, which was created on the basis of stent calculated in these analyses might be underestimated.
anatomical studies. Their results highlighted that, although the Even though this approximation may strongly affect the wall
motion of the coronary artery could significantly affect blood stresses and strains computation, the compliance, which was
particle trajectory, it had slight effect on velocity and WSS. evaluated in this study, is relatively unaffected, depending
The second assumption of this work is related to the basically only on the pressure field.
modeling of the arterial wall, which was considered as a
hyperelastic incompressible isotropic and homogeneous
material using the strain energy function originally proposed 4. Conclusions
by Demiray (1972). A more realistic model for the material
might be implemented taking into account the anisotropy A FSI model of an idealized straight stented coronary artery
related to collagen fiber dispersion in the tissue (Gasser et al., was created and compared to the corresponding rigid-wall
2006) and the three-layered characterization of the arterial model in order to understand the effects of the wall com-
wall (Holzapfel et al., 2005). Moreover, the initial stresses due pliance on the hemodynamics. The effect of two different
to the stent deployment were neglected in this preliminary stent materials, CoCr and PLLA, on the results was evaluated.
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 227

Fig. 9 – WSS distributions at peak flow rate (t¼ 0.59 s, Q¼91.1 mL/min): (left) CoCr and PLLA FSI models; (right) CoCr and PLLA
rigid-wall models. Each bar of the histograms represents the amount of normalized area with a defined range of TAWSS. Bar
widths are 0.1 Pa. The bars with an absolute difference greater than 0.01 between the FSI and rigid-wall models are indicated
by the symbols “xx” and “þþ”, respectively for the CoCr and PLLA cases.

Similar results were found in terms of TAWSS and instan- Engineering “Giulio Natta”, Politecnico di Milano, Italy) for
taneous WSS between compliant and rigid-wall cases: the the initial structural simulation of stent deployment.
contour maps of WSS and also the WSS profiles are qualita- The support of Dr. Jingping Long and Dr. Jianghui Chao
tively similar; the difference of percentage area exposed to of the Adina Support (Adina R&D, Watertown, MA, USA) is
TAWSS lower than 0.4 Pa in the stented region between the highly appreciated.
FSI and the rigid-wall cases is low (about 1.5% and 1.0%, Mauro Malvè and Miguel Angel Martínez are supported
respectively for the CoCr and PLLA cases). The comparison by the research project DPI2010-20746-C03-01 of the Spanish
between compliant and rigid-wall cases showed similar Ministry od Science and Tecnology.
results in terms of WSS although CoCr and PLLA FSI models Francesco Migliavacca and Claudio Chiastra are partially
deforms differently during the cardiac cycle, with higher supported by the project “RT3S- Real Time Simulation for Safer
values of displacement in the stented region for the PLLA vascular Stenting” funded by the European Commission under
FSI case. the 7th Framework Programme, GA FP7-2009-ICT-4-248801.
The results of the present work indicate that, for idealized
models of a stented coronary artery, the rigid-wall assumption
for fluid dynamic simulations appears adequate when the aim of
r e f e r e nc e s
the study is the analysis of near-wall quantities like WSS.

Asanuma, T., Higashikuni, Y., Yamashita, H., Nagai, R., Hisada, T.,
Sugiura, S., 2013. Discordance of the areas of peak wall
Acknowledgments shear stress and tissue stress in coronary artery plaques as
revealed by fluid-structure interaction finite element analysis: a
The authors would like to acknowledge Stefano Morlacchi, case study. Int. Heart J. 54, 54–58, http://dx.doi.org/10.1536/ihj.54.54.
PhD (Laboratory of Biological Structure Mechanics (LaBS), Balossino, R., Gervaso, F., Migliavacca, F., Dubini, G., 2008. Effects
Department of Chemistry, Materials and Chemical of different stent designs on local hemodynamics in stented
228 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

arteries. J. Biomech. 41, 1053–1061, http://dx.doi.org/10.1016/j. blood flow in three-dimensional deformable arteries. Comput.
jbiomech.2007.12.005. Methods Appl. Mech. Eng. 195, 5685–5706, http://dx.doi.org/
Bathe, K.J., 2006a. Theory and Modeling guide, vol. I: ADINA Solids 10.1016/j.cma.2005.11.011.
& Structures, ADINA R&D, Inc., Watertown, MA, USA. Filipovic, N., Teng, Z., Radovic, M., Saveljic, I., Fotiadis, D., Parodi, O.,
Bathe K.J., 2006b. Theory and Modeling guide, vol. II: ADINA CFD 2013. Computer simulation of three-dimensional plaque
& FSI, ADINA R&D, Inc., Watertown, MA, USA. formation and progression in the carotid artery. Med. Biol. Eng.
Bathe, K.J., Zhang, H, 2004. Finite element developments for Comput. 51, 607–616, http://dx.doi.org/10.1007/s11517-012-1031-4.
general fluid flows with structural interactions. Int. J. Numer. Gao, H., Long, Q., Graves, M., Gillard, J.H., Li, Z.Y., 2009. Carotid
Methods Eng. 60, 213–232, http://dx.doi.org/10.1002/nme.959. arterial plaque stress analysis using fluid-structure interactive
Bathe, K.J., Zhang, H., Ji, S., 1999. Finite element analysis of fluid simulation based on in-vivo magnetic resonance images of
flows fully coupled with structural interactions. Comput. four patients. J. Biomech. 42, 1416–1423, http://dx.doi.org/
Struct. 72, 1–16. 10.1016/j.jbiomech.2009.04.010.
Belzacq, T., Avril, S., Leriche, E., Delache, A., 2012. A numerical Gastaldi, D., Morlacchi, S., Nichetti, R., Capelli, C., Dubini, G.,
parametric study of the mechanical action of pulsatile blood Petrini, L., Migliavacca, F., 2010. Modelling of the provisional
flow onto axisymmetric stenosed arteries. Med. Eng. Phys. 34, side-branch stenting approach for the treatment of
1483–1495, http://dx.doi.org/10.1016/j.medengphy.2012.02.010. atherosclerotic coronary bifurcations: effects of stent
Carlier, S.G., van Damme, L.C., Blommerde, C.P., Wentzel, J.J., van positioning. Biomech. Model Mechanobiol. 9, 551–561, http:
Langehove, G., Verheye, S., Kockx, M.M., Knaapen, M.W., //dx.doi.org/
Cheng, C., Gijsen, F., Duncker, D.J., Stergiopulos, N., Slager, C.J., 10.1007/s10237-010-0196-8.
Serruys, P.W., Krams, R., 2003. Augmentation of wall shear Gasser, T.C., Ogden, R.W., Holzapfel, G.A., 2006. Hyperelastic
stress inhibits neointimal hyperplasia after stent modeling of arterial layers with distributed collagen fibre
implantation: inhibition through reduction of inflammation?. orientations. J. R. Soc. Interface 3, 15–35, http://dx.doi.org/
Circulation 107, 2741–2746, http://dx.doi.org/10.1161/01. 10.1098/rsif.2005.0073.
CIR.0000066914.95878.6D. Gerbeau, J.-F., Vidrascu, M., Frey, P., 2005. Fluid-structure
Carmines, D., McElhaney, J., Stack, R., 1991. A piece-wise interaction in blood flows on geometries based on medical
nonlinear elastic stress expression of human and pig imaging. Comput. Struct. 83, 155–165, http://dx.doi.org/
coronary-arteries tested in vitro. J. Biomech. 24, 899–906, http: 10.1016/j.compstruc.2004.03.083.
//dx.doi.org/ Gijsen, F.J., Oortman, R.M., Wentzel, J.J., Schuurbiers, J.C., Tanabe,
10.1016/0021-9290(91)90168-M. K., Degertekin, M., Ligthart, J.M., Thury, A., de Feyter, P.J.,
Chiastra, C., Morlacchi, S., Gallo, D., Morbiducci, U., Cárdenes, R., Serruys, P.W., Slager, C.J., 2003. Usefulness of shear stress
Larrabide, I., Migliavacca, F., 2013. Computational fluid pattern in predicting neointima distribution in sirolimus-
dynamic simulations of image-based stented coronary eluting stents in coronary arteries. Am. J. Cardiol. 92,
bifurcation models. J. R. Soc. Interface 10, 20130193, http://dx. 1325–1328, http://dx.doi.org/10.1016/j.amjcard.2003.08.017.
doi.org/10.1098/rsif.2013.0193. Gundert, T.J., Dholakia, R.J., McMahon, D., LaDisa Jr., J.F., 2013.
Chiastra, C., Morlacchi, S., Pereira, S., Dubini, G., Migliavacca, F., Computational fluid dynamics evaluation of equivalency in
2012. Fluid dynamics of stented coronary bifurcations studied hemodynamic alterations between Driver, Integrity, and
with a hybrid discretization method. Eur. J. Mech. – B/Fluids similar stents implanted into an idealized coronary artery.
35, 76–84, http://dx.doi.org/10.1016/j.euromechflu.2012.01.011. J. Med. Devices 7, 011004, http://dx.doi.org/10.1115/1.4023413.
Davies, J.E., Whinnett, Z.I., Francis, D.P., Manisty, C.H., Aguado- Gundert, T.J., Marsden, A.L., Yang, W., Ladisa, J.F., 2012.
Sierra, J., Willson, K., Foale, R.A., Malik, I.S., Hughes, A.D., Optimization of cardiovascular stent design using
Parker, K.H., Mayer, J., 2006. Evidence of dominant backward- computational fluid dynamics. J. Biomech. Eng. 134, 011002,
propagating suction wave responsible for diastolic coronary http://dx.doi.org/10.1115/1.4005542.
filling in humans, attenuated in left ventricular hypertrophy. Gundert, T.J., Shadden, S.C., Williams, A.R., Koo, B., Feinstein, J.A.,
Circulation 113, 1768–1778, http://dx.doi.org/10.1161/ LaDisa Jr, J.F., 2011. Rapid and computationally inexpensive
CIRCULATIONAHA.105.603050. method to virtually implant current and next-generation
Delfino, A., Stergiopulos, N., Moore Jr., J.E., Meister, J.-J., 1997. stents into subject-specific computational fluid dynamics
Residual strain effects on the stress field in a thick wall finite models. Ann. Biomed. Eng. 39, 1423–1436 http://dx.doi.org/10.
element model of the human carotid bifurcation. J. Biomech. 1007/s10439-010-0238-5.
30, 777–786, http://dx.doi.org/10.1016/S0021-9290(97)00025-0. Haluska, B.A., Jeffriess, L., Mottram, P.M., Carlier, S.G., Marwick, T.H.,
Demiray, H, 1972. A note on the elasticity of biological soft 2007. A new technique for assessing arterial pressure wave
tissues. J. Biomech. 5, 309–311, http://dx.doi.org/10.1016/0021- forms and central pressure with tissue doppler. Cardiovasc.
9290(72)90047-4. Ultrasound 5, 6, http://dx.doi.org/10.1186/1476-7120-5-6.
Donea, J., Giuliani, S., Halleux, J.P., 1982. An arbitrary Lagrangian– Hasan, M., Rubenstein, D., Yin, W., 2013. Effects of cyclic motion
Eulerian finite element method for transient dynamic fluid- on coronary blood flow. J. Biomech. Eng. 135 (12), 121002, http:
structure interaction. Comput. Methods Appl. Mech. Eng. 33, //dx.doi.org/10.1115/1.4025335.
689–723, http://dx.doi.org/10.1016/0045-7825(82)90128-1. Hofer, M., Rappitsch, G., Perktold, K., Trubel, W., Schima, H., 1996.
Duraiswamy, N., Cesar, J.M., Schoephoerster, R.T., Moore Jr., J.E., Numerical study of wall mechanics and fluid dynamics in
2008. Effects of stent geometry on local flow dynamics and end-to-side anastomoses and correlation to intimal
resulting platelet deposition in an in vitro model. Biorheology hyperplasia. J. Biomech. 29, 1297–1308, http://dx.doi.org/
45, 547–561, http://dx.doi.org/10.3233/BIR-2008-0497. 10.1016/0021-9290(96)00036-X.
Ellwein, L.M., Otake, H., Gundert, T.J., Koo, B., Shinke, T., Holzapfel, G.A., Sommer, G., Gasser, C.T., Regitnig, P., 2005.
Honda, Y., Shite, J., LaDisa Jr., J.F., 2011. Optical coherence Determination of layer-specific mechanical properties of
tomography for patient-specific 3D artery reconstruction and human coronary arteries with nonatherosclerotic intimal
evaluation of wall shear stress in a left circumflex coronary thickening and related constitutive modeling. Am. J. Physiol.
artery. Cardiovasc. Eng. Technol. 2, 212–227 http://dx.doi.org/ Heart Circ. Physiol. 289, H2048–H2058, http://dx.doi.org/
10.1007/s13239-011-0047-5. 10.1152/ajpheart.00934.2004.
Figueroa, C.A., Vignon-Clementel, I.E., Jansen, K.E., Hughes, T.J.R., Kelle, S., Hays, A., Hirsch, G., Gerstenblith, G., Miller, J., Steinberg,
Taylor, C.A., 2006. A coupled momentum method for modeling A., Schr, M., Texter, J., Wellnhofer, E., Weiss, R., Stuber, M.,
 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230 229

2011. Coronaryartery distensibility assessed by 3.0 Tesla Papafaklis, M.I., Katsouras, C.S., Theodorakis, P.E., Bourantas, C.V.,
coronary magnetic resonance imaging in subjects with and Fotiadis, D.I., Michalis, L.K., 2007. Coronary dilatation 10 weeks
without coronary artery disease. Am. J. Cardiol. 108, 491–497, after paclitaxel-eluting stent implantation. No role of shear
http://dx.doi.org/10.1016/j.amjcard.2011.03.078. stress in lumen enlargement?. Heart Vessels 22, 268–273, http:
Koshiba, N., Ando, J., Chen, X., Hisada, T., 2007. Multiphysics //dx.doi.org/
simulation of blood flow and LDL transport in a 10.1007/s00380-006-0970-9.
porohyperelastic arterial wall model. J. Biomech. Eng. 129, Park, S.-J., Kang, S.-J., Virmani, R., Nakano, M., Ueda, Y., 2012. In-
374–385, http://dx.doi.org/10.1115/1.2720914. stent neoatherosclerosis: a final common pathway of late
LaDisa Jr., J.F., Olson, L.E., Molthen, R.C., Hettrick, D.A., Pratt, P.F., stent failure. J. Am. Coll. Cardiol. 59, 2051–2057, http://dx.doi.
Hardel, M.D., Kersten, J.R., Warltier, D.C., Pagel, P.S., 2005. org/10.1016/j.jacc.2011.10.909.
Alterations in wall shear stress predict sites of neointimal Perktold, K., Rappitsch, G, 1995. Computer simulation of local
hyperplasia after stent implantation in rabbit iliac arteries. blood flow and vessel mechanics in a compliant carotid artery
Am. J. Physiol. Heart Circ. Physiol. 288, H2465–H2475, http: bifurcation model. J. Biomech. 28, 845–856.
//dx.doi.org/ Peters, R., Kok, W., Rijsterborgh, H., Van Dijk, M., Koch, K., Piek, J.,
10.1152/ajpheart.01107.200. David, G., Visser, C., 1996. Reproducibility of quantitative
Lee, S.H., Kang, S., Hur, N., Jeong, S.-K., 2012. A fluid-structure measurements from intracoronary ultrasound images.
interaction analysis on hemodynamics in carotid artery based beat-to-beat variability and influence of the cardiac cycle.
on patient-specific clinical data. J. Mech. Sci. Technol. 26, Eur. Heart J. 17, 1593–1599.
3821–3831, http://dx.doi.org/10.1007/s12206-012-1008-0. Prosi, M., Perktold, K., Ding, Z., Friedman, M.H., 2004. Influence of
Leuprecht, A., Perktold, K., Prosi, M., Berk, T., Trubel, W., Schima, curvature dynamics on pulsatile coronary artery flow in a
H., 2002. Numerical study of hemodynamics and wall realistic bifurcation model. J. Biomech. 37, 1767–1775, http:
mechanics in distal end-to-side anastomoses of bypass grafts. //dx.doi.org/
J. Biomech. 35, 225–236, http://dx.doi.org/10.1016/S0021-9290 10.1016/j.jbiomech.2004.01.021.
(01)00194-4. Ramaswamy, S.D., Vigmostad, S.C., Wahle, A., Lai, Y.-G., Olszeski,
Leung, J.H., Wright, A.R., Cheshire, N., Crane, J., Thom, S.A., M.E., Braddy, K.C., Brennan, T.M.H., Rossen, J.D., Sonka, M.,
Hughes, A.D., Xu, Y., 2006. Fluid structure interaction of Chandran, K.B., 2004. Fluid dynamic analysis in a human left
patient specific abdominal aortic aneurisms: a comparison anterior descending coronary artery with arterial motion.
with solid stress models. Biomed. Eng. Online 5, 33, http://dx. Ann. Biomed. Eng. 32, 1628–1641.
doi.org/10.1186/1475-925X-5-33. Rikhtegar, F., Pacheco, F., Wyss, C., Stok, K.S., Ge, H., Choo, R.J.,
Malek, A.M., Alper, S.L., Izumo, S., 1999. Hemodynamic shear Ferrari, A., Poulikakos, D., Müller, R., Kurtcuoglu, V., 2013.
stress and its role in atherosclerosis. J. Am. Med. Assoc. 282, Compound ex vivo and in silico method for hemodynamic
2035–2042, http://dx.doi.org/10.1001/jama.282.21.2035. analysis of stented arteries. PLoS One 8, e58147, http://dx.doi.
Malvè, M., Chandra, S., Garcı́a, A., Mena, A., Martı́nez, M.A., Finol, org/10.1371/journal.pone.0058147.
E.A., Doblaré, M., 2014. Impedance-based outflow boundary Rodriguez, J., Ruiz, C., Doblaré, M., Holzapfel, G., 2008. Mechanical
conditions for human carotid haemodynamics. Comput. stresses in abdominal aortic aneurysm: influence of diameter,
Methods Biomech. Biomed. Eng. http://dx.doi.org/10.1080/ asymmetry and material anisotropy. J. Biomech. Eng. 130,
10255842.2012.744396. 1–10, http://dx.doi.org/10.1115/1.2898830.
Malvè, M., Garcı́a, A., Ohayon, J., Martı́nez, M.A., 2012. Blood flow Sanmartı́n, M., Goicolea, J., Garcı́a, C., Garcı́a, J., Crespo, A.,
and mass transfer of a human left coronary artery bifurcation: Rodrı́guez, J., Goicolea, J.M., 2006. Influence of shear stress on
FSI vs. CFD. Int. Commun. Heat Mass 39, 745–751, http://dx. in-stent restenosis: in vivo study using 3D reconstruction and
doi.org/10.1016/j.i.cheatmasstransfer.2012.04.009. computational fluid dynamics. Rev. Esp. Cardiol. 59, 20–27,
Morlacchi, S., Chiastra, C., Gastaldi, D., Pennati, G., Dubini, G., http://dx.doi.org/10.1157/13083645.
Migliavacca, F., 2011. Sequential structural and fluid dynamic Scotti, C.M., Jimenez, J., Muluk, S.C., Finol, E.A., 2008. Wall stress
numerical simulations of a stented bifurcated coronary artery. and flow dynamics in abdominal aortic aneurysms: Finite
J. Biomech. Eng. 133, 121010, http://dx.doi.org/10.1115/ element analysis vs. fluid-structure interaction. Comput.
1.4005476. Methods Biomech. Biomed. Eng 11, 301–322, http://dx.doi.org/
Murphy, E.A., Boyle, F.J, 2012. Reducing in-stent restenosis 10.1080/10255840701827412.
through novel stent flow field augmentation. Cardiovasc. Eng. Scotti, C.M., Finol, E.A., 2007. Compliant biomechanics of
Technol. 3, 353–373, http://dx.doi.org/10.1007/s13239-012- abdominal aortic aneurysms: a fluid-structure interaction
0109-3. study. Comput. Struct. 85, 1097–1113, http://dx.doi.org/
Murphy, J.B., Boyle, F.J., 2010a. Predicting neointimal hyperplasia 10.1016/j.compstruc.2006.08.041.
in stented arteries using time-dependant computational fluid Schaar, J.A., De Korte, C.L., Mastik, F., Van Damme, L.C.A., Krams,
dynamics: a review. Comput. Biol. Med. 40, 408–418, http://dx. R., Serruys, P.W., Van Der Steen, A.F.W., 2005. Three-
doi.org/10.1016/j.compbiomed.2010.02.005. dimensional palpography of human coronary arteries: Ex vivo
Murphy, J.B., Boyle, F.J., 2010b. A numerical methodology to fully validation and in-patient evaluation. Hertz 30, 125–133, http:
elucidate the altered wall shear stress in a stented coronary //dx.doi.org/
artery. Cardiovasc. Eng. Technol. 1, 256–268, http://dx.doi.org/ 10.1007/s00059-005-2642-4.
10.1007/s13239-010-0028-0). Spurk, J.H., Aksel, N, 2008. Fluid Mechanics, second ed. Springer-
Pant, S., Bressloff, N.W., Forrester, A.I.J., Curzen, N., 2010. Verlag, Berlin Heidelberghttp://dx.doi.org/10.1007/978-3-540-
The influence of strut-connectors in stented vessels: a 73537-3.
comparison of pulsatile flow through five coronary stents. Tada, S., Tarbell, J.M., 2005. A computational study of flow in a
Ann. Biomed. Eng. 38, 1893–1907, http://dx.doi.org/10.1007/ compliant carotid bifurcation-stress phase angle correlation
s10439-010-9962-0. with shear stress. Ann. Biomed. Eng. 33, 1202–1212, http://dx.
Papafaklis, M.I., Bourantas, C.V., Theodorakis, P.E., Katsouras, C.S., doi.org/10.1007/s10439-005-5630-1.
Fotiadis, D.I., Michalis, L.K., 2009. Relationship of shear stress Takizawa, K., Brummer, T., Tezduyar, T.E., Chen, P.R., 2012. A
with in-stent restenosis: bare metal stenting and the effect comparative study based on patient-specific fluid-structure
of brachytherapy. Int. J. Cardiol. 134, 25–32, http://dx.doi.org/ interaction modeling of cerebral aneurysms. J. Appl. Mech. 79,
10.1016/j.ijcard.2008.02.006. 010908, http://dx.doi.org/10.1115/1.4005071.
230 journal of the mechanical behavior of biomedical materials 34 (2014) 217 –230

Takizawa, K., Moorman, C., Wright, S., Purdue, J., McPhail, T., conditions with high and normal blood pressures. Comput.
Chen, P.R., Warren, J., Tezduyar, T.E., 2011. Patient-specific Mech. 38, 482–490, http://dx.doi.org/10.1007/s00466-006-0065-
arterial fluid-structure interaction modeling of cerebral 6al. 2010.
aneurysms. Int. J. Numer. Meth. Fluids 65, 308–323, http://dx. Torii, R., Wood, N.B., Hadjiloizou, N., Dowsey, A.W., Wright, A.R.,
doi.org/10.1002/fld.2360. Hughes, A.D., Davies, J., Francis, D.P., Mayet, J., Yang, G.-Z.,
Tang, D., Yang, C., Mondal, S., Liu, F., Canton, G., Hatsukami, T.S., Thom, S.A.McG., Xu, X.Y., 2009. Fluid-structure interaction
Yuan, C., 2008. A negative correlation between human carotid analysis of a patient-specific right coronary artery with
atherosclerotic plaque progression and plaque wall stress: In physiological velocity and pressure waveforms. Commun.
vivo MRI-based 2D/3D FSI models. J. Biomech. 41, 727–736, Numer. Meth. Eng. 25, 565–580, http://dx.doi.org/10.1002/
http://dx.doi.org/10.1016/j.jbiomech.2007.11.026. cnm.1231.
Tang, D., Yang, C., Kobayashi, S., Ku, D.N., 2004. Effect of a lipid Weissman, N., Palacios, I., Weyman, A., 1995. Dynamic expansion
pool on stress/strain distributions in stenotic arteries: 3-D of the coronary arteries: implications for intravascular
fluid-structure interactions (FSI) models. J. Biomech. Eng. 126, ultrasound measurements. Am. Heart J. 130, 46–51.
363–370, http://dx.doi.org/10.1115/1.1762898. Wentzel, J.J., Krams, R., Schuurbiers, J.C.H., Oomen, J.A., Kloet, J.,
Tang, D., Yang, C., Kobayashi, S., Ku, D.N., 2001. Steady flow and Van Der Giessen, W.J., Serruys, P.W., Slager, C.J., 2001.
wall compression in stenotic arteries: a three-dimensional Relationship between neointimal thickness and shear stress
thick-wall model with fluid-wall interactions. J. Biomech. Eng. after wallstent implantation in human coronary arteries.
123, 548–557, http://dx.doi.org/10.1115/1.1406036. Circulation 103, 1740–1745, http://dx.doi.org/10.1161/01.
Tang, D., Yang, C., Ku, D.N., 1999. A 3-D thin wall model with CIR.103.13.1740.
fluid-structure interactions for blood flow in carotid arteries Wentzel, J.J., Gijsen, F.J.H., Schuurbiers, J.C.H., van der Steen, A.F.
with symmetric and asymmetric stenosis. Comput. Struct. 72, W., Serruys, P.W., 2008. The influence of shear stress on in-
357–377, http://dx.doi.org/10.1016/S0045-7949(99)00019-X. stent restenosis and thrombosis. EuroIntervention 4 (Supl. C),
Teng, Z., Canton, G., Yuan, C., Ferguson, M., Yang, C., Huang, X., C27–C32.
Zheng, J., Woodard, P.K., Tang, D., 2010. 3D critical plaque wall Wolters, B.J.B.M., Rutten, M.C.M., Schurink, G.W.H., Kose, U., De
stress is a better predictor of carotid plaque rupture sites than Hart, J., Van De Vosse, F.N., 2005. A patient-specific
flow shear stress: an in vivo MRI-based 3D FSI study. J. computational model of fluid–structure interaction in
Biomech. Eng. 132, 031007, http://dx.doi.org/10.1115/ abdominal aortic aneurysms. Med. Eng. Phys. 27, 871–883,
1.4001028. http://dx.doi.org/10.1016/j.medengphy.2005.06.008.
Tezduyar, T.E., Sathe, S., Cragin, T., Nanna, B., Conklin, B.S., Yang, C., Canton, G., Yuan, C., Ferguson, M., Hatsukami, T.S.,
Pausewang, J., Schwaab, M., 2007. Modelling of fluid-structure Tang, D., 2011. Impact of flow rates in a cardiac cycle on
interactions with the space-time finite elements: arterial fluid correlations between advanced human carotid plaque
mechanics. Int. J. Numer. Meth. Fluids 54, 901–922, http://dx. progression and mechanical flow shear stress and plaque wall
doi.org/10.1002/fld.1443. stress. Biomed. Eng. Online 10, 61, http://dx.doi.org/10.1186/
Theodorakakos, A., Gavaises, M., Andriotis, A., Zifan, A., Liatsis, 1475-925X-10-61.
P., Pantos, I., Efstathopoulos, E., Katritsis, D., 2008. Simulation Yang, C., Canton, G., Yuan, C., Ferguson, M., Hatsukami, T.S.,
of cardiac motion on non-Newtonian, pulsating flow Tang, D., 2010. Advanced human carotid plaque progression
development in the human left anterior descending coronary correlates positively with flow shear stress using follow-up
artery. Phys. Med. Biol. 53 (18), 4875–4892, http://dx.doi.org/ scan data: an in vivo MRI multi-patient 3D FSI study. J.
10.1088/0031-9155/53/18/002. Biomech. 43, 2530–2538, http://dx.doi.org/10.1016/j.
Thury, A., Wentzel, J.J., Vinke, R.V., Gijsen, F.J., Schuurbiers, J.C., jbiomech.2010.05.018.
Krams, R., de Feyter, P.J., Serruys, P.W., Slager, C.J., 2002.
Yang, C., Bach, R.G., Zheng, J., Naqa, I.E., Woodard, P.K., Teng, Z.,
Images in cardiovascular medicine. Focal in-stent restenosis
Billiar, K., Tang, D., 2009. In vivo IVUS-based 3-D fluid-
near step-up: roles of low and oscillating shear stress?.
structure interaction models with cyclic bending and
Circulation 105, e185–e187, http://dx.doi.org/10.1161/01.
anisotropic vessel properties for human atherosclerotic
CIR.0000018282.32332.13.
coronary plaque mechanical analysis. IEEE Trans. Biomed.
Torii, R., Oshima, T., Kobayashi, K., Takagi, T., Tezduyar, T., 2010.
Eng. 56, 2420–2428, http://dx.doi.org/10.1109/
Role of 0D peripheral vasculature model in fluid-structure
TBME.2009.2025658.
interaction modeling of aneurysm. Comput. Mech. 46, 43–52,
Zeng, D., Ding, Z., Friedman, M.H., Ethier, C.R., 2003. Effects of
http://dx.doi.org/10.1007/s00466-009-0439-7al. 2010.
cardiac motion on right coronary artery hemodynamics. Ann.
Torii, R., Oshima, T., Kobayashi, K., Takagi, T., Tezduyar, T., 2006.
Biomed. Eng. 31, 420–429.
Fluid-structure interaction modeling of aneurysmal