Professional Documents
Culture Documents
B. Utter, Et Al
B. Utter, Et Al
Abstract
The governing equations for pulsatile fluid flow were solved in their finite volume formulation in order to simulate blood flow in a
variety of three-dimensional aneurysm geometries. The influence of geometric factors on flow patterns and fluid mechanical forces was
studied with the goal of identifying the risk of aneurysm rupture. Aneurysm morphology was characterized by quantitative shape indices
reflecting the three dimensionality of the vasculature derived from clinical studies. Recirculation zones and secondary flows were
observed in aneurysms and arteries. Regions of extreme and alternating shear stress were observed and identified as sites for potential
aneurysm rupture. The ellipticity of an aneurysm was observed to be strongly correlated with wall shear stress at the aneurysm fundus,
while its non-sphericity, volume, and degree of undulation were more weakly correlated.
r 2007 Elsevier Ltd. All rights reserved.
0021-9290/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jbiomech.2007.01.011
ARTICLE IN PRESS
B. Utter, J.S. Rossmann / Journal of Biomechanics 40 (2007) 2716–2722 2717
UI characterizes the ‘‘lumpiness’’ of an aneurysm, and increases as the A representative volumetric flowrate in the basilar artery was obtained
aneurysm wall becomes more undulatory. using phase-contrast cine magnetic resonance by Marks et al. (1992). The
resulting velocity waveform used in the current study is shown in Fig. 2.
2.3. Computational method Four cycles are simulated to ensure periodic flow conditions, and the
results of the fourth are reported. Flow-through conditions are applied at
the outlets of the daughter vessels. At the walls, which are assumed to be
For the simulations carried out in this study, it is assumed that blood is
rigid, no-slip and no-penetration conditions are imposed. At the basilar
an incompressible, Newtonian fluid and that the flow is laminar and
artery inlet, the velocity profile is assumed to be uniform with the pulsatile
isothermal. The assumption of Newtonian behavior is based on the
waveform as shown in Fig. 2.
findings of Perktold et al. (1989) and others who found minimal changes in
The effect of the inlet velocity profile on the downstream flow was
arterial flow patterns when non-Newtonian effects were included. For such
examined, and the impact of the uniform flow assumption was found to be
a fluid, the continuity and Navier–Stokes equations are as follows:
minimal. As shown in Fig. 3, the velocity profile becomes nearly parabolic
r u ¼ 0, upstream of the bifurcation and aneurysm region whether the inlet flow is
uniform or parabolic, as expressed by the Womersley solution (Zamir,
qu
r þ u r u ¼ rp þ mr2 u , 2000). This is physiologically realistic: although the basilar artery receives
qt
blood from two vertebral arteries, it is sufficiently long to allow adequate
where u is the velocity, r the fluid density, p the pressure, and m the fluid mixing of the streams to achieve a nearly parabolic, pulsatile velocity
viscosity. Solution of these equations in their finite volume form is profile (e.g. Valencia et al., 2006).
accomplished through a commercial software package, CFD–ACE+ As the model was developed, grid refinement studies were performed to
(Singhal, 1998). The discretization is temporally implicit and spatially ensure an optimal balance between resolution and computational expense.
second order. Velocity–pressure coupling is achieved using the SIMPLEC These studies demonstrated that unstructured grids consisting of
algorithm initially suggested by Chorin (1967). 150,000–250,000 cells, depending on the morphology, adequately captured
the details of the flow without requiring unreasonable processing time.
3. Results
Fig. 3. Effect of inlet velocity profile on flow in healthy bifurcation: left, inlet profile is uniform; right, parabolic.
ARTICLE IN PRESS
B. Utter, J.S. Rossmann / Journal of Biomechanics 40 (2007) 2716–2722 2719
Fig. 4. WSS distribution, pressure distribution, and streamlines of intra-aneurysmal flow at systole for 7 mm spherical model. Streamlines are colored by
their distance from the basilar artery midplane.
Fig. 5. Sampling of aneurysm shapes studied: (a) high aspect ratio, (b) distorted 7 mm sphere, and (c) small irregular geometries.
the fundus. These results are in agreement with the findings aneurysmal flow is not decelerated to the same extent as
of other studies of aneurysmal flow, including those of it is in larger models. Despite the increased average flow
Valencia et al. (2006), who also observed low WSS at the and small size of this aneurysm, the average systolic WSS
aneurysm fundus. for this model differs from that in the 7 mm spherical
The instantaneous streamlines in Fig. 4 are colored by model by only 1 %.
their distance from the midplane. The streamlines, which Table 1 shows the spatially averaged systolic wall shear
begin in the basilar artery’s plane of symmetry, extend out stress and pressure, as well as the time-averaged WSS near
of plane by as much as 75% of the aneurysm radius. They the fundus, for several of the geometries examined.
demonstrate the three dimensionality of intra-aneurysmal
flow; recirculation zones and out-of-plane motion are 3.2. Shape indices
present that would not be captured by a two-dimensional
model. Over the 14 aneurysm models studied, the aneurysm
Similar regions of elevated WSS and pressure exist in volume varied from 43 to 517 mm3, the EI from 0.209 to
each model. The high aspect ratio model shown in Fig. 5(a) 0.273, UI from 0.002 to 0.1, and NSI from 0.223 to 0.339.
differs from the spherical models in that a higher average For analysis, all index values are normalized by the largest
pressure acts on the high aspect ratio aneurysm both at value in the set.
systole (7.3% higher than the 7 mm model) and diastole Fig. 6 shows that in general, as the EI increases, the
(6.14% higher). The high aspect ratio model has the highest average WSS acting on the aneurysmal wall decreases.
values of average pressure on the aneurysmal wall at both In both the systolic and diastolic series, the lowest EI
systole (545 N/m2, or 4.09 mmHg) and diastole (242 N/m2, represents the 10 mm geometry, whose volume, more
or 1.82 mmHg). than four times greater than the average volume of the
The aneurysm in Fig. 5(c) has the smallest volume of any thirteen other geometries, is its dominant morphological
aneurysm in this study. Due to its size, the intra- characteristic.
ARTICLE IN PRESS
2720 B. Utter, J.S. Rossmann / Journal of Biomechanics 40 (2007) 2716–2722
Table 1
Spatially averaged aneurysm WSS and pressure at systole, and temporally
averaged WSS at fundus, for several aneurysm models
Fig. 7. WSS near fundus versus time over one cardiac cycle. Systole
occurs 0.2 s into cycle (3.2 s in figure); diastolic or minimum inlet flow
conditions are at 0.75 s (here, 3.75 s).
The WSS at the fundus of each aneurysm model over normalized quantity is referred to as the ‘‘normalized
one cardiac cycle is shown in Fig. 7. The cyclic nature of change in WSS.’’ The WSS at the fundus, averaged over
the WSS is observable, with stress following the inlet the cycle, is also shown.
velocity waveform (systolic 0.2 s into cycle, diastolic at The correlation of normalized change in WSS with EI
0.75 s) with varying closeness. The low-flow conditions in (shown in Fig. 8) is most strongly linear with an r value of
the aneurysm appear to delay both systole and diastole in 0.924, while the linear correlations of Figs. 9 (NSI) and 10
the aneurysm dome. (UI) have less compelling r values of 0.704 and 0.474,
The WSS at the fundus is significantly lower than the respectively. The relative significance of these relationships
spatially averaged WSS throughout the cycle. As Fig. 7 is also demonstrated by p values of 2.29 106 (EI), 0.0049
shows, the fundus WSS values calculated in the larger (NSI), and 0.087 (UI), suggesting a strong relationship of
(10 mm diameter), more irregular (independent of aneurysm normalized change in WSS with EI, a significant but less
volume), and higher AR aneurysms are consistently lower strong one with NSI, and a weak one with UI.
than those in the 7 mm sphere and its undulatory variants.
To illustrate the magnitude of the alternation in WSS at 4. Discussion
the fundus of each aneurysm, the maximum ‘‘peak-to-
peak’’ change in WSS at the fundus is normalized by the Of the shape indices investigated, the EI has the strongest
average fundus WSS over the cycle. In Figs. 8–10, this correlation with wall shear stress in the aneurysm. The EI
ARTICLE IN PRESS
B. Utter, J.S. Rossmann / Journal of Biomechanics 40 (2007) 2716–2722 2721
DBI-0442269, which provided the Linux cluster used for que—evaluation of normal subjects and patients with arteriovenous-
computations. malformations. Radiology 182 (2), 467–476.
Nader-Sepahi, A., Casimiro, M., Sen, J., Kirchen, N., 2004. Is aspect ratio
a reliable predictor of intracranial aneurysm rupture? Neurosurgery 54
References (6), 1343–1347.
Parlea, L., Fahrig, R., Holdsworth, D.W., Lownie, S.P., 1999. An analysis
Beck, J., Rohde, S., el Beltagy, M., Zimmerman, M., Berkefeld, J., Seifert, of the geometry of saccular intracranial aneurysms. American Journal
V., Raabe, A., 2003. Difference in configuration of ruptured and of Neuroradiology 20, 1079–1089.
unruptured intracranial aneurysms determined by biplanar digital Pentimalli, L., Modesti, A., Vignati, A., Marchese, E., Albanese, A., Di
subtraction angiography. Acta Neurochir 145, 861–865. Rocco, F., Coletti, A., Di Nardo, P., 2004. Role of apoptosis in
Chorin, A., 1967. A numerical method for solving incompressible viscous intracranial aneurysm rupture. Journal of Neurosurgery 101 (6),
flow problems. Journal of Computational Physics 2, 12–26. 1018–1025.
Foutrakis, G., Yonas, H., Sclabassi, R., 1999. Saccular aneurysm Perktold, K., Peter, R., Resch, M., 1989. Pulsatile non-Newtonian blood
formation in curved and bifurcating arteries. American Journal of flow simulation through a bifurcation with an aneurysm. Biorheology
Neuroradiology 20 (7), 1309–1317. 26 (6), 1011–1030.
Gobin, Y., Counord, J., Flaud, P., Duffaux, J., 1994. In vitro study of Raghavan, M., Ma, B., Harbaugh, R., 2005. Quantified aneurysm shape
hemodynamics in a giant saccular aneurysm model: influence of flow and rupture risk. Journal of Neurosurgery 102 (2), 355–362.
dynamics in the parent vessel and effects of coil embolization. Rohde, S., Lahmann, K., Beck, J., Nafe, R., Yan, B., Raabe, A.,
Neuroradiology 36, 530–536. Berkefeld, J., 2005. Fourier analysis of intracranial aneurysms:
Gonzalez, C., Cho, Y., Ortega, H., Moret, J., 1992. Intracranial towards an objective and quantitative evaluation of the shape of
aneurysms: flow analysis of their origin and progression. American aneurysms. Neuroradiology 47 (2), 121–126.
Journal of Neuroradiology 13 (1), 181–188. Shojima, M., Oshima, M., Takagi, K., Torii, R., 2004. Magnitude and role
Hademenos, G., Massoud, T., 1998. The Physics of Cerebrovascular of wall shear stress on cerebral aneurysm—computational fluid
Diseases. Springer, New York. dynamic study of 20 middle cerebral artery aneurysms. Stroke 35
Hademenos, G., Massoud, T., Valentino, D., Dickwiler, G., Vinuela, F., (11), 2500–2505.
1994. A nonlinear mathematical model for the development and Singhal, A.K., 1998. Key elements of verification and Validation of CFD
rupture of intracranial saccular aneurysms. Neurological Research 16 Software. In: Proceedings of the 29th AIAA Fluid Dynamics
(5), 376–384. Conference, AIAA 98-2639.
Hoi, Y., Meng, H., Woodward, S., Guterman, ., Bendok, B., Hanel, R., Strother, C.M., 1995. In vitro study of hemodynamics in a giant saccular
Hopkins, L., 2004. Effects of arterial geometry on aneurysm growth: aneurysm model: Influence of flow dynamics in the parent vessel and
three-dimensional computational fluid dynamics study. Journal of effects of coil embolization (letter; comment). Neuroradiology 37,
Neurosurgery 101 (4), 676–681. 159–161.
Humphrey, J., Canham, P., 2000. Structure, mechanical properties, and Toth, M., Nadasy, G.L., Nyary, I., Kerenyi, T., Orosz, M., Molnarka, G.,
mechanics of intracranial saccular aneurysms. Journal of Elasticity 61, Monos, E., 1998. Sterically inhomogeneous viscoelastic behavior of
49–81. human saccular cerebral aneurysms. Journal of Vascular Research 35,
Ingebrigtsen, T., Morgan, M., Faulder, K., Ingebrigsten, L., Sparr, T., 345–355.
Schirmer, H., 2004. Bifurcation geometry and the presence of cerebral Ujiie, H., Tachibana, H., Hiramatsu, O., Hazel, A., Matsumoto, T.,
artery aneurysms. Journal of Neurosurgery 101 (1), 108–113. Ogasawara, Y., Nakajima, H., Hori, T., Takakura, K., Kajiya, F., 1999.
Kassell, N.F., Torner, J.C., Haley, E.C., et al., 1990. the international Effects of size and shape (aspect ratio) on the hemodynamics of saccular
cooperative study on the timing of aneurysm surgery. 1. Overall aneurysms: a possible index for surgical treatment of intracranial
management results. Journal of Neurosurgery 73 (1), 18–36. aneurysms. Neurosurgery 45 (1), 119–129 (discussion 129–30).
Liepsch, D., 2001. An introduction to biofluid mechanics—basic models Valencia, A.A., Guzman, A.M., Finol, E.A., Amon, C.H., 2006. Blood
and applications. Journal of Biomechanics 35, 415–435. flow dynamics in saccular aneurysm models of the basilar artery.
Ma, B., Harbaugh, R., Raghavan, M., 2004. Three-dimensional geome- Journal of Biomechanical Engineering (128), 516–526.
trical characterization of cerebral aneurysms. Annals of Biomedical Vega, C., Kwoon, J.V., Lavine, S.D., 2002. Intracranial aneurysms:
Engineering 32 (2), 264–273. current evidence and clinical practice. American Family Physician 66
Marks, M., Pelc, N., Ross, M., Elzmann, D., 1992. Determination of (4), 601–608.
cerebral blood-flow with a phase-contrast cine MR imaging techni- Zamir, M., 2000. The Physics of Pulsatile Flow. Springer, New York.