S

SU

PEER-REVIEW REPORTS

Hemodynamic Analysis of Growing Intracranial Aneurysms Arising from a PosteriorInferior Cerebellar ArteryShin-Ichiro Sugiyama1, Hui Meng 3,4, Kenichi Funamoto4, Takashi Inoue 2, Miki Fujimura2, Toshio Nakayama4,

hunsuke Omodaka5, Hiroaki Shimizu 2, Akira Takahashi 6, Teiji Tominaga5

rcteo

M

CApghinArtil

rwtzataqdfltafl

otrsar

itwwtpe

�

INTRODUCTION

Hemodynamics is thought to influence theinitiation, growth, and rupture of intracra-nial aneurysms (16). Although there seemsto be general consensus that high wall shearstress (WSS) results in the initiation of in-tracranial aneurysms (13-16), the hemody-namic conditions that drive the growth ofaneurysms after initiation are still not com-pletely clear. High WSS has been postulatedto account for aneurysm progression fromthe distal neck where flow impinges (6, 8),whereas low WSS has been associated withaneurysm growth in the dome (1, 18, 19).

We present a hemodynamic study of twoadjacent aneurysms originating from oneparent artery but growing in different ways.Longitudinal data sets of 3D rotational an-

Key words� Hemodynamics� Intracranial aneurysm� Wall shear stress

Abbreviations and Acronyms3D: Three-dimensionalCFD: Computational fluid dynamicsMR: Magnetic resonanceOSI: Oscillatory shear indexPICA: Posterior inferior cerebellar arteryVA: Vertebral arteryWSS: Wall shear stress

From the Departments of1Neuroanesthesia and

2Neurosurgery, Kohnan Hospital, Sendai, Japan; 3Toshibatroke Research Center, University at Buffalo, The Stateniversity of New York, Buffalo, New York, USA; 4Institute

of Fluid Science, Tohoku University, Sendai, Japan; andDepartments of 5Neurosurgery and 6NeuroendovascularTherapy, Tohoku University Graduate School of Medicine,Sendai, Japan

To whom correspondence should be addressed:Shin-ichiro Sugiyama, M.D., Ph.D.[E-mail: sushi@kohnan-sendai.or.jp]

Citation: World Neurosurg. (2012) 78, 5:462-468.DOI: 10.1016/j.wneu.2011.09.023

Journal homepage: www.WORLDNEUROSURGERY.org

Available online: www.sciencedirect.com

1878-8750/$ - see front matter © 2012 Elsevier Inc.All rights reserved.

giography were used to reconstruct accu- o

462 www.SCIENCEDIRECT.com

ate geometries of the aneurysms and adja-ent arteries. The purpose of this study waso determine whether there were differ-nces in the hemodynamic characteristicsf the two growing aneurysms.

ATERIALS AND METHODS

ase Presentation62-year-old hypertensive female patient

resented with vertigo. Computed tomo-raphic imaging showed no evidence ofemorrhage but the presence of multiple

ntracranial aneurysms. Magnetic reso-ance (MR) angiography performed inpril 2009 revealed six intracranial aneu-

ysms: two in the right middle cerebral ar-ery, two adjacent ones in the right posteriornferior cerebellar artery (PICA), one in theeft middle cerebral artery, and one at the tip

� OBJECTIVE: The role of hemodynarysms is not completely clear. Weadjacent unruptured aneurysms arisindifferent ways. This study aimed to inin hemodynamic characteristics betw

� METHODS: A 62-year-old femalentracranial aneurysms. Catheter angiowo aneurysms located adjacently at th

ere growing over a 1-year period.ere acquired via rotational angiogra

ions were conducted on the 3D aneulsatile flow conditions that were mtry.

RESULTS: The proximal multilobulaphysiological levels of wall shear strethe distal rounded aneurysm had low

� CONCLUSION: Growing aneurysmsand morphologic characteristics and dof an aneurysm could be exposed to eWSS and high oscillatory shear in the

f the basilar artery (Figure 1). The two (

WORLD NEUROSURGE

ight middle cerebral artery aneurysmsere clipped in May 2009, and the basilar

ip aneurysm was treated via coil emboli-ation in August 2009. The presurgicalngiographic data were inaccessible digi-ally because of the unfortunate failure ofdata storage device. The angiograph ac-uired during the coiling was thereforeefined as the baseline for computationaluid dynamic (CFD) analysis. Thereafter,

he patient’s progress was followed up byngiography at 6-month intervals (firstollow-up in February 2010; second fol-ow-up in August 2010).

Between the initial angiograph and thatbtained during embolization of the basilar

ip aneurysm, the three untreated aneu-ysms showed no detectable change inhape or size. However, the two follow-upngiographs revealed that the two aneu-ysms in the right PICA had grown over time

in the growth of intracranial aneu-ent a hemodynamic study with twoom one parent artery but growing inigate whether there were differencesthe two growing aneurysms.

ent presented with six unrupturedhy at 6-month intervals revealed that

ght posterior inferior cerebellar arterye-dimensional aneurysm geometriesComputational fluid dynamic simula-m geometries under patient-specificred by magnetic resonance velocim-

eurysm demonstrated high flow andSS) in the region of growth, whereasand low WSS in the growing sac.

n have heterogeneous hemodynamicnt growing patterns. Growing regions

r high WSS at the inflow zone or lowurysm sac.

micspresg frvesteen

patigrape riThrephy.uryseasu

r anss (Wflow

caiffereitheane

Figure 2). The proximal multilobular aneu-

RY, DOI:10.1016/j.wneu.2011.09.023

tcnar

VDstfGJaafTwscdu

ceVFfwervotvfoiva

rysm

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

rysm at the PICA–vertebral artery (VA) junc-tion (Aneurysm 1) showed partial enlarge-ment of the aneurysm wall, whereas thedistal, more spherical aneurysm (Aneurysm2) showed enlargement of the whole aneu-rysm body.

Figure 1. Magnetic resonance angApril 2009 showing six aneurysmcerebral artery, which were clippebasilar artery, which was coiled inmiddle cerebral artery and two inartery (PICA), referred to as Aneu

Figure 2. Three-dimensional rotational digital angipoints: baseline (August 2009), first follow-up (Fshowing that they grew over time. Both aneurys

PICA, posterior inferior cerebellar artery; Rt., right; VA

WORLD NEUROSURGERY 78 [5]: 462-468

One month after the last angiography,he two growing PICA aneurysms werelipped individually via open occipital cra-iotomy. Aneurysm 1 had yellow blebs withtherosclerosis, whereas Aneurysm 2 waseddish and thin walled (Figure 3).

hy at the initial presentation inis patient: two in the right middleay 2009; one at the tip of the

st 2009; and one in the leftght posterior inferior cerebellar1 and Aneurysm 2.

hy of the two PICA aneurysms at three timery 2010), and second follow-up (August 2010),ere subsequently clipped in September 2010.

t, vertebral artery.

, NOVEMBER 2012 ww

essel and Aneurysm Modelingigital angiography was performed by

tandard transfemoral catheterization ofhe right VA, and digital imaging was per-ormed using a biplanar unit (Innova 3131;eneral Electric Healthcare Japan, Tokyo,

apan). These images were obtained during6-second injection of a contrast agent and200-degree rotation with imaging at 30

rames per second for a total of 8 seconds.he corresponding 150 projection imagesere reconstructed into a 3D data set con-

isting of 512 � 512 � 512 isotropic voxelsovering a field of view of 20 mm in all threeirections. The images were reconstructedsing a dedicated workstation.

The data set was exported to a personalomputer to form 3D isosurfaces of the an-urysm using imaging software (Avizo 6.2;isualization Science Group, Merignac,rance). To provide consistent geometriesrom the data at three different time points,e identified on the baseline data set a ref-

rence vessel segment—the right VA—thatemained unchanged over time. The meanalue of the right VA was calculated, andne-half of the mean value was selected as

he first threshold for defining intraluminalessel volume. The segmented luminal sur-ace was then displayed on native axial, cor-nal, and sagittal slices of 3D angiographic

mages until the displayed luminal surfaceisually matched the luminal boundary ofll regions of interest. To provide a consis-

Figure 3. Intraoperative photograph showingtwo growing aneurysms originating in theright PICA. The proximal aneurysm(Aneurysm 1) had a yellowish wall withatherosclerosis, whereas the distal one(Aneurysm 2) had a reddish and rather thinwall. PICA, posterior inferior cerebellar artery;Rt., right; VA, vertebral artery.

iograps in thd in MAugu

the ri

ograpebruams w

ent threshold from baseline to follow-ups,

w.WORLDNEUROSURGERY.org 463

w

ciOriTt

R

Fotfomrepas(wwiwssesb

fitghfg

aporiA

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

the three geometries were semiautomati-cally coregistered, and thresholds were ad-justed to maintain the same reference diam-eter in subsequent segmentations.

The entire model of the aneurysm com-plex, including both PICA aneurysms, wascreated using commercial software (MagicsRP 13.1; Materialise, Leuven, Belgium).Care was taken to ensure proper length ofthe parent artery (VA). The volumetric grid,composed of tetrahedral elements, wasgenerated using meshing software (Gambit2.4; ANSYS Inc., Lebanon, New Hamp-shire, USA) with a resolution of 0.1 mm,resulting in approximately 1.5 million ele-ment meshes.

Quantitative MR VelocimetryQuantitative MR velocimetry was per-formed on this patient using a 3-Tesla MRimage scanner (Signa HDxt 3.0T: GeneralElectric Healthcare Japan) just before thelast angiography. The protocol entails per-forming a standard axial 2D time-of-flightMR angiography of the cranial vasculatureto select a slice orientation for the arterialblood flow measurements. The optimalperpendicular scan plane was determinedusing the acquired time of flight images.The coordinates obtained specify the posi-tion of an oblique fast 2D phase-contrastsequence, which was then performed basedon these coordinates using a peripheralgated 2D phase-contrast sequence with thefollowing imaging parameters: repetitiontime, 25; echo time, 5.4; number of excita-tions, 1; field of view, 160 � 160 mm; ma-trix, 512 � 512; voxel size, 0.3 � 0.3 mm;velocity encoding, 100 and 50 cm/s for theVA and PICA, respectively; imaging time,approximately 5 minutes; transaxial direc-tion; peripheral gated with electrocardiog-raphy, 30 phases. The phase-contrast im-ages were transferred to the workstation forflow quantification using dedicated soft-ware (CV flow; General Electric HealthcareJapan). A region of interest was semiauto-matically placed on the phase-contrast im-ages over one cardiac cycle. The velocities atall of the pixels inside the vessel border wereintegrated to calculate the flow in millilitersper minute. The quantitative waveform overone cardiac cycle was then drawn usingthese flow rates.

The flow rates were 83.3 and 12.2 mL/min in the right VA and right PICA, respec-

tively. The waveform for the right VA was

464 www.SCIENCEDIRECT.com

used as the inlet condition in numericalsimulations, and the calculated fractionalflow rates were specified for two outletboundaries: 0.15 for the right PICA and 0.85for the right VA.

CFD SimulationsFollowing the conventions for CFD in largevessels (3), blood was treated as an incom-pressible Newtonian fluid; vessel wallswere assumed rigid, and no-slip boundaryconditions were applied at the walls. A fi-nite-volume package, Fluent (ANSYS Inc),was used to solve the governing equations:3D unsteady Navier-Stokes equations andthe equation of continuity. Patient-specificpulsatile flow conditions, derived from theMR velocimetry at the second follow-up,were prescribed at the inlet boundary (theright VA) of the three models. It was as-sumed that the waveform does not changeappreciably over time. The diffusion fluxesin the direction normal to the inlet planewere assumed to be zero, and normal gradi-ents were neglected. For the two outletboundaries, we specified the fractional flowrate through each outlet calculated from theresults of the MR velocimetry. A total ofthree cardiac cycles were computed using100 time-steps per cycle; the results for thethird cycle are presented.

Data AnalysisTo visualize flow patterns, 3D streamlinesof intra-aneurysm flow at peak systole andend diastole exiting the PICA were plottedin the model at each time point (baseline,first follow-up, and second follow-up) us-ing commercial software (ANSYS CFD;ANSYS Inc).

WSS refers to the tangential frictionalstress caused by the action of blood flow onthe vessel wall. For pulsatile flow, the time-averaged WSS was calculated by integratingWSS magnitude over a cardiac cycle for eachtetrahedral element (9):

WSS �1

T �o

T ��→

w�dt,

here �¡ is the instantaneous WSS vectorand T is the duration of the cycle. Becauselow WSS has been associated with aneu-rysm growth and rupture (18, 19, 22), wecalculated WSS for the entire model at each

time point. a

WORLD NEUROSURGE

To describe the temporal disturbance ofintra-aneurysm flow, oscillatory shear in-dex (OSI), a dimensionless measure of di-rectional changes in WSS, was calculatedusing the formula reported by He andKu (7):

OSI �1

2�1 ���o

T�→

wdt��o

T ��→

w�dt�.

Note that 0 � OSI � 0.5, with 0 beingompletely unidirectional shear and 0.5 be-ng completely oscillatory. Because high

SI has been shown to be significant foruptured aneurysms (22), we suspected thatt might also be related to aneurysm growth.herefore, we calculated the OSI for the en-

ire model at each time point.

ESULTS

igure 4A shows reconstructed 3D modelsf the PICA aneurysm complex from the

hree angiographs: the baseline and twoollow-ups. To clearly visualize the regionsf aneurysm growth, we overlapped threeodels by aligning the VA to achieve accu-

ate matching, because this parent arteryxperienced no change over the observationeriod (Figure 4B). Clearly, the proximalneurysm (Aneurysm 1) grew slightly on theuperior side, whereas the distal aneurysmAneurysm 2) showed enlargement of thehole aneurysm body. Aneurysm 1, whichas complex shaped with four blebs, grew

n the region of the small bleb at the top,hereas the round-shaped Aneurysm 2

howed drastic enlargement of the wholeac. The extended tip at the bottom of An-urysm 2 observed at the first follow-upuggests that its growth started from theottom.

Figure 5 shows snapshots of the floweld (3D streamlines) captured at peak sys-

ole (A) and diastole (B). Note that therowing region of Aneurysm 1 experiencedigh-velocity flow. The inlet velocity wave-

orm measured by MR velocimetry is alsoiven.

Figure 6 shows the distribution of WSSnd OSI in two views. Although Figure 6Aresents the lateral view consistent withther figures presented above, the growingegion in Aneurysm 1 cannot be examinedn detail. To observe the growing region ofneurysm 1 en face, we changed the viewing

ngle and plotted the results in Figure 6B.

RY, DOI:10.1016/j.wneu.2011.09.023

gtaflsgWtmtOtw

ei(tWqifi

flaicaw

D

Twmtvi1hwafr

wtvlflflclativrt

radMeasrsibblsaoiordooptgsaw

soeaaTi(ggi

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

From Figures 5 and 6, it is evident that therowing area in Aneurysm 1 (a mild bleb on

he top) was located near the inflow zonend therefore experienced high-velocityow near the wall, as shown by the coloredtreamlines in Figure 5. This growing re-ion was subjected to physiological levels ofSS (0.45–1.20 Pa) (10, 18), much higher

han the WSS (�0.30 Pa) experienced by theore protruded blebs on Aneurysm 1. As

his bleb enlarged, its WSS decreased. TheSI in the growing bleb was initially lower

han that in the other blebs but increasedith time.Aneurysm 2, on the other hand, experi-

nced low intra-aneurysm flow in its grow-ng sac. It started with a rather low WSS0.02– 0.45 Pa) and a high OSI (�0.29) athe baseline; as the whole sac enlarged, the

SS sequentially decreased and the OSI se-uentially increased. Time-resolved visual-

zation (not shown) revealed that the flow

Figure 4. Three-dimensional geometries of thetwo PICA aneurysms reconstructed fromrotational digital angiography at the baseline(0), first follow-up (1), and second follow-up(2). (A) The three images shown in timesequence. (B) Overlapping imagesdemonstrating the areas of growth.Aneurysm 1, being complex shaped,presented with four blebs and gradualenlargement of the small bleb at the top.Aneurysm 2, being round shaped, presentedwith drastic enlargement of the whole sac,probably starting from the bottom. Imageswere aligned to match the parent artery (theright VA) from all three time points. Notethat the right PICA was pulled downward bythe enlarging aneurysm at the first follow-upand remained in that position at the secondfollow-up. PICA, posterior inferior cerebellarartery; Rt., right; VA, vertebral artery.

eld in Aneurysm 2 was unstable, with the r

WORLD NEUROSURGERY 78 [5]: 462-468

ow divisions and vortex structures movinground during the cardiac cycle (12). Thentra-aneurysm flow during the cardiac cy-le became more and more unstable as theneurysm sac enlarged, which is consistentith the increase in the OSI.

ISCUSSION

he purpose of this study was to determinehether there were differences in the he-odynamics of two adjacent aneurysms

hat presented different morphologic andascular characteristics and were growingn different ways. One aneurysm (Aneurysm), which was multilobular in shape, alreadyad three prominent protrusions (lobes)ith an atherosclerotic wall. During the ex-

mination period, it was seen to grow on aourth bleb. The other aneurysm (Aneu-ysm 2) was reddish, thin walled, and

ound. It grew rapidly by enlargement of the g

, NOVEMBER 2012 ww

hole sac during the time interval beforehe next follow-up. Flow simulations re-ealed that the growing bleb of the multi-obular aneurysm was located near the in-ow zone and demonstrated high-velocityow and physiological levels of WSS. Inontrast, the large round aneurysm showedow intra-aneurysm flow, low WSS levels,nd unstable flow patterns (higher OSI) inhe dome. We speculate that the differencesn the hemodynamic, morphological, andascular characteristics of these two aneu-ysms indicate different mechanisms forheir growth.

Regarding the partial growth of Aneu-ysm 1, high WSS has been postulated toccount for aneurysm progression from theistal neck, where flow is impeded (6, 8).eng et al. have shown that in animal mod-

ls, high flow impact leading to high WSSnd steep gradients of stress can cause de-tructive remodeling of the wall and aneu-ysm initiation (12, 14, 15). Cebral et al.howed that high flow impact might be anndicator of bleb formation and that mostlebs progress to low WSS states after thelebs are formed (4). Shojima et al. specu-

ated that such flow impact produces a steephear gradient near the impact zone andccelerates the low shear-mediated processn the aneurysm wall (18). These underly-

ng mechanisms might explain the growthf Aneurysm 1 and suggest that the growingegion of Aneurysm 1 was in the process ofeveloping a bleb like the other three blebsn this aneurysm. The intraoperative viewf Aneurysm 1 at the end of examinationeriod did not expose the growing region;

herefore, we do not know if the wall of therowing region was yellowish with athero-clerotic plaque like the other part of thisneurysm or if it was reddish and thinalled like Aneurysm 2.The growth of Aneurysm 2, in a low shear

tress environment, may be consistent withbservations by Boussel et al. (1). In theirxcellent longitudinal study, seven growingneurysms were followed using 1.5-T MRngiography and patient-specific CFD.hey found that aneurysm growth occurred

n the wall exposed to abnormally low WSS1). Our simulations (with higher resolutioneometries derived from 3D rotational an-iography) produced similar results, show-

ng association of low WSS with aneurysm

rowth. In addition to low WSS, we suggest

w.WORLDNEUROSURGERY.org 465

cganiacedn(dosapfli

qfzi

ioareela

pbctpnupttbpltc

d to r

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

that such aneurysm growth is associatedwith a high OSI, which indicates unstable,disturbed flow with significant directionalchange of WSS during the cardiac cycle.Xiang et al. found that WSS and the OSI arethe significant hemodynamic factors thatdetermine aneurysm rupture (22).

Our findings support those of Boussel etal. (1), suggesting a correlation between lowaneurysmal flow hemodynamics and aneu-rysm growth. Our study did not indicate acausal relationship between the two, al-though such a relationship is entirely possi-ble. Blood vessels can remodel in response toan abnormal hemodynamic environment(10). Aneurysm enlargement could be drivenby biological processes mediated by unfavor-able hemodynamic conditions such as lowand oscillatory fluid shear stress, which havebeen shown to activate atherogenic and pro-inflammatory signaling pathways in endothe-lial cells (5, 11, 17). Activation of such inflam-matory pathways might cause degradation ofthe aneurysm wall and growth of the aneu-

Figure 5. Snapshots of flow field (3D streamlines(A) and late diastole (B) as indicated on the inletmeasured at the right vertebral artery by MR vesecond follow-up. The same waveform was use

rysm. f

466 www.SCIENCEDIRECT.com

Our current study, based on patient-spe-ific geometries, suggests that aneurysmrowth is associated with a decrease in WSSnd an increase in the OSI. Although it isot clear whether low WSS and a high OSI

nfluence aneurysm growth, the reverse islmost certainly true: aneurysm geometryhanges will result in changes of intra-an-urysm flow dynamics because geometryetermines the flow. As demonstrated inumerical experiments by Tremmel et al.

21), aneurysm enlargement is followed by aecreased WSS on the aneurysm wall andften by an increased flow instability (the flowplits into multiple vortices that moveround). Increased flow instability is accom-anied by increased directional changes ofuid shear stress on the aneurysm wall, thus

ncreasing the OSI.The unanswered but highly intriguing

uestion is whether certain hemodynamicactors, such as the high flow in the inflowone of Aneurysm 1 and the disturbed flown the dome of Aneurysm 2, are responsible

ured at peak systoleity waveformtry during the

un all three

simulations. Note thatflow near the entrance3D, three-dimensional;

or aneurysm growth. Further clinical stud- o

WORLD NEUROSURGE

es are required to ascertain the significancef the relationship between hemodynamicsnd aneurysm history. In vivo studies areequired to clarify the mechanobiologicalffects of hemodynamic factors on the an-urysm wall. Such an understanding couldead to exciting opportunities for predictingneurysm growth by image-based CFD.

Our CFD simulations include typical sim-lifications made in model generation andoundary conditions (20). Several biologi-al properties such as the viscoelasticity ofhe vessel wall and the non-Newtonianroperty of the blood are neglected for tech-ical reasons. For the inlet condition, wesed a pulsatile waveform obtained on thisatient before the third angiography for all

hree simulations. It should be pointed outhat this is an improvement over what haseen typically done in CFD, where a non-atient-specific waveform is used as the in-

et condition (2). We believe that despitehese simplifications, our CFD analysis hasaptured the main hemodynamic features

rowing region of Aneurysm 1 experienced high, posterior inferior cerebellar artery; Rt., right;ertebral artery.

) captveloc

locime

the g. PICAVA, v

f growing aneurysms. The region of aneu-

RY, DOI:10.1016/j.wneu.2011.09.023

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

Figure 6. Distribution of time-averaged wall shear stress (WSS) andoscillatory shear index (OSI) in two views: (A) Lateral view consistent withFigures 2–5; (B) En face view of the growing area in Aneurysm 1 (blood in VA comestowards the reader). As evident from B, the growing region of Aneurysm 1experienced higher WSS and lower OSI than the more protruded blebs; however, as

increase (white arrowhead) with time. Aneurysm 2, on the other hand had rather lowWSS (seen from both A and B) and high OSI (seen from A) to start with, and as thewhole sack enlarged, the WSS continued to decrease and OSI continued to increase.Note that the scale of WSS in A is smaller than that in B. PICA, posterior inferiorcerebellar artery; Rt., right; VA, vertebral artery.

this bleb enlarged, the WSS is seen to decrease (black arrow) and OSI to slightly

WORLD NEUROSURGERY 78 [5]: 462-468, NOVEMBER 2012 www.WORLDNEUROSURGERY.org 467

1

1

1

1

1

1

1

1

1

1

2

2

2

Cropt

Rp

CD

J

A

1

PEER-REVIEW REPORTS

SHIN-ICHIRO SUGIYAMA ET AL. HEMODYNAMIC ANALYSIS OF GROWING INTRACRANIAL ANEURYSMS

rysm growth could experience either highWSS at the inflow zone or low WSS and ahigh OSI in the aneurysm sac.

CONCLUSIONS

We have shown two growing aneurysmswith different hemodynamic characteris-tics and growing patterns: a proximalmultilobular aneurysm was growing ableb near the inflow zone with high localvelocity and physiological levels of WSS,whereas a distal rounded aneurysm wasgrowing the entire sac featuring low andunstable intra-aneurysmal flow, with lowWSS and higher OSI. Thus, growing an-eurysms may have heterogeneous hemo-dynamic, morphologic, and vascularcharacteristics associated with differentmechanistic pathways.

REFERENCES

1. Boussel L, Rayz V, McCulloch C, Martin A, Acevedo-Bolton G, Lawton M, Higashida R, Smith WS, YoungWL, Saloner D: Aneurysm growth occurs at regionof low wall shear stress: patient-specific correlationof hemodynamics and growth in a longitudinalstudy. Stroke 39:2997-3002, 2008.

2. Cebral JR, Mut F, Weir J, Putman CM: Association ofhemodynamic characteristics and cerebral aneu-rysm rupture. AJNR Am J Neuroradiol 32:264-270,2011.

3. Cebral JR, Mut F, Weir J, Putman C: Quantitativecharacterization of the hemodynamic environmentin ruptured and unruptured brain aneurysms. AJNRAm J Neuroradiol 32:145-151, 2011.

4. Cebral JR, Sheridan M, Putman CM: Hemodynam-ics and bleb formation in intracranial aneurysms.AJNR Am J Neuroradiol 31:304-310, 2010.

5. Chiu JJ, Chen CN, Lee PL, Yang CT, Chuang HS,

468 www.SCIENCEDIRECT.com

flow on monocytic adhesion to endothelial cells. JBiomech 36:1883-1895, 2003.

6. Hashimoto T, Meng H, Young WL: Intracranial an-eurysms: links among inflammation, hemodynam-ics and vascular remodeling. Neurol Res 28:372-380, 2006.

7. He X, Ku DN: Pulsatile flow in the human left coro-nary artery bifurcation: average conditions. J Bio-mech Eng 118:74-82, 1996.

8. Hoi Y, Meng H, Woodward SH, Bendok BR, HanelRA, Guterman LR, Hopkins LN: Effects of arterialgeometry on aneurysm growth: three-dimensionalcomputational fluid dynamics study. J Neurosurg101:676-681, 2004.

9. Lee SW, Antiga L, Steinman DA: Correlationsamong indicators of disturbed flow at the normalcarotid bifurcation. J Biomech Eng 131:061013,2009.

0. Malek AM, Alper SL, Izumo S: Hemodynamic shearstress and its role in atherosclerosis. J Am Med As-soc 282:2035-2042, 1999.

1. Malek AM, Higashida RT, Halbach VV, PhatourosCC, Meyers PM, Dowd CF: Tandem intracranial stentdeployment for treatment of an iatrogenic, flow-lim-iting, basilar artery dissection: technical case report.Neurosurgery 45:919-924, 1999.

2. Meng H, Metaxa E, Gao L, Liaw N, Natarajan SK,Swartz DD, Siddiqui AH, Kolega J, Mocco J: Progres-sive aneurysm development following hemo-dynamic insult. J Neurosurg 114:1095-1103, 2011.

3. Meng H, Swartz DD, Wang Z, Hoi Y, Kolega J,Metaxa EM, Szymanski MP, Yamamoto J, SauvageauE, Levy EI: A model system for mapping vascularresponses to complex hemodynamics at arterial bi-furcations in vivo. Neurosurgery 59:1094-1100,2006.

4. Meng H, Wang Z, Hoi Y, Gao L, Metaxa E, SwartzDD, Kolega J: Complex hemodynamics at the apex ofan arterial bifurcation induces vascular remodelingresembling cerebral aneurysm initiation. Stroke 38:1924-1931, 2007.

5. Metaxa E, Tremmel M, Natarajan SK, Xiang J, PaluchRA, Mandelbaum M, Siddiqui AH, Kolega J, MoccoJ, Meng H: Characterization of critical hemodynam-

WORLD NEUROSURGE

basilar terminus in a rabbit model. Stroke41:1774-1782, 2010.

6. Nixon AM, Gunel M, Sumpio BE: The critical role ofhemodynamics in the development of cerebral vas-cular disease. J Neurosurg 112:1240-1253, 2010.

7. Passerini AG, Polacek DC, Shi C, Francesco NM,Manduchi E, Grant GR, Pritchard WF, Powell S,Chang GY, Stoeckert CJ Jr, Davies PF: Coexistingproinflammatory and antioxidative endothelial tran-scription profiles in a disturbed flow region of theadult porcine aorta. Proc Natl Acad Sci U S A 101:2482-2487, 2004.

8. Shojima M, Nemoto S, Morita A, Oshima M, Wa-tanabe E, Saito N: Role of shear stress in the blisterformation of cerebral aneurysms. Neurosurgery 67:1268-1274, 2010.

9. Shojima M, Oshima M, Takagi K, Torii R, HayakawaM, Katada K, Morita A, Kirino T: Magnitude and roleof wall shear stress on cerebral aneurysm: computa-tional fluid dynamic study of 20 middle cerebral ar-tery aneurysms. Stroke 35:2500-2505, 2004.

0. Steinman DA, Taylor CA: Flow imaging and com-puting: large artery hemodynamics. Ann BiomedEng 33:1704-1709, 2005.

1. Tremmel M, Dhar S, Levy EI, Mocco J, Meng H:Influence of intracranial aneurysm-to-parent vesselsize ratio on hemodynamics and implication forrupture: results from a virtual experimental study.Neurosurgery 64:622-630, 2009.

2. Xiang J, Natarajan SK, Tremmel M, Ma D, Mocco J,Hopkins LN, Siddiqui AH, Levy EI, Meng H: Hemo-dynamic-morphologic discriminants for intracra-nial aneurysm rupture. Stroke 42:144-152, 2011.

onflict of interest statement: The authors declare that theesearch was conducted in the absence of any commercialr financial relationships that could be construed as aotential conflict of interest. H. M. receives grants fromhe National Institutes of Health.

eceived 04 July 2011; accepted 06 September 2011;ublished online 01 November 2011

itation: World Neurosurg. (2012) 78, 5:462-468.OI: 10.1016/j.wneu.2011.09.023

ournal homepage: www.WORLDNEUROSURGERY.org

vailable online: www.sciencedirect.com

878-8750/$ - see front matter © 2012 Elsevier Inc.

Chien S, Usami S: Analysis of the effect of disturbed ics contributing to aneurysmal remodeling at the All rights reserved.

RY, DOI:10.1016/j.wneu.2011.09.023