(3.1.9)--【2012-radiology】采用IVIM定量评估脑组组灌注.pdf
Original research n Neuroradiology874 radiology.rsna.org n Radiology:Volume 265:Number 3December 2012Quantitative Measurement of Brain Perfusion with intravoxel incoherent Motion Mr imaging1Christian Federau,MD,Dipl Phys ETHPhilippe Maeder,MDKieran OBrien,PhDPatrick Browaeys,MDReto Meuli,MD,PhDPatric Hagmann,MD,PhDPurpose:To evaluate the sensitivity of the perfusion parameters derived from Intravoxel Incoherent Motion(IVIM)MR imaging to hypercapnia-induced vasodilatation and hyper-oxygenation-induced vasoconstriction in the human brain.Materials and Methods:This study was approved by the local ethics committee and informed consent was obtained from all participants.Images were acquired with a standard pulsed-gradient spin-echo sequence(Stejskal-Tanner)in a clinical 3-T system by using 16 b values ranging from 0 to 900 sec/mm2.Seven healthy volunteers were examined while they inhaled four different gas mixtures known to modify brain perfusion(pure oxygen,ambient air,5%CO2 in ambi-ent air,and 8%CO2 in ambient air).Diffusion coefficient(D),pseudodiffusion coefficient(D*),perfusion fraction(f),and blood flowrelated parameter(fD*)maps were calculated on the basis of the IVIM biexponential model,and the parametric maps were compared among the four different gas mixtures.Paired,one-tailed Student t tests were performed to assess for statistically significant differences.Results:Signal decay curves were biexponential in the brain pa-renchyma of all volunteers.When compared with inhaled ambient air,the IVIM perfusion parameters D*,f,and fD*increased as the concentration of inhaled CO2 was in-creased(for the entire brain,P=.01 for f,D*,and fD*for CO2 5%;P=.02 for f,and P=.01 for D*and fD*for CO2 8%),and a trend toward a reduction was observed when participants inhaled pure oxygen(although P.05).D remained globally stable.Conclusion:The IVIM perfusion parameters were reactive to hyper-oxygenation-induced vasoconstriction and hypercapnia-induced vasodilatation.Accordingly,IVIM imaging was found to be a valid and promising method to quantify brain perfusion in humans.q RSNA,20121From the Department of Diagnostic and Interventional Radiology,CHUV,University Hospital Center and University of Lausanne(CHUV-UNIL),Rue du Bugnon 46,1011 Lau-sanne,Switzerland(C.F.,P.M.,P.B.,R.M.,P.H.);and CIBM,University of Geneva,Geneva,Switzerland(K.O.).Received March 10,2012;revision requested May 8;revision received May 24;accepted June 6;final version accepted June 18.Supported in part by the Center for Biomedical Imaging of the Geneva and Lausanne Universities,EPFL.P.H.supported by Leenaards Foundation,Swiss National Science Foundation and Department of Radiology of University of Lausanne.Address correspondence to C.F.(e-mail:christian.federauchuv.ch).q RSNA,2012Note:This copy is for your personal non-commercial use only.To order presentation-ready copies for distribution to your colleagues or clients,contact us at www.rsna.org/rsnarights.Radiology:Volume 265:Number 3December 2012 n radiology.rsna.org 875NEURORADIOLOGY:Intravoxel Incoherent Motion MR Imaging Federau et alimaging to hypercapnia-induced vasodi-latation and hyperoxygenation-induced vasoconstriction.Materials and MethodsThis prospective study was approved by the local ethics committee at the Uni-versity of Lausanne.Informed consent was obtained from all participants.Im-aging was performed in seven healthy volunteers(five men,two women;mean age,28)who were more than 18 years old from September through November 2011.No volunteers were excluded dur-ing this study.Imaging was performed by two radiologists(C.F.and P.H.,with 1 year and 7 years of experience in ra-diology,respectively).Image processing and analysis and statistical analysis were done by C.F.Results analysis and text writing was performed by all authors.IVIM ModelThe IVIM model can be understood as an adaptation of Stejskals and Tanners work(38)on biologic tissue,and was proposed by Le Bihan et al(10,34).The hypothesis is that two compartments ex-ist:a slow moving compartment,where particles diffuse in a Brownian fashion as a consequence of thermal energy,and a fast moving compartment(the vascu-lar compartment),where blood moves changes are often ignored.Finally,dy-namic contrast-enhanced MR imaging and dynamic susceptibility contrast-en-hanced imaging are affected by first-pass extravasation of contrast material(8).A fourth method,which is much less popular,called intravoxel incoher-ent motion(IVIM)imaging,measures perfusion locally and quantitatively(9).IVIM was introduced by Le Bihan et al(10)as a joint method to measure per-fusion and diffusion.Although diffusion imaging has proved to be largely useful in a wide variety of clinical applications(1114)as well as in more advanced ap-plications such as diffusion tensor imag-ing and tractography(1517),the mea-surement of perfusion by using IVIM is not common because of its low signal-to-noise ratio(18),with blood volume in the brain estimated to be in the low single-digit percentage range(19,20).Recently,promising perfusion measurements with IVIM have been achieved in humans in multiple organs(2128).However,to our knowledge,the last attempt to use the technique to measure perfusion in the brain was in the 1990s and was performed mostly in animals(2931)and sporadically in humans(3235).More recently,IVIM has been used in association with blood oxygenation leveldependent(36)and arterial spin labeling(37)techniques.A specific study validating the method in humans is,to our knowledge,lack-ing.Therefore,the purpose of this study was to evaluate the sensitivity of the perfusion parameters derived from Intravoxel Incoherent Motion(IVIM)in the human brain and MR Perfusion is the process of nutritive delivery of arterial blood to the capillary bed of a biologic tissue(1).In the brain,it is classically quanti-fied in terms of cerebral blood flow as a volume of blood per unit of the weight of the brain per unit of time(2).A variety of methods exist to measure brain perfusion by using magnetic resonance(MR)imag-ing.The most common method in clin-ical use,dynamic susceptibility contrast materialenhanced imaging,is based on the measurement of the first-pass T2*effect of a bolus of paramagnetic exoge-nous contrast material(gadolinium che-late)(3)and its volume distribution.A second method that is gaining popularity in recent years because of technical im-provements is arterial spin labeling(4).It uses water in the blood as an endoge-nous contrast agent,which is labeled in the arteries before it enters the brain.A third method is dynamic contrast-enhanced MR imaging,which requires intravenous injection of gadolinium con-trast media and measures the dynamic change in T1 relaxation time.In all three techniques,perfusion quantification is dependent on the ar-terial input function,which is difficult to estimate because of bolus dispersion and delay(57).Furthermore,quan-tification requires many variables that induce additive effects on the error of the perfusion measure.Nonlinear signal Implication for Patient Care nMeasurement of the highly clini-cally relevant cerebral blood flow with IVIM has many theoretical advantages over currently avail-able perfusion imaging methods because it is noninvasive and nonirradiating,requires no intra-venous contrast material injec-tion,is probably mainly depen-dent on capillary flow(little arterial or venous sensitivity),and is intrinsically quantitative.Advances in Knowledge nImaging brain perfusion in humans with a clinical MR im-aging unit is possible with intra-voxel incoherent motion(IVIM)imaging.nWhen compared with air inhala-tion,the IVIM perfusion parame-ters(pseudodiffusion coefficient,the perfusion fraction,and the flow-related parameter)increase gradually with inhalation of in-creasing CO2 concentration(5%and 8%CO2)(P,.05),which is known to vasodilate brain capil-laries,and decrease under inha-lation of pure O2(P.05),which is known for its vasocon-strictive effect.Published online before print10.1148/radiol.12120584 Content code:Radiology 2012;265:874881Abbreviations:D=diffusion coefficientD*=pseudodiffusion coefficientf=perfusion fractionfD*=flow-related parameterIVIM=intravoxel incoherent motionAuthor contributions:Guarantors of integrity of entire study,C.F.,P.M.,R.M.,P.H.;study concepts/study design or data acquisition or data analysis/interpretation,all authors;manuscript drafting or manuscript revision for important intellectual content,all authors;approval of final version of submitted manuscript,all authors;literature research,C.F.,P.M.,K.O.,R.M.,P.H.;clinical studies,C.F.,P.B.,R.M.,P.H.;experimental studies,C.F.,P.M.,P.H.;statistical analysis,C.F.,P.H.;and manu-script editing,C.F.,P.M.,K.O.,P.B.,P.H.Conflicts of interest are listed at the end of this article.876 radiology.rsna.org n Radiology:Volume 265:Number 3December 2012NEURORADIOLOGY:Intravoxel Incoherent Motion MR Imaging Federau et almore robust than a direct biexponential fit.For example,in the case of a voxel with a monoexponential signal decay,a direct biexponential fit would weight the first or second exponential in an aleatory way,giving meaningless values for f.Values less than 0 for f,D,and D*are not physiologic and were set to 0,as were values greater than 1 for f.Values with f greater than 0.3 and D*greater than 0.05 were set(arbitrarily)to zero because they are not physiologic and likely to result either from noise or turbulent cerebrospinal fluid flow.Image AnalysisA whole brain region of interest(ROI)was first drawn on images of each axial section,and the IVIM parameters were calculated on a voxel-by-voxel basis and averaged(more than 100 000 voxels in each participant).Furthermore,we placed small ROIs(approximately 12 cm2)on images of the gray matter of the superior frontal and parietal gyri,in the white matter of the center semiova-le,in the thalamus,and the lenticular nucleus on both sides.The fit was done again on a voxel-by-voxel basis and the result was averaged for each small ROI and then averaged for the gray matter,white matter,thalamus and lenticular nucleus.IVIM parameter maps were generated to demonstrate the voxel-by-voxel percentage variations under the various gases in comparison to those under ambient air.Statistical AnalysisNormal data distribution was assumed and not tested owing to the low number of data.A paired one-tailed Student t test was performed by using Excel(Mi-crosoft,Redmont,Wash)to reject the null hypothesis that the results obtained under hypercapnia and hyperoxygen-ation were similar to those obtained with ambient air.A P value less than.05 was considered to indicate a sta-tistically significant difference.Multiple testing analysis was not performed.ResultsIn all seven volunteers,the signal decay curve as function of b,with b ranging shown to significantly decrease cerebral blood flow(39).We investigated in two healthy volunteers(one man and one woman,mean age,26 years)the varia-tion of the IVIM parameters in the brain after inhalation of ambient air(22%O2,78%N2),and a mixture of 5%CO2 and air(5%CO2,22%O2,73%N2),and in five healthy volunteers(four men and one woman;mean age,28 years)after inha-lation of ambient air,a mixture of 5%CO2 and ambient air,and 8%CO2 and ambient air(8%CO2,22%O2,70%N2),and pure oxygen(100%O2).One experi-ment was interrupted during the 8%CO2 inhalation because of technical problems(participant 3 in Table 1).The gases were kept in bottles outside the Faraday cage,and were provided to the volunteers through an airtight mask over the nose and mouth and a one-way valve system.We waited a fixed time of 6 minutes for equilibrium between gas switches and acquisitions.For safety reasons,expira-tory partial pressure of CO2 was moni-tored by using a Maglife C Plus monitor(Schiller Medical,Wissembourg,France)or capillary partial pressure of CO2 was monitored transcutaneously with a TO-SCA 500 monitor(Radiometer,Thalwil,Switzerland)(40).Image ProcessingGeometric distortion and bulk motion correction were removed with the FSL (http:/www.fmrib.ox.ac.uk/fsl/index.html)eddy-current image registration tool before the IVIM calculation(41).All images were registered to the im-age with a b value of 0.Curve fitting of equation 1 was done on a voxel-by-voxel basis by using the Levenberg-Marquardt algorithm(42)implement-ed with standard Matlab functions(Mathworks,Natick,Massachusetts).In the first step,the curve was fitted for a b value greater than 200 sec/mm2 for the single parameter,D.The assumption for this step was that D*is significantly greater than D,so that the influence of D*on signal decay can be neglected for b factors greater than 200 sec/mm2(43).In the second step,the curve was fitted for f and D*for all the b values,while keeping D constant.This two-step method was found to be as a consequence of pressure gradient.In this second compartment,a pseudo-diffusion term(D*)is introduced that describes on a macroscopic level the displacement of the blood elements in an assumed randomly laid vascular net-work.For the perfusion to have a phys-iologic meaning,one expects that D*is greater than D.Therefore:0()(1)bDbDS bfef eS=+(1)where f is the perfusion fraction;D,the diffusion coefficient;D*,the pseudodif-fusion coefficient;and:=222()3bG(2)The b value regroups the parameters depending on the sequence,namely the gyromagnetic ratio(g),and the dura-tion(d),strength(G),and interval(D)of the magnetic field gradient.Imaging ParametersData were acquired by using a 3-T MR imager(Trio;Siemens,Erlangen,Ger-many)with a 32-channel receiver head coil and a standard monopolar pulsed-gradient spin-echo echo-planar imaging sequence(33,38),which is used routinely for diffusion-weighted imaging.For each participant,we acquired images of nine axial brain sections with the following paramenters:section thickness,4 mm;field of view,297 3 297 mm2;matrix,256 3 256,in-plane resolution,1.2 3 1.2 mm2;repetition time/echo time,4000/99 msec;parallel imaging with an accelera-tion factor of two;and 75%partial Fou-rier encoding.Receiver bandwidth was 1086 Hz/pixel.Fat was suppressed with a frequency-selective saturation routine.We acquired images at multiple b values(0,10,20,40,80,110,140,170,200,300,400,500,600,700,800,and 900 sec/mm2)in three orthogonal directions,averaged four times.Total acquisition time was 12 minutes and 28 seconds.Gas InhalationIncrease in CO2 arterial partial pressure is a well-known potent intracerebral va-sodilator that increases cerebral blood flow;inhalation of pure oxygen has been Radiology:Volume 265:Number 3December 2012 n radiology.rsna.org 877NEURORADIOLOGY:Intravoxel Incoherent Motion MR Imaging Federau et alDiscussionThe results of this study validate IVIM imaging as a method to measure per-fusion in the human brain.To our knowledge,this was not previously well established in the literature.Our re-sults demonstrated that f,D*,and fD*change in a gradual way under hyper-capnia and hyperoxygenation stimuli in the full brain and
