S
N I
L
INTRODUCTION
CONCLUSIONS
REFERENCES & ACKNOWLEDGEMENTS
RESULTS
THEORY & METHODS
Erin K Englund1, Michael C Langham2, Cheng Li1, Emile R Mohler3, Thomas F Floyd4, and Felix W Wehrli2
1 Department of Bioengineering, 2 Department of Radiology, 3 Department of Cardiology, University of Pennsylvania, Philadelphia, PA, United States, 4 Department of Anesthesiology, Stony Brook University School of Medicine, Stony Brook, NY, United States
Simultaneous quantification of perfusion, intravascular venous oxygen saturation, and skeletal muscle T2* during reactive hyperemia in the leg
using an interleaved PASL and multi-echo GRE sequence
We propose to develop and validate an interleaved PASL/multi-echo GRE (PASL/Ox-BOLD) pulse sequence, capable of simultaneously measuring perfusion, venous oxygen saturation, and skeletal muscle T2* during reactive hyperemia in the leg.
• Peripheral arterial disease (PAD), generally a systemic manifestation of atherosclerosis, is a debilitating disease affecting many people, particularly the elderly [1]. • Riskfactorsincludesmoking,diabetes,olderage,andhighbloodpressure. • Patientspresentwithlowankle-brachialindex(ABI)andclaudication. • PADisanindependentriskfactorforheartattackandstroke.• PAD results in structural as well as functional impairments. • Structural-Flow-limitingstenosesinlargearteriesandcompensatorycollateralization[2]. • Functional-Endothelialdysfunction,decreasedperfusion,andbluntedhyperemicresponse[3].• Many current MRI-based methods for assessing PAD require an ischemia/reperfusion paradigm to investigate the functional reserve of the vasculature. • CASL[3]orPASL[4]tomeasureperfusion. • MRoximetrytoassessdynamicparametersofvenousoxygensaturation(SvO2)[5,6]. • MeasurementofskeletalmuscleT2*asamarkerofmuscleoxygenationstatus[7,8].• No previous study has been able to concurrently measure all functional parameters.
Image Analysis• Perfusion • TemporalmatchingofNSandSSimageswasachievedbyaveragingadjacentNStimepoints. • Regionsofinterest(ROIs)weredrawninthegastroc,soleus,peroneus,andtibialisanterior(TA)musclesandperfusionwascalculatedfromEq1. • Parametersextractedfromtheperfusiontimecoursedataincludepeakperfusionandtimetopeakperfusion(TTP).• Ox-BOLD • ForeachOx-BOLDimage,highspatialfrequencydatawasfilledinfromtheaverageofreferenceimagesacquiredimmediatelyafterthescan. • OnlydataacquiredfollowingSSinversionwasusedforanalysis,thoughtheOx-BOLDinterleavewasruneveryPLDtocontrolformagnetizationtransfereffects. • Oximetry • Aphasemapwasgeneratedforeachoftheimagesandthebaselinephaseaccumulationintissuewassubtracted[12]. • ThelargeroftheperonealveinswasselectedforSvO2analysisandanROIwasdrawninthisvein.ThereferencetissuewasselectedinanROIimmediately surroundingtheperonealveinandSvO2wascalculatedfromEq2. • Washouttime(timetominimumSvO2),andovershoot(peakSvO2-baselineSvO2)wererecordedfromtheSvO2timecourse. • T2* calculation • Signalintensityineachofthe5echoesinahomogeneousregionofthesoleusmusclewasextractedandfittedtoamonoexponentialfunction.
Perfusion - Saturation Inversion Recovery • Perfusingbloodislabeledbyalternatingbetweennon-selective (NS)andsliceselective(SS)adiabaticinversionpulses[4].• Perfusioniscalculatedas:
• Wherefisperfusioninclassicalunits(mL·min-1·100g-1) • λisthetissue-partitioncoefficient(0.9mL/g) • Tisthepostlabelingdelay(PLD)+TE(T=960.24ms) • MSS and MNSarethemagnitudesignalintensities • T1=T1tissue=T1blood=1420ms
• Partial-FourierGRE-EPIreadoutfollowingNSandSSinversionswith thefollowingparameters: • FOV=250×250mm2,ST=10mm • Matrix=80×50,Reconstructedmatrix=80×80 • TR/TE=1s/8.05ms,PLD=952.19ms
Figure 2. PulsesequencediagramofPASL/Ox-BOLDsequence.ThePASLsliceislocatedatmidcalf.DuringthePLDofthePASL,akeyholemulti-echoGREwith24phaseencodingstepsacquiresdataataslicelocated3cminferiorly.
[Eq1]
Oximetry and T2* - Multi-echo GREOximetry-MRSusceptometry• Magneticsusceptibilityinduceddifferencesinphaseaccumulation betweenbloodandsurroundingtissueareusedtocalculatehemoglobinoxygen saturation[6,10-11]fromphaseimagesacquiredatTE1 and TE2as:
[Eq2]
• WhereΔχdo isthesusceptibilitydifferencebetweenfullyoxygenatedanddeoxygenated blood(4π·0.27ppm)• Hematocrit(Hct)istakentobe0.45• θismeasuredfromaxialscoutimages
PASL/Ox-BOLDPASL
Experimental Design• To validate PASL/Ox-BOLD, measured perfusion, SvO2, and T2* data were compared to an otherwise identical PASL-only, or a keyhole Ox-BOLD-only version of the sequence. • 5subjectswerescannedontwoseparatedaystoassessaccuracyandreproducibility. • Foreachsubject,twoPASL/Ox-BOLD,onePASL,andoneOx-BOLDsequencewereruninarandomized order.Theprotocolwasrepeatedinthesameorderonthesecondday. • AllimagingwasperformedonaSiemens3Tscannerwiththecalfcenteredinan8chtx/rxkneecoil (Invivo),andacuff(Hokanson)placedaroundthethigh(Figure4). • Eachscanlasted10min,with1minbaseline,3minarterialocclusionwithcuffinflatedto>205mmHg, and6minrecovery.1minofrestwasgivenbeforethenextscan.
Figure 4. Schematicofexperimentalsetup(redslice=PASL,blueslice=Ox-BOLD),thebluecylinderaroundthethighrepresentsthecuff.Thetimelineofeachscanisshownundertheleg.Theperiodofarterialocclusionisalwaysrepresentedbyagraybox.
Base
line
Ar
teria
l O
cclu
sion
(C
uff >
205
mm
Hg)
Reco
very
(Rea
ctive
Hype
rem
ia)
Rest
t=060 240 600660s
SkeletalmuscleT2*• MagnitudeSIfromTE1-TE5arefittedtoamonoexponentialtocalculateT2*.• Multi-echoGREreadoutwithkeyholeacquisition(central¼thofk-spaceacquired everyPLDofPASL)and: • FOV=96×96mm2,ST=10mm,slicelocationlocated3cminthefootdirection. • Matrix=96×24,Reconstructedmatrix=96×96usingreferenceimageacquiredafterdynamicscan. • TR/TE1/TE2/TE3/TE4/TE5=38.12/3.78/6.99/12.32/19.32/26.32ms.
Figure 3. Schematicofphaseaccumulationinreferencetissue(jt)andinthevein(jv).Differenceinphaseaccumulation(jv-jt=Dj)alongwithmeasurementoftheangulardeviationofthevein(q)fromthemainmagneticfield(B0)allowsquantificationofoxygensaturation,givenbyEq2.
Reference Tissue
Vein
VeinReference Tissue
Phase accumulation over DTE
Figure 5 a&b.Anatomicalscout(a)andGRE-EPI(b)oftheperfusionslice.ColoredROIsindicatemuscles(Red=gastroc,Green=soleus,Purple=peroneus,Cyan=TA).
Figure 5 c.Perfusiontimecourseingastroc,averagedoverallsubjects.Grayboxindicatesperiodofischemia.Reactivehyperemiafollowscuffreleaseat240s.ErrorbarsrepresentSD.
Figure 5 d.Zoomedperfusiontimecourseinanindividualsubjectforeachmuscle.Dashedlines=PASL,Solidlines=PASL/Ox-BOLD.ColorscorrespondtomuscleROIsdrawninFigure5(b).
PERFUSION
OXIMETRY
b.
a.
d.
a.
Baselineb.
Figure 6 a.AnatomicalimageoftheOx-BOLDslice,indicatingperonealveins(yellow)andartery(red).
Figure 6 b&c.Zoomedphaseimagesofthemedial(solid)andlateral(dotted)peronealveinsduringbaseline(b)andreactivehyperemia(c).
Figure 6 d. Oximetrytimecourseinthelargerperonealveinaveragedforallsubjects.
Figure 6 e.Comparisonofaveragewashouttime.Averagevaluesarereported,errorbarsindicateSD.Averagedifferencebetweenmethodsis9%.
Figure 6 f. Comparisonofaverageovershootacrossallsubjects.TheaverageovershootdifferenceforPASL/Ox-BOLDcomparedtoOx-BOLDis11%.
ReactiveHyperemiac.
h.
T2*
d.
Ox-BOLD Ox-BOLD
f.
e.
PASL/Ox-BOLDPASL
c.
f.
T 2* (m
s)
Figure 7 e&f. ComparisonofaveragerelativeT2*min and T2*max (e),andtminandtmax (f).Averagevaluesarereported,witherrorbarsindicatingSD.
b.
c.
d.
Figure 1. Flow-limitingstenosesintheperipheralvasculatureisthehallmarkofPAD(NHLBI[9]).
Figure 7 a.AnatomicalimageoftheOx-BOLDslice,withfullOx-BOLDFOVindicatedbythegreenbox.
Figure 7 b&c. T2* images acquiredduringischemia(b)andreactivehyperemia(c).
Figure 7 d. RelativeT2*timecourseinaROIinthesoleusmuscleaveragedforallsubjects.Eachsubject’sT2*wasnormalizedtohisbaselinevalue.
T2*min T2*maxtmin tmax
e. f.
• Perfusion • Goodintra-scanreproducibility,butrelativelypoorinter-scanreproducibilityoftheperfusionparametersmeasuredwithPASL/Ox-BOLD. • Perfusionvariesphysiologicallywithtimeofday,hydration,caffeineintake,hormonalfluctuations,andexercise. • Oximetry and T2* • Itispossibletomeasurevenousoxygensaturationintheperonealveinwithkeyholeoximetryinlessthanonesecond. • T2*min,tmin,andtmaxaresimilarforPASL/Ox-BOLDandOx-BOLD,thoughT2*maxdiffersbetweenPASL/Ox-BOLDandOx-BOLD. • Althoughtheorderofthescanswasrandomized,oneofthePASL/Ox-BOLDscanswasfirstinallbutonesubject.ThiscouldexplainthedecreasedT2*max seeninPASL/Ox-BOLD,howeveralargersamplesizeisnecessarytofullyinvestigatethisbias. • RinginginOx-BOLDimagescouldbeduetokeyholeacquisition;thereforeexplorationofaBRISKacquisitionschemewillfollow.• Conclusions • Simultaneous acquisition of PASL, oximetry, and T2* has no affect on the quantification of perfusion, SvO2, or T2*. • The data demonstrate the feasibility of a combined PASL/Ox-BOLD method for simultaneous measurement of perfusion, venous oxygen saturation, and skeletal muscle T2* during reactive hyperemia. • There are striking differences in the measured parameters between healthy subjects and PAD patients. • Further exploration of these functional parameters in PAD patients could: • HelptobettercharacterizepathophysiologicmechanismsunderlyingthefunctionalimpairmentinPAD. • Provideanew,noninvasivetoolforthediagnosis,evaluation,andmonitoringofPADdiseaseprogressionandresponsetotherapy.
References.[1]Hirsch,etal.JAMA(2001);[2]Mohler.ArchInternMedicine(2003);[3]Wu,etal.JACC(2009);[4]Raynaud,etal.MRM(2001);[5]Langham,etal.JACC(2010);[6]Langham,etal.ISMRM(2011);[7]Ledermann,etal.Circulation(2006);[8]Potthast,etal.FortschrRöntgenstr(2009);[9]NHLBI,WhatisPeripheralArterialDisease?;[10]Haackeetal,HumanBrainMapping(1997);[11]Fernández-Searaetal,MRM(2006);[12]Langham,etal.MRM(2009).
Acknowledgements.NIHGrantsR01HL075649and5T32EB009384.
g.e.
Figure 5 e&f.Comparisonofpeakperfusioninthegastrocacrossallsubjects(e)with101representingsubject1day1,102representingsubject1day2,etc.Theaveragepeakperfusionineachmuscleisshownin(f)witherrorbarsindicatingSD.NosignificantdifferencewasdetectedbetweenPASLandPASL/Ox-BOLD.
Figure 5 g&h. SimilarcomparisonofTTPinthegastrocacrossallsubjects(g)andtheaverageTTPwithPASL/Ox-BOLDandPASL(h)ineachmuscle.
Figure 8. Comparisonofperfusion(a),SvO2(b),andrelativeT2*(c)inarepresentativehealthysubject(lightlines)andPADpatient(darklines).PADpatientexhibitsdecreasedpeakperfusionandincreasedTTPin(a),lowerovershootandincreasedwashouttimein(b),anddecreasedT2*max andincreasedtmaxin(c)comparedtothehealthysubject.
Ischem
iaRe
activ
eHy
peremia
a.
Rela
tive
T 2*
Tim
e (s
)
a. b. c.