Prostate Tissue Composition and MR Measurements.docx

Prostate Tissue Composition and MR Measurements.docx

《Prostate Tissue Composition and MR Measurements.docx》由会员分享，可在线阅读，更多相关《Prostate Tissue Composition and MR Measurements.docx（40页珍藏版）》请在冰豆网上搜索。

ProstateTissueCompositionandMRMeasurements

ProstateTissueCompositionandMRMeasurements:

InvestigatingtheRelationshipsbetweenADC,T2,Ktrans,ve,andCorrespondingHistologicFeatures1

1.DeannaL.Langer,MSc,

2.TheodorusH.vanderKwast,PhD,MD,

3.AndrewJ.Evans,PhD,MD,

4.AnnaPlotkin,MD,

5.JohnTrachtenberg,MD,CM,

6.BrianC.Wilson,PhDand

7.MasoomA.Haider,MD

+AuthorAffiliations

1.1FromtheDepartmentofMedicalImaging,PrincessMargaretHospital,UniversityHealthNetworkandMountSinaiHospital,610UniversityAve,Room3-958,Toronto,ON,CanadaM5G2M9（D.L.L.,J.T.,M.A.H.）;InstituteofMedicalScience,UniversityofToronto,Toronto,Ontario,Canada（D.L.L.,M.A.H.）;DepartmentofPathologyandLaboratoryMedicine,TorontoGeneralHospital,UniversityHealthNetwork,Toronto,Ontario,Canada（T.H.v.d.K.,A.J.E.）;DepartmentofLaboratoryMedicineandPathobiology,UniversityofToronto,BantingInstitute,Toronto,Ontario,Canada（A.P.）;DepartmentofSurgicalOncology,PrincessMargaretHospital,UniversityHealthNetwork,UniversityofToronto,Toronto,Ontario,Canada（J.T.）;andDepartmentofMedicalBiophysics,UniversityofToronto,OntarioCancerInstitute,Toronto,Ontario,Canada（B.C.W.）.

1.Addresscorrespondenceto

M.A.H.（e-mail:

m.haider@utoronto.ca）.

1.Authorcontributions:

Guarantorsofintegrityofentirestudy,J.T.,M.A.H.;studyconcepts/studydesignordataacquisitionordataanalysis/interpretation,allauthors;manuscriptdraftingormanuscriptrevisionforimportantintellectualcontent,allauthors;manuscriptfinalversionapproval,allauthors;literatureresearch,D.L.L.,T.H.v.d.K.,A.J.E.,A.P.,J.T.;clinicalstudies,T.H.v.d.K.,A.P.,J.T.;experimentalstudies,D.L.L.,A.J.E.,A.P.,J.T.;statisticalanalysis,D.L.L.;andmanuscriptediting,allauthors

NextSection

Abstract

Purpose:

Toinvestigaterelationshipsbetweenmagneticresonance（MR）imagingmeasurementsandtheunderlyingcompositionofnormalandmalignantprostatetissue.

MaterialsandMethods:

Twenty-fourpatients（medianage,63years;agerange,44–72years）gaveinformedconsenttobeexaminedforthisresearchethicsboard–approvedstudy.Beforeundergoingprostatectomy,patientswereexaminedwithT2-weighted,diffusion-weighted,T2mapping,anddynamiccontrastmaterial–enhancedMRimagingat1.5T.Mapsofapparentdiffusioncoefficient（ADC）,T2,volumetransferconstant（Ktrans）,andextravascularextracellularspace（ve）werecalculated.Whole-mounthematoxylin-eosin–stainedsectionsweregeneratedanddigitizedathistologicresolution.Percentageareasoftissuecomponents（nuclei,cytoplasm,stroma,luminalspace）weremeasuredbyusingimagesegmentation.CorrespondingregionsonMRimagesandhistologicspecimensweredefinedbyusinganatomicallydefinedsegmentsinperipheralzone（PZ）andcentralglandtissue.CancerandnormalPZregionswereidentifiedathistopathologicanalysis.EachMRparameter–histologictissuecomponentpairwasassessedbyusinglinearmixed-effectsmodels,andcancerversusnormalPZvalueswerecomparedbyusingnonparametrictests.

Results:

ADCandT2wereinverselyrelatedtopercentageareaofnucleiandpercentageareaofcytoplasmandpositivelyrelatedtopercentageareaofluminalspace（P≤.01）.ThesetrendswerereversedforKtrans（P<.001）.Ktranshadasignificantlynegative（P=.01）slopeversuspercentageareaofstroma,andvehadapositive（P=.008）slopeversuspercentageareaofstroma.Thevewasinverselyproportionaltothepercentageareaofnuclei（P=.05）.AllMRimagingparameters（P≤.05）andthepercentageareasofalltissuecomponents（P≤.001）exceptstroma（P>.48）weresignificantlydifferentbetweencancerandnormalPZtissue.

Conclusion:

MRimaging–derivedparametersmeasuredintheprostateweresignificantlyrelatedtotheproportionofspecifichistologiccomponentsthatdifferbetweennormalandmalignantPZtissue.Theserelationshipsmayhelpdefineimaging-relatedhistologicprognosticparametersforprostatecancer.

PreviousSectionNextSection

Introduction

Inprostatecancer,clinicalfactorssuchasGleasonscoreatbiopsy,prostate-specificantigenlevel,andtumorstageareassessedbeforetreatmenttoestimatelong-termpatientoutcome

（1）andinfluencethechoiceoftherapy.MorphologicfeaturessuchasnucleusorlumensizehavebeenshownquantitativelytoberelatedtoprimaryGleasonscorepattern

（2）anddiseaseprogression（3）andhavethepotentialtocontributetotreatmentmanagementstrategies.However,duetosamplingerror,thepropertiesoftissuecollectedfrombiopsymaynotaccuratelyreflectthetissuepropertiesdeterminedafterradicalprostatectomy（4）.Volumetricassessmentofprostatecancerwithmagneticresonance（MR）imagingmayassistinlocalizingthetumor（5–8）,identifyingclinicallynonimportantcancers（9）,and/orpredictingbiochemicaltreatmentfailure（10）.Inaddition,diffusion-weightedimaginganddynamiccontrastmaterial–enhancedMRimagingfacilitateadditionalcharacterizationofprostatecancer,withquantitativedifferencesbetweennormalandmalignantperipheralzone（PZ）tissuereported（11–20）.RecentworkhasalsorevealedacorrelationbetweenMRimagingfindingsandimmunohistochemicalmarkersofproliferation（21,22）.Thus,furtherelucidationoftherelationshipsbetweenMRimagingfindingsandcorrespondinghistologicfeaturescouldhaveatremendouseffectonpatientcare,facilitatingguidedbiopsyatthemostaggressivelocationofprostatecancerand/orrevealingsurrogatesforpropertiesrelatedtooutcome.

HistopathologiccorrelatestoT2-weightedsignalintensityintheprostatehavebeeninvestigated（23,24）,andmorerecently,aninverserelationshipbetweenapparentdiffusioncoefficient（ADC）andcellulardensityhasbeendetermined（22,25）.ThestudyofpropertiesbeyondcellulardensityandnucleusareameasurementsandtherangeofMRimagingparametersinvestigatedhavebeenlimited.Inaddition,histopathologicanalysisofselectedregionsintheavailabletissuehaspotentiallyintroducedbias.Inthecurrentwork,weinvestigatedtherelationshipsbetweenvarioustissuecomponents（percentagesofnuclei,epithelialcytoplasm,stroma,andluminalspace）andinvivoMRimagingparameters（ADC,T2,volumetransferconstant[Ktrans],extravascularextracellularvolumefraction[ve]）byusingwhole-mounthistologicsectionsdigitizedintheirentirety,objectivemethodsofregisteringPZandcentralgland（CG）tissueathistologicanalysisandimaging,andhistopathologicallydefinedPZtumors.ThepurposeofthisstudywastoinvestigatetherelationshipsbetweenMRimagingmeasurementsandtheunderlyingcompositionofnormalandmalignantprostatetissue.

PreviousSectionNextSection

MaterialsandMethods

Patients

TheinstitutionalresearchethicsboardoftheUniversityHealthNetworkapprovedthisstudy,andallpatientsgaveinformedconsent.Thirty-oneconsecutivepatientswereenrolledbetweenJune1,2006,andFebruary15,2008,aspartofanongoingprospectivestudy,andalldatawereretrospectivelyreviewed.Theparticipatingpatientshadundergonenopriortreatmentforprostatecancer,hadnocontraindicationstoMRimagingorcontrastagentadministration,andhadnoevidenceofnodalorbonemetastases,asdeterminedwithpreoperativeimaging.Sevenpatientswereexcludedfromanalysisbecausetheirhistologicslideswereunavailableforreview（n=2）orcouldnotbedigitized（n=5）.Onepatientwasexcludedfromdynamiccontrast-enhancedMRimaginganalysisowingtosubstantialmotionduringimageacquisition.Theclinicaldataforthe24patientsexaminedinthisstudyaresummarizedinTable1.

Viewthistable:

Inthiswindow

Inanewwindow

Table1

Characteristicsof24PatientswithProstateCancer

MRImagingProtocol

Beforeundergoingsurgery,allpatientswereexaminedwithMRimaging,whichwasperformedwitha1.5-Tsystem（ExciteHD;GEHealthcare,Milwaukee,Wis）byusingafour-channelphased-arraysurfacecoilcoupledtoanendorectalcoil（Medrad,Warrendale,Pa）.Inthelatterpatientcohortof11patients,theendorectalcoilwasfilledwitha100%wt/volbariumsulfatesuspension（LiquidPolibar;E-Z-Em,Westbury,NY）（26）toreducemagneticsusceptibilityartifacts.ThemediantimebetweenbiopsyandMRimagingwas70days（range,31–266days）,andthatbetweenMRimagingandsurgerywas13days（range,1–139days）.Axial-obliquefastspin-echoT2-weighted,echo-planardiffusion-weighted（bvalues,0and600sec/mm2）,multiechofastspin-echoforT2-mapping,anddynamiccontrast-enhancedMRimageswereobtainedwiththeimagingplaneperpendiculartotherectalwall–prostateinterface.Dynamiccontrast-enhancedMRimagingdataconsistedoftwodatasets:

dataobtainedwithmultisectionmultiple-flip-anglefastspoiledgradient-recalledacquisitioninthesteadystateforT1mappinganddatasubsequentlyobtainedduring50phasesofmultisectionfastspoiledgradient-recalledacquisitioninthesteadystate（26sections,temporalresolutionof10seconds）.Twobaselineimageswereacquiredbeforetheinjectionof20mLofgadopentetatedimeglumine（Magnevist;BayerScheringPharma,Berlin,Germany）andthena20-mLsalineflush,whichwereadministeredatarateof4mL/secbyusingapowerinjector（SpectrisMRInjectionSystem;Medrad）.

T2-weighted,diffusion-weighted,andmultiechofastspin-echoMRimageswereacquiredwithleft-to-rightphaseencoding.Anterior-posteriorphaseencodingwasperformedforalldynamiccontrast-enhancedMRimagingexaminations.Thesectionlocation,sectionthickness（3mm）,andintersectiongap（0mm）werekeptconsistent.AllotherMRimageacquisitionparametersar