1、Multi-detectorspiralCTstudyoftherelationshipsbetweenpulmonaryground-glassnodulesandbloodvessels1Abstract Objective:Toinvestigatetherelationshipsbetweenpulmo-naryground-glassnodules(GGN)andbloodvesselsandtheirdiagnosticvaluesindifferentiatingGGNs.Conclusion:DifferentGGNshavedifferentrelationshipswith
2、vessels.UnderstandingandrecognisingcharacteristicGGN-vesselrelationshipsmayhelpidentifywhichGGNsaremorelikelytobemalignant.2KeyPoints Multi-detectorCToffersnewinformationaboutground-glassnodules.Differenttypesofground-glassnoduleshavedifferentrelationshipswithvessels.Thismayhelpidentifywhichground-g
3、lassnodulesarelikelytobemalignant.3IntroductionWiththeextensiveacceptanceoflow-dosemulti-detectorspiralCTinlungcancerscreening,thenumberofdetectedGGNsorfocalground-glassopacities(fGGOs)hasdramaticallyincreased.GGNscanresultfromneoplasms,suchaspulmonaryadenocarcinoma,orbenigndiseases,suchasfocalfibro
4、sis,inflammationoralveolarhaemorrhage.4 Inaddition,pre-invasiveabnormalities,includingatypicaladenomatoushyperplasia(AAH)andadenocarcinomainsitu(AIS).IthasbeenreportedthattheproportionofmalignancyinGGNsishigherthaninsolidpulmonarynodules(SPNs)andthemajorityofmalignantGGNsareadenocarcinoma.Duetoimagi
5、ngresemblance,however,itisextremelychallengingtodifferentiatemalignantGGNsfromtheaforementionedbenigncounterparts.5 AccuratedifferentialdiagnosisofGGNswillassistphysicianstomaketreatmentdecisionsandimprovetreatmentoutcomesandprognosis.Severalinvestigatorshavesuggestedthatanalysisofrelationshipsbetwe
6、enSPNsandsurroundingvesselscanhelppredictthelikelihoodofmalignancyinsuchnodules.TherelationshipbetweenGGNsandbloodvesselsremainsunknown.WhetherthisrelationshipcanbeutilisedtofacilitatethediagnosisofmalignantGGNsisaworthyofinvestigation.6MaterialsandmethodsPatients Theimagingdataofpatientswithpulmona
7、ryGGNsreceivingthin-sectionmulti-detectorCTexaminationatourhospitalinJanuary2011throughNovember2012wereretrospectivelyreviewed.Alllesionsweresolitaryandmostofthem(104/108)surgicallyresectedwithin2weeksafterCTscanning.7InclusioncriteriaTheGGNsizewaslessthan3cminthelargestdimension.ground-glassopacity
8、(GGO)comprisedmorethan50%oftheareaofthelesiononCT.-Anareaofover50%GGOwassetasthecutoffvaluetoexcludesolid/semi-solidlesions.-AlthoughsolidnodulesfrequentlyhadGGOcomponentsaroundtheirmargin,probablyrepresentingsurroundingoedemaormerelypooraerationofthesurroundinglungtissuesduetocompressionorretractio
9、nbynodules,thesenoduleshadalreadybeenwellinvestigatedusingCTandthereforewerenotthestudyobjects8Ultimately,108patientswereenrolledintothisstudy,including38malesand70femaleswithmeanageof58.1812.89years(range,22to79years).43patientswereasymptomatic,28hadrespiratorysymptoms,and37hadlungcancerriskfactors
10、,suchassmokingandfamilyhistory.9Accordingtopathologicalfindings,GGNsweredividedintothreegroups:(1)Benigndiseasegroup(10cases),includingfournodulesdiagnosedwithacombinationofclinicalsymptomsandimagingpresentations(nodulesdisappearedorgraduallyreducedinsizeonmultiplefollow-upCTimaging)andsixnodulescon
11、firmedbypathologicalexamination(1caseofsclerosinghaemangiomaand5casesofchronicinflammation).(2)Preinvasivediseasegroup(24cases),including7AAHsand17AISs.10(3)theinvasiveadenocarcinomagroup(74cases),confirmedpathologically,therewere39non-mucinousminimallyinvasiveadenocarcinomas(MIA)and35invasiveadenoc
12、arcinomas(IAC;specifically,13lepidicpredominantadenocarcinomas;19acinus-predominantadenocarcinomas;2papillary-predominantadenocarcinomasand1solidpredominantwithmucin粘蛋白production).11CTimaginganalysis protocolparameters:0.625-mmsectionwidthwitha0.625-mmreconstructioninterval,pitchof0.984,120kVand250m
13、A.Allimageswerereviewedwithahigh-resolution,2,0481,560pixel,standardlungwindow(ww,1,500HU;wl,-500HU)andmediastinalwindow(ww,350HU;wl,50HU)GGNscanbefurthersubdividedintomixedground-glassnodules(mGGNs)andpureground-glassnodules(pGGNs).ThepercentageoftheGGOcomponentwascalculatedasfollows:(DGGO-D)/DGGO1
14、00,whereDGGOisthelargestdiameteroftheentirelesionandDisthelargestdiameterofthesolidcomponentwithinthelesion.12 BloodvesselanalysiswasperformedintermsofvascularmorphologyandvascularrelationshipswithGGNlesions.thediameterofpulmonaryvesselsgraduallydecreasesfromthehilumtowardtheperiphery.Ifthediametero
15、fthevascularsegmentwithinlesionswaslargerthantheproximalsegmentorlesionvesselswereapparentlywiderthanothervesselsatthesamebranchlevel,thevesselwasdeemedasabnormalvascularbroadening.Thevesselswereconsideredtobedistortedorrigidiftravelingastrayfromtheexpectednormalcourse.13 Multiplesupplyingvessels,wi
16、thdifferentoriginatingsources,convergingtowardalesion,wereprobablyindicativeofanincreasedbloodcirculationwithin.Tofurtherclarifyaffiliationsofsupplyingvessels,wetracedvascularcoursesslice-wisebackwardtomajorvesselsinthehilum.TherelationshipsbetweentheGGNsandsupplyingbloodvesselswereanalysedinaxialim
17、ages,MPRimagesCPRimages.14theGGN-vesselrelationshipswerecategorizedintofourtypesaccordingtoimagingfeatures:typeI(pass-by),vesselspassedbyGGNswithoutdetectablesupplyingbranchestolesions.typeI15typeII(pass-through),vesselspassedthroughthelesionswithoutobviousmorphologicalchangesintravelingpathorsize.1
18、6typeIII(distorted/dilated),vesselswithinlesionsweretortuousorrigidwithoutanincreaseinamount17typeIV(complicated),morecomplicatedvasculatureotherthandescribedintheaforementionedtypeswithinGGNs,forinstance,coexistenceofirregularvasculardilationandvascularconvergencefrommultiplesupplyingvessels.18Path
19、ologicalanalysis ThepathologicaldiagnosisandcategorisationofAAH,AIS,MIAandIACweremadebasedonthenewpulmonaryadenocarcinomaclassification,2011edition.GGNswereresectedbyvideo-assistedthoracoscopyorthoracotomysurgery.Allhistologicalpreparationsandanalyseswereperformedbytwoseniorpathologists.Inthecaseofd
20、isagreements,aconsensuswasreachedaftermutualdiscussionand/orconsultationwithathirdpathologist.19StatisticalanalysisSPSS16.0forWindows,SPSS,Chicago,IllIndependentttestwasusedtocomparedifferentpathologicalgroups(benigndiseases,preinvasivediseasesandinvasiveadenocarcinoma)ofGGN.Correlationsbetweenpatho
21、logicalfindingsofGGNsandGGN-vesselrelationshipswereexaminedusingSpearmansranktest.GGN-vesselrelationshipsbetweenMIAandIACdiseaseswerecomparedusingPearsonschi-squaredtest.Whentherewasanexpectedvalue1orapretestprobabilityclosetothetestlevel,Fishersexacttestwasusedinstead.Statisticalresultswereconsider
22、edsignificantwhenthePvaluewaslessthan0.05.20ResultsSize variation of GGN lesions TheaverageGGNsizeinthebenigngroup,preinvasivegroupandadenocarcinomasgroupwas8.12.5mm,9.35.6mmand14.86.0mm,respectively.Nosignificantdifferencesexistedbetweenthepreinvasivegroupandthebenigngroup(t=0.64,p=0.53).However,th
23、ereweresignificantdifferencesbetweenbenignandpreinvasivegroupsandtheinvasiveadenocarcinomagroup(t=6.31,p=0.00;t=3.98,p=0.00).21 Correlations between GGN-vessel relationships and pathological findings Of108GGNs,typeI,II,IIIandIVGGNvessellrelationshipswereobservedin9,58,21and20cases,respectively.thety
24、peIIGGN-vesselrelationshipwasthedominantrelationshipineachpathologicalgroup,seenin9benign(90.0%),16preinvasive(66.7%)and33invasive(44.6%)GGNcases.22comparedwiththelowincidenceoftypeIIIandIVrelationshipsinbenignandpreinvasivegroupsthecombinedincidenceoftypeIII(25.7%)andIV(25.7%)relationshipsintheinva
25、siveadenocarcinomagroupreached51.3%.23MIA couldpresentfourtypes,withtypeIIasthemajortype(48.7%).ThecombinationoftypeIIandIVcomprised about80%oftheMIAsubgroup;forIAC,typeIIandIII hadthesameproportionof40%,hencethecombinationof 80%.StatisticalstudiesshowednodifferenceintypeIIbuta significantdifference
26、wasfoundintypeIIIandIVbetween MIAandIAClesions(p=0.02).24Thevessel(s)travelingthroughGGNcouldbeartery(ies)(categoryA),vein(s)(categoryB),orartery(ies)andvein(s)(categoryC).TherewerenosignificantdifferencesandcorrelationsbetweenvascularcategoriesandGGNgroups(p=0.50and0.96,respectively).25Afurtherexam
27、inationofthecorrelationbetweenvascularcategoriesandGGNswithtypeIIIandIVrelationshipsdidnotgenerateanysignificantresults(p=0.70).26Discussion Solitarypulmonarynodules(SPNs)arecommonfindingsinCTexaminationsandcanbedividedintotwogroupsbasedondensityvariation:solidnodulesandGGNs.In2011,theInternationalA
28、ssociationfortheStudyofLungCancer,theAmericanThoracicSocietyandtheEuropeanRespiratorySocietyproposedanewclassificationforlungadenocarcinomas.Inthenewclassificationsystem,thetermbronchioloalveolarcarcinoma(BAC)isnolongerused.TheformerBACconceptapplicabletomultiplecategoriesinthenewclassificationsyste
29、m,suchasAIS,MIAandthemucinoussubtypeofadenocarcinoma.BothAISandAAHlesionsareclassifiedaspreinvasiveadenocarcinomaunderthenewclassificationsystem27 EarlystagelungcancersoftenpresentasGGNsinCTimages;thus,itisimportanttobefamiliarwiththecharacteristicsofGGNswithmalignantpotential,astimelysurgicalresect
30、ionwillimprovepatientsurvivalandqualityoflife,andforpatientswithbenignGGNs,unnecessarysurgicalprocedurescanbeavoided.28 Clinicaldatahaveshownthatnodulesizeisanindependentpredictivefactorofmalignancy,withsizeincreasingthelikelihoodofmalignancyincreasing,consistentwithourresultsthatthemeansizesofGGNsi
31、nbenign,preinvasiveandadenocarcinomagroupswere8.1mm,9.3mmand14.7mm.Clinicalexperiencehasdemonstratedthatsomecommonimagingfeaturesofmalignantnodules,suchaspleuralindentation,spiculationandlobulation,areseldomseeninveryearlystagemalignantGGNs.29 Thisdemandsfurtherinvestigationofthisparticularabnormali
32、magingfindingtominimisemisdiagnosis.InthemanagementofGGNsinourpatients,clinicalguidelinesfromtheFleischnerSocietyandNationalComprehensiveCancerNetwork(NCCN)werereferenced.Eachindividualcasewasdiscussedbyamultidisciplinaryteam,includingdiagnosticradiologists,thoracicsurgeonsandpathologists,togenerate
33、consequentmanagementstrategies.Allpatientsreceivedadequatefollow-upobservationwith/withoutsupportiveorantiinflammatorytreatment,whichexplainedthefactthatfourGGNsdisappearedpriortothenextscheduledCTexamination.30 Exceptforthesefourcaseswithoutbiopsy,nodularlesionsintheremaining104patientsweresurgical
34、lyremovedbecauseofthecontinuousincreaseinsizeand/ormassonfollow-upimagingstudies.ConsideringthedramaticallyincreasingincidenceoflungcancerinChina,patientsandphysiciansareveryalerttoitandthetreatmentmightbemoreaggressivethaninWesterncountries.31 Tumourbiologystudieshaverevealedthatvasculatureremodell
35、ingorneoangiogenesisisoneoftheinitiatingeventsoccurringintheearlystageoftumourdevelopment.Therefore,analysisofGGNsandrelatedbloodsupplyingvesselscouldprovideinformationonGGNdifferentiation.SmallbloodvesselsandtherelationshipsbetweenvesselsandlesionscanbereadilyrevealedandevaluatedinCTimagesacquiredw
36、ithmodernmulti-detectorscanners,especiallywhenimagingdataarepost-processedusingadvancedcomputertechniques,includingMPRandCPR.ManystudieshavedemonstratedthatrelationshipsbetweenSPNsandvessels,especiallythevascularconvergencesign(VCS),arevaluableforestimationofthemalignancypotentialofSPNs32 Somestudie
37、sindicatedthatdiseaseprogressionfromAAH,AIS,MIAtoIACisacomplicated,polygene-involveddynamicprocess.MIAorIACmaygraduallydevelopfromAAHandAIS.InterstitialfibrehyperplasiawithinlesionsisthemaincontributingfactortotypeIIIandIVvascularmorphologicalchanges.theformationmechanismofVCS,leadingtotheconclusion
38、thatthecourseofadjacentvesselsissubjecttolesions,especiallywhendiseasesinfiltratethebronchiovascularbundleandinterlobularsepta33 Asaresult,involvedvesselsmightappeardistorted,rigidorconcentratedtowardsthelesion.Thus,itisreasonabletopostulate假设thatthevascularconvergencesigncommonlyseeninSPNs.Actually
39、,thetypeIVGGN-vesselrelationshipresemblesVCStosomedegree.Theinvasiveadenocarcinomagroupiscomposedoftwosubgroups,MIAandIAC.SubgroupanalysisshowedMIAandIAChaddifferentpatternsofGGN-vesselrelationships.34 TypeIIIvascularmorphologicalchangeswereobservedmoreoftenintheIACthanMIAsubgroup,indicatingthatwith
40、increasingmalignancy,fibrehyperplasiastimulatedbymalignanttissuesmaybecomemoresevere,andsubsequentlyimpactsonvasculaturebecomeaggravated.Furthermore，tumourmetabolismisfasterthaninnormaltissues;therefore,thebloodsupplydemandedbytumoursismuchhigherthaninnormaltissues.Thesemechanismsindirectlyleadtoves
41、selproliferationandirregularluminaldilation.35 Somestudieshaveshownthatendogenousand/orextrinsictumorangiogenesisandneovascularisationcouldbethedrivingfactorsofvascularabnormalitiesobservedinmalignantearlystage.AsaCTimagingsign,VCSdescribesarelationshipbetweenSPNsandvessels,oneormultiplevesselsconce
42、ntratingtowardsandpassingthroughlesionsorbeingtruncatedattheedgeoflesions.36 Involvedvesselsmayappeartortuous,rigidorirregularlywideningandlinktopulmonaryarteriesorpulmonaryveins.Inthisstudy,theGGN-vesselrelationshipswerecategorizedintofourtypes.Statisticalanalysisindicatedthatwhentherelationshipwas
43、typeIIIorIV,especiallytypeIV,itwashighlylikelythatGGNsweremalignantinvasiveadenocarcinoma,withMIAmorethanIAC.Incontrast,themajorityofbenignandpreinvasivecaseswasseenintypeIortypeIIGGN-vesselrelationships.37 Amajordrawbackofthisstudyisthelimitednumberofcases,especiallyinthebenigngroup,whichmaycomprom
44、isethediagnosticpower.Hence,aprospectiveclinicaltrialwithmoreGGNcasesiswarrantedtofurtherevaluateandvalidatethediagnosticvalueoffindingsinthisstudy.Additionally,thisstudycouldbestrengthenediftheanalysiswereconductedwithacombinationofvesseltypesandotherGGNfeatures,suchassizeandmass.Massmeasurementsca
45、nreflectlesiongrowthearlierwithlessvariabilitythandiametermeasurements.38 Inconclusion,thisstudydemonstratesthatdifferentGGNsmighthavedifferentrelationshipswithvesselsduetovariationindevelopmentalbiologyandbehaviour.UnderstandingandrecognizingGGN-vesselrelationshipsinCTimagingandthestrongcorrelationbetweeninvasiveadenocarcinomaandtypeIIIandIVrelationshipsmayhelpidentifywhichGGNsaremorelikelytobemalignant.39
