Prognostic impact of serum CYFRA 21–1 in patients with advanced lung adenocarcinoma a retrospective

内容需要下载文档才能查看 内容需要下载文档才能查看

Onoetal.BMCCancer2013,13:354

http://wendang.chazidian.com/1471-2407/13/354

PrognosticimpactofserumCYFRA21–1in

patientswithadvancedlungadenocarcinoma:aretrospectivestudy

AkiraOno1*,ToshiakiTakahashi1,KeitaMori4,HiroakiAkamatsu1,TakehitoShukuya1,TetsuhikoTaira1,

HirotsuguKenmotsu1,TateakiNaito1,HaruyasuMurakami1,TakashiNakajima2,MasahiroEndo3

andNobuyukiYamamoto1

*Correspondence:a.ono@scchr.jp1DivisionofThoracicOncology,ShizuokaCancerCenter,1007,

Shimonagakubo,Nagaizumi-cho,Sunto-gun,Shizuoka411-8777,Japan

Fulllistofauthorinformationisavailableattheendofthe

article

©2013Onoetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative

CommonsAttributionLicense(http://wendang.chazidian.com/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and

reproductioninanymedium,providedtheoriginalworkisproperlycited.

Onoetal.BMCCancer2013,13:354

http://wendang.chazidian.com/1471-2407/13/354

Background

Lungcanceristheleadingcauseofcancerdeath,andatpresent,thereexistsnocureofstageIVnon-smallcelllungcancer(NSCLC)[1].Adenocarcinomaandsqua-mouscellcarcinomaarethemostcommonhistologicalsubtypesoflungcancerandaccountforabout70%ofalllungcancers[2].Thefolateantagonistpemetrexedhasbeenshowntoexhibitefficacyagainstnon-squamouscelllungcancers[3],andiscurrentlyusedincombinationwithcisplatinasastandardtreatmentregimenforpatientswithnon-squamouscelllungcarcinoma.Chemotherapywiththeangiogenesisinhibitorbevacizumabadministeredincombinationwithplatinumagentshasalsobeenshowntoexhibitfavorableefficacyagainstnon-squamouscelllungcarcinoma[4,5].Somaticgain-of-functionmutationsinexonsencodingtheEGFRtyrosinekinasedomainhavebeenidentifiedinNSCLC[6,7].SeveralpreviousstudieshavereportedprolongationofthesurvivaltimeinpatientswithEGFR-mutation-positivelungcarcinomastreatedwithEGFR-tyrosinekinaseinhibitors(TKIs)[8-11],there-fore,EGFR-TKIsarewidelyusedinmedicalpractice.EGFRmutationsoccurmorefrequentlyinlungcancerpa-tientswhoareAsians,femalesandnon-smokerswiththehistologicalsubtypeofadenocarcinoma[12-14].Ontheotherhand,whiletherehavealsobeenscatteredreportsofEGFRmutationsamongcasesoflungsquamous-cellcar-cinoma[15-17],arecentreportshowedthattherewerenoEGFRmutation-positivecasesamonglungcancerpa-tientswithpuresquamouscellcarcinoma[18,19].

CYFRA21–1isafragmentofcytokeratin(CK)19.CKs,whicharenowcalledkeratins,aretheprincipalstructuralelementsofthecytoskeleton(keratinfila-ments)ofepithelialcells,includingbronchialepithelialcells,andhavebeenclassifiedinto20subtypesbasedondifferencesinthemolecularmassandisoelectricpointasdeterminedby2-dimensionalelectrophoresis[20,21].CKtypes1–8arecategorizedastypeICKs,andCKs9–20astypeIICKs.MicrofilamentsareheteropolymersformedfromtypeIandtypeIIkeratins,andconstitutethecytoskeleton[22].CK19isasolubletypeICK(acidictype),andhasthelowestmolecularmass(40kDa)amongtheCKs.Itisexpressedintheunstratifiedorpseudostratifiedepitheliumliningthebronchialtree[23],andbeenreportedtobeoverexpressedinmanylungcancertissuespecimens[24].TheCKexpressionpatternsintissuesarewell-maintainedevenduringtheprocessoftransformationofthetissuefromnormaltotumortissue[25].AcceleratedCK19degradationoccursinneoplasticallytransformedepithelialcellsasaresultofincreasedproteaseactivityofcaspase3,aregulatoroftheapoptosiscascade,andfragmentsarereleasedintotheblood.ThisresultsinanincreaseofthebloodCYFRA21–1values,becauseCK19fragmentsarerecognizedbytwomonoclonalantibodies[26].

Page2of10

MeasurementofserumCYFRA21–1levelisausefulauxiliarytestinthediagnosisofNSCLC,andparticularlyhighspecificityofthistesthasbeenreportedforthediagnosisofsquamouscellcarcinomaofthelung[27,28].Ontheotherhand,ameta-analysisalsorevealedthatserumCYFRA21–1maybeausefulprognosticfac-torinNSCLCpatients[29];analysisofthehistologicalbackgroundintheaforementionedmeta-analysisshowedthatnon-adenocarcinomaaccountedforthemajorityofcasesofNSCLC(65%).TherehasalsobeenareportsuggestingthatserumCYFRA21–1levelsmightserveasaprognosticfactorinpatientswithrecurrentNSCLCreceiving3rd-lineorlatergefitinibtherapy[30].SomestudieshavesuggestedthepossibleprognosticvalueofpretreatmentserumCYFRA21–1values(PCV)inpa-tientswithsurgicallytreatedlungadenocarcinoma[31]andadvancedNSCLC[32-34].However,noneofthestudiessuggestingserumCYFRA21–1asaprognosticfactorinpatientswithuntreatedadvancedlungadeno-carcinomahasincludedtheEGFRmutationstatusasavariable.Therefore,inthepresentstudy,weinvestigatedtheimpactofserumCYFRA21–1ontheprognosisofuntreatedadvancedlungadenocarcinomapatients.

Methods

Patients

OfpatientsdiagnosedashavingprimarylungcarcinomabetweenJanuary2003andJune2010attheShizuokaCancerCenter,EGFRmutationanalysiswasperformedon424patientsfromApril2008toJune2010.Ofthese,284lungadenocarcinomapatientshadreceivedinitialtherapy,andweretrospectivelyreviewedthedataofthe163patientswhowerefoundtoharborwild-typeEGFRand121patientswhowerefoundtoharboractivatingEGFRmutations(Figure1).Thefollowinginclusioncri-teriaweresetforthisstudy;patientswithpathologicallyprovenadenocarcinomawhohadreceivedinitialtherapy(includingchemotherapyorchemoradiotherapy)andsurvivedformorethanonemonth;EasternCooperativeOncologyGroupperformancestatus(ECOGPS)of3orless.ThehistologicalandcytologicaldiagnoseswereperformedaccordingtotheWHOclassificationcriteria[35].ThestudywasconductedwiththeapprovaloftheShizuokacancercenterInstitutionalReviewBoard#1(HHSIRBregistrationnumber;IRB00006744).

Weoutsourcedsomeoftheclinicallaboratorytests,suchasmeasurementofthetumormarkersandEGFRmutationanalysis.SerumCYFRA21–1andserumCEAconcentrationsweremeasuredatthebaseline,beforetheinitialtherapy.TheserumCYFRA21–1concentrationwasmeasuredusingaLumipulsePresto®kit(FUJIREBIOInc,Tokyo,Japan),basedonaCLEIA(chemiluminescentenzymeimmunoassay)method,whiletheserumCEAconcentrationsweremeasuredusinganARCHITECT®kit

Onoetal.BMCCancer2013,13:354

http://wendang.chazidian.com/1471-2407/13/354

Page3of10

(AbbottJapan,Tokyo,Japan).EGFRmutationanalysiswasperformedbyfragmentanalysisusingpolymerasechainreaction(PCR)andthecycleavereal-timequantitativePCRtechnique(SRLInc,Tokyo,Japan).

ThereportedupperlimitofnormalforthediagnosisofNSCLCandupperlimitofthepercentilesforhealthyindividualsofserumCYFRA21–1asmeasuredbyEIAare3.5ng/mland2.8ng/ml,respectively[36].Incon-trast,thereportedupperlimitofthepercentilesforhealthyindividualsofserumCYFRA21–1measuredbytheCLEIAmethodis1.6ng/ml[37],alowervalueascomparedtothatsetformeasurementbytheEIAmethod.Therefore,forourstudy,wesetthecutoffvalueforCYFRA21–1at2.2ng/ml,basedonthemeanvalueforhealthysubjects+3SD[37],alowervalueascom-paredtothatsetformeasurementbytheEIAmethod.ThecutoffvalueforserumCEAwassetat5.0ng/ml,whichistheupperlimitofnormal.

Astandardevaluationofthepatients,includingassess-mentofthemedicalhistory,physicalexaminationandroutinelaboratorytests,wasperformedbeforeeachtreat-ment.AllpatientswerestagedbasedontheInternationalAssociationfortheStudyofLungCancer(IASLC)TNM(tumor-node-metastasis)classification,7thedition[38].

Statisticalmethods

survivalwasmeasuredfromthedateofthefirstcourseoftheinitialtherapytothedateofdeathorthatofthelastfollow-upexamination.Alog-ranktestwasperformedtoevaluatethesignificanceofdifferencesintheoverallsurvivalamongthegroups.Pvalues<0.05wereconsid-eredtobeindicativeofstatisticalsignificance.Amultivari-ateanalysisusingtheCoxproportionalhazardsmodelwasusedtoestablishtheassociationbetweentheclinicalvariablesandsurvival.AllstatisticalanalyseswerecarriedoutusingSPSS,version11.0forWindows(SPSSInc.,Chicago,IL,USA).Toreducethepotentialbiasarisingfromsomepatientsdyingtooearlytoreceiveinitialther-apy,thetwopatientswhodiedwithinamonth(30days)ofthestartofinitialtherapywereexcludedfromtheanalysis.

Therewerenomissingdatainourstudy.SurvivalwasestimatedusingtheKaplan-Meiermethod.Overall

Results

Thecohortconsistedof284patientswhowerediag-nosedashavingstageIIIBorIVlungadenocarcinomaandhadreceivedinitialtherapy.

Theclinicalcharacteristicsofthepatientsaresumma-rizedinTable1.Themedianpatientagepriortothestartofinitialtherapywas65years(range,23to87years).Thepatientswerepredominantlyyoungerthan70yearsofage(81%),theECOGPSwas0–2in93%ofpa-tients,and91%ofthepatientshadstageIVdisease.WhilethelungadenocarcinomapatientswithEGFRmutationswerepredominantlyfemale(64%)and

non-smokers

Onoetal.BMCCancer2013,13:354

http://wendang.chazidian.com/1471-2407/13/354

Page4of10

Table1Patientcharacteristics

CharacteristicAge,yearsMedian(range)<70≥70

24

GenderMaleFemaleECOGPS0-1>2SmokingstatusYesNoStageIIIBIV

EGFRmutationExon19deletionExon21L858RExon18G719XWildtypePCV

Median(range)<2.2ng/ml≥2.2ng/mlCEA

Median(range)<5.0ng/ml≥5.0ng/ml

8(0.7-11942)4576

3763

7(0.5-14985)63100

3961

7.4(0.5-14985)108176

3862

1.6(0.1-110.0)7249

6040

2.3(0.1-80.0)7885

4852

2.0(0.1-110.0)150134

5347

59575

49474

163

100

59575163

2120257

6115

595

19144

1288

25259

991

5071

4159

12439

7624

174110

6139

10318

8515

13528

8317

23846

8416

4378

3664

12538

7723

168116

5941

20

29

18

53

19

66(32–87)97

80

65(23–83)134

82

65(23–87)231

81

Mt+(n=121)No.

%

Mt–(n=163)No.

%

All(n=284)No.

%

EGFR:epidermalgrowthfactorreceptor,Mt+:mutantEGFR,Mt-:wild-typeEGFR,PCV:pretreatmentCYFRA21–1value.

(71%),thosewithwild-typeEGFRwerepredominantlymale(77%)andsmokers(76%).

Detailsaboutthefirst-linechemotherapywereavailablefor284patientsincludingbothpatientgroupswithwild-type(Mt-)andmutantEGFR(Mt+)groups(Table2).About40%oftheEGFRmutation-positivepatientsre-ceivedEGFR-TKIsastheinitialtreatment.

Carboplatin-paclitaxel,thetreatmentofchoiceacrossbothgroups,wasadministeredtohalfoftheplatinumdoubletcohortintheMt-patientgroup.Meanwhile,docetaxelwasadministeredtohalfofthemonotherapycohortinthesamepatientgroup.However,cisplatin-pemetrexedwasthemostcommonregimenofsecondchoiceacrossboththeMt+andMt-groups.TheEGFR-TKIusedforeachtreatmentlineintheMt+groupisshowninTable3.Forty-one(58%)patientsre-ceivedgefitinib,while16(22%)receivederlotinibasfirst-orsecond-linetreatmentintheMt+groupwithPCV(<2.2ng/ml).Thirty-seven(73%)patientsreceivedgefitinib,and10(20%)patientsreceivederlotinibasfirst-orsecond-linetreatmentintheMt+groupwithPCV(≥2.2ng/ml).Ofthe121patientsintheMt+group,27didnotreceivegefitinibatanytreatment-linestageoftreat-ment;amongthese27patients,19receivederlotinib(6asfirst-line,10assecond-line,1asthird-lineand2asfurther-linetreatment).IntheMt+group,atotalof113patients(93%)receivedEGFR-TKIs,while8patientsdidnotreceiveEGFR-TKIsatanystageoftreatment.

Onoetal.BMCCancer2013,13:354

http://wendang.chazidian.com/1471-2407/13/354

Table2Summaryofinitialtreatmentdeliveredamong284patients

EGFRmutation

Mt–(n=163)Mt+(n=121)IIIB

IV

IIIB

IV(n=19)(n=144)(n=6)(n=115)

No.%No.

%No.%No.

%TreatmentPlatinumdoublet43

1147022

5445Monotherapy03018

0119EGFR-TKI

0005041

Chemoradiotherapy15

90

4

30SpecificregimensCisplatin-pemetrexed12415197Carboplatin-paclitaxel35232

02722Carboplatin-paclitaxel+bev0

202Otherplatinumdoublets0362211210Gefitinib0004134Erlotinib00076

Docetaxel0161003Vinorelbine0502Others

24

150

6

Mt+:mutantEGFR,Mt-:wild-typeEGFR,bev:bevacizumab.

Furthermore,ofthe160patientsintheMt-group,30pa-tientsreceivedEGFR-TKIs(11assecond-line,7asthird-line,6asfourth-line,3asfifth-line,1assixth-line,1asseventh-line,and1aseighth-linetreatment).Fifty-threepatients(18%)werestillaliveatthetimeoftheanalysis.Themedianfollow-upperiodfordeterminingthesurvivalwas39.3(range;11.8-84.9)monthsafterthestartofinitialtherapy.Theclinicalvariablesidentifiedbyunivariateana-lysistobeassociatedwithsignificantlybettersurvival(Table4)includedfemalegender(MST32.4monthsver-sus20.1monthsinmales:p=.0086),nosmokinghis-tory(33.4monthsversus20.1monthsinsmokers,p=.0012),ECOGPS(0–1)(29.5monthsversus7.9months

Table3SummaryofEGFR-TKIdeliveredamongEGFRmutationpositivepatients

EGFRmutationpositiveLowPCVHighPCV(<2.2ng/ml)(n=72)(≥2.2ng/ml)(n=49)GefitinibErlotinibGefitinibErlotinibNo.

%No.%No.%No.%First-line202857234724Second-line212911151429816Third-line912683648Further-line2314202448Unadministered

20

28

36

50

7

14

31

63

PCV:pretreatmentCYFRA21–1value.

Page5of10

inthosewithaPSof2–3,p<.0001),presenceofEGFRmutation(39.2monthsversus17.8monthsinpatientswithoutEGFRmutations,p<.0001),PCV<2.2ng/ml(38.6monthsversus15.0monthsinthosewithPCV≥2.2ng/ml,p<.0001),serumCEA<5.0ng/ml(32.6monthsversus21.0monthsinthosewithserumCEA≥5.0ng/ml,p=.036),startdateofinitialtherapybeforeApril1,2008(34.1monthsversus19.3monthsinthegroupthatre-ceivedtheinitialtherapyafterApril1,2008,p=.003)andEGFR-TKItreatment(33.7monthsversus15.3monthsinthegroupnottreatedwithEGFR-TKIs,p<.0001).Multi-variateanalysisidentifiedEGFRmutationpositivity(HR0.53;95%CI:0.34-0.84,p=.0069)andPCV<2.2ng/ml(HR0.43;95%CI:0.31-0.59,p<.0001)asindependentfa-vorableprognosticfactors.Anotherfactorthatwasfoundtobeanindependentprognosticindicatorofoverallsur-vivalwasthePS(Table4).Theoverallsurvivalratesofpa-tientswithadvancedlungadenocarcinomawith/withoutEGFRmutationareshowninFigure2.AmongtheMt+patients,theprognosiswasmorefavorableinthegroupwithPCV<2.2ng/ml(n=70)thaninthegroupwithPCV>2.2ng/ml(n=48)(mediansurvivaltime[MST]:67.0vs.21.0months,p<0.0001).AmongthepatientswithMt-also,theprognosiswasmorefavor-ableinthegroupwithPCV<2.2ng/ml(n=78)thaninthegroupwithPCV≥2.2ng/ml(n=86)(MST:24.1vs.10.2months,p<0.0001).

Discussion

Inthepresentstudy,wedemonstratedPCVandEGFRmutationstatusasindependentprognosticfactorsinun-treatedadvancedlungadenocarcinomapatients.WealsoshowedthatPCV<2.2ng/mlwasapredictorofafavor-ableoutcomeinbothadvancedlungadenocarcinomapatientswithwild-typeandmutantEGFR.

SerumCYFRA21–1hasbeenreportedasaprognosticfactorinpatientswithavarietyofcancertypes,includ-ingresectableNSCLC[39,40],biliarytractcancer[41],urothelialcancer[42],headandneckcancer[43],esophagealcancer[44],andcervicalcancer[45].

Ameta-analysisofCYFRA21–1asaprognosticindi-catorinadvancedNSCLCpatientsshowedthatthePCVmaybeareliableprognosticfactor[29],however,sincenon-adenocarcinomaaccountedfor65%ofthecasesandsquamouscellcarcinomafor50%,theroleofserumCYFRA21–1asaprognosticindicatorinthelungadenocarcinomapopulationremainedunclear.More-over,inastudyofPCVasaprognosticindicatorinad-vancedNSCLCpatientsinwhomgefitinibwasusedas3rd-lineorlatertherapy,adenocarcinomaaccountedforfewerthanahalfofthecases(47%)[30].TheEGFRmu-tationstatuswasnotincludedasavariableintheana-lysis,andthetestpopulationwassmall,consistingofonly50patients.