Colorectal and Other Cancer Risks

上传者：李小滨
|
上传时间：2015-05-05
|
密次下载

Colorectal and Other Cancer Risks

结肠癌流行病学

VOLUME30?NUMBER9?MARCH202012

JOURNALOFCLINICALONCOLOGY

ORIGINALREPORT

ColorectalandOtherCancerRisksforCarriersand

NoncarriersFromFamiliesWithaDNAMismatchRepairGeneMutation:AProspectiveCohortStudy

AungKoWin,JoanneP.Young,NoralaneM.Lindor,KatherineM.Tucker,DennisJ.Ahnen,GraemeP.Young,DanielD.Buchanan,MarkClendenning,GrahamG.Giles,IngridWinship,FinlayA.Macrae,JackGoldblatt,MelissaC.Southey,JulieArnold,StephenN.Thibodeau,

ShanakaR.Gunawardena,BharatiBapat,JohnA.Baron,GrahamCasey,StevenGallinger,LoïcLeMarchand,PollyA.Newcomb,RobertW.Haile,JohnL.Hopper,andMarkA.Jenkins

Authoraf?liationsappearattheendofthisarticle.

SubmittedSeptember25,2011;acceptedNovember22,2011;http://wendang.chazidian.comonFebruary13,2012.Thecontentofthisarticledoesnotnecessarilyre?ecttheviewsorpoliciesoftheNationalCancerInstituteoranyofthecollaboratingcentersintheCancerFamilyRegistries(CFRs);inaddition,mentionoftradenames,commercialproducts,ororganizationsdoesnotimplyendorsementbytheUSGovernmentortheCFR.Authorshadfullresponsibilityforthedesignofthestudy,thecollectionofthedata,theanalysisandinterpretationofthedata,thedecisiontosubmitthemanuscriptforpublication,andthewritingofthearticle.

Authors’disclosuresofpotentialcon-?ictsofinterestandauthorcontribu-tionsarefoundattheendofthisarticle.

Correspondingauthor:MarkA.Jenkins,PhD,CentreforMolecular,Environmen-tal,GeneticandAnalyticEpidemiology,MelbourneSchoolofPopulationHealth,207BouverieSt,Level3,TheUniver-sityofMelbourne,Victoria3010,Australia;e-mail:m.jenkins@unimelb.edu.au.

0732-183X/12/3009-958/$20.00DOI:10.1200/JCO.2011.39.5590

ABSTRACT

Purpose

Todeterminewhethercancerrisksforcarriersandnoncarriersfromfamilieswithamismatchrepair(MMR)genemutationareincreasedabovetherisksofthegeneralpopulation.

PatientsandMethods

Weprospectivelyfollowedacohortof446unaffectedcarriersofanMMRgenemutation(MLH1,n?161;MSH2,n?222;MSH6,n?47;andPMS2,n?16)and1,029theirunaffectedrelativeswhodidnotcarryamutationevery5yearsatrecruitmentcentersoftheColonCancerFamilyRegistry.Forcomparisonofcancerriskwiththegeneralpopulation,weestimatedcountry-,age-,andsex-speci?cstandardizedincidenceratios(SIRs)ofcancerforcarriersandnoncarriers.Results

Overamedianfollow-upof5years,mutationcarriershadanincreasedriskofcolorectalcancer(CRC;SIR,20.48;95%CI,11.71to33.27;P?.001),endometrialcancer(SIR,30.62;95%CI,11.24to66.64;P?.001),ovariancancer(SIR,18.81;95%CI,3.88to54.95;P?.001),renalcancer(SIR,11.22;95%CI,2.31to32.79;P?.001),pancreaticcancer(SIR,10.68;95%CI,2.68to47.70;P?.001),gastriccancer(SIR,9.78;95%CI,1.18to35.30;P?.009),urinarybladdercancer(SIR,9.51;95%CI,1.15to34.37;P?.009),andfemalebreastcancer(SIR,3.95;95%CI,1.59to8.13;P?.001).Wefoundnoevidenceoftheirnoncarrierrelativeshavinganincreasedriskofanycancer,includingCRC(SIR,1.02;95%CI,0.33to2.39;P?.97).

Conclusion

Wecon?rmedthatcarriersofanMMRgenemutationwereatincreasedriskofawidevarietyofcancers,includingsomecancersnotpreviouslyrecognizedasbeingaresultofMMRmutations,andfoundnoevidenceofanincreasedriskofcancerfortheirnoncarrierrelatives.JClinOncol30:958-964.©2012byAmericanSocietyofClinicalOncology

INTRODUCTION

Lynchsyndromeisanautosomaldominantinher-iteddisorderofcancersusceptibilitycausedbygermlinemutationsintheDNAmismatchrepair(MMR)genesMLH1,MSH2,MSH6,andPMS2.Estimatesofcarrierfrequencyofgermlinemuta-tionsofthesegenesinthepopulationrangewidelydependingonvariousassumptions,fromapproxi-matelyonein3,010individuals(forMLH1andMSH2combined1)toonein360individuals(forallfourMMRgenescombined2).Mutationcarriersareatsubstantiallyincreasedriskofcancersofthecolon,rectum,endometrium,stomach,ovary,ureter,renalpelvis,brain,smallbowel,andhepatobiliarytract,

andthediagnosesofthesecancersgenerallyoccuratyoungeragesthanforthegeneralpopulation.3Esti-matesofsite-speci?ccancerrisksforMMRgenemutationcarriersinformoptimalclinicalmanage-ment.Screeningcolonoscopy,4,5prophylactichys-terectomy,andbilateralsalpingo-oophorectomy6havethepotentialtodecreasetheriskofcolorectalcancer(CRC),endometrialcancer(EC),andovar-iancancer,respectively,forMMRgenemutationcarriers.AllstudiesestimatingtheriskofcancerforMMRgenemutationcarriershavebeenretrospec-tiveand,therefore,maybebiasedasaresultofdif-ferentialrecalloffamilyhistoryofcancerandfailuretoadjustforascertainmentofsubjectsrecruitedbe-causeofafamilyhistoryofcancer.7Riskestimates

Information downloaded from http://wendang.chazidian.com and provided by at Fudan Univ Lib on September 16, 2014 from

结肠癌流行病学

CancerRisksforMMRGeneMutationCarriersandNoncarriers

usingprospectivedataofmutationcarrierswithnopriordiagnosisofcancerwillnotbebiasedbutarechallengingbecausetheyrequirelong-termfollow-uptoprovidesuf?cienttimeforcancerstobediagnosed.

Genetictestingoffamilymembersofamutationcarrierwillidentifyrelativeswhohaveandwhohavenotinheritedthefamily-speci?cMMRgenemutation.ItisnotknownwhetherthecancerriskfornoncarriersfromfamilieswithMMRgenemutationsisgreaterthanthatofthegeneralpopulation.AfamilyhistoryofCRCincreasesanindividual’srisk1.5-toeight-fold,varyingwiththenumberandagesofaffectedrelativesanddegreeofrelationshiptotheproband.8InLynchsyndromefamilies,itispossiblethatmodi?ergenesparticipateinthemilieuinwhichcancermanifests.9-16NoncarriersofMMRgenemutationsmaysharepredisposinggeneticriskwiththeirmutation-carryingrelativesotherthanthatcausedbytheMMRgenemutationandthereforemaybeatincreasedriskcomparedwiththegeneralpopulation.If,however,theMMRmutationaccountsforalltheexcesscancerrisksinthesefamilies,thennoncarriersshouldbeatpopulationrisk.Inthisstudy,weestimatedcancerrisksformutationcarriersandnoncarriers,whohadnopriordiagnosisofcancer,fromfamilieswithMMRgenemutationsfromtheColonCancerFamilyRegistry,usingaprospectivecohort.

PATIENTSANDMETHODS

ColonCancerFamilyRegistry

Thisstudy,usingtheColonCancerFamilyRegistry,includedcarriersofpathogenicMMRgenemutationsandtheirnoncarrierrelatives.DetailsoftheColonCancerFamilyRegistryhavebeenpublishedpreviously17andcanbefoundattheNationalCancerInstituteWebsite(http://epi.grants.cancer.gov/CFR/).Familieswererecruitedbetween1997and2010andwereascertainedviaCRCcasesidenti?edfrompopulationcancerregistriesintheUnitedStates(Washington,California,Arizona,Minnesota,Colorado,NewHampshire,NorthCarolina,andHawaii),Australia(Victoria),andCanada(Ontario)orfromfamilycancerclinicsintheUnitedStates(MayoClinic,Rochester,MN,andClevelandClinic,Cleveland,OH),Australia(Melbourne,Adelaide,Perth,Brisbane,andSydney),andNewZealand(Auckland).Writteninformedcon-sentwasobtainedfromallstudyparticipants,andthestudyprotocolwasapprovedateachcenter.

DataCollection

Atrecruitment,baselineinformationondemographics,personalchar-acteristics,personalandfamilyhistoryofcancer,cancerscreeninghistory,historyofpolyps,polypectomy,hysterectomy,andothersurgerieswasob-tainedfromallparticipants.Thisparticipantinformationwasupdatedap-proximately5and10yearsafterbaseline.Reportedcancerdiagnosesandagesatwhichtheseoccurredwerecon?rmed,wherepossible,usingpathologyreports,medicalrecords,cancerregistryreports,and/ordeathcerti?cates.Thelocation,histology,andbehaviorofcancerdiagnoseswerecodedusingInter-nationalClassi?cationofDiseasesforOncology(ICD-O).18Bloodandtumortissuesampleswerecollectedforgenetictesting.

MMRGeneMutationTesting

MMRgenemutationtestingwasperformedbySangersequencingordenaturinghigh-performanceliquidchromatography,followedbycon?http://wendang.chazidian.comrgeduplicationsanddeletionsweredetectedbymultiplexligation-dependentprobeampli?cation.17,19-21Mutationclassi?ca-tionandnomenclatureweredeterminedusingtheInSiGHTColonCancerGeneVariantDatabase(http://wendang.chazidian.com/mutations),theMMRgenesvariantdatabaseoftheMemorialUniversityofNewfoundland(http://www.med.mun.ca/MMRvariants/),22andtheMMRGeneUnclassi?edVari-antsDatabase(http://wendang.chazidian.com).23Pathogenicmutationswerede?nedasvariantsresultinginastopcodon,frameshiftmutation,http://wendang.chazidian.com

cationordeletion,ormissensemutationpreviouslyreportedwithinscienti?cliteraturetobepathogenic.

EligibilityCriteria

Forthisstudy,weassembledthreesubcohortstoestimateriskofCRC(CRCrisksubcohort),riskofEC(ECrisksubcohort),andriskofanyothercancer(othercancerrisksubcohort;Fig1).Participantswereeligibleforallsubcohortsiftheyhadundergonegenetictestingfortheirspeci?cfamilygermlineMMRgenemutationandwerecon?rmedtobeeitheracarrierornoncarrierofthispathogenicmutation,hadbeenfollowedupatleastoncesincerecruitment,andhadnocancerdiagnosisbeforeoratthetimeofrecruitment.Atotalof81carriersand126noncarrierswereexcludedfromtheCRCrisksubcohortbecausetheyhadapolypectomybeforeoratthetimeofrecruitment.Atotalof26carriersand46noncarrierswereexcludedfromtheECrisksubcohortbecausetheyhadahysterectomybeforeoratthetimeofrecruitment(Fig1).Forthisanalysiswedidnothaveinformationonwhichparticipantswereawareoftheirmutationstatusorwhentheybecameawareoftheirstatusinpartbecausetestingmayhavebeenconductedclinically,outsidethescopeofthestudy.

StatisticalAnalysis

Observationtimebeganatcompletionofthebaselinequestionnaireandendedatthecancerdiagnosis,death,orlastfollow-up,whicheveroccurred?rst.ForCRCrisk,wecensoredindividualsatthetimeofpolypectomy(exceptwhenitoccurredwithinayearbeforetheCRCdiagnosis,inwhichcaseweassumedpolypectomywasfortheinitialCRCdiagnosis),andforECrisk,wecensoredeachwomanatthetimeofhysterectomy.

Observednumbersofcancerdiagnosesweredividedbytheexpectednumbersofcancerdiagnosestocalculatestandardizedincidenceratios(SIRs).Expectednumbersofcancerdiagnoseswerecalculatedbymultiplyingtheage-,sex-,andcountry-speci?cincidenceforthegeneralpopulationbythecorrespondingobservationtimeinthestudycohort.Country-,age-,andsex-speci?ccancerincidencesforthegeneralpopulationwereobtainedfortheperiodfrom1998to2002fromCancerIncidenceinFiveContinents.24This5-yearperiodwasselectedastheclosestavailabledatasetwithrespecttothemeancalendaryearofcancerdiagnosesofthesample,andgiventhatitaveragedincidenceoverthe5-yearperiod,itprovidesmorestableestimatesofincidence,especiallyforlesscommoncancers.Forcancercasesfromthesamefamily,thejackknifemethodwasusedtocalculate95%CIsbyallowingforanycorrelationofriskbetweenrelativesfromthesamefamily25;otherwise,weusedexactmethodsassumingthatobservedcancercasesfollowedaPois-sondistribution.

WeestimatedSIRsforthefollowingcancersobservedmorethanonceincarriersandnoncarriers:colonand/orrectum(ICD-OC18toC20),pancreas(ICD-OC25),stomach(ICD-OC16),kidneyandrenalpelvis(ICD-OC64andC65),urinarybladder(ICD-OC67),andlung(ICD-OC34)forbothsexes;endometrium(ICD-OC54andC55),ovary(ICD-OC56),andbreast(ICD-OC50)forfemales;andprostate(ICD-OC61)formales.

Kaplan-Meierstatisticswereusedtoestimateage-dependentcumulativerisk(penetrance)at5and10years.Allreportedstatisticaltestsweretwo-sidedandP?.05wasconsideredstatisticallysigni?cant.STATAversion11.0(STATA,CollegeStation,TX)26wasusedforallstatisticalanalyses.

RESULTS

SelectionofeligiblecarriersandnoncarriersofanMMRgenemuta-tionfromtheColonCancerFamilyRegistryforeachcancer-speci?canalysistoestimateriskofCRC,EC,andothercancersisdepictedinFigure1.

CRCRiskSubcohort

Thissubcohortincluded365carriersand903noncarriersfrom284familieswithMMRgenemutations(Table1).Ofthe365carriers,310(85%)were?rst-degreerelatives,50(14%)weresecond-degreerelatives,and?ve(1%)werethird-degreerelativesofpatientswith

959

Information downloaded from http://wendang.chazidian.com and provided by at Fudan Univ Lib on September 16, 2014 from

结肠癌流行病学

Winetal

Fig1.Flowdiagramoftheselectionofcarriersandnoncarriersofamismatchrepairgenemutation.CRC,colorectalcan-cer;EC,endometrialcancer.(*)AlthoughtheECriskcohortwascomposedofwomen,anoteofclari?cationwithre-specttothemenexcludedfromthisgroupisprovided.

CRC.Ofthe903noncarriers,583(64%)were?rst-degreerelatives,169(19%)weresecond-degreerelatives,and151(17%)werethird-degreerelativesofpatientswithCRC.

Overamedianfollow-upof5years,16mutationcarriers(MLH1,n?9;MSH2,n?4;MSH6,n?2;PMS2,n?1)werediagnosedwithCRCatamedianageof49years(incidencerate,8.84;95%CI,5.42to14.43per1,000person-years;SIR,20.48;95%CI,11.71to33.27;P?.001;Table2).TheCRCSIRsforcarriersofspeci?cMMRgenemutationswere39.40(95%CI,18.02to74.80)forMLH1,10.76(95%CI,2.93to27.55)forMSH2,17.19(95%CI,2.08to62.10)forMSH6,and15.47(95%CI,0.39to86.21)forPMS2.

FivenoncarrierswerediagnosedwithCRCatamedianageof60years(incidencerate,1.01;95%CI,0.42to2.44per1,000person-years).TheoverallSIRfornoncarrierswas1.02(95%CI,0.33to2.39;P?.97).AlloftheseCRCswereobservedin?rst-degreerelativesofpatientswithCRC.TheSIRwas1.43(95%CI,0.46to3.34)whenanalysiswasrestrictedtononcarrierswhowere?rst-degreerelativesofpatientswithCRC.

CumulativerisksofCRCwere4.14%(95%CI,2.35%to7.24%)at5yearsand8.05%(95%CI,4.46%to14.29%)at10yearsofprospectivefollow-upforcarriersand0.39%(95%CI,0.12%to1.20%)at5yearsand2.04%(95%CI,0.49%to8.30%)at10yearsfornoncarriers(Fig2andTable3).

Thefrequencyofcolonoscopyscreeningwasnotdifferentbe-tweenindividualswithoutCRC(unaffected)andthosewithCRC(affected).Onaverage,unaffectedcarrierswerescreenedonceevery2.7years(95%CI,2.5to3.0years),whereasaffectedcarrierswerescreenedevery2.5years(95%CI,1.4to3.5years;P?.63).Unaffectednoncarrierswerescreenedevery3.9years(95%CI,3.7to4.1years),whereasaffectednoncarrierswerescreenedevery3.9years(95%CI,0

960

to8.6years;P?.98).Colonoscopyduringthefollow-upperiodforcarriersandnoncarriersisshowninAppendixTableA1(onlineonly).ECRiskSubcohort

Thisfemalesubcohortincluded215carriersand523noncarriersfrom229familieswithMMRgenemutations(Table1).Overame-dianfollow-upof5years,sixcarrierswerediagnosedwithECatamedianageof53years(incidencerate,5.66;95%CI,2.54to12.59per1,000person-years).ThecorrespondingSIRwas30.62(95%CI,11.24to66.64;P?.001)forcarriersofanyMMRgenemutation(Table2).ECwasdiagnosedintwoMLH1mutationcarriers(SIR,27.18;95%CI,6.80to108.66)andfourMSH2mutationcarriers(SIR,44.92;95%CI,16.86to119.68).Cumulativeriskswereestimatedtobe2.84(95%CI,1.06to7.46)at5yearsand9.84(95%CI,3.45to26.33)at10yearsofprospectivefollow-up(Table3).NoECwasobservedinnoncarriers.

OtherCancerRiskSubcohort

Thissubcohortincluded446carriersand1,029noncarriersfrom300familieswithMMRgenemutations(Table1).Overamedianfollow-upof5years,forcarriers,weobservedthreeovariancancers(SIR,18.81;95%CI,3.88to54.95;P?.001),threerenalcancers(twokidneycancersandonerenalpelviscancer;SIR,11.22;95%CI,2.31to32.79;P?.001),twopancreaticcancers(SIR,10.68;95%CI,2.68to47.70;P?.001),twogastriccancers(SIR,9.78;95%CI,1.18to35.30;P?.009),twourinarybladdercancers(SIR,9.51;95%CI,1.15to34.37;P?.009),sevenfemalebreastcancers(SIR,3.95;95%CI,1.59to8.13;P?.001),andthreeprostatecancers(Table2).Cancersobservedinonlyonecarriereachincludedcanceroftheesophagus(ICD-OC15.9),biliarytract(ICD-OC24.0),liver(ICD-O

内容需要下载文档才能查看

C22.1),

JOURNALOFCLINICALONCOLOGY

Information downloaded from http://wendang.chazidian.com and provided by at Fudan Univ Lib on September 16, 2014 from

结肠癌流行病学

CancerRisksforMMRGeneMutationCarriersandNoncarriers

adrenalgland(ICD-OC74.9),smallintestine(ICD-OC17.9),andureter(ICD-OC66.9).

Fornoncarriers,therewasnoevidenceofanincreasedriskoflung,breast,andprostatecancers.Innoncarriers,wealsoobservedonecaseeachofpancreatic(ICD-OC25.9),urinarybladder(ICD-OC67.9),thyroid(ICD-OC73.9),andesophageal(ICD-OC15.9)can-cer(Table2).

DISCUSSION

Wecon?rmedthatMMRgenemutationcarrierswereatincreasedriskofawidevarietyofcancersandfoundthattherewasnoevidenceofanincreasedriskofcancerfortheirnoncarrierrelatives.Amajorstrengthofourstudyistheprospectivenatureofthedesign,becauseobservationtimeforcarriersandnoncarrierscommencedbeforecan-cerdiagnosis.Thisdesign,therefore,avoidedascertainmentbiascom-monlypresentinretrospectiveanalyses,particularlywhenestimatesofcancerriskaremadeusingrelativeswhoareenrolledincancerclinicsbecausetheyhavebeendiagnosedwithcancer.Afurtherstrengthoftheprospectiveanalysisisthatitprovidesanestimate,http://wendang.chazidian.com

piricaldata,offuturecancerriskatthetimetheycomeunderclinicalcare.Ourestimated10-yearriskofCRCforcarrierswithamedianageof49years(8.05%;95%CI,4.46%to14.29%)iscomparabletoourpreviousestimate(usinganascertainment-correctedretrospectiveanalysis)ofthe10-yearriskforMSH6mutationcarriersatage50years(6%[95%CI,3%to9%]formenand3%[95%CI,1%to5%]forwomen).27Itseemstobelowerthanapreviousretrospectiveestimateofthe10-yearestimatesforMLH1andMSH2carriers(26%formenand13%forwomen).28

Anunanticipated?ndingofourstudyisthecon?rmationofanincreasedriskforcancersofthebreastandpancreasforMMRgenemutationcarriers.Althoughriskofpancreaticcancerhasbeeninves-tigatedinLynchsyndrome,29-33theevidencewasinconsistent.ThestudyshowingevidenceforanassociationwasthatofKastrinosetal,34whichrecentlyshowedanincreasedriskofpancreaticcancerinLynchsyndrome(hazardratio,8.6;95%CI,4.5to15.7).Thisresultiscon-sistentwithour?nding(testfordifference,P?.79).However,giventhelimitedevidenceforef?cacyofscreeningforpancreaticcancers,35expertopinionhasnotyetadvocatedpancreaticcancerscreeninginthecontextofLynchsyndrome.

内容需要下载文档才能查看

961

Information downloaded from http://wendang.chazidian.com and provided by at Fudan Univ Lib on September 16, 2014 from

结肠癌流行病学

Winetal

内容需要下载文档才能查看

AmorecontroversialassociationwithLynchsyndromeisthatofbreastcancer.InitiallyraisedbyLynchetal36severaldecadesago,theissueofbreastcancerriskinLynchsyndromehasbeendebatedwithevidenceforandagainstthisassociation.Multiplecasereportsde-scribingMMR-de?cientbreastcancers,includingcasesofmalebreastcancer,havebeenpublished.37-41Walshetal40showedthatabouthalfofbreastcancersobservedinMMRgenemutationcarrieshavelossofexpressioninthetumorofthegenethatismutatedinthegermline.However,thisandotherstudieswerenotabletoaddresswhethertheMMRde?ciencycausedbreastcancerorwasaphenotypeofabreastcancercausedbyanotherfactor.Inthisreport,weobservedanin-creasedriskofbreastcancerforMMRgenemutationcarriersfollowedprospectively.Althoughthenumbersaresmall,theriskwasestimatedtobeincreasedfour-foldabovethatofthegeneralpopulation,suggest-

ingthatMMRmutationcarriersmaybene?tfromenhancedscreen-ing.Furtherclari?cationofage-speci?cmagnitudeofriskisneededtodeterminewhetheragesatscreeningormethodssuchasuseofmag-neticresonanceimagingshouldbemodi?edinLynchsyndrome.

Inadditiontode?ningspectrumandcancerrisksforMMRgenemutationcarriers,itisimportanttoclarifytheriskofCRCandothercancersfornoncarrierrelatives.Identi?cationofmutationnoncarri-ersisconsideredtobeoneofthemajorbene?tsofMMRgenetesting,providingreassurancethatthesepeopleareatsubstantiallylowerriskthantheirmutation-carryingrelativesand,therefore,notsubjecttothesamecostly,frequent,andinvasivescreeningandrisk-reduction

内容需要下载文档才能查看内容需要下载文档才能查看

Fig2.Kaplan-Meiercurvesforcumulativerisksandtheircorresponding95%CIsofcolorectalcancerat5and10years.962

JOURNALOFCLINICALONCOLOGY

Information downloaded from http://wendang.chazidian.com and provided by at Fudan Univ Lib on September 16, 2014 from