TumouredgeincolorectalcancereFig2ColonicadenocarcinomawithpronouncedinflammatoryreactionincludingperivascularlymphocyticcuffsattumouredgeHaematoxylineosinsaffronlocatedatsomedistancefromthetu ID: 947212
Download Pdf The PPT/PDF document "JClinPathol198942162166Associationbetwee..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
JClinPathol1989;42:162-166Associationbetweeninvasiveness,inflammatoryreaction,desmoplasiaandsurvivalincolorectalcancerTBHALVORSEN,*EVASEIMtFromthe*DepartmentofPathology,TrondheimRegionalandUniversityHospital,andthetDepartmentofMathematicsandStatistics,UniversityofTrondheim,NorwaySUMMARYFivehundredandtwentysevencolorectalcarcinomaswerereviewedhistologically.Amultivariateanalysis(Cox)wasusedtotesttheprognosticimportanceofcertainhistologicalfeatures(invasiveness,inflammatoryreaction,andamountoffibroustissue)atthetumouredgeafterallowancehadbeenmadeforclinicopathologicalstage,tumoursite,andhistologicaltypeandgrade.Apoorlydefinedtumourborder,lackofinflammatoryreaction,andapronouncedfibrosis(desmoplasia)atthetumouredgewereassociatedwithunfavourablestagedistributions,buteachofthesefeaturesalsohadanindependenteffectonprognosis.Highlyinvasiveorexpandinggrowthhaverepeatedlybeenshowntoworsentheprognosisincolorectalcancer."Apronouncedinflammatoryreactioninthestromaatthetumouredge,ontheotherhand,hasoftenbeenassociatedwithanimprovedprognosis.2"Sometimesapronouncedperitumouralfibrosis(desmoplasia)isobserved.Thebiologicalroleofthisfeaturehasbeendebated.'2'4Thisstudywasundertakentoevaluatetheprog-nosticimportanceofinvasiveness,inflammatoryreac-tion,anddesmoplasiaincolorectalcancer,andtheassociationbetweenthesemorphologicalfeaturesandotherprognosticfactorssuchastumourstage,grade,andsite.MaterialandmethodsThehistologicalmaterialfrom534surgicallyremovedcolorectaladenocarcinomaswerere-examinedblindbyoneobserver.Thegradingofthesetumoursaccordingtothepredominantdegreeofdifferentiationandtheirclassificationwithregardtomucincontenthaspreviouslybeenreported'516Signetring-cellcar-cinomaswerenotgraded.Thetumourswerecategor-isedbysite-rightcolon,leftcolon,andrectum.'7StagingwasperformedaccordingtoDukes'classification.'8Tumourswithdistantspreadorgrowthbeyondthesurgicalmarginsformedaseparatestagedesignated"advanceddisease",alsoknownas"DukesD".Theinvasivenessofthetumourwasscoredasfollows:score0=welldefinedborderwithexpandinggrowth;andscoreI=poorlydefinedAcceptedforpublication29September1988borderwithpronouncedinvasivegrowth(fig1).Theinflammatoryreactionwasassessedonthebasisofthegeneralimpressionoftheamountofinflam-matorycellsalongtheentiretumouredgeawayfromareasoffrankabscessformation,andscoredasfollows:score0=prominentinflammatoryinfiltrate(fig2);andscore1=inconspicuousinflammatoryinfiltrate.Noattemptwasmadetoevaluatetheproportionsofthevariousinflammatorycellsunlessmanyeosinophilswereobserved,inwhichcasetheinfiltrateswereclassifiedas"eosinophilic".Thepresenceofabscesseswasalsonotedseparately.Finally,thepresenceofperivascularcollections("cuffs")oflymphocytes7wasnoted,eveniftheywere,-~~~~~Fig1Colonicadenocarcinomawithpronouncedinvasivegrowthandextensivefibrosisatthetumouredge.(Haematoxylin-eosin-saffron.)162 Tumouredgeincolorectalcancer'e''Fig2Colonicadenocarcinomawithpronouncedinflammatoryreactionincludingperivascularlymphocyticcuffsattumouredge.(Haematoxylin-eosin-saffron.)locatedatsomedistancefromthetumourmarginitself(fig2).Theamountoffibroustissuewasassessedattheadvancingtumouredgeandwasscoredasfollows:score0=littleormoderateamou
ntoffibrosis;and163score1=desmoplasia,withextensivefibrosisaccount-ingforatleastanequalamountoftissueasthecarcinomaitself(fig1).Sevencasesinwhichhistologicalassessmentcouldnotbeperformedwereexcluded.Fivehundredandtwentyseventumoursfrom271menand256womenremained.Themeanageswere66-0(SEM0-6)and64-7(07)yearsformenandwomen,respectively.SurvivaldataforthepatientsinthisserieswereobtainedfromthefilesoftheCancerRegistryofNorwayuptotheendofSeptember1985.Deathduetocolorectalcancerwasusedastheclinicalendpoint.Cutoffwas85yearsofage.ThesurvivalfunctionwasestimatedbyCox'sproportionalhazardsregressionanalysis'9usingthecomputerprogramBMDP2L,/asexplainedindetailinapreviousreport.'5Cohen'sKstatistic2'wasusedtoestimatethereproducibilityofthehistologicalevaluationsin212randomlyselectedcases.ResultsThestagedistributionsweremorefavourableintumourswithawelldefinedborder,prominentinflam-matoryreaction,eosinophiliaandlackofdefiniteTable1Stagedistributionbytumourbordercharacteristicsin527colorectalcarcinomasClinicopathologicalstage(%)TwnourborderNoofDukes'Dukes'Dukes'AdvancedcharacteristicscasesABCdiseasetWelldefinedborder39318501913Poorlydefinedborder1341344520Inflammation+*18120481913Inflammation-34610443016Lymphocyticcuffs+34513482613Lymphocyticcuffs-18215422617Eosinophilia+263535273Eosinophilia-50112472615Abscesses+9712561319Abscesses-43013442914Desmoplasia+3537463116Desmoplasia-17426461711All52713462615*+signifies"prominent"or"conspicuous"-signifiesabsence.tDenotesdistantspreadortumourgrowthbeyondsurgicalmargins.Table2Growthpatternattumouredgeaccordingtooverallhistologicalgrade(percentagesinparentheses)HistologicalgradeWellModeratelyPoorlyGrowthdifferentiateddifferentiateddifferentiatedAllgradesWelldefinedborder155(89)209(75)21(32)385(75)Poorlydefinedborder19(11)68(25)44(68)131(25)All17427765516**Elevensignetring-cellcarcinomaswereexcludedfromthistable. 164Table3Inflanmatoryreactionanddesmoplasiaaccordingtogrowthpatternatthetumourborderin527colorectalcarcinomasGrowthpatternatthetumourborder:WellPoorlyInflammatoryreactiondefineddefinedanddesmoplasiaborderborderp-Value$Inflammation+*16219Inflammation-231115Lymphocyticcuffs+26679�Ā0-05Lymphocyticcuffs-12755Eosinophilia+2600-005Eosinophilia-367134Abscesses+8314Abscesses-310120Desmoplasia+2401130-0001Desmoplasia-15321*+signifies"prominent"or"conspicuous".-signifiesabsence.$Basedonx2testwithonedegreeoffreedom.desmoplasticreactionthanintumourswith"opposite"characteristics(table1).Tumourswithpoorlydefinedborderwerelessdifferentiatedthantheirwelldefinedcounterparts(X2=80-74;df=2;p0-0001)(table2).Noneoftheothertumourbordercharacteristicsweresignificantlycorrelatedwithhistologicalgrade(datanotshown).Onlyoneof11(9%)signetring-cellcarcinomasexhibitedapronouncedinflammatoryreactioncom-paredwith180of516(35%)oftheothers(p0.01).Desmoplasticreactionwasfoundinonlytwo(18%)ofthesignetring-cellcarcinomascomparedwith351ofTable4FinalmodelinproportionalhazardsregressionanalysisRelativeRegressionrisk(95%coefficientconfidenceExplanatoryvariable(SE)limits)Tumoursite:Leftcolonvrightcolon0-678(0-219)1-97(1-27;3-05)Rectumvrightcolon1-289(0-222)3-63(2-33;5-65)Clinicopathologicalstage:Dukes'BvDuke
s'A0-662(0-282)1-94(1-10;3-41)Dukes'CvDukes'A1-118(0-294)3-06(1-70;5-51)AdvanceddiseasevDukes'A1-782(0-306)5-94(3-22;10-96)Histologicalgrade:Moderatevwelldifferentiation0-406(0-165)1-50(1-08;2-09)Poorvwelldifferentiation1-135(0-226)3-11(198;489)Histologicaltype:Signetring-cellcarcinomav1-735(0-475)5-67(2-19;14-66)adenocarcinomaTumourbordercharacteristics:Poorlydefinedborder0-495(0-152)1-64(1-21;2-23)Lackofprominentinflammatory0-363(0-160)1-44(1-05;1-98)reactionLackofperivascular0-381(0-137)1-46(1-11;1-92)lymphocyticcuffsDesmoplasticreaction0-450(0-166)1-57(1-13;2-19)Halvorsen,SeimYearsafterdiagnosisFig3Estimatedsurvivalfunctiotnforpatientswithtumoursexhibitingawelldefinedborderwithprominentinflammatoryreaction,includingperivascularcuffsandnodesmoplasia(baselinecharacteristics),andpatientsinwhomcorrespondingunfavourablecharacteristicsareoperatinginconcert.StandardisedtomoderatelydifferentiatedDukes'classBrectalcancers.Proportionsurvivingfiveyearsindicated.516(68%)fortheothertumours(p0.001).Theproportionoftumourswithapoorlydefinedmarginandtheprevalenceofperivascular"cuffs"weresimilarinsignetring-cellcarcinomasandothertumours.Sexandtumoursitewerenotassociatedwithanyofthetumourbordercharacteristicsexceptthatperivas-cularcuffswerefoundin180of256(70%)ofthetumoursinwomencomparedwith165of271(61%)ofthoseinmen(p0-05).Aprominentinflammatoryinfiltrate,manyeosino-phils,abscessformation,andlackofdesmoplasiaweresignificantlymorecommonintumourswithwelldefinedbordersthaninthosewithpoorlydefinedborders(table3).Perivascularlymphocyticcuffswerenotrelatedtothegrowthpattern.Suchcuffs,however,werefoundin144of181(80%)ofthetumourswithprominentinflammatoryreactionatthetumourborder,comparedwith201of346(58%)ofthosewithinconspicuousinflammatoryreaction(p0-001).TheKindexesofreproducibilityoftheevaluationofthetumourbordercharacteristicswereintherange0-73to0-86(with95%confidenceintervalsrangingfrom0-63to0-94),exceptfortheevaluationofeosinophilicinfiltrates,forwhichKwas0-39(withthe95%confidenceintervalrangingfrom0-03to0-75).Althoughseveralofthetumourbordercharac-teristicswerecorrelatedwithclinicopathologicalstage(table1)andinsomewayswereinter-related(table3), Tumouredgeincolorectalcancer165theresultspresentedintable4showthatapoorlydefinedtumourborder,desmoplasticreaction,lackofaconspicuousinflammatoryinfiltrateandlackofperivascularlymphocytic"cuffs"eachindependentlyimpliedanincreaseinriskofdeathbyafactorofabout1[5comparedwiththecorrespondingbaselinecharac-teristics.Clinicopathologicalstage,though,hadthestrongestinfluenceonprognosis.Thevariablesofsex,age,amountofmucincontentintheadenocarcinomas,eosinophiliaoftheinflam-matoryinfiltrate,andabscessformationhadnosignificantindependentinfluenceonthesurvival.Theoverallfiveyearsurvival(Kaplan-Meierestimate)intheserieswas537%(withthe95%confidenceintervalrangingfrom49-1to58-3%).Fig3illustratestheinfluenceontheestimatedsurvivalfunctionofthetumourbordercharacteristicsstudiedinthisseries.DiscussionThisstudyshowsthatinvasivenessandotherhisto-logicalvariableswhichmayhaveprognosticimpor-tanceincolorectalcancerareoftenassociatedwitheachother,aswellaswithclinicopathologicalstage.Thecomplexityofthesea
ssociationsrequiresmulti-variateanalysestobecarriedouttodeterminewhetherthevariablesmayhaveanindependentinfluenceonprognosisandtoevaluatetheirimportance.Cox'sproportionalhazardsregressionanalysishascorroboratedthedataobtainedinrecent,largemulti-variateanalysesofrectalcancer4einwhichanirregulargrowthpatternatthetumourborderandascantylymphocyticinfiltrationwereindependentlyassociatedwithapoorprognosis.Despitenumerousstudiesthathavecharacterisedthecellularcompositionoftheinflammatoryinfil-tratessurroundingcolorectalcancers,thebiologicalroleoftheindividualcelltypesremainsunclear.3922-24Jasssuggestedthatthelymphocyticinfiltrationincolorectalcancerreflectedaconservationofthenormalinteractivetrafficbetweenepitheliumandmucosa-associatedlymphoidtissue,andsignalledalowgradegrowth.'"Pihletalregardedperivascularlymphocyticcuffsasabeneficialimmunomorphologicalfeature,7althoughtherearenodirectobservationstoindicatethatthelymphocytesofthecuffsareactivelydefensive.Ourstudysuggeststhatsuchcuffshaveanindependentbeneficialinfluenceontheprognosis,butourfindingsdonotexplainthemechanismsofthiseffect.Itistemptingtohypothesisethatleucocytesandmacro-phagesattractedbychemotaxis(frominfectionandtumournecrosis)mightbenefitthehostresistancethroughnon-immunologicalattackontumourcells.Inthisstudy,however,wewereunabletofindanyprognosticimportanceforabscessformation.Eosinophilicinfiltratesincarcinomasofthecolonhavebeenreportedtobeassociatedwithanimprovedprognosis.25Wewereunabletofindanyindependentprognosticinfluenceofeosinophilicinfiltrates.Thereproducibilityofthispartofthehistologicalevalua-tion,however,waspoor.Tumourdesmoplasiaisacommonfeatureinseveralmalignanthumantumours-breast,bowel,andpros-tatecancer.Jasscouldnotfindanyprognosticinfluencefortumourfibrosisinrectalcancer.'0Bycontrast,theresultsofourstudysuggestthattumourdesmoplasiaisanindependentprognosticvariablewhichmayincreasetheriskofdeathbyafactorofabout1-5comparedwithtumourswithlittleornofibrosis.Martinetalreportedabeneficialeffectoffibro-blasticreactionincoloniccancerinratsandsuggestedthatfibrousencapsulationcouldhavearoleinthemechanismoftumourregression.'4Othershavearguedthatsuchanencapsulationmightbenefitthetumourbyreducingaccessbyhostimmunecells.2Ourresultsfavourthelatterview.JassandMorsonhaverecentlyproposedasystemofprognosticcategorisationforcolorectalcancerwhichtakesheedoftheclinicopathologicalstage,aswellasthecharacteroftheinvasivemargin,andthepresenceorabsenceofaconspicuousperitumourallympho-cyticinfiltrate.26Wesuggestthatthedegreeofperitumouralfibrosis(desmoplasia)andthepresenceorabsenceofperivascularlymphocyticcuffsgiveadditionalinformationaboutprognosisincolorectalcancer.WethankthestaffattheCancerRegistryofNorway,Osloforprovidingsurvivaldata,MrsMetteHeimfortechnicalassistance,andMrsSigrumQrnsjqfortypingthemanuscript.ThisstudywasfinanciallysupportedbygrantsfromKreftfondetvedRegionsykehusetiTrondheim.ReferencesIGrinnellRS.Thegradingandprognosisofcarcinomaofthecolonandrectum.AnnSurg1939;109:500-33.2SprattJS,SpjutHJ.Prevalenceandprognosisofindividualclinicalandpathologicvariablesassociatedwithcolorectalcarcinoma.Cancer1967;20:1976-85.3ThynneGS,WeilandLH,MoertelCG,S
ilversA.Correlationofhistopathologiccharacteristicsofprimarytumourandunin-volvedregionallymphnodesinDukes'classCcoloniccarcin-omawithprognosis.MayoClinProc1980;55:243-5.4JassJR,AtkinWS,CuzickJ,etal.Thegradingofrectalcancer:historicalperspectivesandamultivariateanalysisof447cases.Histopathology1986;10:437-59.5UnderwoodJCE.Lymphoreticularinfiltrationinhumantumours:prognosticandbiologicalimplications:areview.BrJCancer1974;30:538-48. 166Halvorsen,Seim6MurrayD,HrenoA,DuttonJ,HampsonLG.Prognosisincoloncancer.Apathologicreassessment.ArchivesofSurgery1975;110:908-13.7PihIE,MalahyMA,KhankhanianN,HershEM,MavligitGM.ImmunomorphologicalfeaturesofprognosticsignificanceinDukes'classBcolorectalcarcinoma.CancerRes1977;37:4145-9.8ZamcheckN,DoosWG,PrudenteR,LurieBB,GottliebLS.Prognosticfactorsincoloncarcinoma.Correlationofserumcarcinoembryonicantigenlevelandtumorhistopathology.HumPathol1975;6:31-45.9SvennevigJL,LundeOC,HolterJ,BjorgsvikD.Lymphoidinfiltrationandprognosisincolorectalcarcinoma.BrJCancer1984;49:375-7.10JassJR.Lymphocyticinfiltrationandsurvivalinrectalcancer.JClinPathol1986;39:585-9.11BjerkesetT,MorildI,MorkS,SireideO.Tumorcharacteristicsincolorectalcancerandtheirrelationshiptotreatmentandprognosis.DisColonRectum1987;30:934-8.12LiottaLA,RaoCN,BarskySH.Tumorinvasionandtheextracellularmatrix.LabInvest1983;49:636-49.13DvorakHF.Tumours:woundsthatdonotheal.Similaritiesbetweentumourstromagenerationandwoundhealing.NEngiJMed1986;315:1650-8.14MartinMS,CaignardA,HammannA,PelletierH,MartinF.Animmunohistologicalstudyofcellsinfiltratingprogressiveandregressivetumorsinducedbythetwovariantsubpopulationsofaratcoloncancercellline.IntJCancer1987;40:87-93.15HalvorsenTB,SeimE.Degreeofdifferentiationincolorectaladenocarcinomas:amultivariateanalysisoftheinfluenceonsurvival.JClinPathol1988;41:532-7.16HalvorsenTB,SeimE.Influenceofmucinouscomponentsonsurvivalincolorectaladenocarcinomas:amultivariateanalysis.JClinPathol1988;41:1068-72.17HalvorsenTB,SeimE.Tumoursite:aprognosticfactorincolorectalcancer?Amultivariateanalysis.ScandJGastroen-terol1987;22:124-8.18DukesCE.Theclassificationofcanceroftherectum.JPatholBacteriol1932;35:323-32.19CoxDR.Regressionmodelsandlife-tables.JRStatistSoc(B)1972;34:187-220.20DixonWJ.BMDPstatisticalsoftware.Berkeley,California:UniversityofCaliforniaPress,1983.21CohenJ.Acoefficientofagreementfornominalscales.EducationalandPsychologicalMeasurement1960;26:37-46.22VoseBM,GallagherP,MooreM,SchofieldPF.Specificandnon-specificlymphocytecytotoxicityincoloncarcinoma.BrJCancer1981;44:846-55.23CsibaA,WhitwellHL,MooreM.Distributionofhistocom-patibilityandleucocytedifferentiationantigensinnormalhumancolonandinbenignandmalignantcolonicneoplasms.BrJCancer1984;50:699-709.24KochB,GiedlJ,HermanekP,KaldenJR.Theanalysisofmononuclearcellinfiltrationsincolorectaladenocarcinoma.JCancerResClinOncol1985;109:142-51.25PretlowTP,KeithEF,CryarAK,etal.Eosinophilicinfiltrationofhumancoloniccarcinomasasprognosticindicator.CancerRes1983;43:2997-3000.26JassJR,MorsonBC.Reportingcolorectalcancer.JClinPathol1987;40:1016-23.Requestsforreprintsto:DrToreBHalvorsen,DepartmentofPathology,TrondheimRegionalandUniversityHospital,N-7006Trondheim,Norw