32C bcFigure137SpiculatedtumormassisthemostusualmanifestationofinvasivecarcinomasaMacroscopicimagebradiologicimageandchistologicimage UBGROSSMORPHOLOGYOFALTEREDBREASTTISSUEInvasionmosto ID: 242535
Download Pdf The PPT/PDF document "(a)(b)Figure1.36Thicksectionimagesof(a)i..." 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.
32C (a)(b)Figure1.36Thicksectionimagesof(a)invasivecarcinoma,and(b)radialscar. (b)(c)Figure1.37Spiculatedtumormassisthemostusualmanifestationofinvasivecarcinomas.(a)Macroscopicimage,(b)radiologicimage,and(c)histologicimage. UBGROSSMORPHOLOGYOFALTEREDBREASTTISSUEInvasionmostoftenevokesproliferationofbroblastsandmyobroblastsaswellascollagenizationinthestromadamagedbythetumorcellsareactioncalledstromaldesmoplasia,whichpreventsthetumorcellsfrominvadingthestromaevenlyinalldirections.Thisprocessleadstothedevelopmentofaspiculated(stellate)tumormass,themostusualmanifestationofinvasivecarcinoma.Fig-ure1.36(a),demonstratesa3Dimageofsuchatumorandcomparesittothe3Dimageofaradialscar[Fig.1.36(b)].Thetypicalmacroscopicappearanceofastel-latecarcinomaisillustratedinFig1.37.Abouttwo-thirdsofinvasivecarcinomasareofstellateshape.Theshapeoftheinvasivetumorisrelatedtoitshistologictype.Thestellatecarcinomasaremostofteninvasivelobular,tubular,orgradeIorgradeIIinvasiveductalnototherwisespecied(NOS)cancers(seeDiagram1.5);thesetogethercorrespondtomorethan80%ofstellatetumors.Benignstellatelesionsarerare.Onlytheappearanceofelasticbersinthestroma(asinradialscars)orcollagenizationduetoscarformationleadstolesionsimitatingthestellateshapeofbreastcarcinoma.Thus,astellatelesiononamam-mogramismostoften(inmorethan90%ofcases)malignantandwarrantsfur-therdiagnosticandtherapeuticinterventions.Althoughspiculated,radialscarslackawell-formedtumorbody[Fig.1.38(a)]andappearasblackstarsonamammogram[Fig.1.38(b)]incontrasttothewhitestarsincarcinomacases[Fig.1.37(b)].Thehistologicappearanceisalsotypicalastheyexhibitanelasticcoreandacoronaofhyperplasticchanges[Fig.1.38(c)].Inaminorityofcases,radialscarsmaycontainlow-gradeinsitucarcinomafociintheircoronaandrep-resentaspecialtypeofinsitucancer.Ifthedesmoplasticstromalreactiononinvasionofthetumorcellsisweakorabsent,thecellsmayinvadeevenlyinalldirections,givingrisetoacircular(oroval)tumormass.ThesetumormassesaresimilartobenignormalignantlesionsDiagram1.5Distributionof910consecutivestellateinvasivecarcinomasbytumortype,Falun1996 2003. 9.1DETECTIONOFICROCALCIFICATIONSACKGROUNDANDOTIVATIONInWesterncountries,womenhaveahigherthan1-in-8chanceofdevelopingbreastcancerduringtheirlives.Breastcancerrepresentsthemostfrequentlydiagnosedcancerinwomen.TheNationalCancerInstituteofU.S.A.estimatesthat,basedoncurrentrates,13.2%ofwomenborntodaywillbediagnosedwithbreastcanceratsometimeintheirlives.Inordertoreducemortality,earlydetectionofbreastcancerisimportant,be-causetherapeuticactionsaremorelikelytobesuccessfulintheearlystagesofthedisease.Forwomenwhosetumorswerediscoveredearlybymammography,theve-yearsurvivalratewasabout82%asopposedto60%forthecaseswherethetumorswerenotfoundearly.Mammographyiscurrentlythebestradiologicaltechniqueavailableforearlydetectionofnonpalpablebreastcancer.However,itisdifcultforradiologiststoprovidebothaccurateanduniformevaluationsforthelargenumberofmammo-gramsthattheyhavetointerpretinscreeningprogramswheremostofthecasesarenormal;ithasbeenobservedthat10 30%ofbreastlesionsaremissedduringroutinescreening.Thesituationisevenmorechallengingsincetheearlymalig-nancieshavesmallsizeandsubtlecontrastwhencomparedwithnormalbreaststructures.Doublereading(ascarriedout,forexample,bytworadiologists)helpstoreducethenumberoffalsenegativesby5 15%.Digitalimage-processingtechniquesrepresentusefultoolsforhelpingradiol-ogiststoimprovetheirdiagnosiswiththeaidofcomputersystems.Inthissense,differentCAD(computer-aideddiagnosis)toolshavebeendevelopedforimprov-ingimagequality,identifyingmalignantsigns,enhancingmammographicfeatures,etc.Ontheaverage,thereaderssensitivitycanbeincreasedby10%withtheassis-tanceofCADsystems.SomeworkshavestudiedthispotentialofCADtoimproveradiologistsperformanceindetectingclusteredmicrocalcications.Thereareanumberofdifferentclassesofabnormalitythatmaybeobservedinmammograms.Oneofthemostsignicanttypesofmammographicabnormal-ityismicrocalcication.Microcalcicationsaretinygranulelikedepositsofcal-cium.Theyarerelativelybright(dense)incomparisonwiththesurroundingnormalandareuptoabout1mmindiameter,withanaveragediameterof0.3mm.Microcalcicationsareofparticularclinicalsignicancewhenfoundinclustersofthreeormorewithinasquare-centimeterregionofamammogram.Lanyidescribedmicrocalcicationsasthemostimportantleadingsymptominmam-mographicdetectionofpreclinicalcarcinomas.Sicklesnotedthatmorethan50%ofnonpalpablecancershadmammographicallyvisiblecalcications,andin36%ofnonpalpablecancers,calcicationsweretheonlysignofabnormality.Inanimportantstudyofcancersmissedinscreeningmammography,itwasobservedthatthepresenceofmicrocalcicationswasthepredominantfeaturein18%ofthemissedcancers. approachconsistsofusingtheanatomicknowledgeofthestromatodevelopaconstraintbetweentheconrmedportionoftheskinlineandthestromaedge.Thisstepusesthedensityconcepttodevelopthethreshold.Thestromaedgeissmoothedbyttingaspline,andconstraintsareestablishedbetweentheinitialconrmedportionoftheskinlineandthesmoothedstromaedge.Theconstraintsarethenpropagatedtowardtheupperandlowerbreastzonestocorrecttheweakareasoftheboundary,therebygettingclosertothetrueedge.Thepropagationisperformedbyusingagreedyapproachincombinationwithanatomicalconstraints.Comparingtheresultsofthealgorithmwiththoseofoneofthebestmethodsreported,namelythedeformablemodelofFerrarietal.,animprovementinskin-lineestimationisdemonstratedbythisapproach.Notethattheground-truth(GT)boundariesusedforthecomparisonwerepreparedandtracedbyradiologists.Thechapterisorganizedasfollows.Section11.2presentstheoveralldesignofthesystem.Section11.3providesadescriptionofadaptivethresholdingtoobtainaninitialskinline.InSection11.4,ashapeanalysismethodfortheextractionoftheinitialconrmedportionoftheskinlineispresented.InSection11.5,themethodsfortheextractionoftheedgeofthestromaandsplinettingarepresented.Sec-tion11.6presentstheproposeddependencyapproachtoobtainthenalbreastskinline.SeveraltechniquesandmetricsforperformanceevaluationandcomparisonhavebeenimplementedasdescribedinSection11.7,includingthepolylinedis-tancemeasure(PDM)andtheHausdorffdistancemeasure(HDM).TheresultsoftheanalysisareshowninSection11.8,andthepaperisconcludedinSection11.9.11.2ABVERVIEWOFTHEROPOSEDTheconceptofthealgorithmarisesfromthefactthattheedgeofthestromaofthebreasthasausuallyuniformdistancetothebreastskinline(seeFig.11.3).Thestromaedgeisnotaffectedbythesystemnoisebecauseofthehighcontrastbetweenthebright(dense)stromaregionandthelow-intensity(low-density)fattyperipheralregion.Thebackgroundregionandthebreastregionaredistinguishedbyadiscontinuityinthehistogram.Thisdiscontinuitymaybeestimatedusinganadaptivethresholdingtechnique,suchasthemethodproposedbyOjalaetal.stepisrepresentedinFig.11.2asblock(A).Becausethestromaisahigh-densityzone,itcaneasilybemaskedoutbyabi-modalitythresholdingprocedure.Onesuchthresholdingprocedurehasbeenpro-posedbyOtsu,whichiswidelyacceptedinmedicalimageprocessing.TheaddedadvantageofOtsusprocedureforthresholdcomputationisitssimplicity.ThisstepisidentiedinFig.11.2asblock(B).Therefore,aninitialestimateoftheouter-andinner-edgeofthebreastisobtainedusingtwodifferentthresholdingprocesses,yieldinginitialestimatesoftheskin-lineboundaryandthestromaedge.ThisisshownbytheparallelpathsinFig.11.2.Havingextractedaninitialconrmedportionoftheskinlinefromtheinitialskin-lineestimateandtheaccuratestromaedgeinformation,anovelstrategytopropagateorextrapolatethepartialboundarytoafullskin-lineboundarybased IAGNOSISINMAGINGTheexcessivebiopsyofbenignlesionsraisesthecostofmammo-graphicscreeningandresultsinemotionalandphysicalburdentohundredsofthousandsofpatientseveryyear,aswellasnancialburdentosociety.Thus,itwouldbeexceedinglyvaluabletoproducecomputer-aideddiagnosis(CADx)sys-temsthatcouldaidinthedecisiontorecommendbiopsyorrecommendshort-termfollow-upmammography.Thoseverylikelybenigncasesmaybemanagedwithshort-termfollow-up,whichwillavoidahugenumberofunnecessarybiop-sieswhilemaintainingtheveryhighsensitivityofcancerdetection.Moreover,sinceitisestimatedthatabouthalfofmissedcancersaremissedduetomisinter-pretationratherthanoversight,itmaybepossibletoincreasethesensitivityofmammographythroughCADx.27.1.3BRIEFHISTORYOFCADxinbreastimagingstartedinearnestintheearly1990s,andtherehavelit-erallybeenhundredsofpublicationsintheliterature.MostofthemajorpapershavebeeninMedicalPhysicsAcademicRadiology,withoccasionalclinicalevaluationsinRadiologyManyreviewsoftheeldhavebeenwritten.Amongrecenteffortsinpar-ticular,GigerprovidesabroadoverviewofthestateoftheartinCADandCADx,andSampatetal.provideadetailedsummaryofthemanydifferentapproachesem-ployedandperformanceresultsattained.Inthefollowingsections,citationswillbeprovidedtootherrepresentativeand/orrecentpublicationsinthecontextofthemajorchallengesstillfacedinCADxresearch.27.2CADLASSIFIERThereisnouniversalstructureforCADxsystems,butmostdofollowthegen-eralowchartthatisverysimilartothatofCADsystems(seeFig.27.1).Inthisexample,thesystemallowsforinputsfromthreedifferentsources:computer-extractedimage-processingfeatures,radiologist-interpretedndings,andpatienthistoryndings.Thevariousfeaturesareselectedaccordingtosomerationale,thenmergedtogetherusingaclassier.Eachofthesestepsareexplainedinfurtherdetailinthefollowingsections.27.2.1LINEARVERSUSNONLINEARCLASSIFIERSMuchoftheearlyworkinCADxwasbasedonthearticialneuralnetwork(ANN)nonlinearclassier,whichisdescribedinmanyclassictexts.Neuralnetworkswerequicklyembracedbytheeld,andthishistoryisreviewedbelowtomakeanimportantcautionarynote.AnexampleofthemodelarchitectureisshowninFig.27.2.