Abstract
Objective :Theobjectivesofthisstudyweretoexaminethevalidityofasciticfluidcytologyinthedetec -
tionofpathologicalfindings,toexaminethepercentageoffalsepositiveandfalsenegativeresultsinthe
cytologyofasciticfluidandtodeterminethevalidityofperitonealcytologyinrelationtothehistopatho -
logicaltypeoftheovariantumour.
Methods
:Thisretrospectivestudyincluded 170peritonealcytologyfindings.Thestudywasconducted
fromJanuary2010toD ecember2012.Theexperimentalgroupincluded76 cytologyfindingsobtained
frompatientsdiagnosedwithovariancarcinoma,whereasthecontrolgroupwascomposedof94cytol-
ogyfindingsofbenignovariantumoursandlivercirrhosisascites.Thepatientswithovariancarcinoma
had gradesIIIaswellasgradesIandIIcbutonlyincaseswhereoperativeandpathologicalfindingin -
dicatedarupturedorperforated tumourcapsule.
Results
Thesensitivityofperitonealcytologyis68.92%,specificityis 93.61%,positivepredictivevalue
is89.65%,andnegativepredictivevalueis78.57%. In30.02% of patients,theperitonealcytology
showedfalsenegativeresults,whilein6.38% ,theresultswerefalsepositive.Thehighestpercentageof
falsenegativefindingswas77%, foundinendometrioidcarcinoma.
Conclusion
Peritonealcytologyofasciticfluidishighlyspecificbuthasrelativelylowsensitivity,par -
ticularlyinthecaseofendometrioidovariancarcinoma.Inordertoincreasesensitivity,peritonealcy -
tologyshouldbecombinedwithmonoclonalantibodiesandotherbiochemicalandimmunohistochemical
markers.
Keywords
Ascites,asciticfluidcytology,ovariancarcinoma
AscitisenelCarcinomaOvárico–ConfiabilidadyLimitacionesdelAnálisis
Citológico
RŽivadinović 1,2 ,A Petrić 1,2 ,D Krtinić 1,JStevanovićMilosević 1,2 ,SPop TrajkovićDinić 1
RESUMEN
Objetivo: Losobjetivosdeesteestudiofueronexaminarlavalidezdelacitologíadellíquidoascíticoen
ladeteccióndehallazgospatológicos,examinarelporcentajederesultadospositivosfalsosynegativos
falsosenlacitologíadellíquidoascítico ,ydeterminarlavalidezdelacitologíaperitonealenrelación
coneltipohistopatológicodeltumorovárico.
Métodos: Esteestudioretrospectivoincluyó170hallazgosdecitologíaperitoneal.Elestudiosellevóa
cabodesdeenerode2010hastadiciembrede2012.Elgrupoexperimentalincluyó76 resultadosdeci-
tologíaobtenidosdepacientesdiagnosticadosconcarcinomaovárico,mientrasqueelgrupodecontrol
estuvointegradopor94hallazgosdecitologíadetumoresováricosbenignosyascitisdelacirrosishe -
pática.LospacientesconcarcinomaováricoteníangradosIII,asícomogradosIyIIc ,perosóloenlos
casosenlosqueelhallazgooperativoypatológicoindicabaunacápsuladetumorrotaoperforada .
Resultados: La sensibilidaddelacitologíaperitoneales68.92%,laespecificidad93.61% ,elvalorpre -
dictivopositivo89.65%,yelvalorpredictivonegativo78.57%. En30.02%delospacientes,lacitolo -
gíaperitonealmostróhallazgosnegativosfalsos,mientrasqueenel6.38%,los hallazgosfueronpositivos
falsos.Elmayorporcentajedeloshallazgosnegativosfalsosfue77%, encontradosenelcarcinomaen -
dometrioide .
From:1ClinicforGynecology and Obstetrics,ClinicalCenterNišand
2FacultyofMedicine,UniversityofNiš,Niš,Serbia.
DOI: 10.7727/wimj.2014.230
Correspondence:DrD Krtinić,ClinicofGynecologyandObstetrics,Clinical
CenterNiš,18000Niš,Serbia.E-mail:
[email protected]
237
Introduction
Ascitesisalargeamountoffluidaccumulatedintheabdomen.
Undernormalconditions,severallitresofperitonealfluidare
produceddailyand itisnotaccumulated buteffectivelyab -
sorbed.
Ascitesofmalignantaetiologyappearinonly10% ofall
ascitescases(1).Non-malignantaetiologyascitesaremost
commonlycausedbyliverandheartdiseases . Malignant
ascitesmostfrequentlypresentingynaecological,gastroin -
testinalandbreastcarcinomas .A combinationofmalignant
ascitesandcarcinomatosisoftheperitoneumispresentin15–
30%ofcases(2).
Ascitescanbeexudativeandtransudative.Transudates
makeup90% ofasciticfluidsandtheyarecausedbycondi-
tionsofnon-malignantaetiology.Thisfluidisclear,witha
smallnumberofcellsandlowlevelofalbumin.Anexudateis
usuallymalignant,cloudy ,withagreaternumberofcellsand
ahigherlevelofproteinsthantransudate(3).
Thisdifferentiationisenhanced bytheserum-ascites-
albumingradient(SAAG). IfSAAG is> 1.1,thevaluespoint
toatransudatecausedbyportalhypertension,cirrhosis,he -
paticcongestion,portalveinthrombosisetc. IfSAAG is< 1.1,
theexudateismostlikelyofmalignantaetiologyorbyanin -
fectiousprocessintheperitoneum,nephroticsyndromeand
hypoalbuminaemia frommalnourishment(4).
Itisbelievedthatthepathogenesisofmalignantascites
ismultifactorialandthatthemostimportantpathogenetic
mechanismsinclude increased vascularpermeability,lym -
phaticdrainageobstruction,increaseddifferenceinhydraulic
pressureandreduceddifferenceinoncoticpressure(5).
A two-waypermeabilityofbloodvesselsisnecessary
fornormalsupplyofnutrientsandgasesandwasteremoval.
Thepermeabilitycanbebasal,acutevascular(aconsequence
ofshortexposuretovascularendothelialgrowthfactor–
VEGF) andchronic,whichisacharacteristicofpathological
angiogenesis .
Vascularendothelialgrowthfactorcausestheprocessof
neovascularizationandangiogenesisandtheresultishyper -
permeabilityandincreasedporosityoftheendothelialmem -
brane ,whichisfollowedbymigrationandproliferationof
endothelialcellsandcreationofnewcapillaries.Besides
VEGF ,neovascularizationisalsoinfluencedbyfibroblast
growthfactor(bFGF),angiogenin,transforminggrowthfac -
tor(TGFαandβ)andinterleukin8 .
Ascitesisthemostcommoncomplaintofpatientswith
ovariancarcinoma.In54% ofpatientswithperitonealcarci-
nomatosis,asciteswasthefirstdetectablesignofmalignancy
(6).
M orethantwo-thirdsofpatientsthatreporttothedoctor
havegradesIII andIV ofthedisease . Survivalrateinad -
vancedstages(IIIandIV)is5– 20%(7).
Thepresenceofmalignantascitesinmalignanciesofthe
secondarylocalizationisaworseprognosticmarkercompared
toovariancarcinoma,andthesurvivalperiodfromthemoment
ofdetectionis7–13weeks(8).
Thepurposeofthisstudywastotestthevalidityofas -
citicfluidcytologyinthedetectionofpathologicalcytology
results,totestthepercentageoffalsepositiveandfalse
negativeresultsofasciticfluidcytologyandtodeterminethe
validityofperitonealcytologyinrelationtothehistopatho -
logicaltypeofovariantumour.
SUBJECTS AND METHODS
A retrospectiveanalysiswasusedfortheresearchwhichin -
cluded170peritonealcytologyresultsduringtheperiodfrom
January2010toDecember2012.Theexperimentalgroupwas
composedof76 cytologicalfindingsobtainedfrompatients
diagnosedwithovariancarcinoma,whilethecontrolgroupin -
cluded94cytologicalfindingsofbenignovariantumours(fi-
broma,dermoid cysts,endometrioses,serousandmucinous
cysts)andascitesinlivercirrhosis.Thepatientswithovarian
carcinomawereinstagesIIIaswellasinstagesIandIIcbut
onlyincaseswhereoperativeandpathologicalfindingsindi-
catedarupturedorperforatedtumourcapsule.C ytological
findingsofasciticfluidand peritonealcavityeffusionwere
sampledattheGynecologyandObstetricsClinicinNišand
scannedattheInstituteofPathologyattheC linicalCenterNiš .
Allresultswerestatisticallyprocessedbyaformulaformeas -
uringvalidity,χ2testandpresentedintablesandgraphs .
Results
Thehistopathologicaldistributionoffindingsfromtheexperi-
mentalgroupispresentedinFig.1.Thehighestpercentageof
patientshadseroustypeofcarcinoma(71%),whichisdes-
cribedasthehistopathologicallymostfrequenttypeofovarian
carcinomaintheliterature.
C ytologicalfindingsofasciticfluid obtainedfromthe
mentioned76 patientswithovariancarcinomawhowereclas -
sifiedintotheexperimentalgroup(stageIIIandstageIandIIc
withperforatedcapsule)wasnegative-falsenegativein23
(30.2%)patients(Table1).
Conclusión: La citología peritoneal del líquido ascítico es altamente específica , pero su sensibilidad es
relativamente baja, particularmente en el caso del carcinoma ovárico endometrioide. Con el fin de au -
mentar la sensibilidad, la citología peritoneal debe combinarse con anticuerpos monoclonales y otros
marcadores bioquímicos e immunohistoquímicos.
Palabras claves :Ascitis,citologíadellíquidoascítico,carcinomaovárico
West Indian Med J 2015; 64 (3): 237
Živadinović etal
238 A scitesinO varianCarcinoma
Table1: Distributionoffalsenegativecytologicalfindingsofasciticfluidwithrespecttothe
histologicaltypeoftumour
Histologicaltype Totalnumberof Numberofnegative Percentageoffalse
histologicaltype cytologicalfindings negativefindings
Serous 54 15 27.77%
Mucinous 62 33.33%
Endometrioid 43 77%
Clearcell 62 33.33%
Anaplastic 41 25%
Granulo-cellular 20 0%
Total 76 23 30.2%
Table2 : Distributionoffalsepositiveperitonealcytologyfindingswithrespecttothecause and
histologicaltypeoftumour
Histologicaltype Totalnumberof Numberofpositive Percentageoffalse
histologicaltype cytologicalfindingspositivefindings
Fibroma 90 0%
Dermoid 11 0 0%
Endometrioma 13 2 15.38%
Serous 39 2 5.12%
Mucinous 20 2 10%
Livercirrhosis 20 0%
Total 94 6 6.38%
Fig .1: Distributionof histopathologicalfindingsfromtheexperimental
group.
Fig .2: Distributionofhistopathologicalfindingsfromthecontrolgroup.
Histopathologicaldistributionoffindingsfrom thecon -
trolgroupisshow ninFig.2anditalsoshows thehighestfre -
quencyofbenigntumourstobeoftheseroustype(41.98%).
The findingsofperitonealcytologyobtainedfrom thecontrol
group (94patients)withbenignhistopathologicalovariantu -
mourorascitescaused bylivercirrhosiswas positive-false
positiveinsix(6.38%)patients(Table2).
Figure3shows thatthepercentageoffalsenegativere -
sultsinpatientswithovariancarcinomaintheexperimental
group wa s30.2%,whereasthepercentageoffalsepositivere -
sultsinpatientswithbenigntumourwa s6.38%.
Usingtheformula formeasuringsensitivityand speci-
ficity ,thedata forperitonealcytologyvaliditywereobtained.
The measuredsensitivityofperitonealcytologywas 68.92%,
specificitywas 93.61%,positivepredictivevalue(PPV)was
89.65%,negativepredictivevalue (NPV) was 78.57%and
overallvaliditywas 82.35%(Fig.4)
Table1shows thenumberoffalsenegativeresultswith
respecttohistologicaltypeofmalignanttumo ur.Itcanbeseen
that,withrespecttothehistologicaltype,thehighestpercent-
age (77%)offalsenegativeresultswas inendometrioidovar -
iancarcinoma (out of fourendometrioidcarcinoma s,three
239
werefalsenegative results),whilethe lowest percentagewas
withthegranulo-cellulartype.Althou gh thepercentagesshow
thatserous carcinomawas themost frequent(71%ofallfind -
ings),27.77%werefalsenegative.The differenceinfrequency
offalsenegativecytologicalfindingswithrespecttothehis -
tologicaltypeisstatisticallysignificantbecause χ2e=34.75 >
χ20.01 =9.21.
Table2shows the numberoffalsepositiveresultswith
respectto thehistologicaltypeof tumour.The highestper -
centageoffalsepositiveresultsinthe controlgroup wa swith
endometrioidova riancysts–15.38%(out of 13endometri-
omas,two werefalsepositive).The differencewas notstatis -
ticallysignificantbecause χ2e=5.34< χ20.05 =5.99.
DISCUS SION
The maincharacteristicsof malignantascitesareincreased
concentrationofascticfluidproteins,increasedlactatedehy -
drogenase,a largenumberofleukocytesandpositivecytology
forthepresenceofmalignantcells.
A positivecytologicalfindingisimportantinsubclassi-
ficationof stagesIand IIof thediseaseand itrepresentsan
importantpredictivefactorinprognosisandrecurrence.How -
ever,anincreasingnumberofstudiesshows thatmorphologi-
calexaminationofcytologicalsamplesisnota highlysensitive
diagnostictool.
Thereasonforfalsepositivecytologicalresultsisinade-
quate interpretationof reactivelyalteredmesothelialcells.
Thesecellsareenlarged and theyhave a densecytoplasm,a
bignucleuswitha nucleolus and may containvacuoles. En -
dosalpigniosiscanalsopresenta problemindifferentialdiag -
nosis and itshould be carefullydetermined that itis a
well-organizedgroupofuniformcellswithscarcebasophilcy -
toplasm and a nucleuswitha well-definedmembrane,fine
chromatinand a smallnucleolus.
Endometriosisrepresenteda big problemindifferential
diagnosis.Inendometriosis,thereareroundcellsorganized
intothree-dimensionalgroupsandlayers,withroundandbean-
shaped nucleuswhichhasfinechroma tinand scarcevacu o -
latedcytoplasm.The most sensitivefindingsforendometriosis
aremacrophages with haemos iderin.
Thesearethereasonswhy literaturedata statethatperi-
tonealcytology canbefalsepositivein4.5%ofcases(inour
researchitwas 6.38%).Some recentstudiesalso describea
relativelyhighpercentageoffalsenegativefindingswhichex -
ceeds20%(inourresearch itwas 30.2%).Thereasonsfor
such a highpercentageoffalsenegativecytologicalresultsof
asciticfluidmay beinthebad distributionofcellsinthesam -
pledasciticfluid,badpreparation ,orinsufficientcellexfolia -
tion ,andsincecytologyisa subjectivemethod,errorsmay be
due toinadequateinterpretationoffindings(9).
The highestpercentageof falsenegativeresultsinour
researchwas inthegroupofpatientswithendometrialcarci-
noma. Inaddition to thestatedparameterswhichpointtothe
benignnatureofendometrioidcarcinoma,cytologicalelements
whichpointtotheendometrioidcarcinoma should alsobecon -
sidered:theappearanceof three-dimensionalgroupsof cells
withlargepleomorphiceccentricnucleiwithrou gh chromatin,
emphasizednucleolusand abundant cytoplasm. Cytological
findingwhichpointsto serous carcinoma shows cellswhich
areseparateorinirregularincohesivegroups,withlargepleo -
morphicnucleusand emphasizednucleolus.
The sensitivityof peritonealcytologystatedinother
studiesrangesfromonly50to 60%(inourresearch itwas
68.92%),up to 97%dependingon thestudy,diseasestageand
peritonealinclusion(10).
InpatientswithstageIc,cytologywas positivein75%
andwhentheperitoneum was included ,itwas 94%(11).The
sensitivityofcytologywhen peritoneum was included was
82.9%and specificitywas 98.1%[inourresearch itwas
93.61% ](12).The examinationoftotalvalidityofcytologyin
some publications showed somewhat lowersensitivitywhich
was 60%and highspecificityofalmost 100%(13).
The resultofprimarycytologyofasciticfluidisan im -
portantparameterinthe diagnosis,therapeuticapproachand
diseaseprognosis.The resultofsecondarycytologyafterthe
treatmentisalso an importantindependentprognosticmarker
which ishighlycorrelatedwiththeoptimaleffectofsurgical
treatment,recurrenceand overallsurvivalrate.Inpositive
Živadinović etal
Fig .3: Combinedrepresentationofthedistributionofperitonealcytology
findingsfromexperimental(pathologicalfindings –stagesI,IIand
III)and control(benignfindings)groups .
Fig4: Graphicrepresentationofvalidityparametersofperitonealcytology .
PPV: positivepredictivevalue;NPV: negativepredictivevalue
240
secondarycytology,survivalis13to32months,whileinnega-
tivecytology ,itis> 48months(14).
Consideringallaspectsofvalidityofasciticfluidcytol-
ogy,particularlywithcertainhistologicaltypes(endometrioid
ovariancarcinoma),itisbelievedthat furtherresearchis
necessaryaswellastheuseofspecificadditionalimmunohis -
tochemicalmarkersinordertoreduceas much aspossiblethe
percentageofcytologicalerrorswh ich causewronggrading
andinappropriatetreatment.
Inadditionto measuringtheconcentration of alkaline
phosphatase,lactatedehydrogenase,fibronectin,aswellastu -
mou rmarkersCA-125,CEA ,p53,βHCG,thereisalsoa spe -
cific group of panel antibodies,primarilyMOC -31and
Ber-EP4.Theseantibodiesareimportantfordifferentiationof
mesothelialandcancerouscellsand itmay alsocontributeto
thedifferentiationof antibodiesintoadherentcellsand non-
adherentcarcinogenic(15).
Fromtheremainingbiomarkers,telomerasehas lately
beenmostlytested.Telomeraseisan enzymenecessaryfor
normalreplicationof chromosome sandconstantgrowthof
cancercells.Telomeraseactivityisabsentinthemajorityof
somat iccells.On theotherhand,telomerase expressionhas
beenconfirmedinalmost100%ofovariancarcinomacases .
Contraryto24–54%ofcases ,where ,despitethefactthatresi-
dualdiseasewas diagnosed,cytologyandsecond-looksurgery
werenegative,telomerasewas almost 100%positive(16).
Inourstudy ,we concludedthatperitonealcytologyof
asciticfluidishighlyspecific(93.61%)butithas a relatively
low sensitivity(68.92%).In30.02% ,peritonealcytologyhad
falsenegativeresultsand in6.38% ,itshowe dfalsepositive
results.Such distributionof cytologicalfindingsmaycause
inadequategradingofthedisease and inadequate therapeutic
approach (notapplyingthenecessarychemotherapyorappli-
cationofunnecessarychemothe rapy).The highestpercentage
offalsenegativeresultswas withendometrioidovariancarci-
nom a (77%) ,so itisnecessarytobeverycarefulwhen cyto -
logicalscanning isusedwiththistype.Inorderto increase
sensitivity,peritonealcytologyshould becombinedwithother
availablebiochemicaland immunohistochemicalmarkers.
REF ER ENCES
1. RunyonBA. Careofpatientswithascites.N EnglJMed 1994;330: 337–
42.
2. Suma L,Thomas J.Malignantascites:a reviewof prognosticfactors,
pathophysiologyand therapeuticmeasures.WorldJGastrointestSurg
2012;4: 87–95.
3. BeckerG,GalandiD,Blum HE. Malignantascites:systematicreview
and guidelinefortreatment.EurJCancer2006;42: 589–97.
4. Adam RA,Adam YG. Malignantascites:past,present,andfuture.JAm
CollSurg2004;198: 999–1011.
5,StanojevicZ,RancicG, RadicS,Potic-ZecevicN, DjordjevicB,
MarkovicM etal.Pathogenesisofmalignantascitesinovariancancerpa -
tients.ArchOncol 2004;12:115–8.
6. GarrisonRN,KaelinLD, Galloway RH,HeuserLS. Malignantascites.
Clinicaland experimentalobservations.Ann Surg 1986;203:644–51.
7. Shen-GuntherJ ,Mannel RS .Ascitesas a predictorof ovarianmalig -
nancy.GynecolOncol 2002;87: 77–83.
8. Mackey JR,VennerPM. Malignantascites:demographics,therapeuticef -
ficacyand predictorsofsurvival.Can JOncol 1996;6: 474–80.
9. OscarL. Challengesintheinterpretationofperitonealcytologicspeci-
mens.ArchPathol Lab Med 2009;133: 739–42.
10.RunyonBA,HoefsJC,Morgan TR.Asciticfluidanalysisinmalignancy-
relatedascites.Hepatology1988;8: 1104–9.
11.Cheng L ,WolfNG ,Rose PG ,RodriguezM ,Abdul-KarimFW .Peri-
tonealwashing cytologyof ovariantumorsof low malignantpotential:
correlationwithsurfaceovarianinvolvementand peritonealimplants.
Acta Cytol1998;42: 1091–4.
12.Zuna RE ,BehrensAJ. Peritonealwashingcytologyingynecologiccan -
cers:long-termfollow-upof355patients.Acta Cytol1996;88: 980–7.
13.Karoo R,GarcceaG. How valuable isasciticcytologyinthedetection
and management ofmalinancy.PostgradMed J2003;79: 292–4.
14.SiropS,Kanaan M,WieseD,Dutt N,KarlaV,SinghTT etal.A second
peritonealcytologyasa prognosticfactorinepithelialovariancancer[Ab -
stract] .JClinOncol 2011;29(Suppl): e15558 .
15.HechtJL,PinkusJL,PinkusGS. Monoclonal antibodyMOC-31 reac -
tivityas a markerforadenocarcinoma incytologicpreparations.
Cancer 2006;108: 56–9.
16.Duggan BD ,Roman LD ,Muderspach LI.Detectionof ovariancancer
cells:comparisonofa telomeraseassayandcytologicexamination.JNatl
CancerInst1998;90: 238–42.
AscitesinOvarianCarcinoma