O
riginal paperV
alue Of perfusiOnCT
parameTers,
miCrOVessel densiTyand
Vegf
expressiOn in differenTiaTiOn Of benign andmalignanT prOsTaTe TumOurs
ElzbiEta luczynska1, anna Gasinska2, PawEl blEcharz3, andrzEj stElmach4,
barbara alicja jErEczEk-Fossa5,6, marian rEinFuss7
1Radiology Department, Center of Oncology, M. Sklodowska-Curie Memorial Institute, Cracow Branch, Krakow, Poland 2ClinicalRadiobiologyDepartment,CenterofOncology,M.Sklodowska-CurieMemorialInstitute,CracowBranch, Krakow,Poland 3GynecologicOncologyDepartment,CenterofOncology,M.Sklodowska-CurieMemorialInstitute,CracowBranch, Krakow,Poland 4SurgeryDepartment,CenterofOncology,M.Sklodowska-CurieMemorialInstitute,CracowBranch,Krakow,Poland 5RadiotherapyDepartment,EuropeanInstituteofOncology,Milan,Italy 6UniversityofMilan,Milan,Italy 7RadiotherapyDepartment,CenterofOncology,M.Sklodowska-CurieMemorialInstitute,CracowBranch,Krakow, Poland Thepurposeofthisstudywastoassessthecorrelationbetweenparametersevaluat-edusingcomputedtomographyperfusion(CTP)andmicrovesseldensity(MVD), thevascularendothelialgrowthfactorlabellingindex(VEGFLI),aswellasknown clinicopathological indicators of tumour malignancy, in non-advanced prostatic cancer.
We included 110 patients with early stage prostate cancer who were subjected toCTexaminationsfollowedbyradicalprostatectomybetween2007and2011 (inthisanalysisweincludedonlypatientsdiagnosedwithCT).Bothinaffected andinhealthytissuethefollowingperfusionparameterswereassessed:bloodflow (BF),bloodvolume(BV),meantransittime(MTT)andpermeability-surfacearea product(PS).AftersurgeryintheresectedprostatetumourtissuetheMVDand VEGFLIwereassessed. ThemeanBFandPSvaluesweresignificantlyhigherincarcinomaswithhighhis-tologicalgrade(p=0.02).Thesensitivity,specificityandaccuracyofthethreshold BFvalue,forthedistinctionbetweenmalignantandhealthyprostatetissue,were: 67%,54%and59%respectively.ForBVsensitivitywas71%,specificitywas52%, andaccuracywas48%.MicrovesseldensitysignificantlycorrelatedwithBV,MTT andPS(p<0.05),whileVEGFLIdidnotcorrelatewithanyoftheperfusionpa-rameters. OurresultssuggestthatBFandPSmightbehelpfulindiscriminationbetween benignandmalignantprostatetissue,whilethepositivecorrelationbetweenBV, MTT,PSandMVDmightsuggesttheirpotentialutilityinassessmentofcancer angiogenesis.
Key words: perfusioncomputedtomography,prostatecancer,microvessels,vascu-larendothelialgrowthfactor.
Introduction
Prostate cancer is a biologically heterogeneous disease.Manypatientswithlocalized,slow-growing cancerssurviveforalongtimeevenintheabsence of therapy, while others develop metastases despite apparentlyorgan-confineddiseaseandapplicationof localtherapy[1,2].
Prostatebiopsyisstillconsideredthegoldstandard fordiagnosisofprostatecancer.Thequestionarises, whetherconventionalprocedures,suchascomputed tomography (CT) and magnetic resonance imag-ing(MRI)aswellastransrectalultrasound(TRUS), whichhavenotyetbeenprovenreliableinthediag-nosisofprostatecancer[3],mightbeofdiagnostic valueinthefuture.
The development of prostate cancer is a multi-step process, advancing from high-grade prostatic intraepithelial neoplasia (PIN) to focal carcinoma, invasivecarcinoma,andfinallytometastaticdisease. Therefore the planning of future therapies should involve targeting the molecules, that are related to eventsassociatedwitheachstepofprogression.One ofsucheventsisangiogenesis–averycomplicated processrequiringextensiveinteractionsbetweencells (cancer,epithelialandimmunecells),cytokines,and extracellularmatrixcomponents[4-6].Studieshave demonstrated that the expression of factors stim-ulating angiogenesis, such as vascular endothelial growthfactor(VEGF),platelet-derivedgrowthfac-tor(PDGF),andtransforminggrowthfactor(TGF), areincreasedinprostatecarcinoma[7,8].Moreover, ithasbeenshownthatthereisaprogressiveincrease inangiogenesisasprostatecanceradvancesthrough variouspathologicstages.
There are many parameters associated with the process of angiogenesis which may be evaluated in tumours. One of such parameter is microvascular density(MVD)[9].Thisparametercanbeassessed based on the expression of CD34 or CD31 on the endothelial cells of the blood vessels. On the other hand the expression of VEGF might be an indica-torofintensityofneoangiogenesis[10].Apartfrom providinginformationontumourinvasiveness,these markershavebeenusedtochoosetheappropriatean-ti-angiogenictherapy[11]. FunctionalCTisahighspatialresolutiontechnique forassessingtumourneovasculature,butthesignal-to-noiseratioremainspoorascomparedwithMRI[12]. Inoncology,CTPisarecognizedmethodfortheassess-ment of tumour vascularity. This method allows the evaluationofcapillaries’permeabilitybasedontheper-meabilitycoefficient(PS).Thereforeitisbelievedthat, CTPcanreflecttumourmicrovasculardensity[13].
Inmanytypesoftumour,acorrelationhasbeen notedbetweenperfusionparametersinCTandan-giogenesis markers assessed in the tumour [14].
However, these relationships are poorly understood inprostatecancer.Recently,Osimaniet al.[15]found apositivecorrelationbetweenMVDandCTPparame-ters.OntheotherhandthecorrelationbetweenVEGF expressionandtheaforementionedparameterhasnot beenstudiedsofar,andhenceremainsunknown. Therefore,thepurposeofthepresentstudywasto assessofthecorrelationbetweenCTPparametersand MVDorVEGFaswellastherelationshipbetweenall theabove-mentionedparametersandknownclinico-pathological indicators of tumour malignancy. We wouldliketodeterminewhetherCTP,whichisless invasivethantumourbiopsy,couldbehelpfulindis-tinguishingbetweenbenignandmalignantprostate tissue as well as between less and more malignant prostatecarcinomas.
Material and methods
Patients
One hundred and ten patients with early stage prostatecancer,whohadbeendiagnosedusingTRUS/ BGI between 2007 and 2011 were included in the presentstudy.Allpatientsweresubjectedtoradical prostatectomywhichconsistedofremovalofprostate andseminalvesicleswithin2to4weeksaftertheCT test.Themeanageofpatientswas62.7±6.4years.
Clinicalstagingwascarriedoutaccordingtothe American Joint Committee on Cancer (AJCC) and the International Union Against Cancer. Grading wasestablishedbasedontheGleasonscore(GS)dif-ferentiationsystem(range2-10)[16],independently by two experienced pathologists. Patients were di-videdintothreeGleasonscoregroupsfollowingthe criteriadefinedbytheAJCC:welldifferentiatedtu-mours(Gleasonscore≤6),moderatelydifferentiated tumours(Gleasonscoreof7)andpoorlydifferentiat- edtumours(Gleasonscore8-10)[17].Clinicalchar-acteristicsofthegrouparepresentedinTableI.
The protocol was approved by the local bioethical committeeandeverypatientsubmittedwrittenconsent.
Computed tomography examination
Computed tomography examinations were per-formedwitha16-sectionmultidetectorCT(MDCT) scanner (LightSpeed 16; GE Healthcare, Milwau-kee,Wis).Preliminarynon-contrastCTofthepelvis (5-mmthickness)wasperformedtolocatethepros-tate.Thetotalareaofexaminationwas4cm.After detectingthecentreoftheprostategland,50mlof contrast medium was administered and a scan was performed2cmupwards,followedbyasecondinjec-tionof50mlandascanof2cmdownwards.Atotal of100mlofnonioniciodinatedcontrastmaterialwas injected(Ultravist370mgI/ml;BayerScheringPhar-ma,Germany)followedby50mlofsalinesolutionat
arateof5ml/sviaan18-gaugecannula,whichwas placedintherightantecubitalveininallthepatients to exclude any source of variability. Computed to-mographyperfusionscanningstarted5secondsafter contrast administration, with the following param-eters:4contiguous5-mmreconstructedsectionsat a constant table position, 1-second gantry rotation time, 80 kVp and 180 mA. Images were acquired everysecondfor50seconds.
Immediately after completion of CTP scanning, MDCToftheabdomenandpelviswasperformedby using16×0.625mmcollimation;1.25mmsection thicknessandincrement;rotationtime0.6s;speed 9.38mm/rotation;FOV18cm;intravenouscontrast material (Ultravist 370 mg I/ml; Bayer Schering Pharma,Germany);1.5ml/kgataninjectionrateof 2ml/s;acquisitiondelay70s.
Theobtainedimageswereanonymizedandtrans-ferred to an image-processing workstation (Advan-tage Windows 4.2, GE Healthcare). Computed tomography perfusion data was analyzed by two radiologists working separately (EL and STD) with 4and3yearsofexperienceinCTPimagingrespective-ly.Commerciallyavailablesoftware(CTPerfusion4, GE Healthcare) was used for analysis, using a de-convolution-basedtechnique.Thearterialinputwas obtainedfromastandardizedplaceintheregionof theexternaliliacartery(EIA),withselectionofthe
sectionthatallowedforbestvisualizationinorderto avoid partial volume artefacts. A time-attenuation curve,expressedinHU/s,wasautomaticallygenerat-edbythesoftwareforthearterialinput;itsgeometric evaluationallowedreaderstoassessthetimingofthe CTPscansineachpatient,excludinganyearlyen-hancement,identifyingcorrectlytheendofthefirst passofcontrastmaterial,andexcludinganyrecircu-lationeffectintheCTPmeasurements. Functionalmapsofbloodflow(BF),bloodvolume (BV)andmeantransittime(MTT)weregenerated accordingtothecentralvolumeprinciple,whichre-latesBF,BV,andMTTbytheequation:BF=BV/ MTT[18].
Regions of interest (ROI) were manually drawn alongthevisiblemarginsoftheobviousprostatecancer inallsectionsinwhichcancerwasvisibleandsavedfor eachpatient.Meanvaluesofperfusionparameterswere thencalculatedforeachpatientintumourandhealthy tissueseparately.Fordisplaypurposes,thefunctional mapswerepresentedincolouredimages(Fig.1). Immunohistochemical assessment of parameters associated
with tumour angiogenesis
Immunohistochemistry was performed on forma-lin-fixed, paraffin-embedded tissue sections (5 µm
Table I. Therelationshipbetweenevaluatedperfusionparameters,microvesseldensity(MVD)andVEGRLIexpression inrelationtoclinicalandhistologicalfeaturesofprostatecancer parameTer TOTal bf (ml/min/100 g) mean ± se bV (ml/100 g) mean ± se mTT (s) mean ± se ps (ml/min/100g) mean ± se mVd mean ± se Vegrli mean ± se Tumors 110 43.9±1.2 5.0±0.2 8.2±0.2 31.7±1.2 100.1±2.7 14.8±1.4 Normaltissue 110 25.2±0.9 2.4±0.1 6.5±0.1 19.6±0.9 – – p-value1 0.00 0.00 0.00 0.00 – – Degreeofhistologicalmalignancy G1 65 42.0±1.5 4.7±0.2 8.1±0.2 30.2±1.7 99.9±3.6 16.2±2.0 G2 39 45.1±1.8 5.4±0.3 8.4±0.2 32.9±1.7 102.0±4.6 13.1±2.6 G3 6 56.7±8.7 5.0±0.6 8.4±0.2 40.1±6.0 90.4±3.8 15.8±4.6 p-value2 0.02 0.2 0.55 0.14 0.65 0.62 pTNM 1 8 51.9±7.2 4.5±0.4 8.0±0.5 34.9±5.5 99.7±7.1 10.7±1.9 2 66 42.4±1.5 4.9±0.2 8.1±0.2 29.2±35.7 97.6±3.0 14.4±2.1 3 31 44.9±2.1 5.4±0.3 8.4±0.2 35.7±2.2 107.0±6.3 17.0±2.7 4 5 44.2±5.7 4.9±0.6 8.0±0.3 35.0±5.1 90.0±17.8 19.1±6.8 p-value2 0.24 0.45 0.84 0.09 0.91 0.70
1probability of difference between mean value depend on T-test for dependent samples 2probability of difference between mean value depend on ANOVA test
BF – blood flow, BV – blood volume, MTT – mean transit time, PS – permeability-surface area product, MVD – microvessel density, VEGFLI – vascular endotheli-al growth factor labelling index
thick).Forantigenunmasking,afterdeparaffinizatio- nandrehydration,weappliedheatingofslidesinTar- getRetrievalSolution(TRS)(DakoCytomationDen-markA/S,Glostrup,Denmark):(i)forVEGF–pH9, temperature 95-99°C, 20 min., (ii) for CD34 – pH 6, temperature 95-99°C, 40 min. For quenching of the activity of endogenous peroxidases, slides were incubated in 0.3% H2O2 diluted in methanol for 20min.Then,after20min.incubationwith10%nor-malgoatserum,sectionswereincubatedovernightat 4°Cwith:(i)anti-CD34mouseanti-humanmonoclo-nalantibody,diluted1:200,(ii)anti-VEGFamouse monoclonalantibody,diluted1:25(DakoCytomation DenmarkA/S,Glostrup,Denmark).Visualizationwas carriedoutusing:DAKOEnVisionvisualisationsys-tem(37°C,1hincubation),andVECTORImmPRESS ReagentKit,forCD34andVEGFrespectively.Finally, thesectionswereincubatedwith3,3’-diaminobenzi-dine(DAB)andcounterstainedwithhematoxylin.For thenegativecontrolprimaryantibodieswereomitted.
Cytoplasmic VEGF expression was counted as percentageofpositivelyimmunostainedcellsin500-1000tumourcells(VEGFlabellingIndex–VEGFLI). Microvasculardensity(MVD)wasassessedin7-10 tumour fields (0.292 mm2), and expressed as the
meannumberofvesselsper1mm2
.Bothindividu-alCD34-immunopositiveendothelialcellsandlarge vesselswithlumenwereincluded.
Fig. 1. Perfusionmaps:bloodvolumeBV(A),bloodflowBF(B),meantransittimeMTT(C)andpermeabilitysurface PS(D).InallpresentedmapstherightROIrepresentsmanuallyoutlinedenhancement(prostatecancer).TheleftROI representsnormalprostatictissue
A
B
Statistical analysis
A one-sided Student’s t-test (2 groups) and ANOVA(foralargernumberofgroups)wereusedto analysedifferencesinmeanvaluesofcontinuous(per- fusionparameters,MVDandVEGFLI)andcategor-icalvariables(gradeandTNM).Pearson’scorrelation coefficient was applied to measure of the strength anddirectionofthelinearrelationshipbetweentwo continuousvariables.AReceiverOperatingCharac-teristic(ROC)analysiswasconductedtofindwhich perfusion parameters reveal the best diagnostic ac-curacy level in tumour staging differentiation. The sizeoftheareaundertheROCcurvewasexamined usingtheZ-test.Thelimitofstatisticalvariationwas acceptedatthelevelofp<0.05.Thecalculations were performed using the STATISTICA 10.0 soft-ware(StatSoft,Inc.,Tulsa,OK.USA).
Results
Perfusion computed tomography results
Themeanvaluesofperfusionparametersforpros-tatecancerandnormaltissuearepresentedinTableI. Significantly higher values of BF, BV, MTT and PS were found for tumours than for normal tissue
(p<0.05)(TableI).Allperfusionparameterswere correlatedwitheachother(p<0.05),apartfromBF andMTT(TableII).
ThemeanvalueofBFwassignificantlylowerin low grade than in high grade carcinomas (Table I, p<0.05).Similarly,themeanvaluesofPSwerelow-erfortheG1tumoursthanforG3tumours,butthis difference did not reach statistical significance (Ta-bleI).Ontheotherhandnodifferencebetweenthe meanvaluesofBVorMTTandtumourhistological gradingwereobserved(p>0.05)(TableI). TheROCcurveanalysisshowedthatBFandBV maybeusedtodistinguishbetweenwell-,moderate-lyandpoorlydifferentiatedtumours(TableIII).The areasundertheROCcurve(AUC)inthesecaseswere significantlyhigherthan0.5(p=0.05forBFand 0.01forBV)andtherefore0.6ml/min/100gforBF and0.62ml/100gforBVmaybeusedasaclassifiers forprostatetissuesofdifferentdegreeofmalignancy (TableIII).
In the case of BF, the threshold value for more andlessaggressivetumourswas41.7ml/min/100g, withthetestsensitivityof67%,specificity54%and accuracy59%.InthecaseofBV,theoptimalthresh- oldvaluewas4.55ml/100mg,with71%sensitiv-ity, 52% specificity and 48% accuracy (Fig. 2). In
Table II. Thecorrelationbetweenmicrovesseldensity,VEGFexpressionandperfussionparametersinprostatecancer
bV (ml/100 g) mTT (s) ps (ml/min/100 g) mVd Vegrli BF p = 0.000 r = 0.61 p=0.515 r=0.06 p = 0.000 r = 0.56 p=0.290 r=0.10 p=0.795 r=0.03 BV p = 0.000 r = 0.58 p = 0.000 r = 0.55 p = 0.035 r = 0.20 p=0.155 r=0.14 MTT p = 0.001 r= 0.30 p = 0.026 r = 0.21 p=0.288 r=0.10 PS p = 0.022 r = 0.22 p=0.616 r=0.05 MVD p=0.761 r=0.03
r – coefficients of correlation; p – probability
BF – blood flow, BV – blood volume, MTT – mean transit time, PS – permeability-surface area product, MVD – microvessel density, VEGFLI – vascular endotheli-al growth factor labelling index
Table III.TheresultsoftheROCcurveanalysis
bf (ml/min/100 g) bV (ml/100 mg) mTT (s) ps (ml/min/100 g) Theareaunderthecurve AUC±SE 0.60±0.06 0.62±0.05 0.58±0.06 0.57±0.06 p 0.05 0.01 0.09 0.12 Thresholdvalue 41.7 4.55 8.71 20.71
thecaseofMTTandPS,nocut-offvalueshavebeen obtainedbothforeitherwelldifferentiatedorpoorly differentiatedtumours.
The correlation between perfusion
computed tomography results and tumour vascular density and expression of VEGF
Intheanalysedgroupoftumours,meanmicroves-sel density was 100.1 ±2.7 vesIntheanalysedgroupoftumours,meanmicroves-sels/mm2, while the
meanvalueofVEGFLIwas14.8±1.4.Nocorrela-tionwasfoundbetweenMVDandVEGFLI(TableII, p>0.05).
Microvessel density and VEGFLI were correlated neitherwithTNMnorwithgrade(TableI,p>0.05).
ThetumourMVDwassignificantlypositivelycor-relatedwithBV,MTTandPS(p<0.05,TableII). Correlationcoefficientswere0.20,0.21and0.22for BV,MTTandPSrespectively(TableII).Therewas no correlation between microvessel density and BF (p>0.05,TableII).Theimmunohistochemicalex-pressionofVEGFproteindidnotcorrelatewithany oftheperfusionparameters(p>0.05)(TableIII).
Discussion
In the case of prostate cancer MRI is a gold di-agnostic standard. On the other hand computed tomography (CT) was not considered as such, and henceintheliteraturetherearefewreportsreferring toapplicationofthismethodinprostatecancer.
Application of a 64-section MDCT scanner in PCTofpatientswithprostatecarcinoma,byOsimani et al.[15]confirmedthat,PCTparameterscorrelate wellwithmicrovesseldensity.Moreover,byapplying theabove-mentionedtechnique,theauthors[15]ob-tainedthevisualizationofmalignantfociin22from 24tumors,anddemonstratedsubstantialdifferences inmeanvaluesofBV,MTTandPSbetweenprostate cancer, benign prostatic hyperplasia (BPH) chron-ic prostatitis and healthy tissue. The ROC curve showed100%sensitivityandspecificityforBVand MTTtodiscriminatebenignandmalignantlesionsof theprostategland[15].
Similarly to the aforementioned results, but in alargerpatients’group(110patients)weobserved: significant differences between mean values of BV, BF,MTTandPSinnormaltissueandprostatecan-cer, a significant positive correlation between BV, MTTandPSandMVD,andadditionallyarelation-shipbetweenBFandgrade.BloodflowandBVmay beusedtodistinguishbetweenwell-,moderatelyand poorly differentiated tumours. This result suggests that,alltheabove-mentionedCTPparametersmay beofdiagnostic,prognosticandpredictivevalue.
AincreaseinBFvalues,describedbyusinneo-plastic tissue compared to healthy areas, may be explained by the opening of arterio-venous collat-eral circulation within the tumour. These branches are characterized by low resistance to the changes inbloodpressure,whichresultsinanincreaseinthe bloodflowwithinthecapillaries[19,20].
AhigherBVvalueinthetumourthaninhealthy tissuemayreflectincreasedmicrovasculaturedueto formationofnewvessels[21].Thishypothesismight be confirmed by the correlation between BV and MVD,whichwasfoundinthepresentstudy.
Lastly, a higher PS value in the tumour than in healthy tissue as well as its correlation with MVD, couldbeindicatorofgreaterpermeabilityofthewall ofthenewlyformedvesselscomparedwiththenor-malmicrovasculature.In2010,Bellomi[14]wrote that“ifCTPwereabletoreliablyidentifyfociinthe prostate,itwouldtheoreticallybepossibletotarget radiationtherapyandminimizeradiationdosetosur-roundinghealthytissue”[22,23].
As described before, similarly to Osimaniet al. [15]wefoundasignificantcorrelationbetweenmi-crovasculardensityandthreeparametersassessedin CTP: BV, MTT and PS. This result might suggest thatCTPisusefulinassessmentofprostatecancer angiogenesis(theprocessoftumour-inducedgrowth ofnewbloodvessels)andanimportantfactorindicat- ingdisease-specificsurvivalandtheriskofprogres-sionaftertherapy[24-26]. ThepredictivevalueofCTPparametersishighly possible because differences in vascular permeability andarchitecturebetweentumourandnormaltissues (detectedwithCTP)maycontributetodifferencesin oxygenation or gene expression and therefore
influ-Fig. 2. TheROCcurvesforPCperfusionparameters:BF (greencolour);BV(redcolour);referenceline(blackcolour). The relationship between sensitivity and specificity of the parameterswasshown 1.0 0.8 0.6 0.4 0.2 0 Sensitivity 1-Specificity 0 0.2 0.4 0.6 0.8 1.0
encetheresponsetoananti-cancertherapy(radiother-apy,chemotherapy,targetedtherapyetc.)[27-29]. InourstudyBVandPSnotonlyallowedfordis-tinguishingbetweenaffectedandnormaltissue,but alsosignificantlycorrelatedwithMVDandinsignifi-cantlyincreasedwithhighhistologicalgrade.Hence, bothBVandPSseemtobetheimportantperfusion parametersinprostatecancer,whicharenotonlypo-tential diagnostic parameters but also indicators of tumourangiogenesisandmalignancy.
TheMTTistheleastusefulparameter.Thisindi- catesthatcelldivisionentailstheformationofelon-gated vascular shoots, whose ends join together to form capillary loops. The endothelial cells in new-ly created vascular loops of tumour have abnormal shapeandsize.Theyhavewideintercellularconnec- tions,areirregularandhavealeakybasementmem-brane. The above-mentioned hypothesis might be confirmed by correlation between MTT and MVD [30],whichwasnotedinourstudy.
HighBFvalues,whichweresignificantlyrelated to high histological grade, but not correlated with MVD,mighthaveadiagnosticutility,beamarkerof tumourmalignancy,butnotanindicatoroftumour angiogenesis. Conclusions 1.BasedonCTPparametersitispossibletoreveal neoplasticfociwithintheprostate.Hencetheap-plication of this method will allow for diagnosis ofprostatecancerandforfocusingofradiotherapy intothefociandsparinghealthytissue.
2.ExistenceofapositivecorrelationbetweenMVD andBV,MTTandPS,aswellasrelationshipbe- tweenhighBFandhightumourgrademayfacil-itate pretreatment indication of more aggressive carcinomasandhence,applicationofmoreagres-sive treatment schedules/targeted therapies/anti-vasculartherapy.
The authors declare no conflict of interest.
References
1.Albertsen PC, Hanley JA, Penson DF, Fine J. Validation of increasingprostatespecificantigenasapredictorofprostate cancerdeathaftertreatmentoflocalizedprostatecancerwith surgeryorradiation.JUrol2004;171(6Pt1):2221-2225. 2.Bill-Axelson A, Holmberg L, Ruutu M, et al. Radical
pros-tatectomy versus watchful waiting in early prostate cancer. NEnglJMed2005;352:1977-1984.
3.HeidenreichA,AusG,BollaM,etal.EAUguidelinesonpros-tatecancer.EurUrol2008;53:68-80.
4.Sotomayor P, Godoy A, Smith GJ, Huss WJ. Oct4 A is ex-pressed by a subpopulation of prostate neuroendocrine cells. Prostate2009;69:401-410.
5.Folkman J. Angiogenesis-dependent diseases. Semin Oncol 2001;28:536-542. 6.LiekensS,DeClercqE,NeytsJ.Angiogenesis:regulatorsand clinicalapplications.BiochemPharmacol2001;61:253-270. 7.BostwickDG,PacelliA,BluteM,etal.Prostatespecificmem-braneantigenexpressioninprostaticintraepithelialneoplasia andadenocarcinoma:astudyof184cases.Cancer1998;82: 2256-2261. 8.LissbrantIF,LissbrantE,DamberJE,BerghA.Bloodvessels areregulatorsofgrowth,diagnosticmarkersandtherapeutic targets in prostate cancer. Scand J Urol Nephrol 2001; 35: 437-452. 9.delaTailleA,KatzAE,BagiellaE,etal.Microvesseldensity asapredictorofPSArecurrenceafterradicalprostatectomy. AcomparisonofCD34andCD31.AmJClinPathol2000; 113:555-562. 10.StefanouD,BatistatouA,KaminaS,etal.Expressionofvascu- larendothelialgrowthfactor(VEGF)andassociationwithmi-crovesseldensityinbenignprostatichyperplasiaandprostate cancer.InVivo2004;18:155-160. 11.LongoR,SarmientoR,FanelliM,etal.Anti-angiogenicther-apy: rationale, challenges and clinical studies. Angiogenesis 2002;5:237-256.
12.RussoG,MischiM,ScheepensW,etal.Angiogenesisinpros-tate cancer: onset, progression and imaging. BJU Int 2012; 110(11PtC):E794-808.
13.PetraliaG,BonelloL,ViottiS,etal.CTperfusioninoncology: howtodoit.CancerImaging2010;10:8-19.
14.Bellomi M, Viotti S, Preda L, et al. Perfusion CT in solid body-tumours.PartII:Clinicalapplicationsandfuturedevel-opment.RadiolMed2010;115:858-874.
15.OsimaniM,BelliniD,DiCristofanoC,etal.PerfusionMDCT of prostate cancer: correlation of perfusion CT parameters andimmunohistochemicalmarkersofangiogenesis.AJRAm JRoentgenol2012;199:1042-1048.
16.GleasonDF,MellingerGT.Predictionofprognosisforpros-tatic adenocarcinoma by combined histological grading and clinicalstaging.JUrol2002;111:58-64. 17.EditionS.CancerStagingManual.GreeneF,PageD,FlemingI, FritzA,BalchC,HallerD,etal.(eds.).Cancer2010;20-22. 18.PetraliaG,PredaL,D’AndreaG,etal.CTperfusioninsol-id-bodytumours.PartI:Technicalissues.RadiolMed2010; 115:843-857. 19.MilesKA,GriffithsMR.PerfusionCT:aworthwhileenhance-ment?BrJRadiol2003;76:220-231. 20.García-FigueirasR,GohVJ,PadhaniAR,etal.CTperfusion in oncologic imaging: a useful tool? AJR Am J Roentgenol 2013;200:8-19.
21.ChenY,ZhangJ,DaiJ,etal.Angiogenesisofrenalcellcar-cinoma: perfusion CT findings. Abdom Imaging 2010; 35: 622-628.
22.KambadakoneAR,SahaniDV.BodyperfusionCT:technique, clinical applications, and advances. Radiol Clin North Am 2009;47:161-178. 23.HarveyCJ,BlomleyMJ,DawsonP.FunctionalCTimagingof theacutehyperemicresponsetoradiationtherapyofthepros-tategland:earlyexperience.JComputAssistTomogr2001; 25:43-49. 24.SchlemmerHP,MerkleJ,GrobholzR,etal.Canpre-operative contrast-enhanced dynamic MR imaging for prostate cancer predict microvessel density in prostatectomy specimens? Eur Radiol2004;14:309-317.
25.HlatkyL,HahnfeldtP,FolkmanJ.Clinicalapplicationofan-tiangiogenic therapy: microvessel density, what it does and doesn’ttellus.JNatlCancerInst2002;94:883-893. 26.GohV,HalliganS,DaleyF,etal.Colorectaltumorvasculari-ty:quantitativeassessmentwithmultidetectorCT–dotumor perfusionmeasurements.Radiology2008;249:510-517. 27.IzawaJI,DinneyCP.Theroleofangiogenesisinprostateand otherurologiccancers:areview.CMAJ2001;164:662-670.
28.StrohmeyerD,RössingC,StraussF,etal.Tumorangiogene-sisisassociatedwithprogressionafterradicalprostatectomyin pT2/pT3prostatecancer.Prostate2000;42:26-33.
29.Ferrer FA, Miller LJ, Andrawis RI, et al. Angiogenesis and prostatecancer:invivoandinvitroexpressionofangiogenesis factorsbyprostatecancercells.Urology1998;51:161-167. 30.Łuczyńska E, Gasińska A, Wilk W. Microvessel density and
expression of vascular endothelial growth factor in clinically localizedprostatecancer.PolJPathol2013;64:33-38.
Address for correspondence
Elżbieta Łuczyńska MD,PhD RadiologyDepartment CenterofOncology,M.Sklodowska-CurieMemorialInstitute CracowBranch Garncarska11 31-115Krakow,Poland tel.+48124229900 fax+48124212325 e-mail:[email protected]
