Prostate
Cancer
Third-line
Life-prolonging
Drug
Treatment
in
a
Real-world
Metastatic
Castration-resistant
Prostate
Cancer
Population:
Results
from
the
Dutch
Castration-resistant
Prostate
Cancer
Registry
Jessica
C.L.
Notohardjo
a,y,*,
Malou
C.P.
Kuppen
b,y,
Hans
M.
Westgeest
c,
Reindert
J.A.
van
Moorselaar
d,
Niven
Mehra
e,
Jules
L.L.M.
Coenen
f,
Inge
M.
van
Oort
g,
Aad
I.
de
Vos
h,
Walter
L.
Vervenne
i,
Alphons
C.M.
van
den
Bergh
j,
Katja
K.H.
Aben
k,l,
Diederik
M.
Somford
m,
Andries
M.
Bergman
n,
Carin
A.
Uyl-de
Groot
b,
Winald
R.
Gerritsen
e,
Alfons
J.M.
van
den
Eertwegh
aaDepartment of Medical Oncology, Amsterdam UMC, Vrije Universiteit Amsterdam, Cancer Center Amsterdam, Amsterdam, The Netherlands;bInstitute for
Medical Technology Assessment, Erasmus School of Health Policy and Management, Rotterdam, The Netherlands;cDepartment of Internal Medicine, Amphia
Hospital, Breda, The Netherlands;dDepartment of Urology, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands;eDepartment of
Medical Oncology, Radboud University Medical Center, Nijmegen, The Netherlands;fDepartment of Medical Oncology, Isala, Zwolle, The Netherlands;
gDepartment of Urology, Radboud University Medical Center, Nijmegen, The Netherlands;hDepartment of Medical Oncology, Van-Weel-Bethesda Ziekenhuis,
Dirksland, The Netherlands; iDepartment of Medical Oncology, Deventer Ziekenhuis, Deventer, The Netherlands; jDepartment of Radiation Oncology,
University Medical Center Groningen, University of Groningen, Groningen, The Netherlands;kDepartment for Health Evidence, Radboud Institute for Health
Sciences, Radboud University Medical Center, Nijmegen, The Netherlands; lDepartment of Research, Netherlands Comprehensive Cancer Organisation,
Utrecht, The Netherlands;mDepartment of Urology, Canisius Wilhelmina Hospital, Nijmegen, The Netherlands;nDivision of Internal Medicine (MOD) and
Oncogenomics, The Netherlands Cancer Institute Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands
a v ai l a b l e a t w w w . s c i e n c e d i r e c t . c o m
j o u r n al h o m e p a g e : w w w . e u r o p e an u r o l o g y . c o m / e u f o c u s
Articleinfo Articlehistory:
AcceptedMarch24,2020 AssociateEditor:DeryaTilki Keywords: Castration-resistantprostate cancer Life-prolongingdrug Real-worldoutcomes Thirdline Abstract
Background: Evidenceconcerningthird-linelife-prolongingdrugs(LPDs)inthe treat-mentofmetastaticcastration-resistantprostatecancer(mCRPC)patientsisincomplete.
Objective: To evaluate third-line LPD outcomes in a real-world cohort of mCRPC
patients,identifyvariablesassociatedwithoverallsurvival(OS),andestablisha prog-nosticmodel.
Design, setting, and participants: Patients with mCRPC who were progressive on
second-lineLPD before July1, 2017 wereretrospectivelyidentified fromtheDutch Castration-resistant Prostate Cancer Registry (CAPRI) and followed until December 31,2017.
Outcomemeasurementsandstatisticalanalysis: Associationofpotential riskfactors withOSwastestedbyCoxproportionalhazardmodelsaftermultipleimputationof missingbaselinecharacteristics.Apredictivescorewascomputedfromtheregression coefficientandusedtoclassifypatientsintoriskgroups.
Resultsandlimitations: Of1011mCRPCpatientsprogressiveonsecond-lineLPD,602 (60%)receivedthird-lineLPD.Patientsreceivingthird-lineLPDhadamorefavorable
yThese authors shared the firstauthorship.
* Corresponding author. Department of Medical Oncology, Amsterdam UMC, Vrije Universiteit
Amsterdam,CancerCenterAmsterdam,Amsterdam,TheNetherlands.Tel.+31204444881.
E-mail address:j.notohardjo@amsterdamumc.nl(Jessica C.L. Notohardjo).
https://doi.org/10.1016/j.euf.2020.03.009
1. Introduction
Prostatecanceristhemostcommoncanceramongmenin theWesternworld[1].Partofthesepatientswilleventually progressanddevelopmetastaticcastration-resistant pros-tatecancer(mCRPC)[2].In2004, docetaxel,amemberof thetaxanedrugclass,wasthefirsttreatment toimprove overallsurvival(OS)ofmCRPCpatients[3].Inthepastyears, severalnewtherapeuticagents,includingcabazitaxel, abir-ateroneacetate,enzalutamide,andradium-233,havealso been registered for the treatment of mCRPC based on a survival benefit. The outcomes of these life-prolonging drugs(LPDs) asfirst-and/or second-line(post-docetaxel) treatmenthavebeenwellestablished[4–9].
Itiscommon practicetousethesedrugsasathird-lineLPD treatment,afterfirst-andsecond-lineLPDtreatment,inthe hopetoobtainacumulativebenefit[10].Todate,randomized controlled trialsof third-lineLPDs inmCRPCpatients are scarce[11].Thereportson third-lineLPDsareparticularly retrospectiveand basedonsmall cohortsofpatientsreceiving onespecificthird-lineLPD[12–16].PatientswithmCRPCwho areonthird-lineLPDmayhaveworseoutcomes,compared withthoseonfirst-andsecond-lineLPDtreatment,dueto more advanced stages in general, decreased performance status, worse tolerance to treatments [17], and possible cross-resistance[18].
Thus, third-line LPDs might not be appropriate for all patients.SelectionofpatientswithmCRPCwhowillbenefit from third-line LPD treatment iscrucial to improve out-comes,reduceunnecessarytoxicity,improvequalityoflife (QoL),andreducecosts[19].Predictionoftreatment out-comemayallowforbetterpatientselection.Nevertheless, current prognosticmodels for survival using clinical and laboratorybaselinevariablesinmCRPCpatientshavebeen describedonlyinfirst-orsecond-lineLPDs[20–23].
The aim of this retrospective study was to evaluate outcomesofthird-lineLPDtreatmentinareal-worldcohort ofmCRPCpatients,toidentifyclinicalandlaboratory vari-ables associated with survival, and to finally assess the impactofthesevariablesinariskscore.
2. Patientsandmethods 2.1. Studydesignandsetting
Castration-resistantProstateCancerRegistry(CAPRI)isan investigator-initiated, observational, multicenter cohort studyin20hospitalsinTheNetherlands.Thestudydesign has beenpreviouslydescribed[24]. PatientswithmCRPC were includedretrospectively fromJanuary 1,2010 until December31,2015.MetastaticCRPCwasdefinedeitherby thecriteriasetbytheEuropeanAssociationofUrology[25]
orbythetreatingphysician.Thestudyisregisteredinthe DutchTrialRegistryasNL3440(NTR3591).
2.2. Objectives
Theaimofthisstudyistoinvestigatetheoutcomesof third-line LPDtreatment in a real-word population of mCRPC patients,identifyclinicalandlaboratoryvariablesrelatedto survivaloutcomes,andassesstheimpactofthesevariables inariskscore.
2.3. Participants
Metastatic CRPCpatients with progressive disease on or afterasecond-lineLPD,beforeJuly1,2017,wereincludedin the analysis. All patients had received two lines of LPD treatment,ofwhichatleastoneofthetwopreviouslines was docetaxel. They were categorized into two groups: patients receivingathird-line LPDandpatientsreceiving bestsupportivecare(BSC).
Patientspreviouslytreatedwithdocetaxelfor hormone-sensitivemetastaticprostatecancer(n=14)wereexcluded fromtheanalysis.
2.4. Follow-upanddatacollection
Predefinedandreadilyavailabledatafrommedicalrecords were retrospectivelycollected by traineddata managers. Baselinecharacteristicswereincludedintheanalysisifthey prognosticprofileatbaselineandlongermedianOSthanpatientswithbestsupportive care(10.4vs2.4mo,p<0.001).EasternCooperative OncologyGroupperformance status1and2(hazardratio[HR]1.51,p<0.007andHR3.08,p<0.001,respectively), opioiduse(HR1.55,p=0.019),visceralmetastases(HR2.09,p<0.001),hemoglobin <7 mmol/l (HR 1.44, p < 0.002), prostate-specific antigen 130
m
g/l (HR 1.48, p=0.001),alkalinephosphatase170U/l(HR1.52,p<0.001),andlactate dehydroge-nase250U/l(HR1.44;p=0.015) wereassociatedwithshortersurvival.Harrell’s C-indexwas0.74.ThemedianOSvaluesforlow-,low-intermediate-, high-intermedi-ate-,andhigh-riskgroupswere14,7.7,4.7,and1.8mo,respectively.Limitationsinclude theretrospectivedesign.Conclusions: Wedevelopedaprognosticmodelandidentifiedasubgroupofpatientsin whomthird-lineLPD treatmenthasnomeaningful benefit.Our resultsneedtobe confirmedbyprospectiveclinicaltrials.
Patient summary: We reportedoutcomes from third-line life-prolonging drugs in metastaticprostatecancerpatientsanddevelopedaprognosticmodelthatcouldbe usedtoguidetreatmentdecisions.
©2020EuropeanAssociationofUrology.PublishedbyElsevierB.V.Allrights reserved.
weredocumented3wkpriorto3wkaftertheprogression dateafter a second-line LPD. All patients were followed untildeath,losstofollow-up,orDecember31,2017. Follow-upduration wascalculated asthe time fromthedate of progressiononasecond-lineLPDtothelastrecordeddate. 2.5. Outcomes
OutcomeswereOS,treatmentduration(TD),and prostate-specific antigen (PSA) response. OS was calculated in months from the date of progression after second-line LPDtreatmenttothedateofdeathfromanycause.Patients alive at the end of the study or lost to follow-up were censoredatthelastrecordeddate.
TDwasdefinedastheintervalbetweenthestartandstop ofthird-lineLPDtreatment.Ifthestopdatewasunknown, TDwasspecifiedasthetimefromthestartofthird-lineLPD tothestartofnexttreatment,orasthetimefromthestartof third-lineLPDtotheendoffollow-upifthird-linetreatment wasthelasttreatment.Patientsontreatmentattheendof follow-upwerecensoredatthelastrecordeddate.
PSAresponsewasdefinedasthemaximumchangefrom baselinePSAlevels(inpercentages)withoutconfirmation of a second measure. In case no decline was present, responsesweremeasuredat12wk(accordingtoProstate CancerClinicalTrialsWorkingGroup3criteriaforresponse measurement[26])oriftreatmentwas<12wk,attheend oftreatmentorstartofnexttreatment.PSAresponsewas definedasa50%PSAdeclinefrombaseline.
2.6. Statisticalanalysis
Descriptivestatisticswereperformed.Thettest(orMann– WhitneyUtestfornonparametricvariables)wasusedfor continuousvariablesandthePearsonchi-squarewasused forcategoricalvariables. OSandTDwereestimatedusing the Kaplan–Meier method and were compared between groupsusingthelog-ranktest.Awaterfallplotwasmade toindicatePSA response. Missingbaseline characteristics wereimputedusingmultipleimputationwithMonteCarlo MarkovChainmethod.Selectionofprognosticfactorswas basedonclinicalapplicability(routinelycollectedandused by clinicians), previous research, and expert opinion
[27]. Continuous variables were categorized using the mediancutofforclinicallyapplicablecutoffs.Multivariable Coxproportionalhazardanalysisusingabackwardstepwise procedurewasperformedonpooleddataforOS.A simpli-fiedpredictionrulewasobtainedbyroundingthe regres-sioncoefficientstohalf points,whichweremultipliedby two for easier clinical applicability. A risk score for the prediction of OS was then calculated for each patient. Patients could be categorized into different risk groups basedonthesurvivalcurvesofeachriskscore.The prog-nosticperformanceofthepredictionmodelwasevaluated usingHarrell’sconcordanceindex(C-index)intheoriginal dataset. A p value of 0.05 was considered statistically significant.Statisticalanalyseswere performedusingIBM SPSSStatistics,version22.0(SPSSInc.,Chicago,IL,USA).
3. Results
Attheendofthestudy,3616CRPCpatientswereincludedin 20 hospitals. A total of 1011 mCRPC patients (28%) had progressionon orafterasecondLPDtreatmentandwere included in the analysis. At database cutoff, 826 deaths (82%) had occurred, 127 patients (13%) were lost to follow-up,and58patients(6%)werestillalive.
Allpatientswerepreviouslytreatedwithdocetaxeland with abiraterone acetate (n = 525, 52%), enzalutamide (n=282,28%),cabazitaxel(n=155,15%),docetaxel rechal-lenge(n=31,3.0%)orradium-223(n=18,2.0%).
Ofthese1011mCRPCpatients,602(60%)receiveda third-lineLPD.Thethird-lineLPDconsistedofcabazitaxel(n=213, 35%), abiraterone acetate (n = 137, 23%), enzalutamide (n = 129, 21%), radium-223 (n = 78, 13%), and docetaxel (n=45,8.0%).Anoverviewofprevioustreatmentlinesand third-linetreatmentisprovidedinSupplementaryTable 1. 3.1. Baselinecharacteristics
BaselinecharacteristicsofmCRPCpatientsatthe progres-siondateofasecond-lineLPD,accordingtothesubsequent third-lineLPDornot,areshowninTable1.Patients receiv-ingathird-lineLPDhadamorefavorableprognosticprofile (significantlyyounger,betterEasternCooperativeOncology Groupperformancestatus[ECOGPS],lessopioiduse,less visceralmetastases,higherhemoglobin[Hb],loweralkaline phosphatase[ALP],andlowerlactatedehydrogenase[LDH]) comparedwithpatientswhoreceivedBSC.
3.2. OSandrisk-scoringsystem
The medianOS (mOS)fromprogressionon asecond-line LPDwas6.5mo(95%confidenceinterval[CI]5.9–7.2).The mOS was longer for patients receiving a third-line LPD (10.4 mo, 95% CI 9.2–11.6) compared with patients who receivedBSC(2.4mo,95%CI2.1–2.7;Fig.1).
Univariableanalysisrevealedbaseline ECOGPS,opioid use,symptoms,visceralmetastases,lymphnode metasta-ses,Hb,PSA,ALP,LDH,andperiodfromcastrationtoCRPCas beingsignificantvariablesforthepredictionofsurvivalin mCRPCpatientsprogressingonasecond-lineLPD(Table2). ThemultivariableCoxregressionanalysisofpooleddata identifiedsevenvariablesindependentlyassociatedwithOS: ECOGPSof1and2(HR1.51,95%CI1.13–2.00,p=0.007and HR3.08,95%CI2.31–4.10,p<0.001,respectively),opioiduse (HR 1.55, 95% CI 1.10–2.19, p = 0.019), visceral metastases (HR 2.09, 95% CI 1.76–2.49, p < 0.001), Hb <7.0 mmol/l (HR 1.44, 95% CI 1.15–1.84, p = 0.002), PSA 130
m
g/l (HR1.48,95%CI1.20–1.82,p=0.001),ALP170U/l(HR1.52, 95%CI1.26–1.84,p<0.001),andLDH>250U/l(HR1.44,95% CI1.09–1.90,p=0.015);thesewererelatedtoworsesurvivaland includedinthefinalmodel.TheHarrell’sC-indexwas0.74.Based on their regression coefficients, we assigned a score of 1 point to ECOG PS of 1, opioid use, Hb < 7.0 mmol/l, PSA 130
m
g/l, ALP 170 U/l, and LDH >250U/l.Ascoreof2pointswasassignedtoECOGPS2andpresenceofvisceralmetastases(SupplementaryTable 2A). Taking into account the survival curves of the calculated riskscores,patientscouldbecategorizedintodifferentrisk groups:low-risk(score0),low-intermediaterisk(score1–3), high-intermediaterisk(score4–6),andhigh-risk(score7–9; Supplementary Table 2B). The low-risk group included
103patients(10%),thelow-intermediate-riskgroupincluded 467 patients (46%), the high-intermediate-risk group included341patients(34%),andthehigh-riskgroupincluded 56patients(6%).Mediansurvivaltimesfortheselow-, low-intermediate-, high-intermediate-, and high-risk groups were 14.0mo(95%CI10.7–17.3),7.7mo(95%CI6.6–8.9),
Table 1 –Baseline characteristics at the time of progression after a second-line LPD in mCRPC patients according to receiving a third-line LPD
orbestsupportivecare.
Characteristics Total group of patients
progressive after a
second-line LPD
Missing Best supportive
care
Missing Third-line LPD Missing p value
(n = 1011) (n = 409) (n = 602) Age(yr) 71.67.5 21(2.1) 73.07.8 0 71.07.3 21(3.5) 0.032 ECOGPS 508(50) 229(56) 279(46) <0.001 0 93(9) 15(4) 78(13) 1 280(28) 67(16) 213(35) 2 130(13) 98(24) 32(5) Opioiduse 219(22) 605(60) 127(31) 225(55) 92(12) 380(63) <0.001 Symptomatic 704(70) 81(8) 346(85) 13(3) 358(60) 68(11) <0.001 Metastaticsite Bone 871(86) 96(10) 355(87) 41(10) 516(86) 55(9) 0.139 Visceral 169(17) 493(49) 91(22) 202(49) 78(13) 291(48) <0.001 Lymphnodes 469(46) 382(38) 195(48) 163(40) 274(46) 219(36) 0.030 Hemoglobin(mmol/l) 7.11.2 303(30) 6.81.2 111(27) 7.41.1 192(32) <0.001 Platelets(109 /l) 250(193–315) 314(31) 238(167–322) 117(29) 256(205–313) 197(33) 0.032 Prostate-specificantigen (mg/l) 133(42–413) 126(13) 174(42–491) 64(16) 118(42–358) 62(10) 0.058 Alkalinephosphatase(U/l) 170(99–353) 182(18) 260(128–506) 72(18) 139(88–253) 110(18) <0.001 Lactatedehydrogenase(U/l) 289(213–420) 411(41) 389(241–730) 154(38) 251(203–360) 257(43) <0.001 ECOGPS=EasternCooperativeOncologyGroupperformancestatus;LPD=life-prolongingdrug;mCRPC=metastaticcastration-resistantprostatecancer; SD=standarddeviation.
DataarepresentedasmeanSD,median(interquartilerange),orn(%).
Fig.1–Kaplan-Meiercurvesandat-risktablesforoverallsurvival(OS)fromprogressionafterasecond-lineLPD(A)forthetotalgroup(n=1011),(B) classifiedbythird-lineLPD(n=602)orbestsupportivecare(n=409).Thepvalueswereobtainedfromlog-ranktestsforthehomogeneityof Kaplan-Meiercurvesbetweenthird-lineLPDandbestsupportivecare.
Table 2 –Univariable and multivariable analyses of different prognostic variables for overall survival in patients with mCRPC progression
afterasecond-lineLPD.
Variable n Event Censored Missing Univariable analysisa Multivariable analysisa,b
Hazard ratio
(95%CI)
p value Hazard ratio
(95%CI) p value bc Points ECOGPS 503 420 83 508 <0.001 0 93 1 1 0 1 280 1.74(1.33–2.29) 1.51(1.13–2.00) 0.007 0.409 1 2 130 4.55(3.35–6.18) 3.08(2.31–4.10) <0.001 1.123 2 Opioiduse 406 350 56 605 <0.001 0.019 No 187 1 1 0 Yes 219 2.18(1.75–2.73) 1.55(1.10–2.19) 0.438 1 Symptomatic 925 754 171 86 <0.001 No 224 Yes 701 2.07(1.73–2.47) Visceralmetastases 511 409 102 500 <0.001 <0.001 No 344 1 1 0 Yes 167 2.13(1.73–2.62) 2.09(1.76–2.49) 0.738 2 Lymphnodemetastases 622 508 114 389 0.002
No 158 1 Yes 464 1.38(1.12–1.69) Hemoglobin(mmol/l) 708 594 114 303 <0.001 0.002 <7 307 2.22(1.88–2.62) 1.44(1.15–1.84) 0.372 1 7 401 1 1 0 Platelets(109/l) 697 584 113 314 0.535 <250 344 1 250 353 1.05(0.89–1.24) Prostate-specificantigen(mg/l) 885 723 162 126 <0.001 0.001 <130 441 1 1 0 130 444 1.73(1.49–2.00) 1.48(1.20–1.82) 0.393 1 Alkalinephosphatase(U/l) 833 682 151 178 <0.001 <0.001
<170 415 1 1 0
170 418 2.23(1.91–2.60) 1.52(1.26–1.84) 0.421 1 Lactatedehydrogenase(U/l) 600 505 95 411 <0.001 0.015
<ULN 236 1 1 0
ULN 364 2.24(1.86–2.69) 1.44(1.09–1.90) 0.365 1 TimefromcastrationtoCRPC
(mo)
988 806 182 23 0.012 <12 475 1.19(1.04–1.37)
12 513 1
B=betaregression coefficient;CI= confidence interval;CRPC= castration-resistantprostatecancer;ECOGPS=Eastern CooperativeOncologyGroup performancestatus;LPD=life-prolongingdrug;mCRPC=metastaticcastration-resistantprostatecancer;n=numberofpatients;ULN=upperlimitofnormal. a
Coxregressionmodel. b
Thefinalsampleusedinthemultivariateanalysisconsistedof1011patients;826patientsdiedand185werecensored. cThecoefficientofeachvariablewasroundedtohalfpointandthenmultipliedbyaconstant(2)foreasierclinicallyapplicability.
Fig.2–Kaplan-Meiercurvesandrisktablesforoverallsurvival(OS)fromprogressionafterasecond-lineLPDaccordingtothefourriskgroupsfor(A) thetotalgroup(n=1011),(B)patientsreceivingathird-lineLPD(n=602),and(C)patientsreceivingbestsupportivecare(n=409).
4.7mo(95%CI4.0–5.4),and1.8mo(95%CI1.4–2.2), respec-tively(p<0.001;Fig.2A).
Athird-lineLPDwasstartedin69%patients(71outof 103)inthelow-riskgroup,64%patients(299outof467)in thelow-intermediate-riskgroup,53%patients(181outof 341)inthehigh-intermediate-riskgroup,and30%patients (17outof56)inthehigh-riskgroup.ThemOSfortheserisk groups,accordingtowhetherornottreatedwitha third-lineLPD,aredepictedinFig.2.
A nomogram, integrating the significant independent variablesforOS,isprovidedinSupplementaryFigure1. 3.3. TDandPSAresponseofthird-lineLPDtreatment Attheendoffollow-up,26patients(4.3%)withathird-line LPDwere stillon treatment. ThemedianTD (mTD)for a third-lineLPDwas3.3mo(95%CI3.0–3.5).PSAdeclineon thethird-lineLPDwasassessablein560(93%)patientsand observedin130(22%)patients.
ThemTDforthefourriskgroups(low-,low-intermediate-, high-intermediate-,andhigh-riskgroups)were4.6mo(95% CI3.8–5.4),3.4mo(95%CI3.2–3.6),2.7mo(95%CI2.4–3.0), and1.4mo(95%CI1.1–1.7),respectively(p<0.001;Fig.3). PSAresponserates(>50% PSAresponse)were24%(18/76 patients),22%(66/301patients),23%(41/181patients),and 6%(one/17patients),respectively.WaterfallplotofthePSA responsesareshowninFig.4.
4. Discussion
Toourknowledge,thisisthefirstlargemulticenter real-worldcohort,evaluatingtheoutcomesofmCRPCpatients
progressingonasecond-lineLPD,treatedaccordingtothe viewsandopinionsoftheirtreatingphysicians.
We observed the mOSof 6.5 mofrom progression of second-line LPD.The mOS waslonger inpatients with a third-line LPD than in patients receiving BSC (10.4 vs 2.4 mo), but TD was short (3.3 mo) and PSA response waslow(22%).Ourresultsconfirmthepotentialcumulative survival benefit (mOS 7.1–15.8) of previousretrospective studiesonthird-lineLPDtreatment[13–15].
Pivotal phase 3 trials on first- and second-line LPD treatment in mCRPC patients reportedthe mOS of14.0– 34.7mo.ThedifferenceinOScanpartiallybeexplainedby the factthatpatients treatedintrialsnotablydiffer from patientswhoreceivestandardtreatmentoptionsonly[24]
andthemoreadvanceddiseasestateofpatientsaftertwo systemictreatmentlines.Thisisreflectedbypoor perfor-mancescore,highdiseaseburden,andhighALP,LDH,and PSA.AsmCRPCprogresses,diseasecontrolbecomesmore difficult[28].Possiblecross-resistancewithprevious treat-mentscanfurtherdecreasetreatmenteffect[18].Moreover, tolerabilitytonewsystemictreatmentscanbeworse[17], leadingtoearlydiscontinuation.
Evidence concerning optimal sequencing of third-line LPDsislimited,butsuggeststhatpatientsmaynotrespond toandrogenreceptor–targetedtherapies(ARTs;abiraterone or enzalutamide) in third line afterprogression on prior ARTs due to cross-resistance [10,17,29]. This is recently prospectivelyconfirmed bya studyof de Witet al. [11], whichreportedincreasedmOSinpatientsreceiving caba-zitaxel compared with those receiving an ART (13.6 vs 11.0mo)afterpriordocetaxelandtheotherART.Sinceall patients had progression on an alternative ART within 12 mo, they were not comparable with our study
Fig.3–Kaplan-Meiercurvesandat-risktablesforthetreatmentdurationofathird-lineLPD(A)forallpatientsreceivingathird-lineLPD(n=602) and(B)accordingtothefourriskgroups.
population. Our analysis identified seven independent prognosticvariablesassociatedwithsurvival,namelyECOG PS,opioiduse,visceralmetastases,Hb,PSA,ALP,andLDH. Thesevariableswereabletodistinctfourriskgroups(low-, low-intermediate-, high-intermediate-, andhigh-risk) for patients whohad progressivedisease aftera second-line LPD,withcorrespondingmediansurvivaltimesof14.0,7.7, 4.7,and1.8mo,respectively(p<0.001).
Especially,high-riskpatientshadremarkablyshortmOS. Moreover,high-riskpatientstreatedwithathird-lineLPD had worse mOS than patients receiving BSC in low- or low-intermediate-risk groups. These results suggest that high-riskpatientsmayderivenomeaningful benefitfrom third-lineLPDsinclinical practice,whichissupportedby theshortmTDandlowPSAresponses.Therefore,high-risk patientsshouldnotbetreatedwiththird-lineLPDs;instead, theyshouldbetreatedwithBSC.
Ourprognosticmodelallowsforthestratificationoffour riskgroupswithwidelydifferingmOS.Itisimportantfor physicians to consider these different survival times in medical decision making. Proper patient selection for third-line LPDtreatment iscrucial to improveoutcomes, reduceunnecessarytoxicity,andimproveQoL.Also,careful
consideration is warranted considering possiblelowcost effectiveness.
Thisstudyisnotwithoutlimitations.First,ourresultsare limitedbytheabsenceofpreviouslyidentifiedriskfactors such as albumin level [27]. However, albumin is not a routinelyassessedparameterinreal-worldclinicalpractice. Moreover,manypatientshadmissingvaluesofoneormore baselinevariablesatprogressiononsecond-lineLPDdueto theretrospectivenatureofthestudy.Imputationofmissing baselinedataoffersavalidsolutionformultivariable anal-ysis[30].Second,theeffectofthird-lineLPDinother out-comes such as QoL and cost effectiveness could not be included inthis analysis.Lastly, theidentified prognostic modelhasnotyetbeenexternallyvalidatedandistherefore notyetsuitableforclinicaluse.
Nevertheless,ourprognosticmodelwasdevelopedusing alargenumberofpatientswithmCRPCwhowere progres-siveaftersecond-lineLPD,andthenumberofdeathsinthe pooledanalysiswassubstantial,providinggoodstatistical power. Furthermore, this prognostic model is based on readilyavailableclinicalandlaboratoryvariables,andrisk groups can becalculated easily. Although our prognostic modelisbasedonretrospectivedata,itwasabletoidentify
Fig.4–WaterfallplotsofmaximumPSAchangesfrombaselineforpatientstreatedwithathird-lineLPD(n=602). LPD=life-prolongingdrug;PSA=prostate-specificantigen.
four riskgroupswith differing survivaltimes, suggesting thattheidentifiedvariablesmayassistintheselection of patientsforthird-lineLPDtreatmentindailyclinical prac-tice and thereby improving efficacy of these potentially toxicandexpensiveLPD.
5. Conclusions
Third-line LPDs might not be appropriate for all mCRPC patients,whichissupportedbytheshortmTDandlowPSA responsesobserved inour study. Wedevelopeda simple prognostic model, based on routinely used clinical and laboratoryparameters,andidentifiedahigh-risksubgroup in whom no meaningful benefit from third-line LPD is derivedinclinicalpractice.Ourresultsneedtobeconfirmed byfurtherprospectivetrials.
Author contributions: Jessica C.L. Notohardjo had full access to all the
data in the study and takes responsibility for the integrity of the data and
theaccuracyofthedataanalysis.
Study concept and design: Notohardjo, Kuppen, Westgeest, van den
Eertwegh, Uyl-de Groot, Gerritsen.
Acquisitionofdata:Kuppen,Westgeest.
Analysisandinterpretationofdata:Notohardjo,Kuppen,Westgeest,van
den Eertwegh.
Drafting of the manuscript: Notohardjo, Kuppen.
Critical revision of the manuscript for important intellectual content:
Westgeest, van Moorselaar, N. Mehra, Coenen, van Oort, de Vos,
Verv-enne,vandenBergh,Aben,Somford,Bergman,Uyl-deGroot,Gerritsen,
vandenEertwegh.
Statistical analysis: Notohardjo, Kuppen.
Obtaining funding: Westgeest, Uyl-de Groot, Gerritsen.
Administrative, technical, or material support: None.
Supervision: van den Eertwegh, Uyl-de Groot, Gerritsen.
Other:None.
Financial disclosures: Jessica C.L. Notohardjo certifiesthat all conflictsof
interest, including specific financial interests and relationships and
affiliationsrelevanttothesubjectmatterormaterialsdiscussedinthe
manuscript(eg,employment/affiliation,grantsorfunding,
consultan-cies, honoraria, stock ownership or options, expert testimony, royalties,
or patents filed,received, or pending), are the following: J.C.L.
Noto-hardjo: none. M.C.P. Kuppen: travel expenses—Ipsen.H.M. Westgeest:
travel expenses—Ipsen;honoraria—Roche.R.J.A. van Moorselaar:
hono-rariaorconsultationfees—Amgen,Astellas,AstraZeneca,Bayer,Janssen,
and Sanofi-Genzyme. N.Mehra: research funding—Astellas, Janssen,
Pfizer,Roche, and SanofiGenzyme; advisory role—Roche,MSD, BMS,
Bayer, Astellas, and Janssen; travel support—Astellasand MSD. J.L.L.M.
Coenen: advisory board—Sanofi. I.M. van Oort: consulting/advisory
role—Astellas,Janssen, Bayer, Roche, and Mdx Health; research
fund-ing—Astellas,Janssen,andBayer.A.I.deVos:none.W.L.Vervenne:none.
A.C.M.vandenBergh:none.K.K.H.Aben:none.D.M.Somford:research
funding—Astellas;honoraria—Sanofiand Merck Sharp&Dohme; research
funding—Sanofi and Bayer. A.M. Bergman: research funding—Sanofi,
Astellas, and Bayer; consulting/advisory role—Sanofi,Astellas, and Bayer;
travel/accommodations—Sanofi,Astellas, and Bayer. Speakers’bureau—
Sanofi,Astellas, Bayer, and Janssen. C.A. Uyl-de Groot: research funding—
Boehringer Ingelheim, Astellas, Celgene,Sanofi, Janssen-Cilag, Bayer,
Amgen, Genzyme, Merck, Glycostem Therapeutics, Astra Zeneca, Roche,
and Merck. W.R. Gerritsen: speakers’fees—Bayerand MSD; advisory
boards—Bristol-Myers Squibb, Astellas, Bayer, Sanofi, and Amgen;
researchgrant—Bayer,Astellas,andJanssen-Cilag.A.J.M.vanden
Eert-wegh: study grant—Sanofiand Roche; travel expenses—MSDOncology,
Roche, Pfizer, and Sanofi; honoraria—Bristol-MyersSquibb; advisory
board—Bristol-Myers Squibb, MSD Oncology, Amgen, Roche, Sanofi,
Pfizer,Ipsen, and Merck.
Acknowledgments: This research was funded by Sano
fi-AventisNether-lands B.V., Janssen-Cilag B.V., Astellas Pharma B.V., and Bayer B.V. The
fundingorganisationshadnoroleinthedesignandconductofthestudy;
collection, management,analysis,andinterpretation ofthe data;and
preparation,review,orapprovalofthemanuscript.
AppendixA. Supplementarydata
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.
1016/j.euf.2020.03.009.
References
[1]ZhouCK,CheckDP,Lorlet-TieulentJ,etal.Prostatecancerincidence in43populationsworldwide:ananalysisoftimetrendsoveralland byagegroup.IntJCancer2016;138:1388–400.
[2]KarantanosT,CornPG,ThompsonTC.Prostatecancerprogression afterandrogendeprivationtherapy:mechanismsofcastrate resis-tanceandnoveltherapeuticapproaches.Oncogene 2013;32:5501–11.
[3]TannockIF,deWitR,BerryW,etal.Docetaxelplusprednisoneor mitoxantroneplusprednisoneforadvancedprostatecancer.NEngl JMed2004;351:1502–12.
[4]FizaziK,ScherHI,MolinaA,etal.Abirateroneacetatefortreatment of metastatic castration-resistant prostate cancer: final overall survival analysisof the COU-AA-301 randomised, double-blind, placebo-controlledphase3study.LancetOncol2012;13:983–92.
[5]ScherHI,FizaziK,SaadF,etal.Increasedsurvivalwithenzalutamide in prostate cancer after chemotherapy. N Engl J Med 2012;367:1187–97.
[6]deBonoJS,OudardS,OzgurogluM,etal.Prednisoneplus cabazi-taxelormitoxantroneformetastaticcastration-resistantprostate cancerprogressingafterdocetaxeltreatment:arandomised open-labeltrial.Lancet2010;376:1147–54.
[7]ParkerC,NilssonS,HeinrichD,etal.Alphaemitterradium-223and survivalinmetastaticprostatecancer.NEnglJMed2013;369:213–23.
[8]RyanCJ, SmithMR, deBonoJS,etal.Abiraterone inmetastatic prostate cancerwithout previous chemotherapy. N Engl J Med 2013;368:138–48.
[9]Beer TM,ArmstrongDE,RathkopfY.Enzalutamideinmetastatic prostate cancer before chemotherapy. N Engl J Med 2014;371:1755–6.
[10] MainesF,CaffoO,Veccia A,etal.Sequencing newagentsafter docetaxelinpatientswithmetastaticcastration-resistantprostate cancer.CritRevOncolHematol2015;96:498–506.
[11] de Wit R, de Bono CN, Sternberg K, et al. Cabazitaxel versus abirateroneorenzalutamideinmetastaticprostatecancer.NEngl JMed2019;381:2506–18.
[12] BadrisingS,vanderNoortV,vanOortIM,etal.Clinicalactivityand tolerabilityofenzalutamide(MDV3100)inpatientswithmetastatic,
castration-resistantprostatecancerwhoprogressafterdocetaxel andabirateronetreatment.Cancer2014;120:968–75.
[13]BrassoK,ThomsenFB,SchraderAJ,etal.Enzalutamideantitumour activityagainstmetastaticcastration-resistantprostatecancer pre-viously treated with docetaxel and abiraterone: a multicentre analysis.EurUrol2015;68:317–24.
[14]PezaroCJ, OmlinAG,AltavillaA,etal.Activityofcabazitaxelin castration-resistantprostatecancerprogressingafterdocetaxeland next-generationendocrineagents.EurUrol2014;66:459–65.
[15]SchraderAJ,BoegemannM,OhlmannCH,etal.Enzalutamidein castration-resistantprostatecancerpatientsprogressingafter doc-etaxelandabiraterone.EurUrol2014;65:30–6.
[16]CaffoO,deGiorgiU,FratinoL,etal.Clinicaloutcomesof castra-tion-resistantprostatecancertreatmentsadministeredasthirdor fourthlinefollowingfailureofdocetaxelandothersecond-line treatment: results of an Italian multicentre study. Eur Urol 2015;68:147–53.
[17] LoriotY,BianchiniD,IleanaE,etal.Antitumouractivityof abir-aterone acetate against metastatic castration-resistant prostate cancerprogressingafterdocetaxelandenzalutamide(MDV3100). AnnOncol2013;24:1807–12.
[18]vanSoestRJ,VanRoyenME,deMorreeES,etal.Cross-resistance betweentaxanesandnewhormonalagentsabirateroneand enza-lutamidemayaffectdrugsequencechoicesinmetastatic castra-tion-resistantprostatecancer.EurJCancer2013;49:3821–30.
[19]ChanezB,BertucciF,GilabertM,etal.Ascoringsystemtoguidethe decisionforanewsystemictreatmentafteratleasttwolinesof palliative chemotherapy for metastatic cancers: a prospective study.SupportCareCancer2017;25:2715–22.
[20] HalabiS,SmallEJ,KantoffPW,etal.Prognosticmodelforpredicting survivalinmenwithhormone-refractorymetastaticprostate can-cer.JClinOncol2003;21:1232–7.
[21]HalabiS, LinCY,KellyWK,etal.Updated prognosticmodelfor predictingoverallsurvivalinfirst-linechemotherapyforpatients
withmetastaticcastration-resistantprostatecancer.JClinOncol 2014;32:671–7.
[22] ArmstrongAJ, TannockIF, deWitR, GeorgeDJ,EisenbergerM, HalabiS.Thedevelopmentofriskgroupsinmenwithmetastatic castration-resistantprostatecancerbasedonriskfactorsforPSA declineandsurvival.EurJCancer2010;46:517–25.
[23] Chi KN,Kheoh T,Ryan CJ, etal. A prognosticindex modelfor predictingoverallsurvivalinpatientswithmetastatic castration-resistant prostate cancertreated with abirateroneacetate after docetaxel.AnnOncol2016;27:454–60.
[24] WestgeestHM,Uyl-deGrootCA,vanMoorselaarRJA,etal. Differ-encesintrialandreal-worldpopulationsintheDutch Castration-resistantProstateCancerRegistry.EurUrolFocus2018;4:694–701.
[25] CornfordP,BellmuntJ,BollaM,etal.EAU-ESTRO-SIOGguidelines onprostatecancer.PartII:treatmentofrelapsing,metastatic,and castration-resistantprostatecancer.EurUrol2017;71:630–42.
[26] ScherHI,MorrisMJ,StadlerWM,etal.Trialdesignandobjectives forcastration-resistantprostatecancer:updatedrecommendations fromtheProstateCancerClinicalTrialsWorkingGroup3.JClin Oncol2016;34:1402–18.
[27] PinartM,KunathF,LiebV,etal.Prognosticmodelsforpredicting overallsurvivalinmetastaticcastration-resistantprostatecancer:a systematicreview.WorldJUrol2020;38:613–35.
[28] FrielingJS,BasantaD,LynchCC.Currentandemergingtherapiesfor bonemetastaticcastration-resistantprostatecancer.Cancer Con-trol2015;22:109–20.
[29] AttardG,BorreM,GurneyH,etal.Abirateronealoneorin combi-nationwithenzalutamideinmetastaticcastration-resistant pros-tatecancerwithrisingprostate-specificantigenduring enzaluta-midetreatment.JClinOncol2018;36:2639–46.
[30] JakobsenJC,GluudC,WetterslevJ,WinkelP.Whenandhowshould multipleimputationbeusedforhandlingmissingdatain random-isedclinicaltrials-apracticalguidewithflowcharts.BMCMedRes Methodol2017;17:162.
