Documents

j

Description
j
Categories
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Prognostic Value of Health-Related Quality-of-Life Data inPredicting Survival in Patients With AnaplasticOligodendrogliomas, From a Phase III EORTC BrainCancer Group Study   Murielle E.L. Mauer, Martin J.B. Taphoorn, Andrew Bottomley, Corneel Coens, Fabio Efficace, Marc Sanson, Alba A. Brandes, Carin C.D. van der Rijt, Hans J.J.A. Bernsen, Marc Fre´nay, Cees C. Tijssen, Denis Lacombe,and Martin J. van den Bent  A B S T R A C T Purpose This is one of a few studies that have explored the value of baseline symptoms and health-relatedquality of life (HRQOL) in predicting survival in patients with brain cancer. Patients and Methods Baseline HRQOL scores (from the European Organisation for Research and Treatment of Cancer[EORTC] Quality of Life Questionnaire C30 and the EORTC Brain Cancer Module) were examinedin 247 patients with anaplastic oligodendrogliomas to determine the relationship with overallsurvival by using Cox proportional hazards regression models. Refined techniques as the bootstrapresampling procedure and the computation of C indexes and R 2 coefficients were used to explorethe stability of the models as well as better assess the potential benefit of using HRQOL to predictsurvival in clinical practice and research. Results Classical analysis controlled for major clinical prognostic factors selected emotional functioning( P   .0016), communication deficit ( P   .0261), future uncertainty ( P   .0481), and weakness oflegs ( P   .0001) as statistically significant prognostic factors of survival. However, several issuesquestion the validity of these findings and no single model was found to be preferable over allothers. C indexes, which estimate the probability of a model to correctly predict which patientamong a randomly chosen pair of patients will survive longer, and R 2 coefficients, which measurethe proportion of variability explained by the model, did not exhibit major improvement whenadding selected or all HRQOL scores to clinical factors. Conclusion While classical techniques lead to positive results, more refined analyses suggest that baselineHRQOL scores add relatively little to clinical factors to predict survival. These results may haveimplications for future use of HRQOL as a prognostic factor for patients with cancer. J Clin Oncol 25:5731-5737. © 2007 by American Society of Clinical Oncology  INTRODUCTION Ithasbecomeincreasinglyacceptedthat,inadditionto the traditional assessment of clinical outcomes,health-related quality-of-life (HRQOL) informa-tion can play a key role in cancer research and helpindividualpatientcare. 1,2 Patients’selfassessmentof HRQOL is now an established end point for treat-mentcomparisonsinmanycancerdiseasesites 3 par-ticularlyinadvanceddisease. 4,5 TheuseofHRQOLoutcomescouldhavevari-ousapplications,includingsupportingclinicaldeci-sion making by providing the patient’s perspectiveorprovidingprognosticinformation.Recentstudieshave shown that HRQOL parameters can be inde-pendent prognostic factors for survival in cancerpatients. 6 If HRQOL parameters are independentpredictors of survival, they could be used in daily practice to identify patients who will benefit from aspecificintervention;further,itcouldpreventover-treatment of patients who will gain no benefitfrom often toxic and aggressive therapies or to setup more tailored psychosocial intervention pro-grams aimed at improving patients’ HRQOL.Furthermore, they could be used to better stratify patients in future cancer clinical trials, hence bet-ter interpreting study outcomes, or to identify  From the European Organisation forResearch and Treatment of Cancer,Quality of Life Unit, Data Center, andthe Data Center Brain Group, Brussels,Belgium; Medical Center Haaglanden/ Westeinde Hospital, Den Haag; Eras-mus University Medical Center,Rotterdam; Canisius Wilhelmina Zieken-huis, Nijmegen; St Elisabeth Zieken-huis, Tilburg, the Netherlands; CentreHospitalier Universitaire Pitié-Salpeˆtrière, Paris; Centre AntoineLacassagne, Nice, France; and AziendaSanitaria Locale Ospedale Bellaria-Maggiore, Bologna, Italy.Submitted February 16, 2007; acceptedAugust 13, 2007.Supported in part by Grant Nos.5U10CA11488-30 through5U10CA11488-34 from the NationalCancer Institute and by the EuropeanOrganisation for Research and Treat-ment of Cancer (EORTC) Brain CancerGroup; also supported by the EORTCCharitable Trust.Authors’ disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.Address reprint requests to MurielleMauer, PhD, European Organisation forResearch and Treatment of CancerData Center, Quality of Life Unit, AveMounier 83/11, Brussels, Belgium1200; e-mail: murielle.mauer@eortc.be.© 2007 by American Society of ClinicalOncology0732-183X/07/2536-5731/$20.00DOI: 10.1200/JCO.2007.11.1476  J OURNAL OF C LINICAL O NCOLOGY   O R I G I N A L R E P O R T VOLUME 25    NUMBER 36    DECEMBER 20 2007 5731 Downloaded from ascopubs.org by 200.68.138.39 on May 27, 2017 from 200.068.138.039Copyright © 2017 American Society of Clinical Oncology. All rights reserved.  critical areas that could help in the selection of key end points forfuture clinical trials.HRQOL prognostic factor analyses have been carried out onseveral different cancer populations including, among others,lung, 7-10 esophageal, 11,12 advanced breast, 4,13 metastatic malignantmelanoma, 14 metastatic bladder cancer, 15 and head and neck  16 can-cers,highlightingtheimportanceHRQOLscoresmayhaveinpredict-ing survival. Only three studies have examined HRQOL and/orcognitive functioning as a prognostic factor in brain cancer. 17-19 ThisstudyevaluatestheprognosticvalueofHRQOLdatacollectedfromaprospective, large scale international randomized controlled trial, by using various statistical techniques in an attempt to provide robustconclusionsontheprognosticvalueofHRQOLinpatientswithana-plasticoligodendrogliomas. PATIENTS AND METHODS Treatment  Inthisinternational,multicenterstudy(EuropeanOrganisationforRe-search and Treatment of Cancer [EORTC] trial 26951), adult patients withnewly diagnosed and histologically proven anaplastic oligodendroglioma oroligoastrocytomawererandomlyassignedtotreatmentwithradiationtherapy (RT)onlyorradiationtherapyfollowedbysixcyclesofprocarbazine,lomus-tine,andvincristine(PCV)chemotherapy.Patientswerestratifiedforinstitu-tion, performance status (WHO performance status 0 or 1  v   2), age (  40 or  40),theextentoftheresectionatsurgery(biopsyonly  v  debulkingsurgery/resection),andpossiblepriorsurgeryforalow-gradeoligodendroglioma(yesor no). The details on trial conduct and clinical outcome have already beenreported. 20 A central pathologist (J.M. Kros) reviewed all tumor samples andscoredthepresenceorabsenceoftumornecrosis.Thepresenceorabsenceof loss of the short arm of chromosome 1 (1p loss) was assessed in a subset of patientswithfluorescenceinsituhybridization.ThetrialwasapprovedbytheEORTCprotocolreviewcommitteeandtheinstitutionalreviewboardofeachparticipatingcenter.Allpatientsprovidedwritteninformedconsent. Procedures  The primary end point of the clinical trial was survival, with HRQOLbeingasecondaryendpoint.TwoHRQOLmeasureswereselectedforthetrial:the EORTC Quality of Life Questionnaire C30 (EORTC QLQ-C30, version2) 21 and the EORTC QLQ-Brain Cancer Module (QLQ-BN20). 22 Both toolshave robust psychometric properties resulting from rigorous testing and de-velopment from their use in several international cancer clinical trials. 3 TheEORTC QLQ-C30 is a core measure designed to be supplemented withdisease-specific questionnaires. The EORTC QLQ-BN20 was developed spe-cificallyforpatientswithbraincancer.Bothinstrumentswereavailableinthelanguageofallparticipatingpatients. 23 TheEORTCQLQ-C30measurecomprisesfivefunctioningscales:phys-ical, role, emotional, cognitive, and social; three symptom scales: fatigue,nausea/vomiting,andpain;sixsingleitemscales:dyspnoea,insomnia,appetiteloss,constipation,diarrhea,andfinancialimpact;andtheoverallhealth/globalQOLscale.The EORTC QLQ-BN20, designed for use with patients undergoingchemotherapy or RT, includes 20 items assessing visual disorders, motordysfunction,communicationdeficit,variousdiseasesymptoms(eg,headachesandseizures),treatmenttoxicities(eg,hairloss),andfutureuncertainty.The items on both measures were scaled and scored using the recom-mended EORTC procedures. 24 Raw scores were transformed to a linear scaleranging from 0 to 100, with a higher score representing a higher level of functioningorhigherlevelofsymptoms.Providedatleastonehalfoftheitemsin the scale were completed, the scale score was calculated using only thoseitemsforwhichvaluesexisted.PatientswererandomlyassignedaftersurgeryandbeforethestartoftheRT. RT started within 4 to 6 weeks after surgery. Valid HRQOL assessmentswere performed at baseline, not more than 6 weeks before or after randomassignment, more than 7 days after surgery, and before the start of RT.Follow-up assessments were performed at regular intervals (ie, at the end of RT, then every 3 months for the first year after RT, and then at 6-monthly intervalsuntilrecurrenceofthedisease).HRQOL was a mandatory aspect of this clinical trial protocol. Theprotocolstipulatedthataresponsiblenurse,clinician,ordatamanageradmin-isteredthequestionnaire,requestingcompletionanditsreturntotheEORTCData Center. EORTC guidelines for administering questionnaires were pro-vided,ensuringastandardapproachandoptimalcomplianceofHRQOLdatacollectionbyallpersonnel. Statistical Analysis  For this analysis, only baseline HRQOL scores were used. To minimizethe risk of false-positive results, we excluded several HRQOL scales (sevenfrom the QLQ-C30 and two from QLQ-BN20) from the analyses. The vari-ables decided by the group to be excluded were not expected to have any prognosticvalueoralternativelywereknowntobehighlyintercorrelatedwithotherscales,therebynotcontributingtothemodel.Inthevalidationprocess,itwas nevertheless checked that the results do not change when including thenonpreselectedscores.Thefinalanalysisincluded17HRQOLvariables:eightfrom the core questionnaire (appetite loss, cognitive functioning, emotionalfunctioning,fatigue,physicalfunctioning,globalhealthstatus,socialfunction-ing, and insomnia); and nine from QLQ-BN20 (bladder control, communi-cation deficit, drowsiness, future uncertainty, headaches, motor dysfunction,seizures,visualdisorder,andweaknessoflegs).The Cox proportional hazards regression model 25 with overall survivalmeasured from time of random assignment as dependent outcome was usedfor both univariate and multivariate analyses. A Collett’s Model selectionapproach 26 was used with a level of significance of 0.15 for the univariatescreening and stay and entry criterions of 0.05. The HRQOL scales wereincluded as continuous factors. The model was controlled for the majorestablishedprognosticbaselineclinicalfactors—age(  45;45to54;and  55 years),tumorlocation(frontal v  other),performancestatus(0 v  1 v  2),tumornecrosis (yes  v   no), and 1p loss (yes  v   no) — as well as for treatment type by forcingtheirinclusioninallmodels.Validation of the final model was undertaken by use of several refinedstatistical techniques. The stability of the final model was investigated using abootstrap resampling procedure as proposed by Sauerbrei and colleagues, 27 applied in the context of HRQOL. 28 This technique generates a number of samples (each the same size as the srcinal data set), by randomly selectingpatients and replacing them before selecting the next patient (ie, bootstrapresampling). The frequency of inclusion of the HRQOL scores in the Cox proportional hazards regression models, including all the selected clinicalfactorsandtreatment,fittedtoeachofthesedatasetsusingautomaticforwardstepwise selection (entry level of    .05), can be considered to be indicativefortheimportanceoftheHRQOLfactors.Thistechniquewasappliedto1,000bootstrapsamples.Deviance residuals from the model with clinical factors were plottedversus the HRQOL scores to explore the relationship between each HRQOLscore and the remaining part of the hazard not already explained by clinicalfactors. Finally, discrimination C indexes and Nagelkerke’s R  2 coefficientswerecomputedtoquantifythepredictiveaccuracyofamodel.Alldataanaly-seswereperformedusingSAS,version9(SASInstitute,Cary,NC). RESULTS BetweenAugust1996,andMarch2002,368patientsfrom40institu-tions in 10 countries were randomly assigned to receive either RTalone (n  183) or RT plus chemotherapy (n  185; Fig 1). Of thesepatients, 288 had baseline HRQOL measures completed (150 in RTand 138 in RT plus adjuvant chemotherapy). Genetic status (loss of theshortarmofchromosome1(1p)andthelongarmofchromosome19 (19q)) was assessed in 311 patients (156 in RT and 155 in RT plus Mauer et al 5732  J OURNAL OF C LINICAL O NCOLOGY  Downloaded from ascopubs.org by 200.68.138.39 on May 27, 2017 from 200.068.138.039Copyright © 2017 American Society of Clinical Oncology. All rights reserved.  chemotherapy). Only the 247 patients with both baseline HRQOLmeasuresandwithgeneticstatusassessedwereincludedintheanalysis(128 in RT and 119 in RT plus chemotherapy), which represents67.1%oftheoriginalsamplesize. Clinical Results  The clinical results have recently been published. 20 In brief, thestudydemonstratedthatimmediatepostradiotherapytreatmentwithPCV leads to an increase in progression-free survival compared with Population analysedEfficacy Safety Safety ITT population (n = 185)Safety population for RT (n = 180)Safety population for CT (n = 161)Population analysedEfficacy ITT population (n = 183)Safety population for RT (n = 182)  PD (n = 39)Hematological toxicity (n = 53)Nonhematological toxicity (n = 8)Patient’s refusal (n = 8)Death (n = 2)Medical decision (n = 4)Unclear (n = 1)  Lost to follow-up (LFU) (n = 0)Lost to follow-up (LFU) (n = 0)Discontinued RT for other reason than normal completionand LFU (n = 4) PD (n = 0)Other (n = 4)Discontinued RT for other reason than normal completionand LFU (n = 5) PD (n = 3)Other (n = 2)   Allocated to RT (n = 183)RT treatment did not start (n = 1)   Reasons:   Random assignment in a wrong trial (n = 1)Allocated to RT/PCV (n = 185)RT did not start (n = 5)   Reasons:   Early progression (n = 1)   Patient’s refusal (n = 2)   Other (n = 2)   PCV did not start (n = 19)   Discontinued PCV for other reason than normal completionand LFU (n = 115) AllocationFollow-upAnalysis Fig 1.  CONSORT diagram. RT, radiationtherapy; PCV, procarbazine, lomustine, andvincristine; PD, progression of disease;ITT, intention to treat; CT, chemotherapy. Prognostic Value of Quality of Life in Brain Cancer www.jco.org   5733 Downloaded from ascopubs.org by 200.68.138.39 on May 27, 2017 from 200.068.138.039Copyright © 2017 American Society of Clinical Oncology. All rights reserved.  RT alone (median progression-free survival of 23  v   13.2 months; P   .002bythelog-ranktest).Themediansurvivalwas40.3monthsintheRTplusPCVgroupcomparedwith30.6monthsintheRTgroup( P     .23). Eighty-two percent of patients in the RT arm receivedchemotherapyatthetimeofprogression,mostlyPCV.Combinedlossof1p/19qidentifiedaprognosticfavorablesubgroupofoligodendro-glialtumors,regardlessoftreatment.Nogeneticsubgroupcouldbeidentifiedthatbenefitedwithrespecttooverallsurvivalfromadju-vant PCV. Baseline clinical characteristics for patients with a validbaseline HRQOL questionnaire and with 1p/19q loss assessed aredepicted in Table 1.The distribution of the baseline clinical characteristics was very similar between patients with a valid baseline HRQOL questionnaireandthosewithoutandbetweenpatientswith1p/19qlossassessedandthosewithout. Prognostic Factor Analysis Results  Univariate analysis.  When controlled for the major prognosticclinicalfactors,theHRQOLscoreswhichpassedthe15%significancelevel were emotional functioning, bladder control, communicationdeficit, seizures, and weakness of legs. Table 2 presents the results of thisanalysis.  Multivariate analysis.  The Cox multivariate model selected by Collett’smodelselectionapproachcontainedemotionalfunctioning,communication deficit, future uncertainty, and weakness of legs, inadditiontotheselectedclinicalfactors(Table3).However,thesignsof the coefficients related to emotional functioning and communica-tion deficit were opposite to what was expected (ie, worse emo-tional functioning and more communication deficit were relatedto better survival). Bootstrap Resampling Procedure  Table 4 presents the results of the bootstrap resampling proce-dure. The inclusion frequencies higher than 50% were related toemotional functioning (58.1%) and weakness of legs (89%). The fre-quencies of selection of each possible set of HRQOL scores were very low. The most frequently selected model (emotional functioning,futureuncertainty,andweaknessoflegs)wasselectedonly3.4%ofthetime. This indicates a high degree of model instability with no singlemodeltobeuniformlypreferableoverallothers. Plots of the Residuals to Explore Trends  The graphical exploration of the residuals suggested several is-sues.First,duetotheverylownumberofpatientswithextremevaluesforsomeoftheHRQOLscores,theresultscanbeinfluencedbyafew outliers (eg, for emotional functioning, the positive trend could bedriven by the only four patients with a score of 0 but who actually survived considerably longer than most patients). It also suggestedthatfutureuncertaintywasenteredinthemodelasacorrectionforemotional functioning. Finally, the plots indicate a high variability in terms of survival between patients with fixed levels of theHRQOL scores. Discrimination Index C and Nagelkerke’s R  2  coefficient  Discrimination C-indexes and Nagelkerke’s R  2 coefficients werecomputed for the model with clinical prognostic factors only and forvariousmodelswithaddedHRQOLscores.TheCindexofthemodelwith only clinical baseline characteristics is C    0.743 and theNagelkerke’s R  2 coefficient is R  2   0.348. When adding HRQOLscores,C  0.766andR  2  0.412forthefourselectedHRQOLscoresand C  0.769 and R  2  0.436 for all HRQOL scores. Even whenadding all HRQOL scores including the ones which were excludedupfront, C  0.777 and R  2  0.462. For the model without clinicalfactorsbutwithallHRQOLscores,C  0.653andR  2  0.191. Table 1.  Baseline Clinical Characteristics for Patients With HRQOL and1p/19q Loss AssessedVariableTreatmentTotalRT/PCV RTNo. % No. % No. %No. of patients 119 128 247Age, years  45 46 38.7 50 39.1 96 38.945-54 37 31.1 38 29.7 75 30.4  54 36 30.3 40 31.3 76 30.8SexMale 68 57.1 77 60.2 145 58.7Female 51 42.9 51 39.8 102 41.3MMSE  27 33 27.7 41 32.0 74 30.027-30 80 67.2 80 62.5 160 64.8Missing 6 5.0 7 5.5 13 5.3WHO PS0 51 42.9 43 33.6 94 38.11 50 42.0 62 48.4 112 45.32 18 15.1 21 16.4 39 15.8Missing 0 0.0 2 1.6 2 0.8Histologic diagnosisOligodendroglioma 92 77.3 89 69.5 181 73.3Oligoastrocytoma with  25% oligodendralelements27 22.7 38 29.7 65 26.3Missing 0 0.0 1 0.8 1 0.4Previous resectionNo 100 84.0 112 87.5 212 85.8Yes 19 16.0 15 11.7 34 13.8Missing 0 0.0 1 0.8 1 0.4Tumor enhancementNo 21 17.6 22 17.2 43 17.4Yes 92 77.3 98 76.6 190 76.9Unknown 5 4.2 8 6.3 13 5.3Missing 1 0.8 0 0.0 1 0.4Tumor locationFrontal 56 47.1 61 47.7 117 47.4Elsewhere 63 52.9 67 52.3 130 52.6Type of surgeryBiopsy 17 14.3 15 11.7 32 13.0Partial remission 63 52.9 57 44.5 120 48.6Total remission 39 32.8 56 43.8 95 38.5Endothelial abnormalitiesNo 20 16.8 26 20.3 46 18.6Yes 98 82.4 99 77.3 197 79.8Missing 1 0.8 3 2.3 4 1.6NecrosisNo 54 45.4 47 36.7 101 40.9Yes 64 53.8 78 60.9 142 57.5Missing 1 0.8 3 2.3 4 1.61p and19q lossno loss 55 46.2 61 47.7 116 47.019q loss 13 10.9 18 14.1 31 12.61p loss 21 17.6 20 15.6 41 16.61p/19q loss 30 25.2 29 22.7 59 23.9Abbreviations: HRQOL, health-related quality of life; RT, radiation therapy;PCV, procarbazine, lomustine, and vincristine; MMSE, Mini Mental StatusEvaluation; PS, performance status. Mauer et al 5734  J OURNAL OF C LINICAL O NCOLOGY  Downloaded from ascopubs.org by 200.68.138.39 on May 27, 2017 from 200.068.138.039Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

Aptitude 1

Jun 15, 2018

AMOR INTENSO

Jun 15, 2018
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks