EPLY: We agree with Dr Atkin’s description description of the concluIN R EPLY sions fro sions from m therepor thereportt of th thee Int Intern ernat ation ional al Wo Works rkshopon hopon Scr Screen eeningfor ingfor 1 Breast Cancer. The Thestat statemen ementt repo reportin rtingg a mor mortali tality ty bene benefit fit desc describe ribed d forwomen forwome n ol olderthanage derthanage 40year 40yearss is issup suppo porte rted d bya me metata-ana analy lysisthat sisthat included several randomized trials and specific specifically ally evaluated benefits forr wo fo wome men n ag agee 40to 49yea 49years rs..2 We regr regret et thatthis addi addition tional al refe referenc rencee wasnot att attac achedto hedto thi thiss sta state teme ment.We nt.We agr agree ee wit with h Dr At Atkin kinss tha thatt the there re remains uncertainty regarding long-term radiation exposure from mammogr mamm ography aphy,, part particu icularl larlyy in hig high-ri h-risk sk wome women n who whobegi begin n scre screenin eningg earl ea rlie ierr th than an ag agee 40 ye year ars. s. As we ho hope ped d to co conv nvey ey,, we fe feel el th that at al alte tern rnat ativ ivee imaging imagi ngmodal modalities ities,, such suchas asbreastmagneticresonanceimaging,which breastmagneticresonanceimaging,whichdo do not invol involve ve radiat radiation ionexpos exposure uremay mayulti ultimatel matelyy replac replacee mamm mammograp ography hy in individual indiv idualized ized progr programs ams for breas breastt cance cancerr preve prevention ntion..3
Angela R. Bradbury and Olufunmilyo I. Olopade Department of Medicine, Section of Hematology/Oncology, University of Chicago, Chicago, IL AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The authors indicated no potential conflicts of interest. REFERENCES 1. Fletcher SW, Vlack W, Harris R, et al: Report of the International Workshop on Screening for Breast Cancer. J Natl Cancer Inst 85:1644-1656, 1993 2. Hend Hendric rick k RE, Smith RA, Rut Rutled ledge ge JH III III,, et al: Benefit of scr screen eening ing
mammography mammograph y in women aged 40-49: A new meta-analysis meta-analysis of random randomized ized controlled trials. J Natl Cancer Inst Monogr 22:87-92, 1997 Bradbury ry A, Olopa Olopade de OI: The case for individualized individualized screening recommenrecommen3. Bradbu dations for breast cancer. J Clin Oncol 24:3328-3329, 2006
■ ■ ■
Should Subgroup Analysis of Randomized Clinical Trials Have a Direct Impact on Clinical Practice? Thee rec recentpubl entpublic icati ation on in the Jour Journal nal of Clin Clinical ical TO THE EDITOR : Th Oncology of of the randomized clinical trial performed performed by the European Organisation Organisat ion for Research and Treatment of Cancer Chronotherapy Group did raise the issue of the clinical reliability of sex subgroup analysis anal ysis in cli clinica nicall tria trials. ls.1 Th Thee pri prima mary ry end poi point nt of thi thiss lar large ge co coope operrative multicentric phase III trial was 2-year survival, and the sample sizee was the siz then n cal calcu culat lated ed to de deter termi mine ne a 10 10% % sur survi vival val di diffe fferen rence ce (fr (from om 30%to 30% to 40% 40%)) in favo favorr of the chro chronomo nomodula dulated ted arm arm(sup (superio eriorit rityy tria trial). l). Patients Pati entswere werestra stratifie tified d acc accordi ording ng to perf performa ormance ncestat status, us,live liverr invo involvelvement, men t, andcent andcenter.In er.In theabst theabstrac ractt (wh (whic ich h hasto be becon consid sidere ered d th thee mo most st rapid source for a take-home message), the authors conclude that “both “bo th reg regim imens ens ac achie hieved ved sim simil ilar ar me medi dian an su survi rviva vall ti time mess mo more re tha than n 18 months with an acceptable toxicity” and that “chronomodulated schedul sche dulee prod produceda uceda surv survivaladvant ivaladvantage agein inmen. men.””1 A controve controversy rsy exists regarding the correct interpretation of subgroup analysis of randomized clinical trials, given that the risk to provide misinterpretation could lead to wrong treatment guidelines. Although the trial is adequately powered for its main end point, it is not able to detect difference ences s inanalysis efficacy effic acylead between bet ween subgroup subg s wit with h adeq adequate uateeffect power. powe r. Moreover More over,, the sex to both aroups significantly better for chronotherapy in men, but also a significantly better effect for standard chemotherapy in women, and this is not mentioned in the abstract conclusion.1 Actually, in the article a balanced discussion about the subgroup issues is present. Although Although the multivariate analysis should protect prot ectoverbaselin overbaselinee prog prognost nostic ic fact factors’ ors’inte interact raction ion(whe (when n adeq adequate uately ly studied), the list of the patients’ characteristics, which the authors decided to consider for subgroup analysis, should be reported in the article as well. A number of pieces of specific literature have been published regarding the risk of the erroneous interpretation of subgroup gro up ana analys lyses es and the mo most st com common mon co concl nclusi usions ons ar aree tho those se ana analy lyses ses 2 can be overinterpreted and may lead to suboptimal patient care. A sort of methodologic recommendation for subgroup analyses and their interpretation has been recently produced by Brookes et al3,4— only onlyprev previous iously planned pla nned subgroup subg analyses yses shou should ld be performe perf ormed; d; any triall shou tria should ldbe belypowe powered redtaki taking ngroupanal intoaccountthe prespec pres pecifiedsubgrou ifiedsubgroup p analyses anal yses;; notw notwiths ithstand tanding ing all thes thesee cons consider ideratio ations, ns, any spec specific ific anal analyywww.jco.org
sis is aff affect ected ed by sev severa erall bi bias, as, and the co concl nclusi usions ons nee need d to be sof softe tened ned.. In abs absenc encee of str stron ongg com commo mon n ev evid idenc ence, e, thi thiss is an hyp hypoth othesi esissgenerati gene rating ng exe exercis rcisee for forfurt further herspec specifica ifically llyaddr addresse essed d rand randomi omized zedtri trial. al. The risk in subgroup analysis to provide errors has been also quantified, fie d, an and d se seem emss to be no nott re rela late ted d to th thee pa pati tien entt sa samp mple le si size ze..2 1 Correctl Corr ectly, y,Giac Giacchet chetti ti et etal al performe performed d a sex-treat sex-treatment mentinteract interaction ion analysis, and this interacti interaction on resulted significantly at the multivariate analysis. This result is interesting. However, the observed survival differences differen ces by sex could be due to imbalanc imbalancee in unmeasure unmeasured d baseline prognostic factors (demographic and/orcharacteristics molecular characteristics). According to article’s results, “Patients’ according to sexwerewell bal balanc anced edexc excep eptt forage andPS. Ag Agee 50yearsandPSof 1 or 2 were more frequent in women than in men (21% v 14%; 14%; and 59% v 46.5%, respectively). After stratification for age and PS, the interaction tests were not found significant (P .1).” If not due to imbalance, the difference of outcome between males and females could be a true difference or could be due to statistical chance alone, especi esp ecial ally ly be becau cause se not onl onlyy sex bu butt a lo long ng li list st of pa patie tient nt cha charac racter terist istic icss has been analyzed. So, in our opinion, more caution should be used when authors decide to put the results of secondary, exploratory analy ana lyses ses in the abs abstr tract act of an art artic icle le..
Emilio Bria Department of Medical Oncology, Regina Elena National Cancer Institute, Rome, Italy
Massimo Di Maio
National Cancer Institute, Naples, Italy
Federica Cuppone, Cecilia Nistico`, `, and Francesco Cognetti Department of Medical Oncology, Regina Elena National Cancer Institute, Rome, Italy
Diana Giannarelli Biostatistics, Regina Elena National Cancer Institute, Rome, Italy AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The authors indicated no potential conflicts of interest. REFERENCES 1. Giacchetti S, Bjarnason G, Garufi C, et al: Phase III trial comparing 4-day chronomodulated chronomodul ated therap therapy y versus 2-day conven conventional tional delivery of fluorou fluorouracil, racil, leucovorin, leucov orin, and oxalip oxaliplatin latin as first-l first-line ine chemo chemotherap therapy y of metas metastatic tatic colorectal cancer can cer:: The Eur Europe opean an Org Organi anisat sation ion for Res Resear earch ch and Tre Treatm atment ent of Can Cancer cer Chronotherap Chrono therapy y Group. J Clin Oncol 24:356 24:3562-3569 2-3569,, 2006 2. Lagakos SW: The challenge of subgroup analyses–reporting without distorting. N Engl J Med 354:1667-1669, 2006
3. Brookes ST, Whitley E, Peters TJ, et al: Subgroup analyses in randomized controlled trials controlled trials:: Quant Quantifying ifying the risks of false false-posit -positives ives and falsefalse-negati negatives. ves. Health Technol Assess 5:1-56, 2001
4. Brookes ST, Whitely E, Egger M, et al: Subgroup analyses in randomized trials:: Ris trials Risks ks of sub subgro groupup-spe specifi cific c ana analys lyses; es; pow power er and sam sample ple siz size e for the interaction intera ction test. J Clin Epidemiol 57:229-236, 57:229-236, 2004
■ ■ ■
Bria et al emphasize that more cauti cau tion on is ne neede eded d in rep report ortin ingg the eff effec ects ts of sex on sur survi vival val and to toler ler-abilitythat abil itythat werederiv werederived edfromthe fromthe Eur Europea opean n Orga Organisa nisationfor tionfor Rese Research arch and Treatment of Cancer phase III trial 05963. 1 Their comment is based base d on the cont controve roversy rsy rega regardin rdingg the gene generati ration on and the inte interpre rpretata-
cause of thei cause theirr hist history oryof of bei being ng over overint interpr erpretedwould etedwouldhavebeen havebeen a stee steep p price to pay. Rather than advocating for a ready change in clinical practi pra ctice ce whi which ch cou could ld im impl plyy th thee de deli liver veryy of FOL FOLFOX FOX (2 (2-ho -hour ur inf infusi usion on of oxaliplatin and leucovorin on day 1 and 2-hour infusion of leucovorin vori n on day 2, whic which h inte intercal rcalcul culated ated 22-hour 22-hour cons constant tant inf infusio usion n
tion of subgroups analyses. Indeed, we strongly believe that the abstract serves the purpose of delivering the most clinically relevant findin find ings gs wi with th hon honest esty. y. Thi Thiss is iswhywe whywe wro wrote,“In te,“In wom women,the en,the ris risk k of an earlier death on ChronoFLO4 was increased by 38% compared with FOLFOX2. . . .In men, the risk of death was decreased by 25% on ChronoFLO4. . . .” So, we agree with Bria et al that the abstract is the most rapid source for a take home message, but we mean the whole abstract abst ract,, not the conc conclus lusion ion only only.. For tre treatm atment ent int intera eracti ction on tes tests, ts, we con consid sidere ered d fac factor torss rel relate ated d to patient characteristics (age, sex, WHO performance status [PS] 0 v 0) 0),, baseline hematologic and biochemical data (leukocyte, granulocyte and platelet plate let count counts, s, alkal alkaline ine phosph phosphatases atases,, gamm gammaa glut glutamylt amyltransfe ransferase, rase, transaminases, carcinoembryonic antigen, and CA 19-9), and tumor characteristics (primary site, Duke’s stage, liver involvement, number of metastatic sites). At the exception of sex, no other factor displayed any
rate of fluorouracil on days 1 and 2) in women and that of chronomodulated FLO4 in men as standard best practice, we proposed to investigate the mechanisms through which optimal scheduling of chemotherapy differs between men and women. Indeed, increased attenti atte ntion on is cur current rently lybein beingg paidby the themed medicaland icaland sci scienti entific fic comm commuunity to the role of sex for treatment toxicities and patient outcome in cancer can cer and oth other er di disea seases ses.. Ourr co Ou corr rres espo pond nden ence ce (B (Bri riaa et al an and d ou ourr re repl ply) y) em emph phas asiz izes es th thee ne need ed for studies testing new chemotherapy and chronotherapy associations with targeted molecules and a priori planning of effects of sex or other factor fac torss ide identi ntified fiedin in sub subgro group up ana analys lyses. es. Alr Alread eadyy a sex dep depend endenc encyy of tox tox-icit ic ityy an and d su surv rviv ival al ha hass be been en re rece cent ntly ly fo foun und d by ou ourr gr grou oup p fo forr tw two o se sepa para rate te 7,8 studies, including one with cetuximab. We are currently developing newpro new protoc tocols olsaim aimed ed at the thedet determ ermina inatio tion n of the thebes bestt sch schedu edule lein in wom women en with metastatic colorectal cancers. Moreover, to better understand the
strong str ongly ly sig signifi nifican cantt int intera eracti ction on wit with h the thedel delive ivery ry sch schedu edule le (P .01). Asmen As mentio tionedin nedin theartic thearticle,an le,an imb imbala alancebetwe ncebetween enwom women enandmen andmen wass fo wa foun und d fo forr PS an and d ag age. e. Af Afte terr st stra rati tific ficat atio ion n fo forr ag agee an and d PS PS,, th thee in inte tera racction ti on te test stss we were re no nott fo foun und d si sign gnifi ifica cant nt (P .1 .1)) be beca caus usee of a la lack ck of po powe wer, r, but presented the same trend for a higher efficacy in men treated by infusion of fluorouracil, leucovorin, and oxaliplatin for 4 days (chronomodula mod ulatedFLO4 tedFLO4)) in inthediffe thedifferen rentt ageand ageandPS PSstr strata ata.. Ind Indeed eed,, thediffe thedifferen rence ce betwee bet ween n men menand andwom women en cou could ld be a tru truee dif differ ferenc encee or sta statis tistic tical al cha chance nce.. However, in a Cox proportional hazards model a sex x schedule interaction term was found significant in presence of PS, age, number of metastatic sites, and percentage of liver involvement (Table (Tab le 4).1 Thi Thiss su suppo pports rts tha thatt se sexx ef effec fectt is a tru truee di diff ffere erence nce.. Lagakos recommends being cautious with subgroups analysis. He advises to perform planned subgroups analysis, appropriate appropriate anal ysis with inte interact raction ion test tests, s, and to giv givee mag magnitu nitude de of the trea treatme tment nt
mechanism mechan ism of the sex eff effect ect,, tra transl nslati ationa onall res resear earch ch pro projec jects ts of mol molecu ecular lar clock clo ck det determ ermina inants ntsof of opt optima imall sch schedu edule le are areals also o ong ongoin oing. g. The cha chall lleng engee ahe ahead ad of us is the adm admini inist strat ratio ion n of tai tailor lored ed ch chroronotherapy (ie, the determination of the right drugs at the right doses and an d at th thee ri righ ghtt ti time me in th thee ri righ ghtt pa pati tien ent) t)..
difference with corresponding confidence interval instead of the P value.2 Bria et al acknowledge that we followed these recommendations, including interaction test and Forrest plot of interaction between schedule and sex. However, we could not plan the subgroups analy ana lyses ses si since nce suc such h ef effec fectt was lar largel gelyy unk unknow nown n whe when n we ini initi tiate ated d thi thiss trial. Thus, a consistent impact of sex on chemotherapy toxicity or patient outcome started to arise in the literature once accrual to our triall was comp tria complete leted. d.3 We feel that clinical research should aim at discovering new prognostic and predictive factors, new treatments, better treatment schedules, and new strategies. Through the use of chronomodulated deliveryof deli veryof chem chemothe otherapy rapy,, our ourgrou group p first firstdemo demonstr nstratedthe atedthe acti activit vityy of oxaliplatin oxalipl atin against colorectal cancer,4 then the relevance of liver metastasessurger tast asessurgeryy forthe surv surviva ivall of pati patientswith entswithdown downstag staged edprev previou iously sly 5,6 unresectable unresecta blemetastase metastases. s. Bothof thes thesee findi findingsinitial ngsinitially lyenco encounte untered red
INSERM U 776, University Paris Sud, and Hôpital Paul Brousse, Assistance Publique-Hôpitaux de Paris, Paris, France
EPLY: In their Conclusion, Conclusion, IN R EPLY
great scepticism, scepticism, yet they now have profound profoundly ly modified the current treatment of colorectal cancer worldwide. worldwide. Thus we share the opinion of Lagakos that avoiding any presentation of subgroup analysis be606
Sylvie Giacchetti INSERM U 776 and Hôpital Paul Brousse, Assistance Publique-Hôpitaux de Paris, Paris, France
Thierryy Gorlia Thierr Data Center, European Organisation Organisation for Research and Treatm Treatment ent of Cancer Cancer,, Brussels, Belgium
Carlo Garufi Istituto Istitu to Regina Elena, Roma, Italy
Francis Fra ncis Le´vi ´vi
AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The authors indicated no potential conflicts of interest. REFERENCES 1. Giacchetti S, Bjarnason G, Garufi C, et al: Phase III trial comparing 4-day chronomodulated chronomodul ated therap therapy y versus 2-day conven conventional tional delivery of fluorou fluorouracil, racil, leucovorin, leucov orin, and oxalip oxaliplatin latin as first-l first-line ine chemo chemotherap therapy y of metas metastatic tatic colorectal cancer can cer:: The Eur Europe opean an Org Organi anisat sation ion for Res Resear earch ch and Tre Treatm atment ent of Can Cancer cer Chronotherap Chrono therapy y Group. J Clin Oncol 24:356 24:3562-3569 2-3569,, 2006 2. Lagakos SW: The challenge of subgroup analyses–reporting without distorting. N Engl J Med 354:1667-1669, 2006 3. Neugut AI, Jacobson JS: Women and lung cancer: Gender equality at a crossroad? JAMA 296:218-219, 2006 chronopharm pharmacolog acologic ic phase II clinic clinical al 4. Lévi F, Misset JL, Brienza S, et al: A chrono trial tri al wit with h 5-fl 5-fluor uorour ouraci acil, l, fol folini inic c aci acid d and oxa oxalip liplat latinu inum m usi using ng an amb ambula ulator tory y multichannel programmable pump: High antitumor effectiveness against metastatic colorectal cancer. Cancer 69:893-900, 1992