design of phase i combination trials: recommendations of ... · these include the selection of...

Review

Design of Phase I Combination Trials: Recommendations ofthe Clinical Trial Design Task Force of the NCI InvestigationalDrug Steering Committee

Channing J. Paller1, Penelope A. Bradbury11, S. Percy Ivy2, Lesley Seymour11, Patricia M. LoRusso3,Laurence Baker4, Larry Rubinstein2, Erich Huang2, Deborah Collyar5, Susan Groshen6, Steven Reeves2,Lee M. Ellis7, Daniel J. Sargent8, Gary L. Rosner1, Michael L. LeBlanc9, and Mark J. Ratain10

AbstractAnticancer drugs are combined in an effort to treat a heterogeneous tumor or to maximize the

pharmacodynamic effect. The development of combination regimens, while desirable, poses unique

challenges. These include the selection of agents for combination therapy that may lead to improved

efficacywhilemaintaining acceptable toxicity, the design of clinical trials that provide informative results for

individual agents and combinations, and logistic and regulatory challenges. The phase I trial is often the

initial step in the clinical evaluation of a combination regimen. In view of the importance of combination

regimens and the challenges associated with developing them, the Clinical Trial Design (CTD) Task Force of

theNationalCancer Institute InvestigationalDrug SteeringCommittee developed a set of recommendations

for the phase I development of a combination regimen. The first two recommendations focus on the

scientific rationale and development plans for the combination regimen; subsequent recommendations

encompass clinical design aspects. TheCTDTask Force recommends that selection of theproposed regimens

be based on a biologic or pharmacologic rationale supported by clinical and/or robust and validated

preclinical evidence, and accompanied by a plan for subsequent development of the combination. The

design of the phase I clinical trial should take into consideration the potential pharmacokinetic and

pharmacodynamic interactions as well as overlapping toxicity. Depending on the specific hypothesized

interaction, the primary endpoint may be dose optimization, pharmacokinetics, and/or pharmacodynam-

ics (i.e., biomarker). Clin Cancer Res; 20(16); 4210–7. �2014 AACR.

Disclosure of Potential Conflicts of InterestL. Baker is a consultant/advisory board member for Cytrx, Immune Design, and Morphotek. L.M. Ellis is a consultant/advisory board

member for Amgen, Genentech/Roche, and ImClone Systems/Eli Lilly. G. Rosner reports receiving speakers bureau honoraria from

Novartis. M.J. Ratain reports receiving a commercial research grant from Bristol-Myers Squibb and is a consultant/advisory boardmember

for AbbVie, Daiichi Sankyo, and Genentech. No potential conflicts of interest were disclosed by the other authors.

CME Staff Planners' DisclosuresThe members of the planning committee have no real or apparent conflicts of interest to disclose.

Learning ObjectivesUpon completion of this activity, the participant should be able to describe what is required to justify phase I evaluation of a combination

of drugs. In addition, the participant should be able to determine whether a combination of drugs needs a formal phase I evaluation or a

formal drug–drug interaction study, or no formal phase I study is needed.

Acknowledgment of Financial or Other SupportThis activity does not receive commercial support.

Authors' Affiliations: 1Sidney Kimmel Comprehensive Cancer Center,Johns Hopkins Hospital, Baltimore; 2National Cancer Institute, Bethesda,Maryland; 3Karmanos Cancer Institute, Detroit; 4University of Michigan, AnnArbor, Michigan; 5Patient Advocates in Research, Danville; 6Norris Compre-hensive Cancer Center, University of Southern California, Los Angeles,California; 7University of Texas MD Anderson Cancer Center, Houston,Texas; 8Mayo Clinic, Rochester, Minnesota; 9Fred Hutchinson CancerResearch Center, Cancer Research and Biostatistics, Seattle, Washington;10The University of Chicago, Department of Medicine, Chicago, Illinois; and11NCIC Clinical Trials Group, Queen's University, Kingston, Ontario, Canada

Note: C.J. Paller and P.A. Bradbury contributed equally to this article.

Corresponding Author: Mark J. Ratain, University of Chicago Medicine,MC2115, 5841 SouthMaryland Street, Chicago, IL 60637. Phone: 773-702-4400; Fax: 773-702-3969; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-14-0521

�2014 American Association for Cancer Research.

ClinicalCancer

Research

Clin Cancer Res; 20(16) August 15, 20144210

Research. on September 30, 2014. © 2014 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

IntroductionIn most tumors, no single pathway has been identified

that uniquely drives the malignant process. A more favor-able therapeutic response may be obtained by combiningdrugs that target multiple pathways and/or inhibit resis-tance mechanisms (e.g., pharmacodynamic modulation).The past decade has seen the development of a vast arrayof new drugs focusing predominantly on specific molec-ular targets or pathways of interest. Perhaps the greatestclinical benefit from this approach has been demonstrat-ed in malignancies driven predominantly by an identifi-able molecular aberration (1–5). However, resistanceusually develops (6–10). Conversely, in tumors that aregenetically diverse (multiple "driver" mutations/altera-tions), focusing on a single target in an unselected pop-ulation has had modest results (11–16). Combiningmolecularly targeted and/or cytotoxic drugs may be onestrategy to overcome these limitations and improve effi-cacy (17).The design and conduct of the phase I combination trial

present specific challenges, such as the optimum selectionof agents to combine among the range of possible combi-nations; the selection of the appropriate dose and schedule(includingwhich drug or drugs to dose escalate); drug–druginteractions; overlapping toxicities; and logistic and regu-latory challenges. To address these challenges, the Investi-gational Drug Steering Committee (IDSC) of the NationalCancer Institute (NCI) appointed a Clinical Trial Design(CTD) Task Force composed of academics, pharmaceuticalindustry representatives, and patient advocates, to developrecommendations (Table 1) similar to those developedpreviously for phase I and II clinical trials (18, 19). TheCTD Task Force focused on development of combinationsof systemic agents (marketed or investigational), with con-sideration of the proposed mechanism of action, pharma-cokinetics, and expected toxicities. The recommendationsprovide pragmatic clinical guidelines rather than a rigid setof rules and do not encompass in-depth details of studydesigns or regulatory or logistic challenges of combinationregimens. The consensus recommendations were reviewed

and approved by the IDSC on March 13, 2012 (Fig. 1;ref. 20).

Consensus RecommendationsRecommendation 1

All phase I combination trials should state an explicit orimplicit hypothesis justifying the combination, including a phar-macologic or biologic rationale that includes at least one of thefollowing: in vitro data, in vivo data, or clinical data. Therationalemay extrapolate from results with similar drugs andmaybe based on in silico analyses. The hypothesis supporting thecombination should be clearly stated in the protocol.

Given the vast number of combinations of anticancerdrugs that could be evaluated (21), priority should begiven to those combinations that are based on the stron-gest rationale and are most likely to result in clinicallysignificant therapeutic advances. The phase I studyshould, therefore, have a clearly referenced rationalejustifying evaluation of the combination. The overarchinghypothesis to combine anticancer drugs is to enhanceantitumor effects (Table 2). The underlying hypothesisshould include a pharmacologic and biologic rationalesupported by at least one of the following: in vitro data, invivo data, or clinical data.

The level of preclinical data required topredict a benefit inclinical trials is currently unknown, because preclinicalstudies do not generally predict success of clinical trials(22–24). Preclinical data may help in the prioritization ofcombinations to advance to clinical trials and in the designof the subsequent clinical trial (see recommendations 3 and4; ref. 25). Poor therapeutic indices may be associated withthe high attrition rates found in oncology drug develop-ment, yet seem tobe infrequently evaluated (26). Preclinicalmodels have been shown to predict some non hematologictoxicities in humans, including skin toxicity and gastroin-testinal toxicities associated with EGFR inhibitors (27).Toxicities such as myalgias, arthralgias, and headaches arenot detectable in preclinical studies.

The additive or synergistic effects of a drug combi-nation can be evaluated in cell line assays; however,

Table 1. Past and present members of the of the IDSC Clinical Trial Design Task Force

Position Name

Co-chairs Mark J. Ratain, Michael L. LeBlancMembers Laurence H. Baker, Penelope A. Bradbury, Lee M. Ellis, Elizabeth Garrett-Mayer,

Richard Gaynor, Gary Gordon, Susan Groshen, Robert Iannone, Patricia Keegan,Patricia M. LoRusso, Stuart Lutzker, Channing J. Paller, Gary L. Rosner, Larry Rubinstein,Daniel J. Sargent, Lalitha K. Shankar, Manish Sharma, Anthony Shields, David R. Spriggs

NCI liaison S. Percy IvyPast chair and co-chair Lesley Seymour, Donald BerryNonvoting members Rajeev Agarwal, Lori Minasian, Peter UjhazyIDSC patient advocate Deborah CollyarPast members John Crowley, Afshin Dowlati, Jeffrey Humphrey, Mario Sznol,

Miguel Villalona-Calero, Siu-Long Yao

Recommendations for Design of Phase I Combination Trials

www.aacrjournals.org Clin Cancer Res; 20(16) August 15, 2014 4211



determining synergy in vivo is complex, standard defi-nitions of synergy in vivo do not exist, and such standarddefinitions are used exclusively with in vitro models(28). Another consideration is the potential for antag-onistic effects of agents when combined and the effectof sequence, rarely tested preclinically. The addition ofgefitinib or erlotinib to chemotherapy has not confer-red a demonstrable clinical benefit. Subsequent preclin-ical studies demonstrated that concurrent administra-tion of an EGFR TKI with standard chemotherapy innon–small cell lung cancer may be antagonistic, whereas

sequential administration may have improved activity(29–31). This potential can be explored using preclinicalmodels and may inform the design of schedules in theclinical trials. The IDSC has outlined some considera-tions about the selection of agents to take forward intoclinical trials (18). However, negative preclinical resultsmay not be reported because of publication bias, lim-iting relevant evidence, and possibly contributing tohigh failure rates in oncology drug development. Pub-lication of negative results through journals such asthe Journal of Negative Results in BioMedicine (32, 33)

CCR Reviews

© 2014 American Association for Cancer Research

4a. Use a formal phase I

evaluation with success/

failure criteria stated a priori.

4b. Use a formal drug–drug

interaction design—primary

endpoint PK.

4c. No formal phase I required.

Consider “tolerability” as first

phase (run-in) of the phase II study.

3c. If no plausible basis

for PK or PD interaction

3b. If no overlapping DLT and

no plausible PD, but plausible

basis for PK interaction

3a. If overlapping DLT(s)

or plausible basis for

PD leading to DLT(s)

2. State potential clinical results from

phase I and hypotheses for future

studies.

1. Provide explicit or implicit hypothesis

justifying the combination, including a

pharmacologic or biologic rationale.

Figure 1. Process fordetermination of the phase Icombination trial design. PD,pharmacodynamic; PK,pharmacokinetic.

Table 2. Hypotheses justifying combination trials

Target Tumor type Examples

Target multiple mechanismsof action

Gastric, colorectal,and NSCLC

Addition of HER-2 (22), EGFR (23), or VEGF (46–48, 60)pathway inhibitors broadened anticancer activity ofstandard chemotherapy while minimizing cross-resistance

Optimize the inhibition of aspecific target or pathway

Melanoma Combination of CTLA-4 and PD-1 receptor inhibitors resultedin tumor regression beyond that expected from monotherapy (61)

Target a potential resistancemechanism (bypass pathway)

Breast, melanoma,and NSCLC

Addition of an mTOR inhibitor to antiestrogens (62), a MEKinhibitor to a BRAF kinase inhibitor (34), or an EGFR inhibitor toa MET inhibitor (49, 63) results in restoration of sensitivity anddecreased proliferation in cell lines and, for patients with NSCLC,to increased progression-free survival

Abbreviations: CTLA-4, cytotoxic T lymphocyte–associated antigen 4; NSCLC, non–small cell lung cancer; PD-1, programmeddeath 1.

Paller et al.

Clin Cancer Res; 20(16) August 15, 2014 Clinical Cancer Research4212



may provide important information for the combina-tion trial design.

Recommendation 2The potential results and next steps of the development plan of

the combination should be clearly described. The descriptionshould include two parts: the rationale for why the biologic orpharmacologic interactions should translate into clinical effects,and one ormore examples of phase II studies to test the hypothesis.The phase II example(s) should follow the guidelines of the PhaseII Consensus Recommendations of the NCI’s IDSC CTD TaskForce.In addition to a robust underlying hypothesis outlining

why specific agents should be combined (recommendation1), a clearly defined plan of how the combination will beevaluated in phase I and II clinical studies (recommenda-tions 3 and 4), and the anticipated outcome of those trials,should be outlined. The trial design may aim to optimizea toxicity endpoint, pharmacodynamic biomarker or bedescriptive and exploratory. As there are an infinite numberofmaximally tolerated doses for a drug combination, whichmay also be the case when optimizing a pharmacodynamicbiomarker, the recommendation is not intended to berestrictive in nature. However, the rationale for the designshould consider future trials. For example, a three-partphase I/II trial explored the potential for an MEK inhibitorto delay the resistance to BRAF inhibition and the safety ofthe combination: (part A) determined potential pharma-cokinetic interactions; (part B) evaluated toxicity, safety,and pharmacokinetics of escalating doses of both agents;and (part C) proof-of-principle in the randomized phase IIstudy evaluating progression-free survival (PFS; ref. 34). Ifthe development plan for a combination is unclear or notfeasible, it calls into question the rationale for undertakingthe phase I trial in the first place. Furthermore, specificcriteria, including decision rules for success (e.g., a regimenthat can be moved forward in development) and failure

(e.g., a regimen that is too toxic for further evaluation),should be defined and fully developed in the clinical pro-tocol (Table 3). Trametinib and dabrafenib were success-fully combined with no significant incremental toxicity andfewer squamous cell carcinomas than for patients receivingmonotherapy, and proof-of-principle was demonstratedwith an improvement in PFS with the combination (34).In contrast, when low-dose sorafenib and bevacizumabwere combined, the tolerated doses of both agents were aquarter to half the single-agent dose used in other solidtumor studies because of unexpectedly severe toxicities,including hand–foot syndrome, hypertension, proteinuria,and thrombocytopenia (35).

Recommendation 3The design of combination phase I studies should address the

following three factors: overlapping dose-limiting toxicities(DLT); a plausible mechanistic basis for a pharmacodynamicinteraction leading to DLTs; and a plausible mechanistic basis fora pharmacokinetic interaction.

Because of the differences in pharmacology, mechanismof action, and toxicity of individual agents, the design of thephase I combination study should be tailored to the specificdrugs to be combined. This may involve a formal phase Idose escalation trial (36), a pharmacokinetic endpoint (37),or a safety run-in to a phase II study. Three importantconsiderations are (i) the potential for overlapping DLTs,(ii) pharmacodynamic interactions, and (iii) pharmacoki-netic interactions.

Overlapping DLTs may limit escalation of doses to levelsrequired for optimal activity ormay affect dose intensity dueto dose reductions when a regimen is administered chron-ically. Even when overlapping DLTs do not exist, pharma-codynamic interactions may result in toxicity and affectdose. Combining bevacizumab and sorafenib (35, 38) orsunitinib (39) resulted in proteinuria and thrombocytope-nia, requiring modification of dosing and scheduling. The

Table 3. Selected pharmacokinetic interactions

Drug–drug or drug–food combination PK interaction Mechanism

Gefitinib with bexarotene Plasma levels of gefitinib significantlyreduced (64, 65)

Gefitinib is metabolized by multiplecytochrome P450 enzymes,including bexarotene

Temsirolimus with lenalidomide Administration of temsirolimusincreased maximum concentrationand area under the concentration-timecurve of lenalidomide (43)

Lenalidomide is P-glycoprotein substrate

Imatinib, dasatinib, and nilotinibwith high-fat meals

AUC increased by 82% when nilotinibwas given 30 minutes after ahigh-fat meal (66)

Oral TKIs have a high risk of PK interactionswhen administered in conjunctionwith high-fat meals

Imatinib, dasatinib, and nilotinibwith ketoconazole, levothyroxine,and verapamil

Imatinib exposure increased followingketoconazole coadministration (67)

Oral TKIs such as have a high risk ofdrug interactions when administeredwith drugs affecting CYP3A4

Abbreviations: AUC, area under the curve; CYP3A4, cytochrome P450 3A4; PK, pharmacokinetic; TKI, tyrosine kinase inhibitors.





additive effects of mild overlapping adverse events mayimpair tolerance, particularly for drugs that are intendedto be administered chronically (28).

Pharmacokinetic interactions may alter the absorption,distribution, metabolism, and excretion of one or bothdrugs (Table 3). However, pharmacokinetic assessments ofdrug–drug interaction should be routinely included in thephase I trial design onlywhen scientific justification for suchinteractions, at pharmacologically achievable drug concen-trations, has been identified (40). We recommend com-pleting an initial pharmacokinetic analysis of the individualdrug before initiating the next drug to increase the reliabilityof evidence for or against an interaction (37, 41, 42), ratherthan performing pharmacokinetic studies of all drugs in thecombination and comparing the results to historical con-trols (43).

Combining agents with conflicting requirements withrespect to the timing ofmeals adds challenges. Understand-ing these potential interactions will help generate hypoth-eses to be addressed in the phase I clinical trial, which willthen influence the subsequent clinical trial design. Explicitinstructions to trial participants are critical to ensure carefultreatment selection for future trials.

Drug scheduling of combinations may affect additive orsynergistic effects on efficacy, toxicity, or both. Combiningdrugswith on-off scheduling, such as sunitinib (4weeks on,2 weeks off is standard) and capecitabine (2 weeks on, 1week off), will require the evaluation of sequences differentfrom those used when the drugs are administered as singleagents. If drug–drug interactions affect exposure to eitherdrug or their metabolites, then applying common single-agent, on-off schedulesmay result in variable drug exposure(21). Preclinical studies could identify alternate schedulesthat result in more consistent drug exposures. Similarscheduling and sequencing issues may arise when intrave-nously administered drugs are combined with oral agents.Drug sequencing may also be a design issue when combin-ing drugs with pharmacokinetic or pharmacodynamicinteractions and short half-lives. Scheduling and dosingdecisions also affect toxicities, and novel schedules such asalternating administration of the drugsmay allow extendedadministration when concurrent administration is tootoxic (44).

Recommendation 4Selection of the clinical trial design should be based on the

scientific rationale, underlying data and hypothesis for the com-bination, and the intended development plan for the combination(recommendations 1–3). Recommendation 4 is not intended tobe prescriptive, but to provide pragmatic guidelines on the selec-tion of the phase I trial design for a drug combination.

No single standard exists for the phase I trial design forcombinations, but the design should address the specifichypothesis (recommendation 1) and subsequent devel-opment plans for the combination (recommendation 2).The classic DLT-driven cohort expansion design (3þ3)has most commonly been used (45). To date, mostcombinations have added one or more investigational

agents to a standard backbone already in clinical practice(34, 46–49). One challenge is the need to distinguish theincremental toxicity of the combination. Hamberg andcolleagues propose some solutions, including a 3þ3þ3design, which may reduce the chance of falsely declaringthat the maximum tolerated dose (MTD) has beenreached (when in fact observed DLTs might be due tothe standard therapy alone; ref. 21). Another alternative isto include controls in phase I trials, either an intrapatientcontrol (e.g., by introducing the novel agent after thestandard backbone has been started) or the randomiza-tion of patients to commence the combination up frontor in a staggered approach (21). Mathematical modelingcan also be used to refine dose escalation based on datathat emerge during the trial. A Bayesian approach hasbeen proposed to address background toxicity that ariseswhen a new agent is added to standard treatment (21).In addition, if the combination is intended to be devel-oped in several disease settings with different backbones,multiarm studies may also be more efficient than a seriesof separate single-agent trials (50–52). Furthermore, aphase I trial may fail to reflect optimal dose relationships.Thus, investigators may need to compare alternate dosesand schedules in subsequent trials (21). Although theremay be multiple appropriate phase I design options, theproposed design should be fully specified in the protocolas an algorithm, including any stage-wise decision rules,so that the statistical properties of the proposed phase Idesign can be evaluated under hypothesized outcomeprobability models.

Recommendation 4ACombination therapies with overlapping DLTs or a plausible

basis for a pharmacodynamic interaction leading to DLTs requireformal phase I evaluation. The selected doses to be studied shouldbe justified on the basis of the specific phase II plans.

Where overlapping DLTs exist or where pharmacody-namic interaction may be anticipated, a formal phase Idesign is required to evaluate toxicity, and to explore therecommended phase II doses and possibly the optimalschedule of the combination. The selection of doses to beevaluated should be justified on the basis of the preclin-ical data, nature of interaction of the study drugs, anddependence of the target for DLTs versus efficacy. Inaddition, the criteria for success and failure of a combi-nation should be defined. An inability to escalate one ormore drugs such that the combined dose would beexpected to have greater efficacy than the single-agentcounterparts may mean the combination is not suitablefor progressing to a phase II study. In addition, DLTsoccurring after the first cycle can prevent administration ofsubsequent cycles and should be factored into determi-nation of recommended phase II dosing. Finally, toxicitymay be unacceptable in the population for which thecombination is planned to be developed.

When considering dose escalation for both agents, amodel-based approach can be very helpful, particularlywhen it considers both toxicity and efficacy (53–55). When

Paller et al.




one of the agents hasmarkedly less single-agent activity or isbeing added primarily as a modulator, the agent withgreater single-agent activity should, in general, be main-tained at or near its single-agent dose (MTD), while grad-ually titrating the second agent.

Recommendation 4BCombinations without overlapping DLTs and without a plau-

sible basis for pharmacodynamic interaction, but with a plausiblepharmacokinetic interaction, should be studied using a formaldrug–drug interaction design. The primary endpoint is pharma-cokinetics. The crossover design is often optimal.In phase I trials in which pharmacokinetic interactions

are anticipated (e.g., drugs metabolized by the same path-way; ref. 56), a drug–drug interaction design should beconsidered. This would facilitate pharmacokinetic analysesfor both single agents and the combination. A commonmethod of assessing pharmacokinetic interaction usescrossover study designs that limit the number of patientsrequired and allow testing for different effects based upondifferent sequential ordering of the agents. Statistical anal-ysis using repeated measures over the same patientcan account for interpatient variability for some of theendpoints and thereby increases statistical power. Combi-nations involving drugs with long half-lives, such as vismo-degib, require phase I designs with pharmacokinetic wash-out periods (57). Crossover studies may use randomizeddesigns in which drug 1 is followed by drug 1 and drug 2, ordrug 1 and drug 2 are followed by drug 1, with a washoutperiod, with extensive pharmacokinetic sampling duringboth the single-agent and combination phases. An alterna-tive is the single sequence crossover design in which drug 1plus drug 2 always follows drug 1, or the reverse (58). Inboth of these designs, both interpatient and intrapatientvariability may need to be considered.Where pharmacokinetic interactions could lead to drug

accumulation, washout periods and/or low initial dosesshould be considered. The decision to escalate to the nextdose would be based on interim pharmacokinetic resultsin which the predefined dose-escalation rule might be "Ifthe drug A level at steady-state increases less than x% overthe previous level, and no DLT is present, then use dose B,otherwise use dose C." In situations in which CYP3A4interactions are expected, for example, this pharmacoki-netic-informed dose escalation method is appropriate.The hypothesis-driven development plan performed inrecommendation 2, in case of clinically significant phar-macokinetic interactions, must be based upon a reason-able assurance that any dose reductions in individualagents, necessitated by the need to maintain acceptabletoxicity, preserves the expectation of superior efficacy forthe combination (59).

Recommendation 4CCombinations without overlapping toxicities, without a plau-

sible basis for a pharmacodynamic interaction leading to a DLT,andwithout a plausible basis for a pharmacokinetic interaction donot require a formal phase I study. A pilot or safety run-in fortolerability can be conducted as an initial step of a phase II study.

If a combination regimen is not anticipated to haveoverlapping toxicity, pharmacodynamic or pharmacokinet-ic interaction, a formal phase I trial may not be required.Instead, a short pilot or safety run-in period may be under-taken as the initial part of a phase II trial to explore a limitednumber of dose levels (for example, clinicaltrials.gov iden-tifier NCT01839487 or NCT01708993), or to evaluate ananticipated recommended phase II dose directly beforeenrolling patients in the phase II portion of the trial. If thisapproach is chosen, the phase II design could include anearly safety interim analysis to ensure the regimen is toler-able (21). However, before determining that a formal phaseI trial is not required, consideration should be given toinvolving pharmacokinetics, pharmacodynamics, and datasafety experts.

SummaryThe CTD Task Force formulated recommendations for

the design of early-phase combination trials of anticanceragents based on consensus developed with members of theIDSC, the Task Force, and external experts. The selection ofthe proposed regimens is based on a biologic or pharma-cologic rationale supported by clinical and preclinical data,accompanied by a plan for subsequent development of thecombination trials. The potential pharmacokinetic andpharmacodynamic interactions, as well as overlapping toxi-cities, should be considered. Depending on the specifichypothesized interaction, the primary endpoint may bedose optimization, pharmacokinetics, and/or pharmaco-dynamics. The rationale and design of combination clinicaltrials should be carefully evaluated. Additional guidancemay be obtained by consulting with regulatory authorities(e.g., the FDA and/or the European Medicines Agency)before trial initiation. These recommendations comple-ment consensus guidelines on the design of phase I and IIclinical trials testing cancer therapeutics (18, 19), and werereviewed and formally approved by the IDSC.

AcknowledgmentsThe authors thank Amy Gravell.

Grant SupportP.A. Bradbury was supported by a Cancer Care Ontario Research Chair in

Experimental Therapeutics.

Received March 5, 2014; revised May 4, 2014; accepted May 9, 2014;published online August 14, 2014.

References1. Demetri GD, von Mehren M, Blanke CD, Van den Abbeele AD, Eisen-

berg B, Roberts PJ, et al. Efficacy and safety of imatinib mesylate inadvanced gastrointestinal stromal tumors. N Engl J Med 2002;347:472–80.





2. Druker BJ, Talpaz M, Resta DJ, Peng B, Buchdunger E, Ford JM, et al.Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosinekinase in chronic myeloid leukemia. N Engl J Med 2001;344:1031–7.

3. Mok TS, Wu YL, Thongprasert S, Yang CH, Chu DT, Saijo N, et al.Gefitinib or carboplatin–paclitaxel in pulmonary adenocarcinoma.N Engl J Med 2009;361:947–57.

4. Sosman JA, Kim KB, Schuchter L, Gonzalez R, Pavlick AC, Weber JS,et al. Survival in BRAF V600-mutant advanced melanoma treated withvemurafenib. N Engl J Med 2012;366:707–14.

5. Kwak EL, Bang YJ, Camidge DR, ShawAT, Solomon B,Maki RG, et al.Anaplastic lymphoma kinase inhibition in non–small cell lung cancer.N Engl J Med 2010;363:1693–703.

6. Kantarjian HM, Talpaz M, O'Brien S, Giles F, Garcia-Manero G, FaderlS, et al. Dose escalation of imatinib mesylate can overcome resistanceto standard-dose therapy in patients with chronic myelogenous leu-kemia. Blood 2003;101:473–5.

7. Kobayashi S, Boggon TJ, Dayaram T, Janne PA, Kocher O, MeyersonM, et al. EGFRmutation and resistance of non–small cell lung cancer togefitinib. N Engl J Med 2005;352:786–92.

8. Kosaka T, Yatabe Y, Endoh H, Yoshida K, Hida T, Tsuboi M, et al.Analysis of epidermal growth factor receptor genemutation in patientswith non–small cell lung cancer and acquired resistance to gefitinib.Clin Cancer Res 2006;12:5764–9.

9. Mehra R, Camidge DR, Sharma S, Felip E, Tan DS, Vansteenkiste JF,et al. First-in-human phase I study of the ALK inhibitor LDK378 inadvanced solid tumors. J Clin Oncol 2012;30 Suppl:174s, abstr 3007.

10. Soverini S, Martinelli G, Rosti G, Bassi S, Amabile M, Poerio A, et al.ABLmutations in late chronic phase chronicmyeloid leukemia patientswith up-front cytogenetic resistance to imatinib are associated with agreater likelihood of progression to blast crisis and shorter survival: astudy by the GIMEMA Working Party on Chronic Myeloid Leukemia.J Clin Oncol 2005;23:4100–9.

11. Chee KG, Longmate J, Quinn DI, Chatta G, Pinski J, Twardowski P,et al. The AKT inhibitor perifosine in biochemically recurrent prostatecancer: a phase II California/Pittsburgh Cancer Consortium trial. ClinGenitourin Cancer 2007;5:433–7.

12. Jonker DJ, O'Callaghan CJ, Karapetis CS, Zalcberg JR, Tu D, Au HJ,et al. Cetuximab for the treatment of colorectal cancer. N Engl J Med2007;357:2040–8.

13. Karapetis CS, Khambata-Ford S, Jonker DJ, O'Callaghan CJ, Tu D,Tebbutt NC, et al. K-ras mutations and benefit from cetuximab inadvanced colorectal cancer. N Engl J Med 2008;359:1757–65.

14. Krug LM, Kindler H, Calvert H, Manegold C, Tsao AS, Fennell D, et al.Vorinostat in patients with advanced malignant pleural mesotheliomawho have failed prior pemetrexed and either cisplatin or carboplatintherapy: a phase III, randomized, double-blind, placebo-controlledtrial. Eur J Cancer 2011; 47 Suppl 2:2–3.

15. Miller VA, Hirsh V, Cadranel J, Chen YM, Park K, Kim SW, et al. Afatinibversus placebo for patients with advanced, metastatic non–small celllung cancer after failure of erlotinib, gefitinib, or both, and one or twolines of chemotherapy (LUX-Lung 1): a phase 2b/3 randomised trial.Lancet Oncol 2012;13:528–38.

16. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, Tan EH, Hirsh V,Thongprasert S, et al. Erlotinib in previously treated non–small celllung cancer. N Engl J Med 2005;353:123–32.

17. Lorusso PM, Eder JP. Therapeutic potential of novel selective-spec-trum kinase inhibitors in oncology. Expert Opin Investig Drugs 2008;17:1013–28.

18. LoRusso PM, Boerner SA, Seymour L. An overview of the optimalplanning, design, and conduct of phase I studies of new therapeutics.Clin Cancer Res 2010;16:1710–8.

19. Seymour L, Ivy SP, Sargent D, Spriggs D, Baker L, Rubinstein L, et al.The design of phase II clinical trials testing cancer therapeutics:consensus recommendations from the clinical trial design task forceof the national cancer institute investigational drug steering commit-tee. Clin Cancer Res 2010;16:1764–9.

20. National Cancer Institute: Transforming the NCI Clinical Trials Enter-prise. Roster of Investigational Drug Steering Committee members.[Accessed 2012 March 25]. Available from: http://www.cancer.gov/aboutnci/organization/ccct/steering-committees/investigational-drug.

21. Hamberg P, Ratain MJ, Lesaffre E, Verweij J. Dose-escalation modelsfor combination phase I trials in oncology. Eur J Cancer 2010;46:2870–8.

22. Dancey JE, Chen HX. Strategies for optimizing combinations ofmolecularly targeted anticancer agents. Nat Rev Drug Discov 2006;5:649–59.

23. Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM,Schepartz S, et al. Relationships between drug activity in NCI preclin-ical in vitro and in vivomodels and early clinical trials. Br JCancer 2001;84:1424–31.

24. Voskoglou-Nomikos T, Pater JL, Seymour L. Clinical predictive valueof the in vitro cell line, human xenograft, andmouse allograft preclinicalcancer models. Clin Cancer Res 2003;9:4227–39.

25. Holbeck S, Collins JM, Doroshow JH. NCI-60 combination screeningmatrix of approved anticancer drugs. Eur J Cancer 2012;48:11.

26. Ocana A, Amir E, Yeung C, Seruga B, Tannock IF. How valid areclaims for synergy in published clinical studies? Ann Oncol 2012;23:2161–6.

27. Park SR, Davis M, Doroshow JH, Kummar S. Safety and feasibility oftargeted agent combinations in solid tumours. Nat Rev Clin Oncol2013;10:154–68.

28. Chou TC. Theoretical basis, experimental design, and computerizedsimulation of synergism and antagonism in drug combination studies.Pharmacol Rev 2006;58:621–81.

29. Gatzemeier U, Pluzanska A, Szczesna A, Kaukel E, Roubec J, DeRosa F, et al. Phase III study of erlotinib in combination withcisplatin and gemcitabine in advanced non–small cell lung cancer:the Tarceva Lung Cancer Investigation Trial. J Clin Oncol 2007;25:1545–52.

30. Giaccone G, Herbst RS, Manegold C, Scagliotti G, Rosell R, Miller V,et al. Gefitinib in combination with gemcitabine and cisplatin inadvanced non–small cell lung cancer: a phase III trial–INTACT 1.J Clin Oncol 2004;22:777–84.

31. MokTW,WuYL, Thongprasert S, YuCJ, Zhang L, LadreranGE, et al. Arandomized placebo-controlled phase III study of intercalated erlotinibwith gemcitabine/platinum in first-line advanced non–small cell lungcancer (NSCLC): FASTACT-II. J Clin Oncol 30, 2012 (suppl; abstr7519).

32. Begley CG, Ellis LM. Drug development: raise standards for preclinicalcancer research. Nature 2012;483:531–3.

33. Journal of Negative Results in BioMedicine. [Accessed 2014 Feb 10].Available from: www.jnrbm.com.

34. Flaherty KT, Infante JR, Daud A, Gonzalez R, Kefford RF, Sosman J,et al. Combined BRAF and MEK inhibition in melanoma with BRAFV600 mutations. N Engl J Med 2012;367:1694–703.

35. Azad NS, Posadas EM, Kwitkowski VE, Steinberg SM, Jain L, Annun-ziata CM, et al. Combination targeted therapy with sorafenib andbevacizumab results in enhanced toxicity and antitumor activity.J Clin Oncol 2008;26:3709–14.

36. Kummar S, Ji J, Morgan R, Lenz HJ, Puhalla SL, Belani CP, et al. Aphase I study of veliparib in combination with metronomic cyclophos-phamide in adults with refractory solid tumors and lymphomas. ClinCancer Res 2012;18:1726–34.

37. Gangadhar TC, Cohen EE,WuK, Janisch L, Geary D, KocherginskyM,et al. Two drug interaction studies of sirolimus in combination withsorafenib or sunitinib in patients with advanced malignancies. ClinCancer Res 2011;17:1956–63.

38. Lee JM, Sarosy GA, Annunziata CM, Azad N, Minasian L, Kotz H, et al.Combination therapy: intermittent sorafenib with bevacizumab yieldsactivity and decreased toxicity. Br J Cancer 2010;102:495–9.

39. Rini BI,Garcia JA,CooneyMM,ElsonP,Tyler A,BeattyK, et al. A phaseI study of sunitinib plus bevacizumab in advanced solid tumors. ClinCancer Res 2009;15:6277–83.

40. Wu K, House L, Ramirez J, Seminerio MJ, Ratain MJ. Evaluation ofutility of pharmacokinetic studies in phase I trials of two oncologydrugs. Clin Cancer Res 2013;19:6039–4.

41. Connolly RM, Rudek MA, Garrett-Mayer E, Jeter SC, Donehower MG,Wright LA, et al. Docetaxel metabolism is not altered by imatinib:findings from an early phase study in metastatic breast cancer. BreastCancer Res Treat 2011;127:153–62.


Paller et al.



42. Zhang X, Huang Y, Navarro MT, Hisoire G, Caulfield JP. A proof-of-concept and drug–drug interaction study of pamapimod, a novel p38MAP kinase inhibitor, with methotrexate in patients with rheumatoidarthritis. J Clin Pharmacol 2010;50:1031–8.

43. Hofmeister CC, Yang X, Pichiorri F, Chen P, Rozewski DM, JohnsonAJ, et al. Phase I trial of lenalidomide and CCI-779 in patients withrelapsed multiple myeloma: evidence for lenalidomide-CCI-779 inter-action via P-glycoprotein. J Clin Oncol 2011;29:3427–34.

44. RibasA,Hodi FS,CallahanM,KontoC,Wolchok J.Hepatotoxicitywithcombination of vemurafenib and ipilimumab. N Engl J Med 2013;368:1365–6.

45. Le Tourneau C, Lee JJ, Siu LL. Dose escalation methods in phase Icancer clinical trials. J Natl Cancer Inst 2009;101:708–20.

46. Bang YJ, Van Cutsem E, Feyereislova A, ChungHC, Shen L, Sawaki A,et al. Trastuzumab in combination with chemotherapy versus chemo-therapy alone for treatment of HER2-positive advanced gastric orgastro-oesophageal junction cancer (ToGA): a phase 3, open-label,randomised controlled trial. Lancet 2010;376:687–97.

47. Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, Hainsworth J,Heim W, et al. Bevacizumab plus irinotecan, fluorouracil, andleucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–42.

48. Pirker R, Pereira JR, Szczesna A, von Pawel J, Krzakowski M, RamlauR, et al. Cetuximab plus chemotherapy in patients with advanced non–small cell lung cancer (FLEX): an open-label randomised phase III trial.Lancet 2009;373:1525–31.

49. ScagliottiGV,NovelloS,Schiller JH,HirshV,Sequist LV, Soria JC, et al.Rationale and design of MARQUEE: a phase III, randomized, double-blind study of tivantinib plus erlotinib versus placebo plus erlotinib inpreviously treated patients with locally advanced or metastatic, non-squamous, non–small cell lung cancer. Clin Lung Cancer 2012;13:391–5.

50. Mita AC, Takimoto CH, Mita M, Tolcher A, Sankhala K, SarantopoulosJ, et al. Phase 1 study of AMG 386, a selective angiopoietin 1/2-neutralizing peptibody, in combination with chemotherapy in adultswith advanced solid tumors. Clin Cancer Res 2010;16:3044–56.

51. Goss G, Shepherd FA, Laurie S, Gauthier I, Leighl N, Chen E, et al. Aphase I and pharmacokinetic study of daily oral cediranib, aninhibitor of vascular endothelial growth factor tyrosine kinases, incombination with cisplatin and gemcitabine in patients withadvanced non–small cell lung cancer: a study of the NationalCancer Institute of Canada Clinical Trials Group. Eur J Cancer 2009;45:782–8.

52. Laurie SA, Gauthier I, Arnold A, Shepherd FA, Ellis PM, Chen E, et al.Phase I and pharmacokinetic study of daily oral AZD2171, an inhibitorof vascular endothelial growth factor tyrosine kinases, in combinationwith carboplatin and paclitaxel in patients with advanced non–small

cell lung cancer: the National Cancer Institute of Canada clinical trialsgroup. J Clin Oncol 2008;26:1871–8.

53. Harrington JA, Wheeler GM, Sweeting MJ, Mander AP, Jodrell DI.Adaptive designs for dual-agent phase I dose-escalation studies. NatRev Clin Oncol 2013;10:277–88.

54. Mandrekar SJ, Cui Y, Sargent DJ. An adaptive phase I design foridentifying a biologically optimal dose for dual agent drug combina-tions. Stat Med 2007;26:2317–30.

55. Thall PF, Cook JD. Dose-finding based on efficacy-toxicity trade-offs.Biometrics 2004;60:684–93.

56. Reardon DA, Vredenburgh JJ, Desjardins A, Peters KB, Sathornsu-metee S, Threatt S, et al. Phase 1 trial of dasatinib plus erlotinib inadults with recurrent malignant glioma. J Neurooncol 2012;108:499–506.

57. Rudin CM. Vismodegib. Clin Cancer Res 2012;18:3218–22.58. Lewis LD. Drug–drug interactions: is there an optimal way to study

them? Br J Clin Pharmacol 2010;70:781–3.59. ReardonDA,GrovesMD,WenPY,Nabors L,MikkelsenT,RosenfeldS,

et al. A Phase I/II Trial of pazopanib in combination with lapatinib inadult patients with relapsed malignant glioma. Clin Cancer Res 2013;19:900–8.

60. Sandler A, Gray R, Perry MC, Brahmer J, Schiller JH, Dowlati A, et al.Paclitaxel-carboplatin alone or with bevacizumab for non–small celllung cancer. N Engl J Med 2006;355:2542–50.

61. Wolchok JD, Kluger H, Callahan MK, Postow MA, Rizvi NA, LesokhinAM, et al. Nivolumab plus ipilimumab in advancedmelanoma. N Engl JMed 2013;369:122–33.

62. Boulay A, Rudloff J, Ye J, Zumstein-Mecker S, O'Reilly T, Evans DB,et al. Dual inhibition of mTOR and estrogen receptor signaling in vitroinduces cell death in models of breast cancer. Clin Cancer Res2005;11:5319–28.

63. Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO,et al. MET amplification leads to gefitinib resistance in lung cancer byactivating ERBB3 signaling. Science 2007;316:1039–43.

64. Li J, Zhao M, He P, Hidalgo M, Baker SD. Differential metabolism ofgefitinib and erlotinib by human cytochrome P450 enzymes. ClinCancer Res 2007;13:3731–7.

65. Padda SK,Chhatwani L, Zhou L, JacobsCD, Lopez-AnayaA,WakeleeHA. Phase I and pharmacokinetic study of bexarotene in combinationwith gefitinib in the third-line treatment of non–small cell lung cancer:brief report. Anticancer Drugs 2013;24:731–5.

66. Nilotinib: Highlights of Prescribing Information, 2008. [2013 March26]. Available from: http://dailymed.nlm.nih.gov/dailymed/archives/fdaDrugInfo.cfm?archiveid¼6775.

67. Haouala A, Widmer N, Duchosal MA, Montemurro M, Buclin T, Deco-sterd LA. Drug interactions with the tyrosine kinase inhibitors imatinib,dasatinib, and nilotinib. Blood 2011;117:e75–87.





CCR FOCUS
Guidelines for the Development and Incorporation ofBiomarker Studies in Early Clinical Trials of Novel Agents
Janet E. Dancey3, Kevin K. Dobbin3, Susan Groshen2, J. Milburn Jessup3, Andrew H. Hruszkewycz3, Maria Koehler4,Ralph Parchment5, Mark J. Ratain6, Lalitha K. Shankar3, Walter M. Stadler6, Lawrence D. True7, Amy Gravell8, andMichael R. Grever1 On behalf of the Biomarkers Task Force of the NCI Investigational Drug Steering Committee

Abstract

Authors' Aof SouthernBethesda,Frederick, IChicago, IllEmmes Co

Note: Janesearch, Tor

The contenpolicies ofmention ofendorseme

CorresponOntario InsCollege Str673-6532;

doi: 10.115

©2010 Am

www.aacr

The National Cancer Institute (NCI) Investigational Drug Steering Committee (IDSC) charged theBiomarker Task Force to develop recommendations to improve the decisions about incorporation ofbiomarker studies in early investigational drug trials. The Task Force members reviewed biomarkertrials, the peer-reviewed literature, NCI and U.S. Food and Drug Administration (FDA) guidance docu-ments, and conducted a survey of investigators to determine practices and challenges to executingbiomarker studies in clinical trials of new drugs in early development. This document provides stan-dard definitions and categories of biomarkers, and lists recommendations to sponsors and investiga-tors for biomarker incorporation into such trials. Our recommendations for sponsors focus on theidentification and prioritization of biomarkers and assays, the coordination of activities for the devel-opment and use of assays, and for operational activities. We also provide recommendations for in-vestigators developing clinical trials with biomarker studies for scientific rationale, assay criteria, trialdesign, and analysis. The incorporation of biomarker studies into early drug trials is complex. Thusthe decision to proceed with studies of biomarkers should be based on balancing the strength ofscience, assay robustness, feasibility, and resources with the burden of proper sample collection onthe patient and potential impact of the results on drug development. The Task Force provides theseguidelines in the hopes that improvements in biomarker studies will enhance the efficiency of inves-tigational drug development. Clin Cancer Res; 16(6); 1745–55. ©2010 AACR.

In an era of increasing number of identified cancer tar-gets and agents affecting those targets, efforts to determinetarget modulation, and to find subpopulations of patientsmost likely to benefit or experience harm from therapy, arepotentially valuable for rational prioritization of therapiesfor clinical research evaluation and for individual patientmanagement. Although nearly half of the recently ap-

ffiliations: 1Ohio State University, Columbus, Ohio; 2UniversityCalifornia, Los Angeles, California; 3National Cancer Institute,

Maryland; 4GlaxoSmithKline, Narberth, Pennsylvania; 5SAIC-nc., NCI-Frederick, Frederick, Maryland; 6University of Chicago,inois; 7University of Washington, Seattle, Washington; and 8Therporation, Rockville, Maryland

t E. Dancey, is currently at the Ontario Institute for Cancer Re-onto, Canada.

t of this publication does not necessarily reflect the views orthe Department of Health and Human Services, nor doestrade names, commercial products, or organizations implynt by the U.S. Government.

ding Author: Janet E. Dancey, High Impact Clinical Trials, Thetitute for Cancer Research, MaRS Centre, South Tower, 101eet, Suite 800, Toronto, ON Canada M5G 0A3. Phone: 416-Fax: 416-977-7446; E-mail: [email protected].

8/1078-0432.CCR-09-2167

erican Association for Cancer Research.

journals.org

proved oncology therapies have predictive markers, thequalification of putative biomarkers remains limited andthe practical realization of successful biomarker use in ear-ly clinical drug development remains to be more fully de-veloped. A recent extensive review of the use of biomarkerevaluations in early phase I trials of anticancer agentsshows the limited contributions resulting from these ef-forts (1). In addition to the expense and additional timerequired to incorporate these correlative studies, there arerisks to patients who agree to invasive procedures fortissue-based biomarker studies, thus raising ethical issuesif assay limitations are poorly understood or the utility ofthe study poorly justified.Although the potential for biomarker studies to affect

the clinical development strategy may be greatest in ap-propriately designed early trials by assisting in dose de-termination or identifying patients likely to benefit or atrisk for toxicity, practical measures to achieve the goalsare slow to appear, often uncoordinated and, at times,misdirected. Recognizing that biomarker studies havethe potential to improve the value of clinical trialsand can be integral to the design of a trial, but that thispotential has been largely unrealized, the National Can-cer Institute (NCI) Investigational Drug Steering Com-mittee (IDSC) created the Biomarker Task Force in

1745

Cl1746

CCR FOCUS
Table 1. Definitions
Term

in Cancer Res; 16(6) Marc

Definition E

h 15, 2010

xample

Biomarker nomenclature
Biomarker A characteristic that is objectively measured
and evaluated as an indicator of normal biologicprocesses, pathogenic processes, orpharmacologic responses to a therapeuticintervention (25)

Pharmacodynamic
Provide evidence that there is a directpharmacological effect of a drug
D
rug-related toxicity, modulation of target proteinphosphorylation, alteration of vascularpermeability, tumor response
Prognostic
Provide evidence about the patient's overalldisease outcome independent of any specificintervention.
P
erformance status, OncotypeDx test
Predictive
Provide evidence about the probability ofbenefit or toxicity from a specific intervention.
H
ER2 amplification and effectiveness oftrastuzumab; KRAS mutation andineffectiveness of cetuximab
Surrogate
Subsets of biomarkers that are intended to serveas a substitute for a clinically meaningfulendpoint (26)
P
rogression-free survival
Validation and qualification
Assay methodvalidity
Extent of the process by whichbiomarker assay characteristics are tested

In
cludes but not limited to analytic sensitivity,analytic specificity, precision, and inter- aswell as intrapatient variability undertypical clinical scenarios
Biomarkerqualification

The extent of the process by which a biomarkeris linked to a clinical significance.

Known validbiomarker*

A biomarker that is measured in an analytical testsystem with well-established performancecharacteristics and for which there is widespreadagreement in the medical or scientific communityabout the physiologic, toxicologic, or clinicalsignificance of the results

H
ercept(TM) test
Probable validbiomarker

A biomarker that is measured in an analytical testsystem with well-established performancecharacteristics and for which there is a scientificframework or body of evidence that appears toelucidate the physiologic, toxicologic, or clinicalsignificance of the results.

E
pidermal growth factor receptor mutationsin lung carcinoma
Exploratorybiomarker

A biomarker that does not meet the criteria forprobable or known valid biomarker

M
ost biomarkers in early development
Role within the clinical trial†

Integral role
Tests that are done for the trial to proceed;Integral studies are inherent in the design of thetrial from the onset and are done in real timefor the conduct of the trial
B
iomarker is used to determine eligibility or tostratify to different arms of the trial.
Of note, if integral markers are to be used tomake individual patient decisions, then CLIAregulations apply

P
hase I trial: the biomarker is used to decidewhether to dose escalate or expand a dose level.
http://wwwn.cdc.gov/clia/regs/toc.aspx P
hase II trial: the biomarker is used to decidewhether to continue enrollment or terminatepatient enrollment because of lack of activity inbiomarker-defined population.
(Continued on the following page)

Clinical Cancer Research

Guidelines for Biomarker Studies in Early Trials

2007 to develop recommendations for the conduct ofbiomarker studies in early clinical trials. Although ini-tially developed for NCI Cancer Therapy Evaluation Pro-gram (CTEP)-sponsored studies, the recommendationsare applicable to early investigational drug studies con-ducted by academic investigators for NCI or other spon-sors. Background on the IDSC, the Biomarker TaskForce, and the development of the recommendationsis provided in Text Box 1.This document is divided into five sections. It begins

with standard definitions and categories of biomarkersbased on intended use and level of validation and clinicalqualification. Subsequently, recommendations to spon-sors and investigators for biomarker incorporation intosuch trials are provided. Our recommendations for spon-sors focus on the identification and prioritization of bio-markers and assays, the coordination of activities for thedevelopment and use of assays, and for operational activ-ities. We also provide recommendations for investigatorsdeveloping clinical trials with biomarker studies for scien-tific rationale, assay criteria, trial design, and analysis. Thefinal sections provide commentary on the trial design andanalytical issues that arise with biomarker studies in earlyclinical trials.

Definitions

The Biomarker Task Force recommendations use specificnomenclature, on the basis of a number of recent manu-scripts and recommendations to define and describe spe-cific types of biomarkers, assay validation and clinicalqualification, and the role the biomarker may have withina clinical trial. The terms, definitions, and examples arefound in Table 1.

www.aacrjournals.org

Recommendations for biomarker studies insupport of new therapeutic agents

Biomarker studies in early phase trials require a care-ful coordination of expertise, infrastructure, and fundingresources as well as a careful consideration of scientificrationale, assay characteristics, and trial design Text Box2. On the basis of the occasional successful integrationof biomarker studies in early phase clinical trials and themany examples of less successful ventures in this arena,the following recommendations that are specific forNCI supported trials but modifiable for other applications,are made:

Recommendations to the sponsorRecommendations for the identification and prioritization

of biomarkers and assays

A review of potential biomarkers by an expert or expertpanel should be completed prior to a solicitation forclinical trial proposals of a novel agent. Preferred bio-markers and assays should be prioritized on the basisof the availability of well-characterized assays in hu-mans and/or human specimens as well as on putativemechanism-based considerations.

Investigational agents should be more readily andrapidly available to aid assay development.

Biomarker studies conducted in conjunction withpharmacokinetic studies should be encouraged.

Identification and prioritization of biomarkers thatmight inform clinical evaluations of a novel therapeuticshould be based on strength of science supporting themechanism of action of the agent and the determinantsof its activity or toxicity, as well as the robustness of

Table 1. Definitions (Cont'd)

Term
Definition E xample
Integrated role
Tests are intended to identify or validate assaysor markers that are planned for use in futuretrials. Trials are designed to test a hypothesisand include complete plans for specimencollection, laboratory measurements, andanalysis. Statistical design and analysis shouldbe prespecified.
P

P

hase I trial: the biomarker is studied at the MTDor at selected doses to confirm effect on target.hase II trial: predictive marker is measured on allcases but the result is not used for eligibility,treatment assignment, or treatment managementin the current trial. Statistical design and samplesize are prespecified for the marker analysis.

Ancillary and/orexploratory role

Trial data are used to develop biomarkers and/orassays or to better understand therapeuticagent potential; biomarker data are notfundamental to the successful completion ofthe phase I or II trial.

R
etrospective biomarker assays, pilot or feasibilitybiopsies, or methodological assessment;exploratory and/or hypothesis generatinganalyses.
*http://www.fda.gov/downloads/RegulatoryInformation/Guidances/ucm126957.pdf.†http://restructuringtrials.cancer.gov/files/BIQSFP_Announcement_12_12_08.pdf.

Clin Cancer Res; 16(6) March 15, 2010 1747

Cl1748

CCR FOCUS
Table 2. Analytical and performance requirements of laboratory assays of biomarkers with integral orintegrated roles in clinical trials
Types of laboratory issues

Clinical trial useof the assay

in Cancer Res; 16(6) M

Primary drug effect assay

arch 15, 2010

Secondary drug effect assay
Integral diagnostic assay PD assay PD, predictive, or response assay
Role and/orpurposeof the assay

Show a directpharmacological orbiological effect of a drugor prove feasibility of a newassay for an establisheddrug effect

Associate a pharmacologicalor biological effect of a drugto a known therapeutic outcome

Clinical decision-making, e.g.,patient eligibility, stratification,or assignment to treatment

May be exploratory biomarkerand serve ancillary role

May be exploratory biomarkerand serve ancillary role

May be exploratory biomarkerbut serves an integral roleand needs to be done in aCLIA-certifiedlaboratory

Assay clinical readiness

Validation of assay analytical performance (using reference and calibration standards)

Accuracy,precision

Required
Required Required
Quantitationof assayendpoint

Required
Required Required Spike-recovery and dilution
linearity using intendedclinical matrix or linearresponse with appropriatecalibrators for IHC-ISH ornonliquid-based assays

Spike-recovery and dilutionlinearity using intendedclinical matrix or linearresponse with appropriatecalibrators for IHC-ISH ornonliquid-based assays

Spike-recovery and dilutionlinearity using intendedclinical matrix or linearresponse with appropriatecalibrators for IHC-ISH ornonliquid-based assays

Specificity
Required Required Required Assay results parallel
modulation of analyteconcentration

Assay results parallel modulationof analyte concentration

Assay results parallel modulationof analyte concentration

Robustness
Required Not required Required Assay performance is
independent of operatorand based on a SOP

Assay performance isindependent of operator andbased on a SOP

Specimen acquisition, processing, and storage SOP that yields valid assay results

Type of specimen
Using intended specimentypes from preclinicalmodels and from relevantclinical donors
Using intended specimentypes from preclinical modelsor from relevant clinical donors

Using intended types ofspecimens from relevantclinical donors

Specimen storageprocedure

Required
Required Required Quantify how storage time
and conditions of intendedspecimen types affectassay measurements

Quantify how much storagetime and conditions ofintended specimen typesaffect assay measurements

Quantify how much storagetime and conditions ofintended specimen typesaffect assay measurements.

Clinical readiness of combining specimen and assay SOPS (modeled using intended specimen types from preclinical modelsand/or intended specimen types from clinical donor protocols)

Dosediscrimination

Required
Required, limited to detectchanges in assay results from
Required
Proven ability to detect Proven ability to detect changes
(Continued on the following page)



available technologies and assays to measure candidatebiomarkers in humans or human tissues. Ideally, pro-posed biomarkers should have strong preliminary datalinking biomarker changes to agent effects, a well-definedassay and interpretation suitable for scientific and clinicalinference, and a clearly defined process for specimencollection (including timing and method of sampleprocurement) and processing with feasibility data. Ifperformance of the assay will be delayed, stability ofthe samples must be assured. In general, early access tothe investigational agent is needed for laboratory-basedexperiments to accomplish these goals. Such experiments


should identify candidate biomarkers that reflect themechanism of action, toxicity, pharmacokinetics, and ac-tivity of the therapeutic agent. The experiments shouldalso analytically validate assays to measure these markersand correlate the marker with dose or exposures as wellas the aforementioned details about sample collectionand storage (2–4).Recommendations for the coordination of activitiesDialogue between the sponsor and institutional inves-tigators is encouraged to identify the best assaysand opportunities for biomarker developmentand use.

Table 2. Analytical and performance requirements of laboratory assays of biomarkers with integral orintegrated roles in clinical trials (Cont'd)

Types of laboratory issues

Clinical trial useof the assay

Primary drug effect assay
Secondary drug effect assay
Cli

Integral diagnostic assay
PD assay PD, predictive, or response assay
baseline when the assay error,patient biological variability, andclinical trial design areconsidered

changes in assay resultsneeded by protocol whenthe assay error, patientbiological variability, andclinical trial design areconsidered

in assay results needed byprotocol when the assay error,patient biological variability,and clinical trial design areconsidered

LLOQ, ULOQ, anddynamic range

Required
Required Required Dynamic range must cover
planned dose or exposurelevels, and desired drugeffect level

Dynamic range mustcover planned drug effectlevel

Dynamic range must cover theclinically expected levels ofresponse or expression

Absence ofpre-analyticvariables

Required
Required Required Consistency of baseline
values; understandingof influence of plannedprior sampling onbiological variability inthe assay

Consistency of baseline values;understanding of influenceof planned prior sampling onbiological variability in theassay

Consistency of baseline values;understanding of influence ofplanned prior sampling onbiological variability in theassay

Timing of specimencollection to hitassay dynamicrange andsensitivity

Required
Required Required Experimental evidence to
guide timing of collection
Experimental evidence to
guide timing of collection
Experimental (early phase trials)
or clinical (late phase trials)evidence to guide timing ofcollection

Need forCLIA-certifiedlaboratory

No
No Yes
NOTE: This table describes laboratory assays of biomarker issues that determine “clinical readiness.” The concept of “clinical read-iness” combines proving analytical performance of a validated assay and fit-for-purpose when replicating the clinical situation inwhich the assay is intended to be used. It is recommended that the type, purpose, assay analytical performance, and clinicalreadiness as outlined in this table be described in proposals or protocols to sponsors. Although there are no mandated standardson analytical performance or “clinical readiness” of an assay prior to trial initiation, the greater the rigor of assay optimization andimplementation, the greater the potential scientific value of biomarker data derived from the clinical trials. Only “integral” biomarkertests need to be done in a CLIA-certified laboratory; however Good Laboratory Practice (27) standards as well as the performanceand reporting standards as outlined in STARD and REMARK (28–31) should be considered for nonintegral biomarkers.

n Cancer Res; 16(6) March 15, 2010 1749

1750

CCR FOCUS
Collaboration between institutions should be en-couraged and central reference laboratories shouldbe considered to standardize assay methodologies, re-agents, and calibrators.
Clinical trial proposals should not be penalized for lackof biomarker studies and instead sponsors shouldfacilitate collaboration with institutions and investi-gators with applicable biomarker capabilities.

The number of potential biomarkers that might beevaluated for a specific agent and the expertise and re-sources required for assay development are significant.Such expertise and resources are available within NCI'sDivision of Cancer Treatment and Diagnosis (DCTD),industry, and at select institutions. Collaborations be-tween stakeholders for the selection, evaluation, and op-timization for biomarkers and assays could facilitate theeffective use of the resources and expertise. The use ofstandardized assays at reference laboratories would as-sure a greater degree of consistency across studies, and

Clin Cancer Res; 16(6) March 15, 2010

avoid unnecessary duplication of efforts to develop andevaluate assays at individual sites. The centralization oflaboratory activities also allows the rapid accrual to well-designed clinical trials at sites that lack in-house re-sources to run biomarker assays but have the capabilitiesfor sample collection and processing.Recommendations for operational activities

▪ Funding should be made available for resources that canform the foundation for biomarker-based research withspecial attention to:
(1) investigational imaging procedure;(2) collection and storage of biospecimens that can be secured at low cost and low patient risk.To date, most biomarker studies have assessed pharma-
codynamic effects or sought prognostic or predictive mar-kers in tumor specimens even if the state of science, assaymethodology, or trial design were insufficient to lead to in-formative results. Biomarkers that can be measured nonin-vasively through functional imaging studies or on normal



tissues or blood may be preferable to subjecting patients toadditional biopsies. Although biomarker studies of readilyaccessible normal tissue are encouraged, the results of suchstudies should be used with caution in consideration of


decisions about dosing and further development, unlessthere are convincing preclinical or early phase trial datarelating such biomarkers to antitumor activity. Additionalresources to support these alternatives are required.


1752

CCR FOCUS
Scientific and design recommendations for theinvestigatorTo comprehensively review proposals for clinical trials
with biomarkers, a clear description of the scientificrationale and supporting data for the proposed bio-marker, assay as well as the study design, and analysisare needed.Scientific recommendations. Proposals should provide:

A hypothesis for biomarker utility and a description ofthe impact on therapeutic agent development basedon the following considerations: (1) biological and/or mechanistic rationale with data to support relation-ship between biomarker and agent effects; (2)intended use within the proposed study (refer to clas-sification, descriptions, and roles as described inTable 1); (3) preclinical in vitro, in vivo, and clinicalresults if available.

The assay method validity and appropriateness for thestudy (see Table 2).

The investigator's experience and competence with theproposed assays.

The data supporting the degree of biomarker “fit for pur-pose” and clinical qualification.

To evaluate the clinical trial and prioritize the biomarkerstudies requires information on the likelihood that thebiomarker is (1) relevant to the trial population and inter-vention; (2) can be measured within the study populationand the study samples; and (3) that the assay is reliableand thus the results are likely to be correct and, ideally,contribute meaningful scientific knowledge to the devel-opment of the agent. The trial proposal should thus con-tain the purpose of incorporating a biomarker into aclinical trial and a strong scientific hypothesis that is basedon data. The proposed contribution of these measure-ments to the development of the anticancer agent shouldalso be explained. The expertise and practical experience ofthe investigator in working with human samples relevantto the proposed assay should be clearly documented. If the“fit-for-purpose” principles outlined by Lee (5), Wagner(6, 7), DeSilva (8), and their colleagues are followed, assayperformance within their anticipated clinical trial contextwill be known (9). Only when the analytical performanceof the assay within its clinical study context is character-ized will the results of the assay on trial specimens be in-terpreted with confidence.Trial design recommendations

Justification of numbers of patients and specimens: (1)to show feasibility; and (2) to show that studiesare likely to produce interpretable and meaningfulresults.

Consideration about the risk to the patient of obtainingsamples, specimens, or data for biomarker studies inthe context of data on biomarker validity and degreeof clinical qualification.

Biomarker correlations are frequently done on resultsobtained from the analysis of samples from a subset of pa-tients. In addition, biomarker results often classify patients


into subsets that are then correlated with patient out-comes. In both cases, sample sizes in early clinical trialsare generally too small to generate statistically robust con-clusions, although results may still be informative. It is im-portant, however, to consider not only how manyspecimens are likely to be obtained and analyzed but alsowhether the data will yield a meaningful conclusion thatjustifies the potential risk to patients, as well as the addi-tional resources and expense required to do the study and/or studies. If the biomarker specimen is to be obtainedfrom patients using invasive procedures, there shouldbe some justification of the potential risks associated withthe procedures (10, 11).

Trial Design and Statistical Considerations

The incorporation of biomarkers in early clinical trialsshould be based on specific biomarker hypotheses, avail-ability of an adequate assay, and appropriate design ofthe study. Although the specifics may vary, the general es-tablished objectives for phase I and II trials remain thesame. The phase I trial primary goal is to establish safetyand to select one or more doses and schedules for furtherstudy. The phase II trial explores one or a small numberof doses and/or schedules for safety and determineswhether the tested agent or regimen has sufficient evidenceof clinical activity to warrant further investigation. In theseearly trials, patient numbers are generally limited, and inphase I studies, patients are generally heterogeneousin terms of tumor histology and prior treatments. Bio-markers tested in early phase trials are often exploratory(Table 1), i.e., to collect preliminary data that may aid inthe understanding of the agent and mechanisms, andassays may not be well characterized. These considerationsmay limit the types of biomarker hypotheses that can beevaluated and the strength of the conclusions that can bedrawn.

Biomarkers in phase I trialsIn phase I studies, pharmacodynamic biomarkers are of-

ten of interest based on assumptions that modulation ofthese markers may provide proof of drug target inhibitionand support the selection of drug and dose for furtherevaluation. These are almost always exploratory biomar-kers (Table 1). The design and statistical discussion thatfollows assumes that (1) the proposed biomarker is rele-vant for the drug's purported mechanism of action; (2) theassays have been established for the types of samples thatwill be studied; (3) the correct timing of the sampling hasbeen established; (4) if normal tissues are to be used, thatthere is a correlation between changes in the normal tissuewith changes in the tumor tissue; and (5) that the acqui-sition of tissue appropriately considers the risks and/or in-convenience to the patient and resources of the institution.In early drug trials, these ideal conditions may not beachieved prior to the initiation of the trial. However, pre-clinical pharmacokinetic-pharmacodynamic-drug activitymodeling and laboratory assay analytical evaluation can


(www.clsi.org)


provide supportive information for the initial selection ofbiomarker(s), sampling, and analyses for early studies.The goal in these pharmacodynamic marker studies

(e.g., changes in tumors noted during tumor imaging, mo-lecular changes in biospecimen samples) in phase I trialsis to provide evidence that the agent reaches or modulatesthe putative target. These studies can be conducted byanalysis of samples and/or images obtained prior to andafter treatment, or by comparison to an untreated control.Samples may be collected and analyzed from patientstreated at the recommended phase II dose (RP2D) orat several dose levels to construct a preliminary dose-pharmacodynamic response curve.The required degree of scientific and analytical valida-

tion of a biomarker and assay depends on whether thebiomarkers are exploratory or to be used to make deci-sions within the trial. If a biomarker has an integral role(Table 1) to guide the dose escalation decisions for subse-quent patients in a phase I trial, key challenges are deter-mining magnitude of biomarker effect related to drug;detecting the effect within patients or samples indepen-dent of the variability because of specimen or assay perfor-mance; and establishing the methodology for completingthe assay within a short turnaround time. In addition,Clinical Laboratory Improvement Amendments (CLIA)regulations imply that if the test results are used for med-ical decisions for the patient who has the test, the testshould be done in a CLIA facility. The rules for decidingwhether to dose escalate, expand, or de-escalate trial de-sign will reflect the specifics of the agent tested, the bio-marker characteristics, and whether toxicity or otherparameters will also be used to guide the decisions. Thereis no standard or established design, although several ap-proaches have been proposed (12, 13). If a biomarker willplay a nonintegral role (i.e., is an exploratory biomarkerand has an ancillary role; Table 1) in a phase I trial, thenumber of patients in the expanded cohort at the MTDor RP2D may be based on estimating the patient-to-pa-tient variability. When the biomarker levels are approxi-mately normally distributed, the number of patients willdetermine the precision with which the true standard de-viation can be estimated (14).

Biomarkers in phase II trialsIn phase II trials biomarkers can be used to provide ev-

idence that the agent modulates the putative target orpathway in a pharmacodynamic assessment similar tothe phase I setting or to evaluate the association betweenthe biomarker and clinical outcome. Less commonly, bio-markers may be used to determine patient eligibility (forexample, HER2 status for trastuzumab trials). Larger pa-tient numbers in phase II allow the determination of thedose-response relationship of a pharmacodynamic markeracross a narrow set of dose cohorts (generally one or two)and more homogenous patient population. However,phase II trials to show correlations between clinical out-comes with an investigational agent and biomarkers pres-ent at baseline, or changes in markers prior and on


treatment, are difficult to design rigorously; there mustbe a sufficient number of patients with the clinical andbiomarker outcomes of interest. When markers are usedfor medical decision-making such as assignment to thera-py or stratification then the assays must be done in a CLIA-certified laboratory. Although a phase II trial is unlikely todefinitively establish whether a marker can be used topredict clinical benefit, investigators may identify an asso-ciation that can then be further explored in definitivephase III trials (15).

Pre-analytical and Analytical Considerations

The precision and accuracy of biomarker measurements,the standardization of specimen or sample collection, stor-age, and analysis are key to the successful interpretation ofclinical trial results. Although exploratory biomarkers maybe used in an early clinical trial, it is preferable that bothpre-analytical and analytic issues be addressed before clin-ical trials are initiated to ensure valid scientific inference oftrial results.Analytic measurements can be confounded by biologic

and pharmacologic variations among study participants,and by variations in specimen collection and processing.Reliability of biomarker values is dependent on the anal-ytic validation of biomarker assays. Although there arecurrently no mandated standards for ensuring methodo-logical quality of biomarker measurements for clinical re-search, there are reference sources that provide specificrequirements for quality standards for biomarker testsused for clinical laboratory and imaging diagnostic tests.This is exemplified by the CLIA 1988 standards definedby Clinical and Laboratory Standards Institute8 and anumber of Food and Drug Administration (FDA) and In-ternational Committee on Harmonization (ICH) guidancedocuments (16, 17). Basic analytic validation of biomar-kers includes availability of standard reference materials,calibration of equipment, reagent checks, optimizationof test protocols, acceptance criteria for standard curves,and assessment of matrix effects and interferences. Qualitycontrol measures include determinations of intra-assayand interassay precision and establishment of criteria ofacceptability for results obtained with quality control sam-ples at high and low concentrations. Similar principles ap-ply to the evaluation of imaging assessment modalities(18–20). If samples are collected and stored for futureuse, additional resources and activities are needed suchas a sample repository and database with careful defini-tions of data fields and nomenclature, and quality controlprocedures to ensure the integrity of the sample repositoryand data (21, 22).A number of resources are available to assist investiga-

tors in their considerations of imaging- and biospecimen-based biomarkers in early trials of therapeutics. TheNCI Translational Research Working Group has outlined

8


1754

CCR FOCUS
developmental pathways for both biospecimen and imag-ing assessment modalities (23, 24). Table 2 outlines theconsiderations for integral pharmacodynamic marker as-says developed by DCTD Clinical Pharmacodynamic Pro-gram that have been adapted to describe analytical issuesfor biomarkers that have integral, integrated, or exploratoryroles. The NCI Office of Biorepositories and BiospecimenResearch provides guidelines for specimen collection andpreservation, which are available through its website (2).Although there are no mandated standards, the greaterthe rigor of laboratory and imaging practices for biomarkermeasurements and sample repositories linked to clinicaltrials, the greater the potential scientific value of biomarkerdata derived in the course of clinical trials. Support of assayvalidation, specimen repositories, and accompanying data-bases (across phase I and II trials) is encouraged.
Concluding Remarks

Biomarker studies have the potential to contribute toour understanding of human cancer biology and optimalpatient treatment. Biomarkers in early phase studies cantheoretically be used to enhance development of noveltherapeutics, and occasionally to prematurely terminatethe investigation of a new drug. However, to realize thispotential requires careful consideration of study hypothe-sis, biomarker, technologies, assay development and per-formance, and patient and sample selection in the designof trials of novel agents that incorporate biomarkers.Prioritization based on solid science and the needs ofpatients, better coordination of biomarker and assay de-velopment evaluation, standardized tools and procedures,and improved operational efficiency and resources are re-quired. In early drug trials, these ideal conditions may notbe achievable without considerable coordination of activ-ities for biomarker development prior to initiation of the


clinical trials. The Biomarker Task Force submits these re-commendations in a genuine effort to enhance the utilityof these studies to cancer drug development. Our hope isto encourage the next generation of scientists to fully ex-plore and optimize the use of laboratory and imaging cor-relates to expedite the identification of exciting novelagents. Perhaps the most important consideration for in-vestigators and reviewers is to clearly understand the goalof incorporating a specific biomarker into a clinical trialand to balance this goal with the risk of sample procure-ment to the patient. Cost and time implications for devel-oping and including biomarkers within a clinical trial mayoutweigh gain. Thus, the decision to proceed with bio-marker studies within early clinical trials should be basedon the strength of the scientific rationale, the feasibility ofsuccessfully conducting the studies in the clinical trial, andthe potential impact of the results of such studies on drugdevelopment and lives of patients.

Disclosure of Potential Conflicts of Interest

M. Koehler, employment, stock holder, Pfizer.

Acknowledgments

Biomarkers Task Force of the NCI Investigational Drug SteeringCommittee members: Richard Bold, Deborah Collyar, David Gandara,Michael Maitland (Angiogenesis Task Force liaison), Mary Relling, LarryRubinstein, and Joseph Tomaszewski.

NCI Investigational Drug Steering Committee cochair: CharlesErlichman.

Grant Support

This project has been funded in whole or in part by federal funds fromthe National Cancer Institute, National Institutes of Health, under ContractNo. HHSN261200800001E.

Received 11/02/2009; revised 01/11/2010; accepted 01/18/2010;published OnlineFirst 03/09/2010.

References
1. Parulekar WR, Eisenhauer EA. Phase I trial design for solid tumor
studies of targeted, non-cytotoxic agents: Theory and practice.J Natl Cancer Inst 2004;96:990–7.

2. NCI Best Practices for Biospecimen Resources. Office of Bioreposi-tories and Biospecimen Research, National Cancer Institute, Bethes-da (MD); June 2007. Available from: http://biospecimens.cancer.gov/bestpractices/.

3. Holland NT, Smith MT, Eskenazi B, Bastaki M. Biological sample col-lection and processing for molecular epidemiological studies. MutatRes 2003;543:217–34.

4. International Society for Biological and Environmental Repositories(ISBER). Best practices for repositories I: Collection, storage, andretrieval of human biological materials for research. Cell PreservTechnol 2005;3:5–48.

5. Lee JW, Hall M. Method validation of protein biomarkers in supportof drug development or clinical diagnosis/prognosis. J Chromatogr BAnalyt Technol Biomed Life Sci 2009;877:1259–71.

6. Wagner JA. Strategic approach to fit-for-purpose biomarkers in drugdevelopment. Annu Rev Pharmacol Toxicol 2008;48:631–51.

7. Wagner JA, Williams SA, Webster CJ. Biomarkers and surrogate endpoints for fit-for-purpose development and regulatory evaluation ofnew drugs. Clin Pharmacol Ther 2007;81:104–7.

8. Ray CA, Patel V, Shih J, et al. Application of multi-factorial design ofexperiments to successfully optimize immunoassays for robustmeasurements of therapeutic proteins. J Pharm Biomed Anal 2009;49:311–8.

9. Smolec J, DeSilva B, Smith W, et al. Bioanalytical method validationfor macromolecules in support of pharmacokinetic studies. PharmRes 2005;22:1425–31.

10. Chau CH, Rixe O, McLeod H, Figg WD. Validation of analyticmethods for biomarkers used in drug development. Clin CancerRes 2008;14:5967–76.

11. Taylor JM, Ankerst DP, Andridge RR. Validation of biomarker-basedrisk prediction models. Clin Cancer Res 2008;14:5977–83.

12. Bekele BN, Shen Y. A Bayesian approach to jointly modelingtoxicity and biomarker expression in a phase I/II dose-finding trial.Biometrics 2005;61:343–54.

13. Hunsberger S, Rubinstein LV, Dancey J, Korn EL. Dose escalationtrial designs based on a molecularly targeted endpoint. Stat Med2005;24:2171–81.

14. Gehan EA. The determination of the number of patients required in apreliminary and a follow-up trial of a new chemotherapeutic agent.J Chronic Dis 1961;13:346–53.

15. Hahn OM, Yang C, Medved M, et al. Dynamic contrast-enhanced



magnetic resonance imaging pharmacodynamic biomarker study ofsorafenib inmetastatic renal carcinoma. JClin Oncol 2008;26:4572–8.

16. International Conference on Harmonisation. Validation of AnalyticalProcedures: Text and Methodology, Harmonised Tripartite Guideline,Document Q2(R1), Current Step 4 version. Fed Regist 1995;60:11260–2, Available from: http://www.ich.org/LOB/media/MEDIA417.pdf.

17. Guidance for Industry. Bioanalytical Method Validation. Center for Vet-erinary Medicine (CVM) and Center for Drug Evaluation and Research(CDER), Food and Drug Administration, U.S. Dept of Health & HumanServices, Rockville (MD); May 2001. Available from: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM070107.pdf.

18. Recommendations for Magnetic Resonance (MR) MeasurementMethods at 1.5-Tesla and Endpoints for Use in Phase 1/2a Trials ofAnti-Cancer Therapeutics Affecting Tumor Vascular Function. In:NCI Cancer Imaging Program MR Workshop on Translational Re-search in Cancer - Tumor Response, Bethesda (MD); November 22–23, 2004. Available from: http://imaging.cancer.gov/clinicaltrials/guidelines.

19. Recommendations for Magnetic Resonance Spectroscopy (MRS)Methods and Endpoints for Use in Multicenter Trials of Anti-CancerTherapeutics. In: NCI Cancer Imaging Program MR Workshop onTranslational Research in Cancer - Tumor Response, Bethesda(MD); November 22–23, 2004. Available from: http://imaging.can-cer.gov/clinicaltrials/guidelines.

20. Shankar LK, Hoffman JM, Bacharach S, et al. Consensus recom-mendations for the use of 18F-FDG PET as an indicator of therapeu-tic response in patients in National Cancer Institute Trials. J NuclMed 2006;47:1059–66.

21. Merz JF, Sankar P, Taube SE, Livolsi V. Use of human tissues in re-search: clarifying clinician and researcher roles and informationflows. J Investig Med 1997;45:252–7.

22. Stoneburner G, Goguen A, Feringa A. Risk Management Guidefor Information Technology Systems [online]. NIST Special


Publication 800–30, National Institute of Standards and Technology,Gaithersburg (MD). 2002. Available from: http://csrc.nist.gov/publications/nistpubs/800–30/sp800–30.pdf.

23. Dorfman GS, Sullivan DC, Schnall MD, Matrisian LM. The Transla-tional Research Working Group developmental pathway for image-based assessment modalities. Clin Cancer Res 2008;14:5678–84.

24. Srivastava S, Gray JW, Reid BJ, et al. Translational Research Work-ing Group developmental pathway for biospecimen-based assess-ment modalities. Clin Cancer Res 2008;14:5672–7.

25. Biomarkers Definitions Working Group. Biomarkers and surrogateendpoints: Preferred definitions and conceptual framework. ClinPharmacol Ther 2001;69:89–95.

26. Lesko LJ, Atkinson AJ, Jr. Use of biomarkers and surrogate end-points in drug development and regulatory decision making: Criteria,validation, strategies. Annu Rev Pharmacol Toxicol 2001;41:347–66.

27. Ezzelle J, Rodriguez-Chavez IR, Darden JM, et al. Guidelines ongood clinical laboratory practice: Bridging operations between re-search and clinical research laboratories. J Pharm Biomed Anal2008;46:18–29.

28. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete andaccurate reporting of studies of diagnostic accuracy: The STARDinitiative. Standards for Reporting of Diagnostic Accuracy. Clin Chem2003;49:1–6.

29. Bossuyt PM, Reitsma JB, Bruns DE, et al. The STARD statement forreporting studies of diagnostic accuracy: Explanation and elabora-tion. Ann Intern Med 2003;138:W1–12.

30. Lumbreras B, Porta M, Marquez S, et al. QUADOMICS: An adapta-tion of the Quality Assessment of Diagnostic Accuracy Assessment(QUADAS) for the evaluation of the methodological quality of studieson the diagnostic accuracy of '-omics'-based technologies. ClinBiochem 2008;41:1316–25.

31. McShane LM, Altman DG, Sauerbrei W, et al. Reporting recommen-dations for tumor marker prognostic studies (REMARK). J NatlCancer Inst 2005;97:1180–4.


1764

CCR FOCUS
The Design of Phase II Clinical Trials Testing CancerTherapeutics: Consensus Recommendations from the ClinicalTrial Design Task Force of the National Cancer InstituteInvestigational Drug Steering Committee
Lesley Seymour1, S. Percy Ivy2, Daniel Sargent3, David Spriggs4, Laurence Baker5, Larry Rubinstein2,Mark J. Ratain6, Michael Le Blanc7, David Stewart8, John Crowley7, Susan Groshen9,Jeffrey S. Humphrey10, Pamela West11, and Donald Berry8

Abstract

uthors' Aanada; 2Nivision ofemorial

ork, Newharmacoepartmenlinois; 7Frnd Biostanderson Corris Comos Angelersey; ande Clinical

orrespon0 Stuart Sax: 613-53

oi: 10.115

2010 Am

Clin Canc

The optimal design of phase II studies continues to be the subject of vigorous debate, especiallystudies of newer molecularly targeted agents. The observations that many new therapeutics “fail” indefinitive phase III studies, coupled with the numbers of new agents to be tested as well as theincreasing costs and complexity of clinical trials, further emphasize the critical importance of robustand efficient phase II design. The Clinical Trial Design Task Force (CTD-TF) of the National CancerInstitute (NCI) Investigational Drug Steering Committee (IDSC) has published a series of discussionpapers on phase II trial design in Clinical Cancer Research. The IDSC has developed formal recommen-dations about aspects of phase II trial design that are the subject of frequent debate, such asendpoints (response versus progression-free survival), randomization (single-arm designs versusrandomization), inclusion of biomarkers, biomarker-based patient enrichment strategies, and statisticaldesign (e.g., two-stage designs versus multiple-group adaptive designs). Although these recommendationsin general encourage the use of progression-free survival as the primary endpoint, randomization, inclu-sion of biomarkers, and incorporation of newer designs, we acknowledge that objective response asan endpoint and single-arm designs remain relevant in certain situations. The design of any clinical trialshould always be carefully evaluated and justified based on characteristic specific to the situation. ClinCancer Res; 16(6); 1764–9. ©2010 AACR.

Background

Many new drugs targeting molecular pathways are readyfor clinical development, necessitating the use of efficienttrial designs to quickly and accurately identify promisingagents, while also identifying those for which all furtherdevelopment should be stopped. Although the develop-

ffiliations: 1NCIC Clinical Trials Group, Kingston, Ontario,ational Cancer Institute, Bethesda, Maryland; 3Mayo Clinic,Biomedical Statistics and Informatics, Rochester, Minnesota;Sloan-Kettering Cancer Center, Department of Medicine, NewYork; 5University of Michigan, Department of Medicine andlogy, Ann Arbor, Michigan; 6University of Chicago,t of Medicine, Section of Hematology/Oncology, Chicago,ed Hutchinson Cancer Research Center, Cancer Researchtistics, Seattle, Washington; 8University of Texas M. D.ancer Center, Houston, Texas; 9Keck School of Medicine,prehensive Cancer Center, University of Southern California,es, California; 10Bristol-Myers Squibb, Plainsboro, New11The Emmes Corporation, Rockville, Maryland on behalf ofTrial Design Task Force

ding Author: Lesley Seymour, NCIC Clinical Trials Group,treet, Kingston, ON, Canada K7L 3N6. Phone: 613-533-6430;3-2941; E-mail: [email protected].

8/1078-0432.CCR-09-3287

erican Association for Cancer Research.

ACD4MYPDIlaANLJth

C1F

d

©

er Res; 16(6) March 15, 2010

ment of some drugs is discontinued after phase I, themajor drug development decision is generally made onthe basis of phase II results. Although traditional oncologytrial designs using the endpoint of response and a singlearm design seem to have done this task reasonably wellfor cytotoxic agents, the same does not seem to be truefor newer agents in which high rates of tumor shrinkagemay not be expected, nor for combinations of agents (suchas a new drug combined with standard treatments). Cer-tainly, success rates for phase III trials seem to be decreas-ing (1). This decrease has led to considerable scientificdiscussion, debating the advantages and disadvantages ofusing response versus progression (2) or other imagingendpoints (3), single arm versus randomized designs(4), patient enrichment and biomarker endpoints (5),and optimal statistical designs, such as adaptive designor phase I–II designs.The Investigational Drug Steering Committee (IDSC) of

the National Cancer Institute Cancer Therapy and Evalua-tion Program (NCI CTEP) appointed a Clinical Trial De-sign Task Force to advise on the design of early (phase Iand II) clinical trials (Table 1). In keeping with its broadmandate, Task Force members include IDSC members, aswell as external representation from academia and the

Design of Phase II Clinical Trials Testing Cancer Therapeutics

pharmaceutical industry (Table 2). Members have exper-tise in early clinical trial design, conduct, and analyses,and the Task Force is composed of statisticians, clinicians,imaging specialists, pharmacologists, biomarker experts,radiation and systems biologists, as well as patient advo-cates. One of the first initiatives of the Task Force was tocoordinate a Phase II Workshop attended by Task Forcemembers, other IDSC members, and a number of invitedexperts in the field. This Phase II Workshop formed thebasis of a series of discussion documents on all aspectsof phase II design and conduct, published in this journal(1–5). Subsequently, at the request of the IDSC, the TaskForce formulated specific recommendations for the designof phase II clinical trials (Figs. 1 and 2). These recommen-dations, although based on the original Phase II Work-shop and subsequent publications, underwent extensivediscussion and revision to ensure broad applicability andacceptance and were formally approved by the IDSC.We report here on the IDSC's recommendations about

phase II trial design.

Review and Discussion

Types of phase II trials. Although the most commongrouping of phase II trials is by design, i.e., single armand randomized trials (Fig. 3), another conceptual group-ing is the ultimate aim of the trial. Is the trial to be used toscreen for any evidence of activity of a new drug that hasrecently completed phase I testing, to look for preliminaryhints of activity, and to guide selection of tumor types forfurther study? Or, is the trial designed to provide a go–nogo answer to allow the conduct of a definitive registra-tional trial in a specific disease? The Task Force considereda number of ways to categorize phase II trials, includingthe concept of screening (e.g., IIa) versus decision-making(i.e., IIb, sufficiently robust to support progression tophase III). However, the consensus opinion was that formost drugs, conducting two phase II trials in sequencewould be inefficient. We recognized that there are circum-stances in which such a “proof of concept” approachmight be reasonable, such as seeking a signal about theselection of tumor types for further study (e.g., when notreadily apparent from preclinical or phase I studies) or forbiomarker-based studies to validate a proposed mecha-nism of action. Ideally, however, these concepts would


be embedded, possibly adaptively, in a single phase II trial.Thus, we used single-arm versus randomized studies as ourprimary categorization of phase II trials (Fig. 3).Selection of the appropriate primary endpoint. The

advantages and limitations of objective tumor responseas the primary endpoint in phase II trials, and alternativeendpoints such as multinomial endpoints (6), response asa continuous variable, progression-free survival, imaging,patient reported outcomes (7), and biomarker endpointswere reviewed by Dhani (2), Shankar (3), and McShane(5), and will not be further addressed here. We believethat the decision on the most appropriate endpoint iscritical to inform the appropriate design. Although theTask Force accepted that response-based endpoints arestill relevant for some agents (when tumor shrinkage andclinically relevant response rates are expected) and sometrials, the recommendations emphasize the need to consid-er the inclusion of a progression-free survival primary end-point as more informative (8, 9). Overall survival is notrecommended as an endpoint, as subsequent therapymay confound conclusions, and progression is usuallysubstantially earlier, thus shortening the duration of thetrial and follow up.Randomization, blinding, and crossover. The Task Force

agreed that randomization was generally required to eval-uate the efficacy of combinations of agents (e.g., for ap-proved drugs and investigational agents). Randomizationis usually essential for a phase II trial in which progres-sion-free survival is the most appropriate endpoint.Nonetheless, single arm designs are still appropriate forthe evaluation of a monotherapy or when a well-definedhistorical control database is available (10). As for anytrial the design and the selected null and alternative hy-potheses must be carefully justified. If a randomized de-sign is selected, blinding of the agents (against placebo,other doses of the same agent, or other active agents)should be considered. When the primary endpoint is pro-gression-free survival or response based, designs that al-low crossover after progression maintain the integrity ofthe study and can provide additional data that could in-form the future development of the agent.Biomarkers. Biomarkers are of considerable interest in

the setting of the phase II study, but they present signif-icant challenges in their incorporation, measurement,and interpretation. In most instances, the biomarker is

Table 1. Objectives of the clinical trial design task force

Objectives

Provide guidance to the IDSC on current best practice for all aspects of early clinical trial designIdentify areas inwhich further research and investigation are needed to improve the quality and/or efficiency of early clinical trial designAssist NCI and the IDSC or as needed with the implementation of recommendations for further research and investigationCollaborate with other IDSC task forces to minimize duplication of effortsDevelop and implement a plan to disseminate and publish the recommendations and guidance formulated by the task force and

approved by the IDSC


1766

CCR FOCUS
not clinically validated as a predictive marker (of efficacy)early in the development of a new agent (i.e., at the time ofthe phase II trial). The IDSC's recommendations about bio-marker use in early clinical trials are detailed in the currentFocus Series (11–16). Because of these limitations, theserecommendations encourage the prospective inclusion ofmolecular markers in phase II trials to evaluate predictivemarkers, but discourage prospective patient selection onthe basis of a biomarker (unless already clinically vali-dated), except in the setting of an appropriate (and ex-plicit) adaptive design. Phase II trials including patientswith a specific biomarker but with multiple histologicalsubtypes were considered of particular interest and maybe a more efficient screening tool, especially when com-bined with an adaptive design.Statistical designs. Improved efficiencies in clinical trial
design with associated shortening of development timesfor effective agents are highly desirable. Numerous designshave been proposed, including randomized selection de-signs (pick-the-winner), adaptive designs (17), random-ized discontinuation designs (18), and other randomizeddesigns (19). Prospectively specified adaptive designs areof particular interest in the context of phase II studies ofmolecularly targeted agents in which biomarker identifica-tion and validation may be emergent during the conduct ofthe trial, limiting the ability to select patients or identifyoptimal doses and/or schedules at the trial outset. Suchadaptive designs are also particularly useful for trials in-cluding patients with a range of histologic subtypes butwith biomarkers of interest. Adaptive designs in such set-tings should be efficient and may result in improved preci-sion. Despite the multiplicity of new designs that havebeen proposed, their inclusion in new trials has in generalbeen modest at best (20, 21). Reasons postulated includerequirements for statistical support as well as concernsabout robustness, accrual, and cost.The Task Force is strongly supportive of designs that im-

prove efficiency and shorten development time, such asadaptive designs, but recognized the need to continue toformally evaluate these designs to encourage wider accep-tance and implementation. An ongoing initiative is the cre-


ation of a database to allow the formal testing, in silico ofnewer designs, in order to validate their use in future trials.Interestingly, although formulated prior to the publica-

tion of the editorial, these recommendations are congru-ent with a review of phase II trials published in theJournal of Clinical Oncology (22), as well as with otherreviews and recommendations (23).

Consensus Recommendations

Choosing the appropriate primary endpointThe first and critical decision point for the design of a

phase II trial is based on the choice of the most appropri-ate primary endpoint, which should be tailored to the dis-ease and drug(s) under investigation.• Response-based endpoints such as those defined by

Response Evaluation Criteria In Solid Tumors (RE-CIST), are standard, especially in early phase II trials.Other qualified biomarkers, such as molecular imag-ing or tumor markers, may be appropriate in selectcircumstances. Response-based endpoints are appro-priate primary endpoints if unambiguous and clini-cally relevant direct antitumor activity (such astumor shrinkage) is hypothesized.

• If a response-based endpoint is not appropriate, espe-cially in later phase II trials, progression-free survival isrecommended as the primary endpoint. Other bio-marker endpoints (such as tumor burden, tumor mar-kers, novel imaging, tumor response, molecularbiomarkers) and patient-reported outcomes are al-ways encouraged as secondary endpoints, especiallyin the context of studies that aim to qualify such end-points. It is acknowledged that once qualified, thesebiomarker endpoints will become appropriate prima-ry endpoints.

Study design: primary endpoint is tumor responseMonotherapy trials. Single arm designs are acceptable.

However, randomization should be encouraged to opti-mize dose and schedule or to benchmark activity againstknown active therapies.

Table 2. Past and present members of the clinical trial design task force

Position
Member
Chair
Lesley Seymour Co-chair Donald Berry CTEP S. Percy Ivy Clinical and/or
pharmacology
A. Adjei, S. Yao, L. Baker, S. Lutzker, J. Humphrey, D. Stewart, A. Dowlati, P. Keegan, P.
LoRusso, M. Ratain, D. Spriggs, J. Collins, M. Grever, C. Erlichmann
Statistics J. Crowley, S. Groshen, M. Le Blanc, L. Rubinstein, D. Sargent Imaging L. Shankar, A. Shields Advocate D. Collyar Nonvoting R. Agarwal, L. Minasian, P. Ujhazy, L. Jensen, P. West Past members G. Eckhart, S. Arbuck, M. Christian, G. Fyfe, R. Humphrey, M. Sznol, M. Villalona-Calero, M. Weinblatt


Combination trials. With some exceptions (e.g.,availability of a well-validated robust control database),randomization is usually required for trials testing combi-nations of agents to establish efficacy. An example is stan-dard therapy ± novel agent or combinations of novelagents.

Study design: primary endpoint is progression-freesurvivalMonotherapy or combination trials1. With some exceptions (e.g., availability of a robust

control database), randomization is required.2. For randomized trials, blinded designs are encour-

aged when feasible. Although placebo controlledtrials are challenging, they are encouraged wheneverpossible. Alternatives include dose ranging, random-ization versus active controls or other novel agents,and randomized discontinuation and other cross-over designs.

3. It may be informative to prospectively incorporatecrossover to the standard therapy + novel agent forthose patients initially assigned to the standard ther-


apy alone, although careful consideration should begiven to the timing of crossover (e.g., only after theprimary endpoint has been observed). Such cross-over designs increase the access of patients to inves-tigational agents, and also provide additionalinformation about the activity of the study arms.

Patient selection and enrichment strategiesMonotherapy or combination trials1. A goal of phase (I and) II development should be

to define biomarkers predictive of efficacy and/ortoxicity. When feasible and appropriate, molecularbiomarkers should be explored in order to identifysubsets of patients of interest for future study.

2. Enrollment should, in general, not be limited bybiomarker status unless there are strong confirmatoryand supportive clinical data justifying the enrichmentstrategy. Adaptive statistical designs may be used toallow modification of enrollment if data suggest abiomarker is predictive.

3. In an unselected trial (i.e., patients not defined bya biomarker), the patient population of primary

Fig. 1. Process for development of recommendations.


C1768

CCR FOCUS

interest (i.e., a cohort defined by a biomarker)should be predefined and the study powered accord-ingly to detect an effect in that subset.

4. Multidisease phase II designs should be considered,especially if the objective is to test a biomarker-focused hypothesis.

lin Cancer Res; 16(6) March 15, 2010

Statistical designsProspective designs that adapt to what is learned dur-

ing the trial can improve the efficiency of drug develop-ment and provide greater precision. Available adaptationsinclude stopping early, continuing longer than antici-pated, dropping arms (or doses), adding arms, focusing

C

Fig. 2. Process and/or flowor approaches for determinationof phase II trial designrecommendations. PRO, patientrelated outcomes; PFS,progression-free survival.

Fig. 3. Types of phase II studies.

linical Cancer Research


on patient subsets, assignment of better performingtreatment arms with greater probability, and seamlesslymoving from phase I to II or phase II to III during asingle trial.

Conclusions

The Task Force formulated recommendations (Fig. 2)for the design of phase II trials of anticancer agents onthe basis of consensus gained during a workshop and ex-tensive discussions with members of the IDSC, the TaskForce, and external experts. These recommendations weresubsequently approved by the IDSC. Although theserecommendations in general encouraged the use ofprogression-free survival as the primary endpoint, the useof randomization, the inclusion of biomarkers, and theuse of newer designs, they acknowledge that objective


response and single-arm design remain relevant in appro-priate circumstances. The design of any clinical trialshould always be carefully evaluated and justified.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The Clinical Trial Design Task Force would like to acknowledge thesupport of the NCI and the IDSC members and chairs, Amy Gravell,LeeAnn Jensen, the attendees of the original Phase II Workshop, as wellas all Task Force members, past and present, especially Drs. Alex Adjeiand Michaele Christian who originally led the Task Force (Table 2).

Received 12/15/2009; revised 01/08/2010; accepted 01/13/2010;published OnlineFirst 03/09/2010.

References
1. Adjei AA, Christian M, Ivy P. Novel designs and end points for phase
II clinical trials. Clin Cancer Res 2009;15:1866–72.2. Dhani N, Tu D, Sargent DJ, Seymour L, Moore MJ. Alternate end-

points for screening phase II studies. Clin Cancer Res 2009;15:1873–82.

3. Shankar LK, Van den Abbeele A, Yap J, Benjamin R, Scheutze S,FitzGerald TJ. Considerations for the use of imaging tools for phaseII treatment trials in oncology. Clin Cancer Res 2009;15:1891–7.

4. Rubinstein L, Crowley J, Ivy P, LeBlanc M, Sargent D. Randomizedphase II designs. Clin Cancer Res 2009;15:1883–90.

5. McShane LM, Hunsberger S, Adjei AA. Effective incorporation of bio-markers into phase II trials. Clin Cancer Res 2009;15:1898–905.

6. Goffin JR, Tu D. Phase II Stopping rules that employ response ratesand early progression. J Clin Oncol 2008;26:3715–20.

7. Wagner LI, Wenzel L, Shaw E, Cella D. Patient-reported outcomesin phase II cancer clinical trials: lessons learned and future directions.J Clin Oncol 2007;25:5058–62.

8. Fleming TR, Rothmann MD, Lu HL. Issues in using progression-freesurvival when evaluating oncology products. J Clin Oncol 2009;27:2874–80.

9. Freidlin B, Korn EL, Hunsberger S, et al. Proposal for the use of pro-gression-free survival in unblinded randomized trials. J Clin Oncol2007;25:2122–6.

10. Korn E, Liu P-Y, Lee SJ, et al. Meta-analysis of phase II cooperativegroup trials in metastatic stage IV melanoma to determine progres-sion-free and overall survival benchmarks for future phase II trials.J Clin Oncol 2008;26:527–34.

11. LoRusso PM, Boerner SA, Seymour LK. An overview of the optimalplanning, design and conduct of phase I studies of new therapeutics.Clin Cancer Res 2010;16:1710–8.

12. Dancey JE, Dobbin KK, Groshen S, et al. Guidelines for the develop-ment and incorporation of biomarker studies in early clinical trials ofnovel agents. Clin Cancer Res 2010;16:1745–55.

13. Senderowicz A. Information needed to conduct first-in-human oncol-ogy trials in the U.S. A view from a former FDA Medical Reviewer.Clin Cancer Res 2010;16:1719–25.

14. Rock EP, Molloy VJ, Humphrey JS. Good clinical practice data qual-ity in early drug development. Clin Cancer Res 2010;16:1756–63.

15. Forster MD, Saijo N, Seymour L, Calvert AH. Performing phase I clin-ical trials of anticancer agents: perspectives from within the Europe-an Union and Japan. Clin Cancer Res 2010;16:1737–44.

16. Ivy SP, Siu L, Garrett-Mayer E, Rubinstein L. Approaches to phase 1clinical trial design focused on safety, efficiency and selected patientpopulations: A report from the Clinical Trial Design Task Force of theNational Cancer Institute Investigational Drug Steering Committee.Clin Cancer Res 2010;16:1726–36.

17. Coffey CS, Kairalla JA. Adaptive clinical trials: progress and chal-lenges. Drugs R D 2008;9:229–42.

18. Rosner GL, Stadler W, Ratain M. Randomized discontinuation de-sign: application to cytostatic antineoplastic agents. J Clin Oncol2002;20:4478–84.

19. Rubinstein LV, Korn EL, Friedlin B, Hunsberger S, Ivy SP, Smith NA.Design issues of randomized phase II trials and a proposal for phaseII screening trials. J Clin Oncol 2005;23:7199–206.

20. El-Maraghi RH, Eisenhauer EA. Review of phase II trial designs usedin studies of molecular targeted agents: outcomes and predictors ofsuccess in phase III. J Clin Oncol 2008;26:1346–54.

21. Lee JJ, Feng L. Randomized phase II designs in cancer clinicaltrials: current status and future directions. J Clin Oncol 2005;23:4450–445.

22. Cannistra S. Phase II trials in Journal of Clinical Oncology. J ClinOncol 2009;27:3073–6.

23. Booth CM, Calvert HA, Giaccone G, et al. Design and conduct ofphase II studies of targeted anticancer therapy: recommendationsfrom the task force on methodology for the development of innova-tive cancer therapies (MDICT). Eur J Cancer 2008;44:25–9.


design of phase i combination trials: recommendations of ... · these include the selection of...

Documents