Validation of Analytic Methods for Biomarkers Used in Drug Development

2008;14:5967-5976. Clin Cancer Res Cindy H. Chau, Olivier Rixe, Howard McLeod, et al. DevelopmentValidation of Analytic Methods for Biomarkers Used in Drug

Updated version

http://clincancerres.aacrjournals.org/content/14/19/5967

Access the most recent version of this article at:

Cited Articles

http://clincancerres.aacrjournals.org/content/14/19/5967.full.html#ref-list-1

This article cites by 47 articles, 14 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/14/19/5967.full.html#related-urls

This article has been cited by 11 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications

Research. on April 27, 2014. © 2008 American Association for Cancerclincancerres.aacrjournals.org Downloaded from Research. on April 27, 2014. © 2008 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/content/14/19/5967

http://clincancerres.aacrjournals.org/content/14/19/5967.full.html#ref-list-1

http://clincancerres.aacrjournals.org/content/14/19/5967.full.html#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]

http://clincancerres.aacrjournals.org/


Validation of Analytic Methods for Biomarkers Used inDrug DevelopmentCindy H. Chau,1Olivier Rixe,1Howard McLeod,2 and William D. Figg1

Abstract The role of biomarkers in drug discovery and development has gained precedence over the years.As biomarkers become integrated into drug development and clinical trials, quality assurance and,in particular, assay validation become essential with the need to establish standardized guidelinesfor analytic methods used in biomarker measurements. New biomarkers can revolutionize boththe development and use of therapeutics but are contingent on the establishment of a concretevalidation process that addresses technology integration and method validation as well asregulatory pathways for efficient biomarker development.This perspective focuses on the generalprinciples of the biomarker validation process with an emphasis on assay validation andthe collaborative efforts undertaken by various sectors to promote the standardization of thisprocedure for efficient biomarker development.

Biomarkers are playing an increasingly important role indrug discovery and development from target identificationand validation to clinical application, thereby making theoverall process a more rational approach. The potential useof biomarkers in each phase of the drug development processis summarized in Table 1 (1). The incorporation ofbiomarkers in drug development has clinical benefits thatlie in the screening, diagnosing, or monitoring of the activityof diseases or in assessing therapeutic response. Thedevelopment and validation of these mechanism-basedbiomarkers serve as novel surrogate end points in early-phasedrug trials. This has created a much appreciated environmentfor protein biomarker discovery efforts and the developmentof a biomarker pipeline that resembles the various phasesof drug development. The components of the biomarkerdevelopment process include discovery, qualification, verifi-cation, research assay optimization, clinical validation, andcommercialization (2).The role of biomarkers in rational drug development has

been a major focus of the Food and Drug Administration (FDA)critical path initiative and the NIH roadmap (3). Althoughthe overwhelming majority of biomarkers are proteins used assurrogate end points for drug development, diagnostic bio-

markers may also prove useful for understanding the biologyof the disease. Successful biomarker development depends ona series of pathway approach that originates from the discoveryphase and culminates in the clinical validation of an appro-priately targeted biomarker. Much emphasis has been placedon the paradigm of biomarker translation specifically on theprinciples of biomarker validation in clinical trials, and thearticles in this edition of CCR Focus will address various issuesalong the validation pathway, including the analysis of micro-array data sets (4), the validation of predictive models (5), thedesign of clinical trials using genomics (6), and the overallstatistical challenges that exist (7). New biomarkers canrevolutionize both the development and use of therapeuticsbut are contingent on the establishment of a concretevalidation process that addresses technology integration andmethod validation as well as regulatory pathways for efficientbiomarker development. This perspective will feature highlightson the biomarker validation process and includes a discussionon analytic method validation.

Biomarker Definitions

Numerous publications have described the application ofbiomarkers in drug development using various nomencla-tures to describe distinct aspects of this process. We beginwith the standardization of terminology for ease of under-standing the biomarker literature. A consensus definition ofa biomarker is a factor that is objectively measured andevaluated as an indicator of normal biological processes,pathogenic processes, or pharmacologic responses to atherapeutic intervention (8). A clinical end point is definedas a variable that measures how patients feel, function, orsurvive, whereas a surrogate end point is a biomarker that isintended to substitute for a clinical end point. In this case, asurrogate end point is expected to predict clinical benefit.Examples of surrogate end points and clinical end points areprovided in Table 2.

Authors’Affiliations: 1Medical Oncology Branch, Center for Cancer Research,National Cancer Institute, Bethesda, Maryland and 2Institute for Pharmaco-genomics and Individualized Therapy, University of North Carolina, Chapel Hill,North CarolinaReceived 2/20/08; revised 6/26/08; accepted 7/2/08.Requests for reprints: Howard McLeod, Institute for Pharmacogenomics andIndividualized Therapy, University of North Carolina, Campus Box 7360, 3203Kerr Hall, Chapel Hill, NC 27599-7360. Phone: 919-966-0512; E-mail: [email protected] orWilliam D. Figg, Medical Oncology Branch, Center for Cancer Research,National Cancer Institute, Building 10/Room 5A01, 9000 Rockville Pike, Bethesda,MD 20892. Phone: 301-402-3623; Fax: 301-402-8606; E-mail: wdfigg@

helix.nih.gov.F2008 American Association for Cancer Research.doi:10.1158/1078-0432.CCR-07-4535

www.aacrjournals.org Clin Cancer Res 2008;14(19) October1, 20085967

CCR FOCUS

Research. on April 27, 2014. © 2008 American Association for Cancerclincancerres.aacrjournals.org Downloaded from


Another critical distinction should be made when a bio-marker undergoes analytic method validation versus clinicalqualification. Analytic method validation is the process ofassessing the assay, its performance characteristics, and theoptimal conditions that will generate the reproducibility andaccuracy of the assay. Clinical qualification is the evidentiaryprocess of linking a biomarker with biological processesand clinical endpoints (9). Although validation, qualification,or evaluation has been used interchangeably in the literature,the distinction should be made to properly describe theparticular phase the biomarker is transitioning through inthe drug development process. As such, the term validationis reserved for analytic methods, and qualification forbiomarker clinical evaluation to determine surrogate endpointcandidacy (8, 9). Both validation and qualification processesare intertwined, and hence, their integration guides biomarkerdevelopment with the principle of linking the biomarkerwith its intended use (see Fit-for-Purpose Method Validation;ref. 10).

Biomarker Qualification Process Map

The FDA has issued guidance for industry on pharmacoge-nomic data submissions and in classifying the various types ofgenomic biomarkers and their degree of validity: exploratory

biomarkers, probable valid biomarkers, and known validbiomarkers.3 Exploratory biomarkers lay the groundwork forprobable or known valid biomarkers and can be used to fill ingaps of uncertainty about disease targets or variability in drugresponse and bridge the results of animal model studies toclinical expectation or used for the selection of new compounds(11). Examples of exploratory biomarkers include the use ofgene panels used for preclinical safety evaluation or theevaluation of vascular endothelial growth factor as a target toassess the efficacy of angiogenesis inhibitors. For an exploratorybiomarker to achieve the status of probable valid biomarker, itneeds to be ‘‘measured in an analytic test system with well-established performance characteristics and for which there isan established scientific framework or body of evidence thatelucidates the physiologic, toxicologic, pharmacologic, orclinical significance of the test results.’’ A probable validbiomarker seems to have predictive value for clinical outcomesbut has not been independently replicated or widely accepted.The advancement from probable valid to known valid lies inthe achievement of a broad consensus in cross-validation

Table 1. Potential uses of biomarkers to facilitate the drug development process

Phases of drug development process Potential uses of biomarkers during drug development

Target discovery and validation Biomarkers used to identify and justify targets for therapy, suchas cellular growth factor receptors and signaling molecules[e.g., HER2 proto-oncogene frequently amplified in breast cancerand associated with poor prognosis (this correlation provided therationale for anti-HER2 therapeutic strategies leading to thedevelopment of trastuzumab)]

Lead discovery and optimization Biomarkers used to determine target effects with target-associatedassays to identify leads and evaluate the effects of molecular-targeted drugs in preclinical development

Lead agents developed against given target further optimized basedon biomarker endpoints in model systems or animal studies

Preclinical studies Development of appropriate animal models of cancer that featurebiomarker properties comparable with those seen in patientpopulations to enhance their utility as predictive models

Biomarkers can play an essential role in the validation of new diseasemodels (e.g., transgenic mouse models of breast cancer thatoverexpress HER2)

Biomarkers are used to assess toxicity and safety of the drug

Clinical trials Biomarker-based studies can provide early evaluations of whetherthe drug is hitting the target (success or failure)

Mechanism-based biomarkers can help guide rational selection ofeffective drug combinations

Optimization of dose and schedule can be based on pharmacologiceffects on biomarker-based endpoints rather than on maximumtolerated dose

Biomarkers can serve as tools for the selection of appropriate patientpopulations or used to stratify patients based on differential clinicalresponse or to identify responders in a subpopulation

Development and validation of mechanism-based biomarkers thatreflect disease activity or the interactions between disease targetsand targeted therapy may lead to new surrogate endpoints ofclinical benefit

Selected biomarkers may have the potential to predict clinical outcome

3 U.S. Food and Drug Administration. Guidance for industry: pharmacogenomicdata submissions, 2005. Available from: http://www.fda.gov/cder/guidance/6400fnl.pdf.

www.aacrjournals.orgClin Cancer Res 2008;14(19) October1, 2008 5968

CCR FOCUS



experiments, which include the independent validation of thebiomarker by replicating the outcome at different sites. Thus,a known valid biomarker is defined as ‘‘a biomarker that ismeasured in an analytic test system with well-establishedperformance characteristics and for which there is widespreadagreement in the medical or scientific community about thephysiologic, toxicologic, pharmacologic, or clinical significanceof the test results.’’ That is, the scientific community at largeaccepts these known valid biomarkers to predict clinical orpreclinical outcomes. Examples of such valid genomic bio-markers that are listed on the labels of FDA-approved drugproducts are presented in Table 3.The qualification process is introduced to bridge the gap

from an exploratory biomarker to a known valid biomarkerstatus, keeping in mind that a validation process requires aconsensus on an efficient and transparent process map forgenomic biomarker validation. A qualification process map hasbeen proposed by the FDA that evaluates exploratory genomicbiomarkers of preclinical drug safety to assess the potential ofgenomic technologies in mock submission (12, 13) andidentify key variables that can be used to determine the successof these biomarkers in voluntary genomic data submission(14). The proposal transitions an exploratory biomarker to aknown valid genomic biomarker through a series of phasesfrom discovery to method development to validation studiesand cross-validation consortium (11). In the case of a processmap that involves the validation of genomic biomarkers inclinical trials, the regulatory agency will review the biomarkervalidation package in terms of the usefulness of the biomarkerin predicting clinical benefit (11). Although this particularqualification process addresses genomic biomarkers, its appli-cation can be further extended to other types of biomarkers(e.g., protein or diagnostic biomarkers) granted the qualifica-tion approach remains intact. Figure 1 shows the integration ofthe FDA biomarker qualification process along the phases ofthe drug development process.The inclusion of biomarkers in drug development and

regulatory review will improve the efficiency of the biomarkerdevelopment process. Biomarker qualification is also observedin the codevelopment of biomarkers (in the form of diagnostictests) and drugs with the use of these biomarkers limited tothe application of the drug (11).4 Codevelopment imposes thenecessity to generate specific guidelines describing analytic test

validation (sensitivity and specificity of the assays), clinical testvalidation (ability of the assays to detect and predict diseases),and clinical utility.5 Furthermore, biomarkers play a significantrole in adaptive trial designs in which patient populationstratification and efficacy determination is based on biomarkerreadouts (15). Adaptive trials attempt to maximize thestatistical power of the study with a small sample size. Theintegration of biomarkers into drug development is furtherdiscussed in the section below.

Fit-for-PurposeMethodValidation

It is important to point out that biomarker methodvalidation is distinct from pharmacokinetic validation androutine laboratory validation. The FDA has issued guidance forindustry on bioanalytic method validation for assays thatsupport pharmacokinetic studies that are specific for small-molecule drugs and that are not directly related to thevalidation of biomarker assays.6 Whereas routine laboratoryvalidation refers to laboratories that do testing on humanspecimens for diagnosis, prevention, or treatment of anydisease and falls under the jurisdiction of the ClinicalLaboratory Improvement Amendments of 1988, there is littleregulatory guidance on biomarker assay validation. In October2003, the American Association of Pharmaceutical Scientists(AAPS) and the U.S. Clinical Ligand Society (CLAS) cospon-sored a Biomarker Method Validation Workshop to address thevalidation challenges of biomarker assays in support of drugdevelopment (16). At this meeting, it was concluded thatbiomarker methods should not be validated by the sameguiding principles developed for drug analysis used inbioanalytic method validation. Hence, a ‘‘fit-for-purpose’’approach for biomarker method development and validationis derived with the idea that assay validation should be tailoredto meet the intended purpose of the biomarker study.Method validation should show the reliability of the assay for

the intended application with the rigor of the validation processincreasing from the initial validation proposed for exploratorypurposes to the more advanced validation dependent on theevidentiary status of the biomarker (10). The fit-for-purposemethod validation is an umbrella terminology that is used todescribe distinct stages of the validation process, including

4 U.S. Department of Health and Human Services, Food and Drug Administration.Table of valid genomic biomarkers in the context of approved drug labels, 2007.Available from: http://www.fda.gov/cder/genomics/genomic_biomarkers_table.htm.

Table 2. Examples of surrogate end points and clinical end points

Disease Surrogate end points Clinical end points

Hypertension Blood pressure StrokeDyslipidemia Cholesterol, LDL Coronary artery diseaseDiabetes Glycosylated hemoglobin (HbA1c) Retinopathy, nephropathy, neuropathy, heart diseaseGlaucoma Intraocular pressure Loss of visionCancer Biomarkers Progression-free survival

Tumor shrinkage, response rate Overall survival

5 U.S. Department of Health and Human Services, Food and Drug Administration.Drug-diagnostic co-development concept paper, 2005. Available from: http://www.fda.gov/cder/genomics/pharmacoconceptfn.pdf.6 U.S. Food and Drug Administration. Guidance for industry: bioanalytical methodvalidation, 2001. Available from: http://www.fda.gov/cder/guidance/4252fnl.pdf.


BiomarkerMethodValidation



Table 3. Valid genomic biomarkers in the context of FDA-approved drug labels

Genomic biomarker Context in label for which biomarker is valid Other drugs associatedwith this biomarker

Representative label Test* Drug

c-KIT mutationsc Presence and type of GIST c-KITmutations (exon 11 vs exon 9)predicts sensitivity to imatinib

3 Imatinib mesylate

CCR5—chemokineC-C motif receptor

Drug blocks CCR5 receptor on T cellthat HIV binds to for entry(use only in patients withCCR5-tropic HIV-1 detectable)

1 Maraviroc

CYP2C9 variants CYP2C9 PM variants " and EMvariants # drug exposureand risk

3 Celecoxib

CYP2C9 mutations CYP2C9 mutations " bleeding riskthus requiring lower drug dose

2 Warfarin

CYP2C19 variants CYP2C19 variants with geneticdefect leads to change in drugexposure (PM " drugexposure and toxicity)

3 Voriconazole Omeprazole, pantoprazole,esomeprazole,diazepam, nelfinavir,rabeprazole

CYP2D6 variants CYP2D6 PM variants " and EMvariants # drug exposureand toxicity

3 Fluoxetinehydrochloride(HCl)

Fluoxetine HCl and olanzapine,cevimeline HCl, tolterodine,terbinafine, tramadol andacetamophen, clozapine,aripipraxole, metoprolol,propanolol, carvedilol,propafenone, thioridazine,protriptyline HCl, atomoxetine,venlafaxine, risperidone,tiotropium bromide inhalation,tamoxifen, timolol maleate

Deletion ofchromosome 5q

Cytogenetic abnormality inmanagement of low- orintermediate-1 riskmyelodysplastic syndromes

3 Lenalidomide

DPD deficiency DPD deficiency " risk of toxicity 3 Fluorouracilb Capecitabine, fluorouracilcream, fluorouraciltopical solution and cream

EGFR mutationsx EGFR mutations " response inNSCLC

3 Gefitinib Cetuximab

EGFR expression EGFR (+) expression requiredfor CRC

1k Cetuximab Panitumab, gefitinib

Her2/neu overexpressionor amplification

Detection of Her2/neu overexpressionor amplification required to selectpatients for therapy in breast cancer

1 Trastuzumab Lapatinib

K-RAS mutations{ K-RAS mutations confer resistance tocetuximab in CRC

3 Cetuximab

N-acetyltransferase(NAT) variants

NAT variants slow and fast acetylators(slow acetylation " drug exposureand toxicity)

3 Isoniazed Hydralazine HCl

Philadelphia (Ph1)chromosome-positiveresponders

Ph1 presence predict response—busulfan less effective inpatients with (Ph1-) chronicmyelogenous leukemia

3 Busulfan

Philadelphia (Ph1)chromosome-positiveresponders

Ph1 presence predict response—dasatinib is indicated foradults with (Ph1+) acutelymphoblastic leukemia

1 Dasatinib

PML/RARa geneexpression

Presence of PML/RAR (a) fusiongene predicts response to drug

3 Tretinoin Arsenic oxide

Protein C deficiencies Hereditary or acquired deficienciesof protein C may " risk of tissuenecrosis

2 Warfarin

TPMT variants TPMT deficiency or mutation " riskof myelotoxicity

2 Azathioprine

(Continued on the following page)


CCR FOCUS



pre-validation, exploratory and advanced method validationand in-study method validation (Fig. 1). Method validation isthus a continuous and iterative process of assay refinement withvalidation criteria that is driven by the application of thebiomarkers with increasing rigor at each successive validationstep and focusing on method robustness, cross-validation, anddocumentation control.

Technology Integration in BiomarkerValidation:Choice of Assays

Before addressing the elements of biomarker assay develop-ment and method validation, it is important to recognize thatthe biomarker method validation process begins with choosingthe right assay followed by developing this assay into a validatedmethod. Indeed, the integration of various technologies provespivotal to not only biomarker identification and characteriza-tion but also validation. In fact, the platform applied inbiomarker discovery can also be further developed and usedas an analytic platform. Biomarker measurement can be assessedat different biological levels with different technologies; thus,the appropriate choice of assay depends on the application ofthe biomarker and the limitations of the respective technology.Various types of assays can be used in the biomarker methodvalidation process and range from the relatively low technologyend, such as immunohistochemistry to immunoassays, to thehigh technology end, including platforms for genomics,proteomics, and multiplex ligand-binding assays.A genomics approach consists of various methods that

measure gene expression analysis, such as in microarrays,which has become the standard technology used for target

identification and validation. Reverse transcription-PCR is avery sensitive, reproducible technology and often time used tovalidate microarray-generated data. Comparative genomichybridization can be used to detect chromosomal alterationsassociated with certain cancers. Proteomics involves globalprotein profiling to provide information about proteinabundance, location, modification, and protein-protein inter-actions. Whereas proteomics is a discovery technology,immunoassays are routinely used for protein biomarkerassessments due to its straightforward clinical application andtranslation into a potential diagnostic assay. The multiplexingof protein assays can increase the throughput for simultaneousanalysis of several proteins; however, it is limited by the need tostandardize assay conditions, the lost of sensitivity over singleassays, and the quality control (QC) of each analyte in thecomplete multiplex panel (17).Metabonomics (or metabolomics) is the profiling of endog-

enous metabolites in biofluids or tissue for characterization ofthe metabolic phenotype. The analytic platforms used are basedon nuclear magnetic resonance spectroscopy and the combi-nation of liquid chromatography with mass spectroscopy. It isprincipally used in biomarker discovery, although by definitionit is the ultimate end point measurement of biological events.Yet the technology is limited by the lack of comprehensivemetabolite databases and throughput both of which affects dataanalysis and interpretation. The integration of these technolo-gies lends to the field of bioinformatics where linkingexpression data derived from genomic/proteomic approachesto target biological pathways can provide a comprehensiveunderstanding of the disease biology and further validating theapplication of the biomarker (18).

Table 3. Valid genomic biomarkers in the context of FDA-approved drug labels (Cont’d)

Genomic biomarker Context in label for which biomarker is valid Other drugs associatedwith this biomarker

Representative label Test* Drug

UCD deficiency Contraindicated in UCD patients;evaluation for UCD beforestart of therapy

2 Valproic acid Sodium phenylacetate andsodium benzoate, sodiumphenyl buterate

UGT1A1 mutations UGT1A1 mutation " drug exposureand toxicity

2 Irinotecan

UGT1A1 mutations UGT1A1 mutation " bilirubin levels 3 NilotinibVKORC1 variants VKORC1 variants confer sensitivity

to warfarin thus # dose of warfarin2 Warfarin

NOTE: Data from http://www.fda.gov/cder/genomics/genomic_biomarkers_table.htm.Abbreviations: CRC, colorectal cancer; CYP, cytochrome P450; DPD, dihydropyrimidine dehydrogenase; EGFR, epidermal growth factorreceptor; EM, extensive metabolizer; G6PD, glucose-6-phosphate dehydrogenase; HIV, human immunodeficiency virus; NADH, nicotinamideadenine dinucleotide; NSCLC, non–small cell lung cancer; PM, poor metabolizer; PML/RAR, promyelocytic leukemia/retinoic acid receptor;TMPT, thiopurine methyltransferase; UCD, urea cycle disorders; UGT, UGD glucuronosyltransferase; VKORC1, vitamin K epoxide reductasecomplex.*Reference is made to the requirement of testing for the biomarker (1 = test required, 2 = test recommended, 3 = information only). The testrecommendation listed above is current and up to date at the time this article is written.cRecent studies have shown that c-KITexon 11 mutations are most common for gastric GISTs and these mutants respond well to imatinib. Theless common c-KIT exon 9 mutations occur in intestinal GISTs and are less sensitive to imatinib (47, 48). The current FDA-approved drug labelfor imatinib does not contain this information.bRecent studies have shown that higher fluorouracil plasma levels correlated with acute grade 3 toxicity (49).xRecent studies have shown that EGFR mutations and/or amplifications correlate with tyrosine kinase inhibitor activity (50, 51). The currentFDA-approved drug label for gefitinib and cetuximab does not contain this information.kAlthough this is a required test, it is not a predictive marker of activity.{Reference (52). The current FDA-approved drug label for cetuximab does not contain this information.





Furthermore, advances in novel imaging approaches alsohave profound implications for biomarker development.Molecular and functional imaging technologies are used toassess cell proliferation and apoptosis (e.g., 18F-fluoro-L-thymidine and 99mTc-Annexin imaging), cellular metabolism(e.g., 18F-fluorodeoxyglucose positron emission tomography),and angiogenesis and vascular dynamics (e.g., dynamiccontrast-enhanced computed tomography and magnetic reso-nance imaging). Therefore, making the right choice of assay isan important first step to successful biomarker methodvalidation. Validating the developed method or assay to reliablymeasure the biomarker depends on a series of variables that isaddressed below.

BiomarkerAnalytic MethodValidation

The key variable assay elements of biomarker methodvalidation are more complicated than for the typical bioanalyticassay that follows good laboratory practice (GLP) guidelines.Table 4 compares these two validation paradigms and high-lights some of the validation challenges encountered withbiomarker assays. Biomarker assay development and methodvalidation is a complex process that depends on severalvariables from the choice of the matrix to maintaining sampleintegrity to assay standardization and accuracy.Specifications for biological matrices need to be determined

taking into consideration the site of biomarker production andthe physiology and distribution of the biomarkers. The firstchallenge is to identify and select a meaningful sample matrixthat can be readily accessible, such as whole blood, plasma,serum, or urine. Sources of analytes can influence thevalidation process as evidenced by feasibility in the acquisitionof biological material during the study, such as in the collectionof noninvasive (sputum, urine, feces, and saliva) versusminimally invasive (blood or plasma) samples. If the methodhas sufficient sensitivity, then the preferred matrix choice is

based on ease of sample collection and analysis. However, ifsensitivity is a factor and measurement of the biomarker in thespecified matrix poses as a challenge, then the preferred matrixis chosen based on sample concentration even if this presents asa greater challenge for sample collection and preparation.In addition to the influence of sample sources, material

collection and processing should be examined to maintainsample integrity. It is important for researchers to realize thatbiospecimen collection varies across populations; how they arehandled and differences in sample processing variables candramatically affect the results of a trial. Thus, appropriateconditions for collecting, handling, and storing study samplesneed to be standardized along with adequate training of theclinical trial management personnel to preserve the stabilityand integrity of the analyte. Sample integrity can be affected byrepeated cycles of freeze-thawing specimens or by long-termstorage, and hence, the stability of the sample becomescompromised. Depending on the type of samples (biologicalfluids or tissues), minimization in variability at each step of thisprocedure from collection to processing is critical to ensureconsistent and valid analyte measurement at subsequentbiomarker assays. For biological fluids, differences in thehandling of urine versus whole blood samples can affect theanalytic assay, and thus, optimization of a processing protocolis necessary and should be based on the specific biomarker inaddition to the source of analyte. Thus, standardization ofsample collection procedures and appreciation of the associat-ed limitations allow this variability to be minimized. Moreover,the integrity of reagents is another variable that can also affectbiomarker analysis. Reagents such as antibodies are subjectto their own problems of supply, stability, and QC as theythemselves are derived from biological sources.QC measures should be undertaken to document analytic

performance during clinical studies and to determine theacceptance or rejection of an analytic run during sampleanalysis (19). Similar to bioanalytic method validation,

Fig. 1. Integration of the biomarker assayvalidation and the qualification process withdrug development.


CCR FOCUS



biomarker analysis requires a systematic review of the analytestability in calibration standards, QCs, and study samples (20).In general, QCs are prepared to evaluate the lower, middle, andupper limits of standard curve ranges. Whereas QC samples areused during study sample analysis to judge the acceptability ofassay runs, validation samples (VS) are used in assay validationexperiments to estimate intra-run and inter-run accuracy/precision and stability. Whereas only three VS concentrationsare required in GLP bioanalytic assays (21, 22), at least fivedifferent concentrations of VS should be analyzed in duplicateon at least six different runs during the prestudy validationbecause quantitative biomarker assays often exhibit nonlinearcalibration curves; thus, more VS are required (10, 23).Because biomarkers are endogenous substances, difficulty

may arise in obtaining biomarker/analyte-free matrices either to

do specificity studies on or to prepare for the calibration curve.Most of the time, the target biomarker molecule is not availableto act as a certified calibration standard (19). Researchers thenmay rely on the use of a noncertified standard, a recombinantprotein, or a surrogate matrix to construct the calibration curve(10). If assay standards are prepared in a nonauthentic matrix,QC samples should be prepared and tested in the samematrix asthe study samples to show that the assay performance is similarbetween authentic and nonauthentic matrices (24). Parallelismstudies should be conducted when surrogate standards andmatrices are used for calibration purposes. Dilution linearity canalso be problematic, as antibody and ligand-binding affinitiescan vary significantly in different media. Other importantcomponents of biomarker assay validation include the referencematerials, precision and accuracy, dynamic range, sample

Table 4. Comparison of bioanalytic assay and biomarker assay validation variables

Variable Bioanalytic (GLP) assay Biomarker assay

Assay method category Most are definitive quantitative Most are relative or quasi-quantitativeRegulatory requirement GLP No specific guidelinesNature of analyte Exogenous EndogenousStability Drug standards, QCs, sample analyte

stability often goodStability of standards and matrixanalytes often poor

Stability testing Freeze/thaw, bench top, long term measuredby spiking biological matrix with drug

Freeze/thaw, bench top, storage stabilitywith study samples

Standards/calibrators Standards prepared in study matrix; certifiedstandard readily available

Standards/calibrators made in matrixdifferent than study samples; certifiedstandards not available

Calibration model Mostly linear Choose appropriate calibration model fittingmethod and tools

QCs Certified standard and blank patientsample matrix available

Certified standard or blank matrix usually notavailable; substitute with surrogate matrices

VS and QC measurements Made in study matrix. 4-5 VS levelsand 3 QC levels

Made in study matrix. At least 5 VS levels and3 QC levels. If study matrix is limited mayuse surrogate matrix

Assay acceptance criteria 4-6-15 rule (for small molecules) 4-6-X rule or establish confidence intervalPrecision/accuracy Robust technology with acceptance criteria Variable; no acceptance criteriaSpecificity/selectivity Drugs not present in sample matrix;

samples are subject to cleanup andanalyte recovery

Specificity issues: biomarkers present insample matrix; samples not subject tocleanup; assess matrix effects and minimize;investigate sources of interference

Sensitivity LLOQ defined by acceptance criteria Limited sensitivity and dynamic range;LLOQ and LOD defined based on working criteria

Abbreviation: LOD, limit of detection.

Table 5. Summary of validation variables applicable to each category of biomarker assay

Definitive quantitative Relative quantitative Quasi-quantitative Qualitative

Validation variablesSample stability + + + +Reagent stability + + - -Assay range + + + -Parallelism + + - -Dilution linearity + + - -Accuracy + + - -Precision + + + -Sensitivity + + + +Specificity + + + +Example of assay Mass spectrometry ELISAs Immunogenicity immunoassays Immunohistochemistry





recovery, sample volumes, and instrument validation. Addi-tionally, the variability in method validation can also be affectedby assay results at different locations and the correct calibrationof the assay at different test sites.The technical validation of biomarkers depends on all aspects

of the analytic method, including assay sensitivity, specificity,reliability, and reproducibility (19, 25–27). Specificity refers tothe ability of the assay to clearly distinguish the analyte ofinterest from structurally similar substances. Selectivity meas-ures the degree to which unrelated matrix components causeanalytic interference. Precision is determined by the repeatabil-ity and reproducibility of the assay, which are factors used toquantitatively express the closeness of agreement among resultsof measurements done under specific conditions (28). Repeat-ability describes the measurements that are done under thesame conditions, whereas reproducibility addresses measure-ments done under different conditions. The reproducibility ofthe assay relies on its variability, levels of technical/instrumen-tal and biological noise, as well as different validation phases ofthe method (pre-study and in-study validation of the method).The acceptance criteria of the assay performance are estab-

lished based on the study objectives and the known assayvariability (29). The nature of the assay methodology and thedata generated using that particular assay can influence theestablishment of assay acceptance criteria. The categories ofbiomarker data that reflect the type of assay used have beendefined at the AAPS/CLAS workshop (16). To aid in theformulation of a method validation plan, a biomarker assay canbe placed into various functional categories, each requiring adistinct level of validation. A definitive quantitative assay makesuses of calibrators and a regression model to calculate absolutequantitative values for unknown samples. The referencestandard must be well defined and fully representative of thebiomarker. This type of assay can be validated to be accurate andprecise. A relative quantitative assay uses a response-concentra-tion calibration with reference standards that are not fullyrepresentative of the biomarker. Because the calibration curvemay use either a noncertified standard or surrogate matrix orboth, studies on parallelism and dilution linearity are necessary.Precision can be validated but accuracy can only be estimated. Aquasi-quantitative assay (possesses certain attributes) does notuse a calibration standard but has a continuous response that isexpressed in terms of a characteristic of the test sample.Precision can be validated, but not accuracy (16).A qualitative assay generates categorical data that lack

proportionality to the amount of analyte in a sample. The datamay be ordinal in that the assay relies on discrete scoring scaleslike those often used for immunohistochemistry or nominalsuch as the presence or absence of a gene product (10, 16).Qualitative assays are only required to show that they aresufficiently sensitive and specific to detect the target analyte. Inaddition to assay functionality, ensuring that the degree ofvalidation done reflects the level of importance of the biomarkeritself is equally important. Table 5 summarizes the validationvariables for each category of biomarker assay as recommendedby the AAPS/CLAS workshop.What should be the acceptance criteria of the assay

performance? Rather than setting the acceptance criteria forprecision and accuracy at a fixed value, as in GLP assays,

biomarker assays should but evaluated on a case per case basis,with F25% acting as default value [F30% at the lower limit ofquantitation (LLOQ); ref. 30]. In determining acceptance limitsfor QCs during sample analysis, either a 4-6-X rule orestablishing confidence intervals should be considered (10,30). Such is the case for bioanalytic assays of small moleculeswhere the analytic run is accepted as valid when at least 67%(4/6) of the QCs fall within 15% of their nominal values (the4:6:15 rule; refs. 21, 22, 31). Because the target molecule isoften present in predose samples or in the QC matrix,limitations are often placed on LLOQ.In summary, there are numerous factors that affect the

biomarker method validation process, including the sources ofvariability in measurements, the intended application of abiomarker, patient selection, sample collection and processing,and analytic validation. As such proper method validationshould be carried out in early clinical studies so that theseanalytic results can be used to assess whether the method affordsthe sensitivity, precision, and robustness of the assay. Duringearly exploratory phases of drug development, it is not necessaryto do full validation of biomarkers as long as the methodsprovide reliable data, information, and knowledge (24). As drugdevelopment progresses, validation should keep pace with therequired precision and reliability needed to achieve the studyobjectives (24, 26). Pre-study validation should be completedbefore clinical studies are begun and should set the foundationfor establishing method acceptance criteria. In-study sampleanalyses and validation must use QCs to document analyticperformance during clinical studies (19). As we advance towardthe later stages of drug development, the effect of the biomarkerdata on decisions around critical safety, efficacy, pharmacody-namic, or surrogate information increases. Thus, an increasedrigor in advanced method validation is undertaken as describedin the scaled, fit-for-purpose approach (Fig. 1).

Standardization andValidation throughCollaboration

Recognizing the importance and impact of biomarkerscoupled together with the complexity that exists in the drugdevelopment process, researchers realized that there is muchneeded standardization for biomarker development and havetherefore joined forces in an effort to integrate biomarkers intodrug development. The most recent of these alliances is aconsortium called the Cancer Biomarkers Collaborative (CBC)composed of the AACR, the FDA, and the National CancerInstitute with an initiative focused on facilitating the use ofvalidated biomarkers in clinical trials (32). The goal of the CBCis to develop guidelines in the areas of biospecimens, assayvalidation, bioinformatics, and information sharing. The CBCwill recommend standards and specifications on how to collectbiospecimens and integrate them into drug trials such thatthe desired endpoint of the biomarker measurement is reachedand these endpoints can then be compared among clinicaltrials. Effort in terms of validation is aimed at identifyingand defining how to validate a biomarker assay and make iteligible for inclusion into clinical trials. Hence, the need for awell-defined process with consensus standards and guidelinesfor biomarker development, validation, qualification, and use


CCR FOCUS



is apparent and an important priority of the collaboration. TheCBC intends to pave a regulatory pathway for biomarkers, asthey transition from the development phase through the FDAapproval process and then on to clinical utility.In addition to the CBC, other alliances in existence include

partnerships with government, industry, patient advocacygroups, and other nonprofit private sector organizations. TheBiomarkers Consortium is a public-private biomedical researchpartnership formed by the Foundation for the NIH, the NIH,FDA, and the Pharmaceutical Research and Manufacturers ofAmerica. The Biomarkers Consortium aims to rapidly identifyand qualify biomarkers, verify their individual value, andformalize their use in research and regulatory approval to guideclinical practice. Therefore, these collaborative efforts amongvarious sectors have arisen to address the lack of standardizedguidelines in biomarker validation, particularly for methodvalidation, and in doing so hope to promote an efficientbiomarker development process.

Integrating Biomarkers into Drug Development

Up to this point, we have discussed the importance ofbiomarker method standardization and validation. However,another challenge that remains to be addressed is in under-standing how to effectively and efficiently integrate biomarkersinto the drug development process. Biomarkers can play apivotal role in facilitating drug development (Table 1),particularly in oncology drug development where tumormarkers seem to correlate with prognosis and potentially be avaluable measure of treatment outcome. The incorporation ofbiomarkers as surrogate endpoints of clinical efficacy and safetyassessment is being intensely evaluated and pursued in rationaldrug development, especially in the case of biomarker and drugcodevelopments such as HER2 (also called ErbB2 and Neu) andthe development of trastuzumab.HER2 is a proto-oncogene that became a potential biomarker

when studies showed that its overexpression in breast cancer isassociated with poor prognosis (33). The role of HER2 as aclinically relevant biomarker led directly to the development oftrastuzumab, a specific targeted therapy of a recombinantmonoclonal antibody directed against the extracellular domainof HER2. HER2 as a biomarker was used for the selection of theappropriate patient populations (that express or overexpress theoncogenic protein) in clinical trials and to evaluate potentialefficacy after therapeutic intervention. Indeed, several clinicalstudies have confirmed that patients with high levels of HER2receptor overexpression (via immunohistochemical staining)are likely to receive clinical benefit from therapy and that HER2gene amplification (by fluorescence in situ hybridization) ismost predictive (34). As such, trastuzumab was subsequentlyapproved in 1998 by the FDA as second-/third-line mono-therapy or first-line therapy in combination with paclitaxel forthe treatment of HER2-overexpressing metastatic breast cancer(35–38). Thus, biomarker-based patient selection in the earlystages of the clinical trial process proves to be critical to theevaluation of a targeted agent as shown by the successfuldevelopment of trastuzumab.Another example of the successful use of biomarkers in cancer

drug development is the development of imatinibmesylate. This

molecular-targeted drug is highly efficacious in chronic myeloidleukemia (CML; ref. 39) and gastrointestinal stromal tumor(GIST; refs. 40, 41). The BCR-ABL fusion protein translocation inCML provided a biomarker and a therapeutic target for thisrationally designed small molecule. In clinical trials of imatinib,assessment of biomarker-based responses facilitated proof of itsclinical benefit in CML. Clinical benefit was determined fromevaluation of the molecular target in trials using conventionalcytogenetics, fluorescence in situ hybridization assays of theBCR-ABL translocation, and reverse transcription-PCR detectionof BCR-ABL transcripts, which are now used to guide treatmentdecisions (42). Treatment response and dose optimization canbe based on measuring the level of BCR-ABL kinase inhibitionachieved in vivo as determined by calculating the reduction inprotein levels of phospho-CRKL in mononuclear blood cellstaken from CML patients (43).Imatinib also inhibited the receptor tyrosine kinase encoded

by the oncogene c-KIT and expression of this oncogenicbiomarker provided a rationale for its use in patients withGIST. In addition, the response to imatinib is closely correlatedwith the presence and type of c-KIT mutation, creating anotherrole for these mutations to serve as biomarkers in treatmentselection for individuals with GIST (44–46). GISTs with themost common c-KIT exon 11 mutations are associated withincreased imatinib sensitivity, whereas the less common c-KITexon 9 mutations are less sensitive to imatinib (47, 48). GISTslacking mutations in c-KIT or the alternative receptor tyrosinekinase PDGFRA show much lower rates of response to imatinib(47, 48). Thus, the example of imatinib illustrates howbiomarkers derived from BCR-ABL can both stimulate initialdrug discovery efforts and serve as a useful end pointassessment of treatment effect in biomarker-based patient anddose selection for imatinib therapy.With these examples of biomarkers providing key rationale

and end points in the development of molecular-targetedagents, biomarker-based drug development can be regarded asa proven, successful strategy for developing novel anticancerdrugs. The scope of evaluation and validation of the biomarkerwill depend on its intended use as a marker of toxicity, safety,or efficacy. These biomarkers may serve as surrogate end pointsthat predict clinical outcomes. As biomarkers are incorporatedin drug development, regulatory approaches will be developedand will be more stringent than that needed to guide early drugdevelopment to ensure that the development, validation, andimplementation of biomarkers into clinical trials is a safe,effective, and efficient process.

Conclusions

The need for a standardized pathway approach toward thebiomarker validation process is becoming increasingly impor-tant given the recent surge in the biomarker developmentpipeline. As biomarker research progresses toward establishingthe fundamentals of personalized medicine, the future of drugand biomarker codevelopment resides in identifying the rightpopulation that would benefit from that drug. The significanteffort and resources that are invested in the development ofthese biomarkers will expand their roles as surrogate endpointsand diagnostic indicators for disease screening, monitoring





disease progression and treatment efficacy, and in assessingpatient outcome or identifying potential side effects such as intoxicity. The emphasis now would be placed on biomarkerassay development and method validation to eliminate thefailure of biomarkers that occur in the clinic as a result of poorassay choice and the lack of robust validation. Futureadvancements in biomarker research will be heavily focused

on transitioning biomarkers from the development andvalidation phases to their clinical applications incorporatedinto drug trials.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

References1. Park JW, Kerbel RS, Kelloff GJ, et al. Rationale for

biomarkers and surrogate end points in mechanism-driven oncology drug development. Clin Cancer Res2004;10:3885^96.2. Rifai N, Gillette MA, Carr SA. Protein biomarker dis-

covery and validation: the long and uncertain path toclinical utility. Nat Biotechnol 2006;24:971^83.3. Zerhouni E. Medicine. The NIH roadmap. Science

2003;302:63^72.4. Owzar K, Barry WT, Jung S-H, Sohn I, George

SL. Statistical challenges in preprocessing inmicroarray experiments in cancer. Clin CancerRes 2008;19:5959^66.5. Taylor JMG, Ankerst DP, Andridge RR.Validation of

biomarker-based risk prediction models. Clin CancerRes 2008;19:5977^83.6. Simon R.The use of genomics in clinical trial design.

Clin Cancer Res 2008;19:5984^93.7. George SL. Statistical issues in translational cancer

research. Clin Cancer Res 2008;19:5954^8.8. BiomarkersDefinitionsWorkingGroup.Biomarkersand

surrogate endpoints: preferred definitions and concep-tual framework.ClinPharmacolTher2001;69:89^95.9.Wagner JA. Overview of biomarkers and surrogate

endpoints in drug development. Dis Markers 2002;18:41^6.10. Lee JW, Devanarayan V, Barrett YC, et al. Fit-for-

purpose method development and validation forsuccessful biomarker measurement. Pharm Res2006;23:312^28.11. Goodsaid F, Frueh F. Process map proposal for the

validation of genomic biomarkers. Pharmacogenomics2006;7:773^82.12. LeightonJK,BrownP,EllisA,etal.Workgroupreport:

Review of genomics data based on experience withmock submissions-view of the CDER PharmacologyToxicology Nonclinical Pharmacogenomics Subcom-mittee.EnvironHealthPerspect 2006;114:573^8.13. Goodsaid F, Frueh F. Biomarker qualification pilot

process at the US Food and Drug Administration.AAPS J 2007;9:E105^8.14. Salerno RA, Lesko LJ. Pharmacogenomic data:

FDA voluntary and required submission guidance.Pharmacogenomics 2004;5:503^5.15. Freidlin B, Simon R. Adaptive signature design: an

adaptive clinical trial design for generating and pro-spectively testing a gene expression signature for sen-sitive patients. Clin Cancer Res 2005;11:7872^8.16. Lee JW,Weiner RS, Sailstad JM, et al. Method vali-

dation and measurement of biomarkers in nonclinicaland clinical samples in drug development: a confer-ence report. Pharm Res 2005;22:499^511.17. Kingsmore SF. Multiplexed protein measurement:

technologies and applications of protein and antibodyarrays. Nat Rev Drug Discov 2006;5:310^20.18. Ilyin SE, Belkowski SM, Plata-Salaman CR. Bio-

marker discovery and validation: technologies and in-tegrative approaches. Trends Biotechnol 2004;22:411^6.19. Miller KJ, Bowsher RR, Celniker A, et al.Workshop

on bioanalytical methods validation for macromole-cules: summary report. Pharm Res 2001;18:1373^83.20.NowatzkeW,Woolf E. Best practices duringbioana-

lytical method validation for the characterizationof as-say reagents and the evaluation of analyte stability inassay standards, quality controls, and study samples.AAPS J 2007;9:E117^22.21. ShahVP, Midha KK, Dighe S, et al. Analytical meth-

ods validation: bioavailability, bioequivalence andpharmacokinetic studies. Conference report. Eur JDrug Metab Pharmacokinet 1991;16:249^55.22. ShahVP, Midha KK, FindlayJW, et al. Bioanalytical

method validationDa revisit with a decade of prog-ress. Pharm Res 2000;17:1551^7.23. Smolec J, DeSilva B, SmithW, et al. Bioanalytical

method validation for macromolecules in support ofpharmacokineticstudies.PharmRes2005;22:1425^31.24. ColburnWA. Biomarkers in drug discovery and de-

velopment: from target identification through drugmarketing. JClin Pharmacol 2003;43:329^41.25. Findlay JW, Smith WC, Lee JW, et al.Validation of

immunoassays for bioanalysis: a pharmaceutical in-dustry perspective. J Pharm Biomed Anal 2000;21:1249^73.26. LeeJW, HulseJD, ColburnWA. Surrogatebiochem-

ical markers: precise measurement for strategic drugand biologics development. J Clin Pharmacol 1995;35:464^70.27. Lee JW, Nordblom GD, Smith WC, Bowsher RR.

Validation of bioanalytical assays for novel bio-markers: practical recommendations for clinical inves-tigation of new drug entities. In: Bloom J, editor.Biomarkers in clinical drug development. New York:Marcel Dekker; 2002.28. Dybkaer R.Vocabulary for use in measurement pro-

cedures and description of reference materials in labo-ratory medicine. Eur J Clin Chem Clin Biochem 1997;35:141^73.29.Veltri RW, Miller MC, An G. Standardization, analyt-

ical validation, and quality control of intermediate end-point biomarkers. Urology 2001;57:164^70.30. Cummings J, Ward TH, Greystoke A, Ranson M,

Dive C. Biomarker method validation in anticancer drugdevelopment. BrJPharmacol 2008;153:646^56.31. Bansal S, DeStefano A. Key elements of bioanalyti-

cal method validation for small molecules. AAPS J2007;9:E109^14.32. Yu LR, Veenstra TD. AACR-FDA-NCI cancer bio-

markers collaborative. Expert Rev Mol Diagn 2007;7:507^9.33. Slamon DJ, Clark GM,Wong SG, Levin WJ, Ullrich

A, McGuire WL. Human breast cancer: correlation ofrelapse and survival with amplification of the HER-2/neu oncogene. Science1987;235:177^82.34. Smith IE. Efficacyand safety of Herceptin in women

withmetastaticbreast cancer: results frompivotal clin-ical studies.AnticancerDrugs2001;12Suppl4:S3^10.35. Baselga J,Tripathy D, Mendelsohn J, et al. Phase II

study of weekly intravenous recombinant humanizedanti-p185HER2 monoclonal antibody in patients withHER2/neu-overexpressing metastatic breast cancer.JClin Oncol 1996;14:737^44.36. Cobleigh MA,Vogel CL,Tripathy D, et al. Multina-

tional study of the efficacy and safety of humanizedanti-HER2 monoclonal antibody in women who haveHER2-overexpressing metastatic breast cancer that

has progressed after chemotherapy for metastatic dis-ease. JClin Oncol 1999;17:2639^48.37. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of

chemotherapy plus a monoclonal antibody againstHER2 for metastatic breast cancer that overexpressesHER2. NEngl JMed 2001;344:783^92.38.Vogel CL, Cobleigh MA,Tripathy D, et al. Efficacy

and safety of trastuzumab as a single agent in first-linetreatment of HER2-overexpressing metastatic breastcancer. JClin Oncol 2002;20:719^26.39. DrukerBJ. Imatinib aloneandincombinationforchro-

nicmyeloidleukemia.SeminHematol2003;40:50^8.40. Demetri GD, von Mehren M, Blanke CD, et al. Effi-

cacy and safety of imatinib mesylate in advanced gas-trointestinal stromal tumors. N Engl J Med 2002;347:472^80.41. Cassier PA, Dufresne A, Arifi S, et al. Imatinib mesi-

late for the treatment of gastrointestinal stromal tu-mour. Expert Opin Pharmacother 2008;9:1211^22.42. Deininger MW, O’Brien SG, Ford JM, Druker BJ.

Practical management of patients with chronic mye-loid leukemia receiving imatinib. J Clin Oncol 2003;21:1637^47.43.White D, SaundersV, Grigg A, et al. Measurement ofin vivo BCR-ABL kinase inhibition to monitor imatinib-induced target blockade and predict response inchronic myeloid leukemia. J Clin Oncol 2007;25:4445^51.44.Wardelmann E, Buttner R, Merkelbach-Bruse S,

Schildhaus HU. Mutation analysis of gastrointestinalstromal tumors: increasing significance for risk as-sessment and effective targeted therapy. VirchowsArch 2007;451:743^9.45. Heinrich MC, Corless CL, Blanke CD, et al. Molecu-

lar correlates of imatinib resistance in gastrointestinalstromal tumors. JClin Oncol 2006;24:4764^74.46. Heinrich MC, Corless CL, Demetri GD, et al. Kinase

mutations and imatinib response in patients withmetastatic gastrointestinal stromal tumor. JClin Oncol2003;21:4342^9.47. Hornick JL, Fletcher CD.The role of KIT in the man-

agement of patients with gastrointestinal stromaltumors. Hum Pathol 2007;38:679^87.48. LasotaJ, Miettinen M. Clinical significance of onco-

genic KITand PDGFRA mutations in gastrointestinalstromal tumours. Histopathology 2008;53:245^66.49.GamelinE,DelvaR,JacobJ,etal. Individualfluoroura-

cil dose adjustment based on pharmacokinetic follow-up compared with conventional dosage: results of amulticenter randomized trial of patients with metastaticcolorectalcancer. JClinOncol 2008;26:2099^105.50. Sequist LV, Martins RG, Spigel D, et al. First-line

gefitinib in patients with advanced non-small-cell lungcancer harboring somatic EGFR mutations. J ClinOncol 2008;26:2442^9.51. Goncalves A, Esteyries S,Taylor-Smedra B, et al. A

polymorphism of EGFR extracellular domain is associ-ated with progression free-survival in metastatic colo-rectal cancer patients receiving cetuximab-basedtreatment. BMC Cancer 2008;8:169.52. LievreA, BachetJB, BoigeV, et al. KRAS mutations

as an independent prognostic factor in patients withadvanced colorectal cancer treated with cetuximab.JClin Oncol 2008;26:374^9.


CCR FOCUS



Documents

Validation of Analytic Methods for Biomarkers Used in Drug Development