BridgingStudiesAGenomicApproach專題培訓課件

2023/10/101BridgingStudies-AGenomicApproach

atThe34TrainingCourseonClinicalTrialsFoundationofMedicalProfessionalsAllianceinTaiwan

October14,2023TheviewsexpressedinthispaperareprofessionalopinionsofthepresenterandmaynotnecessarilyrepresentthepositionoftheNationalTaiwanUniversityandNationalHealthResearchInstitutes,Taiwan.2023/10/102AcknowledgementsandThanks

Herng-DerChen,MD,PhD,CenterforDrugEvaluation

MeyWang,PhD,CenterforDrugEvaluationChin-FuHsiao,PhD,NationalHealthResearchInstitutes2023/10/103OutlineI.StatisticalInterpretationofICHE5II.ImplementationofBridgingStudiesIII.ExamplesofBridgingStudiesIV.CurrentStatisticalApproachesV.AStatisticalGenomicApproachVI.Summary2023/10/104StatisticalInterpretationofICHE5ICHE5

ICHHarmonisedTripartiteGuideline(Feb.5,1998)EthnicFactorsintheAcceptabilityofForeignClinicalData

AUSFDAGuidance(FederalRegister,June10,1998)2023/10/105StatisticalInterpretationofICHE5ICHE5EthnicFactorsintheAcceptabilityofForeignClinicalDataThepurposeofthisguidanceistofacilitatetheregistrationofmedicinesamongICHregionsbyrecommendingaframeworkforevaluatingtheimpactofethnicfactorsuponamedicine’seffect,i.e.,itsefficacyandsafetyataparticulardosageanddoseregimen.2023/10/106StatisticalInterpretationofICHE5ObjectivesofICHE5(Section1.1)Todescribethecharacteristicsofforeignclinicaldatathatwillfacilitatetheirextrapolationtodifferentpopulationandsupporttheiracceptanceasabasisforregistrationofamedicineinanewregion.Todescriberegulatorystrategiesthatminimizeduplicationofclinicaldataandfacilitateacceptanceofforeignclinicaldatainthenewregion. Todescribetheuseofbridgingstudy,whennecessary,toallowextrapolationofforeignclinicaldatatoanewregion.Todescribedevelopmentstrategiescapableofcharacterizingethnicfactorinfluencesonsafety,efficacyanddoseregion2023/10/107StatisticalInterpretationofICHE5ICHE5EthnicFactorsintheAcceptabilityofForeignDataBRIDGINGDATAPACKAGE(Section3.2)

Abridgingdatapackageconsistsof 1)selectedinformationfromtheCompleteDataClinicalPackage

thatisrelevanttothepopulationofthenewregion,including pharmacokineticdata,andanypreliminarypharmacodynamic anddose-responsedata,

and 2)ifneeded,abridgingstudytoextrapolatetheforeignefficacy

and/orsafetydatatothenewregion.2023/10/108StatisticalInterpretationofICHE5CompleteClinicalDataPackage(CCDP)

Aclinicaldatapackageintendedforregistrationcontainingclinicaldatathatfulfillsregulatoryrequirementsofthenewregionandpharmacokineticdatarelevanttothepopulationinthenewregion

2023/10/109StatisticalInterpretationofICHE5BridgingStudy

Abridgingstudyisdefinedasastudyperformedinthenewregiontoprovidepharmacodynamicorclinicaldataonefficacy,safety,dosageanddoseregimeninthenewregionthatwillallowextrapolationoftheforeignclinicaldatatothepopulationinthenewregion2023/10/1010

ExtrapolationandSimilarity

ICHE5EthnicFactorsintheAcceptabilityofForeignDataIfthebridgingstudyshowsthatdoseresponse,safetyandefficacyinthenewregionaresimilar,thenthestudyisreadilyinterpretedascapableof"bridging"theforeigndataIfabridgingstudy,properlyexecuted,indicatesthatadifferentdoseinthenewregionresultsinasafetyandefficacyprofilethatisnotsubstantiallydifferent

fromthatderivedintheoriginalregion,itwilloftenbepossibletoextrapolatetheforeigndatatothenewregion,withappropriatedoseadjustment,ifthiscanbeadequatelyjustified(e.g.,bypharmacokineticand/orpharmacodynamicdata).2023/10/1011EthnicFactors

IntrinsicEthnicFactors

aremoregeneticandphysiologicinnature e.g.,geneticpolymorphism,age,gender,height,weight,leanbodymass,bodycomposition,anddiseaseconditions,etc.ExtrinsicEthnicFactors

aremoresocialandculturalinnature e.g.,environment,culture,medicalpractice,healthinsurance,practicesinclinicaltrialsorconduct2023/10/1012Anapprovedreportofalocalclinicaltrialstudy

is

requiredforthenewdrugapplicationinTaiwan—July7.1993:Double7AnnouncementDisadvantages:

Aminimalsamplesizeof40asrequiredcouldbe

difficulttoprovideconclusiveandsubstantialevidenceofefficacyandsafetyThestudydesignofthelocaltrialusuallyonly

repeatedastudythathasbeendoneintheforeigncountriesbutinasmallersamplesize;ThestudyhasnotbeendesignedbasedonthemedicalsituationinTaiwan

Taiwan’sStrategytoImplementBridgingStudy2023/10/1013Taiwan’sStrategytoImplementBridgingStudySmoothlyconvertcompulsoryLocalClinicalTrials(LCT)tomeaningfulbridgingstudiesGradually,stepwiseannouncewaivedlocalclinicaltrialsCreateanenvironment:(1)meetinternationalregulations,ICH(2)requireoptimizeddosageforTaiwanesepatientsCommunicatewithlocalandinternationalpharmaceuticalindustryAnnouncenewregulationsaccordingtotheinternationalnormandtheconsensusfromcommunicationsCreateaninternationalplatform“APEC–Taipei”ImplementDoubleTwelveAnnouncement–BridgingStudy

2023/10/1014StepwiseImplementationTwoyearsafterthe1998announcement,switchfromLCTtobridgingstudyManycommunicationsandnegotiationswithlocalandinternationalpharmaceuticalindustryDec.12,

2023:(DoubleTwelveAnnouncement)–publicannouncementofthebridgingstudyregulations1998:FiveannouncementsofLCTwavierAtwo-yeartransitionperiod:bothLCTandbridgingstudiesconcurrentlyacceptablefrom2023~2023ManyinternationalconferencesheldinTaipeiandotherAsiancountries,regardingBS,throughtheAPECplatformConsultwithCDEtocompletethepracticalissuesrelatedtoimplementationofBSJan.1,

2023:Bridgingevaluation

2023/10/1015

ProductsRequiringNoVerificationofEthnicInsensitivity

DrugsfortreatmentofAIDSDrugsfororgantransplantationTopicalagentsNutritionsupplementsCatharticspriortosurgeryRadio-labelleddiagnosticpharmaceuticalsThedrugistheonlychoiceoftreatmentforagivensevere

diseaseDrugsforlife-threateningdiseasehavedemonstratedabreakthroughefficacyLackingadequatetrialsubjectsforanydrugusedforraredisease2023/10/1016ProductsRequiringVerificationofEthnicInsensitivityAnticancerdrugsDrugswithbreakthroughefficacyDrugsofsingleuseDrugswithdifferentsaltofthesamecompositionandthesameadministeredroutehavebeenapprovedinternalDrugsforchronicpsychologicalorimmunologicaldiseasesandconductingclinicaltrailsinternaldifficultlyEachcompoundsofnewcombinationdrughavebeenprovedinternal,andtheefficacyisthesameasthesinglecompoundDrugswiththemechanism,administeredroute,efficacyandadverseeffect,similartotheapproveddrugsNewcombinationcomposedofsinglecompoundofapprovedcombinationorcompoundsofapprovedcombinationhasthesameefficacyasapprovedcombination2023/10/1017BridgingDataPackageSummaryfortheConsiderationofBridgingStudyAcceptsubmissionCheckingListTechnicalReview(Designatereviewer)ReviewmeetingSponsormeetingSupplementClinicalReviewCommitteeReviewreportandRecommendation:1.NoBridgingstudyrequired2.Bridgingstudyisrequired–TypeofBridgingstudyResultofEvaluation:1.NoBridgingstudyrequired2.Bridgingstudyisrequired–TypeofBridgingstudyNotificationSponsorBoPACDECDEacceptanceverificationExpertConsultants(Statistical,Clinical,Pharmacokineticsreviewers)ScheduleSponsormeetingEvaluationProcess2023/10/10182023/10/10192023/10/1020

ExamplesExampleIDrugAisafixedcombinationoftwoanti-plateletagentswithindicationforsecondarypreventionofthromboembolicstrokeAfterthestandardprocessofBSE,wedecidedtorequestabridgingstudyduetoanethnicdifferenceinmedicalpractice(muchlowerdoseforoneofthecomponentsinTaiwan)andhigherheadache-associateddropoutrateinpreviousPhilippinestudy2023/10/1021Examples

CaseIFixedcombination200mgdipyridamole/25mgaspirin1bidforpreventionofrecurrentstrokeHeadachedropoutrate:Phillipino>CaucasianLocalBridgingStudyResult:first4weeks

ReducedDose2wkFullDose

PlaceboFullDose2wk4wk

Headache8.7%6.7%16.3%

RiskManagement:Changelabeling’sinstructionforuse2023/10/1022ExamplesCaseIIDrugBisanewpotentlipid-loweringagentThePKstudyinJapaneseshowsthatCmaxofJapaneseis1.9~2.5timesofthatforCaucasianwhileAUCis2~2.5timesofthatforCaucasianAlthoughthemeaninterracialdifferenceisnotsubstantial,Taiwanapprovedthedrugwithreducedmaximaldosageduetothedose-dependent,drug-relatedrareSAEofrhabdomyolysis2023/10/1023Examples

CaseIIThedecisionwasfurtherechoedbyUSFDAAfterreviewingtheresultsofaPhaseIVPKstudyinAsian-Americans,FDAurgedthephysicianstoreducethestartingdoseandprescribehighdosewithcautionforAsiansinLabelinginMarch,20232023/10/1024CurrentStatisticalApproaches

“Similarity”

PositiveRxeffectEquivalenceNon-inferiority2023/10/1025ExtrapolationandSimilarityPositivetreatmenteffect(betterthancontrol)

Theefficacyorsafetyofthetestdrugisbetterthancontrolinthenewregion Ho:NT-NC

0vs.Ha:NT-NC>0CurrentStatisticalApproaches2023/10/1026

CurrentStatisticalApproaches

PositiveRxEffect

0

NDOD

2023/10/1027

CurrentStatisticalApproaches

Similarity(Nosubstantialdifference)

Two-sidedequivalence

Therelativeefficacyorsafety(test-control)ofthenewregioniswithinsomeclinicallyacceptancelimitofthatoftheoriginalregion

Let=(NT-NC)-(OT-OC)

Ho:

-or

vs. Ha:-<<

whereissomeclinicallyacceptablelimit(equivalencelimit).

2023/10/1028

CurrentStatisticalApproaches

Equivalence

0

NDOD

2023/10/1029

CurrentStatisticalApproaches

Similarity(Nosubstantialdifference)

One-sidednon-inferiority

Therelativeefficacyorsafety(test-control)ofthenewregionisnotinferiortotheoriginalregionbysomeclinicallyacceptancelimit.

Ho:

-vs.Ha:>-.2023/10/1030

CurrentStatisticalApproaches

Non-inferiority

0

NDOD

2023/10/1031

CurrentStatisticalApproaches

Between-studyAnalysis:Equivalenceornon-inferiority

HierarchicalModel(Liu,Hsueh,andChen2023,BiometricalJ.)

Step1:Fromthecompleteclinicaldatapackage,underthe hierarchicalmodel,usetheclinicaldatafromtheoriginalregionto obtaintheestimateofrelativeefficacyanditsestimatedstandard error. Step2:Fromthedataofthebridgingstudy,obtaintheestimateof relativeefficacyanditsestimatedstandarderrorinthenewregion. Step3:Basedontheestimatedrelativeefficacyanditsstandard errorfromboththenewandoriginalregionsandequivalencelimit ,performtheusualtwoone-sidedtestsprocedureorone-sided non-inferioritytestprocedure(orconfidenceinterval).2023/10/1032EmpiricalBayesianApproach

Bridgingstudies

Smallsamplesize

Needtoborrow“strength”fromCCDPoforiginalregion.

Informationondoseresponse,efficacyandsafetyofthe originalregioncanandshouldbeincorporatedina statisticallysoundmannertoevaluatebridgingevidencebythebridgingstudiesinthenewregion.

Positivetreatmenteffect:(Liu,Hsiao,andHsueh,2023,JBS)

Noninferiorityapproach:(Liu,Hsueh,andHsiao2023,JBS)CurrentStatisticalApproaches2023/10/1033

Between-studyAnalysis:BayesianApproach(Liu,Hsiao,andHsueh,2023)

UsetheestimateoftreatmenteffectfromtheoriginalregionformulatedasanormalpriorComputetheposteriortreatmenteffectwiththedatafromthenewregionComputetheposteriorprobabilityofsimilarity,Pspastheposteriorprobabilityofapositivetreatmenteffect

Concludetheresultsoftheforeignregioncanbeextrapolatedtothenewregionif

Pspis

sufficientlylarge.

SamplesizemightbedeterminedbasedonthedifferencebetweentheposteriorandpriortreatmenteffectCurrentStatisticalApproaches2023/10/1034

CurrentStatisticalApproaches ForapositivetreatmenteffectThemodel(Liu,Hsiao,andHsueh,2023)Yi=P(1-Xi)+{Nzi+O(1-Zi)}Xi+i,I~N(0,2),where

P:controleffectN:treatmenteffectofthenewregionO:treatmenteffectoftheoriginalregionX=1(0)treatment(control)groupZ=1(0)new(originalregion)2023/10/1035CurrentStatisticalApproachesForapositivetreatmenteffectEmpiricalBayesianApproach

(Liu,Hsiao,andHsueh,2023)GivenP,N,theestimatesofP,N,saypandaNfollowsabivariatenormaldistributionwithmeanvector(P,N)’andthediagonalcovariancematrixdiag(VP,VN).Inaddition,thepriordistributionof(P,N,O)’followsatrivariatenormaldistributionwithmeanvector(p,N,O)anddiagonalcovariancematrixdiag(2p,2N,2O).Concludeapositivetreatmenteffectifposteriorprobabilityofsimilarity

P(N-P>0dataandprior)>1-,forsome>0.5.(Simon,1999)

2023/10/1036CurrentStatisticalApproachesForapositivetreatmenteffectEmpiricalBayesianApproach

(Liu,Hsiao,Hsueh,2023)UnderassumptionthatN=O,forP(N-P>0dataandprior)>1-,thefollowingequationmustbesatisfied (O-p)/2O+2p

=-1(1-q),where-1(1-q)representstheevidenceforpositivetreatmenteffectintheoriginalregion. 2023/10/1037

CurrentStatisticalApproaches

ResultsfromOriginalRegionChangefrombaselineinsittingDBPatweek12

RegionStatistics Test PlaceboI n 138 132 Mean -18 -3 SD 11 12II n 185 179Mean -17 -2 SD 10 11III n 141 143 Mean -15 -5 SD 13 14

2023/10/1038CurrentStatisticalApproaches

ResultsfromNewRegion:ChangefrombaselineinSittingDBPatweek12

RegionStatistics Test PlaceboNew n 64 65 Mean -4.5 -3.8 SD 11 11Posteriorprobabilityofsimilarity:Psp12023/10/1039

CurrentStatisticalApproaches

Originalregion:Efficacyofthetestdrugissuperiortotheplacebo.NewRegion:ReductionofsittingBPofthetestdrugissameasthatoftheplacebo.Conclusion:Theresultsoftheoriginalregioncanbeextrapolatedtothenewregiondespiteofinconsistentresultsbetweenoriginalandnewregions.Evaluationofbridgingstudiesisoverwhelminglybytheresultsoforiginalregionduetoimbalanceofinformationprovidedbythetworegions.

.2023/10/1040

CurrentStatisticalApproaches

AMixturePriorBayesianApproach(Hsiao,HsuandLiu,2023)Define

N=NT-NC

Amixture

priormodel:(N)=1(N)+(1-)2(N)

1(N)isanoninformativepriorandissettobe1.2(N)isanormalpriorthatsummarizestheresultsoforiginalregionisaweighingfactor;01=0:thesamepriorusedinLiu,HsuehandHsiao(2023)=1:noresultsoforiginalregionisused

2023/10/1041

CurrentStatisticalApproaches

PosteriorProbabilityofSimilarity2023/10/1042

CurrentStatisticalApproaches

MarginalDistribution2023/10/1043

CurrentStatisticalApproaches

PosteriorDistribution2023/10/1044

CurrentStatisticalApproaches

FindthesmallestnNthat2023/10/1045

CurrentStatisticalApproaches

GroupSequentialMethod(Hsiao,Xu,Liu,2023)ATwo-stageDesign(Hsiao,Xu,Liu,2023)

Reason: UnderthehierarchicalmodelorBayesianapproach, evaluationofsimilarityornon-inferioritybasedonthe differenceofrelativeefficacymightstillrequirelarge samplesizeforthebridgingstudyinthenewregion. Thesearebetween-studyanalysiswithoutinternalvalidityandmayprovidebiasedinference

Criterionforsimilarity Maintenanceofthesimilartrendinthenewregion2023/10/1046

CurrentStatisticalApproaches

Step1:Whendesigningtheadequateandwell-controlled studiesforsubmissiontotheoriginalregion,includethe patientsinthenewregionaspartofrecruitmentforthe wholestudy(Thebridgingstudyisasub-study).Step2:Thestudyshouldhaveastructureofgroupsequential design.Usetheregionasgroupsequencetoenrollthe patientsfromtheoriginalregionfirstandthentoenroll patientsfromthenewregionsubsequently.Step3:Pre-specifytheboundariesintheprotocol,sayalpha- spendingfunction.Becausetheprimaryobjectiveofthe trialisforsubmissiontotheoriginalregion,mostoftypeI errorrateshouldbespentfortheinterimanalysisbasedon theresultsfromtheoriginalregion.2023/10/1047

CurrentStatisticalApproaches

Step4:Whentherecruitmentofpatientsintheoriginalregionis completed,performtheinterimanalysisuptoresultsofthe originalregion.Step5:Enrollthepatientsinthenewregion.Afterthe recruitmentofthepatientsiscompleted,performthefinal analysiswithadditionaldatafromthenewregionand adjustmentoftheinterimanalysis.Ifsimilarresults(i.e., similarsignificanceleveltomeetrequirementofcrossing boundary)areobtainedforthefinalanalysis,thentheresults ofthenewregioncanbedeclaredsimilaritytotheoriginal region.2023/10/1048TargetedClinicalTrialsHER2(thehumanepidermalgrowthfactorreceptor2)geneinmetastaticbreastcancer-Herceptin-requirementofscreeningthepatientswithover-expressedHER2level(Slamon,2023).Estrogenreceptorpolymorphism-EstrogenReplacementAtherosclerosistrial(ERA,Herrington,etal,2023):atotalof9SNPswereidentifiedandinteractionbetweentreatmentofHRTandsomeofSNPsinelevationoflipidlevelsissuggestedSamplesizedetermination:Fijal,etal.(2023)andMaitournamandSimon(2023).AStatisticalGenomicApproach

2023/10/1049TargetedClinicalTrialsandEGFRIressa(gefitnib)andTarceva(Erlotinib)aretargtedattheEGFRpathway.Efficacyiscorrelatedto race numberofgenecopiesproteinexpressionEGFRmutationGappuzzoetal.(JNCI,2023),Tsao,etal(NEJM,2023)AStatisticalGenomicApproach2023/10/1050AStatisticalGenomicApproachAsian(n=342)

HR=0.66(0.48,0.91),P=.011RR=12.0%Non-Asian(n=1350)

HR=0.93(0.81,1.08),P=.364RR=6.5% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 151.00.90.80.70.60.50.40.30.20.10.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15——

IRESSA?------

PlaceboTime,moPatientssurviving(%)2023/10/1051

From:Tsao,etal(2023,NEJM)AStatisticalGenomicApproach2023/10/1052From:Tsao,etal(2023,NEJM)AStatisticalGenomicApproach2023/10/1053From:Tsao,etal(2023,NEJM)AStatisticalGenomicApproach2023/10/1054From:Tsao,etal(2023,NEJM)AStatisticalGenomicApproach2023/10/1055

AStatisticalGenomicApproach

CurrentstatisticalmethodsforbridgingstudiesdonotreallytakeethnicfactorsintoconsiderationsAfterHumanGenomeProject,theavailabilityofgenomicdatacanprovidethenecessaryquantitativeinformationforintrinsicethnicfactorGenomicinformationshouldbeincorporatedintoevaluationofbridgingstudiesBridgingstudiesmaybeconsideredasonetypeoftargetedclinicaltrialswithgenomicdataasthebio-targetsforintrinsicethnicfactors2023/10/1056

AStatisticalGenomicApproach

StratifiedApproach

OriginalRegion Genetic

Polymorphism Proportion Test Control

1 PO1

1T 1C

2 PO2

2T 2C . . . .. . . .

K POK

KT KC

POiiT POiiC

2023/10/1057

AStatisticalGenomicApproach

StratifiedApproach

NewRegion

Genetic

Polymorphism Proportion Test Control

1 PN1

1T 1C

2 PN2

2T 2C . . . .. . . .

K PNK

KT

KC

PNiiT PNiiC

2023/10/1058

AStatisticalGenomicApproachPoi:theproportionoftypeiforsomegeneticpolymorphismintheoriginalregionPNi:theproportionoftypeiforsomegeneticpolymorphisminthenewregioniT:themeanresponseofapatientwithtypeipolymorphismintreatmentgroup2C:themeanresponseofapatientwithtypeipolymorphismincontrolgroupOT=POiiT:themeanresponseofthetreatmentgroupintheoriginalregion(meanofamixturedistribution)OC=POiiC:themeanresponseofthecontrolgroupintheoriginalregionNT=P