establish the maximum tolerated dose in phase-i trials ... · aim maximum tolerated dose (mtd) the...

EstablishtheMaximumToleratedDoseinPhase-ITrialsusing3+3Method

Anup PillaiCytel,Pune,India

Vienna11th - 14th October2015

21October2015 1SP04PhUSE 2015

Ø Introduction to Phase-1 trials

Ø Dose Escalation Studies

Ø 3+3 design for findingMaximum Tolerated Dose

Ø Case Study for findingMaximum Tolerated Dose

Ø SAS® Macro for Simulating 3+3 Design

Ø Limitations of 3+3 Design

21October2015 2SP04PhUSE 2015

Agenda

PhasesofaClinicalTrialInitialtestingdone inlaborwithanimals

Phase- IV

Find thesafestdoseFindmosteffectivewaytoadministeradose

ConfirmatoryPhase

CheckingforSafetyandEffectiveness

PostMarketingSurveillance

21October2015 3SP04PhUSE 2015

Phase- III

Phase- II

Phase- I

PreClinical

Phase-ITrials

Aim MaximumToleratedDose(MTD)

Thehighestdoseofatreatmentthatdoesnotcauseunacceptablesideeffects.

MTD

DoseLimitingToxicity.Unacceptablesideeffectsortoxicity.

DLT

21October2015 4SP04PhUSE 2015

Phase-ITrials

• Phase-Itrialsarefirsttrialsconductedonhumans.

• Usuallythesetrialsincludehealthyvolunteers.Buttherearecircumstanceswhenrealpatientsareused,suchasoncologytrials.

21October2015 5SP04PhUSE 2015

• Anincreaseddoseisassociatedwithincreasedchanceofclinicalefficacy.

• PhaseItrialsaredesignedasadose-escalationstudytodeterminetheMTD.

Phase-ITrials

Dose

Respon

se

Efficacy

Toxicity

21October2015 6SP04PhUSE 2015

DoseEscalationStudies• Minimize the number of patients exposed to toxic doses,

while identifyingthe MTD.

• Dose escalation methods fall into two broad classes:– Rule Based Design– Model Based Design

• Rule-based designs allow dose escalation and de-escalationdepending on the absence or presence of DLTs in the previouscohort of treated subjects.

• The most widely used rule based design is the 3+3 design.

21October2015 7SP04PhUSE 2015

Treat3SubjectsonStartingDosei

Enroll3moreSubjectsonDosei

0DLT >1DLT1DLT

EscalatetoDosei+1

De-escalatetoDosei-1

1/6DLT >1/6DLT

3+3DesignAlgorithmofatraditional3+3Design

21October2015 8SP04PhUSE 2015

CaseStudy• A study was conducted to determine the MTD of HB-110, a

vaccine administered by Electroporation in chronic hepatitis Bpatients.

• The 3+3 design was used to reach the MTD.– Subjects were observed for a minimumof 28 days.– Each subject was administered HB-110 per day.

• The dose-levels of HB-110 used were 1mg, 2mg, 4mg & 6mg.

21October2015 9SP04PhUSE 2015

SubjectID

DoseLevel(m

g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

6

4

2

1

DLT

MTD

PerSubjectResponseinthetrial

21October2015 10SP04PhUSE 2015

CaseStudy

NoDLT

SubjectID

DoseLevel(m

g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

6

4

2

1

DLT

UnabletofindtheMTD

21October2015 11SP04PhUSE 2015

CaseStudy

NoDLT

MTDbelowLowestDose

SubjectID

DoseLevel(m

g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

6

4

2

1

DLT

UnabletofindtheMTD

21October2015 12SP04PhUSE 2015

CaseStudy

NoDLT

MTDaboveHighestDose

%MTD_3x3(treatment=1,no_sim =1000,sample_siz =30);

21October2015 SP04PhUSE 2015 13

SASMacroThemacrosimulatesa3+3Design

SerialNumber

DoseValue Probabilityofobserving aDLT

StartingDose Numberof simulations MaximumSampleSize

Inputdataset:‘dose_escalation’

21October2015 SP04PhUSE 2015 14

SASMacro

Outputdataset:‘Simulation_summary’

• Thedesignisinflexible.

• Decisionsarenotbasedonoutcomesfromallrecruitedsubjects.

• ManysubjectsaretreatedatdoseslowerthanMTDwhilefewsubjectsactuallyreceivetheMTD.

TheselimitationsareovercomebymodelbaseddesignslikeCRM(ContinualReassessmentMethod)BLRM(BayesianLogisticRegressionMethod)

21October2015 15SP04PhUSE 2015

Limitations

• 3+3 remains the most popular method because of its simpleconcept and operational ease.

• It can be implemented without any complex statisticalconsiderationsand computations.

• 3+3 design is used as a starting step for carrying out morecomplex designs.

Summary

21October2015 16SP04PhUSE 2015

21October2015 SP04PhUSE 2015 17

ReferencesLeTourneau,Christophe,J.JackLee,andLillianL.Siu."Doseescalationmethods inphaseIcancerclinicaltrials."JournaloftheNationalCancerInstitute (2009).

Storer,BarryE."DesignandanalysisofphaseIclinicaltrials."Biometrics(1989):925-937.

Neuenschwander,Beat,MichaelBranson,andThomasGsponer."CriticalaspectsoftheBayesianapproachtophaseIcancertrials." Statisticsinmedicine 27.13(2008):2420-2439.

Thall,P.F.,andS-J.Lee."Practicalmodel-baseddose-finding inphaseIclinicaltrials:Methodsbasedontoxicity." InternationalJournalofGynecologicalCancer 13.3(2003):251-261.

Zohar,Sarah,andSylvieChevret."Thecontinualreassessmentmethod:comparisonofBayesianstoppingrulesfordose-rangingstudies." Statistics inmedicine 20.19(2001):2827-2843.

O'Quigley,John,MargaretPepe,andLloydFisher."Continualreassessmentmethod:apracticaldesignforphase1clinicaltrialsincancer."Biometrics(1990):33-48.

CaseStudy:“Tolerability,ImmunogenicityandEfficacyofHB-110AdministeredbyElectroporationinChronicHepatitisBPatients.”https://clinicaltrials.gov

21October2015 18SP04PhUSE 2015

THANKYOU

QUESTIONS