Endovascular Therapy Following Imaging Evaluation for ... · Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3 Protocol Version/Version Date Version 2.4 April

Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3

Study Protocol

NCT02586415

April 20, 2017

Gregory W. Albers, MD, Principal Investigator Stanford University

Stanford, California 94305

DEFUSE 3 clinical protocol Version 2.4; 20 APR 2017

PROTOCOLTITLE

DEFUSE3:EndovascularTherapyFollowingImagingEvaluationforIschemicStroke3

ProtocolVersion/VersionDate

Version2.4

April20,2017

ProtocolDirectors

GregoryAlbers,MDMichaelMarks,MD

MaartenLansberg,MD,PhD

StanfordUniversityStanfordStrokeCenter

780WelchRd.SuiteCJ350Stanford,CA94304

Supportedby

TheNationalInstituteofNeurologicalDisordersandStroke(NINDS)

U01NS092076

IDENumber


DEFUSE3

AGREEMENTONTHEPROTOCOL

BysigningbelowIconfirmthat:

1) IhavereadthisprotocolanditcontainsallnecessarydetailsforconductingthisstudyAND

2) Iagreetoconductthetrialincompliancewiththisprotocolandtoadheretoallregulationsthatgoverntheconductofthestudy.

________________________________________________________ _______________________PrincipalInvestigator’sSignature Date________________________________________________________PrincipalInvestigator’sName

_________________________________________________________SiteName:


1

TABLEOFCONTENTS

TableofContents…………………………………………………………………………………………… P.1

1. SummaryofTrial……….…………………………………………………………… P.2

2.ScientificBackground…………………………………………………………………………. P.2

2.1StateoftheScienceonEndovascularStrokeTherapyforstroke…….. P.2

2.2PriorStudiesandrationalefordevelopment…………………………………. P.3

3. InvestigationalPlan………………………….……………………………………………….. P.6

3.1Purpose………………………………………………………………………….. P.6

3.2ProtocolDesign………………………………………………………………. P.6

3.3EnrollmentCriteria………………………………………………………….. P.7

3.3.1ClinicalInclusionCriteria…………………………………. P.7

3.3.2ClinicalExclusionCriteria………………………………… P.7

3.3.3NeuroimagingInclusionCriteria………………………. P.8

3.3.4NeuroimagingExclusionCriteria……………………… P.8

3.4EnrollmentandRandomization……………………………………….. P.9

3.4.1Enrollment………………………………………………………. P.9

3.4.2Randomization………………………………………………… P.10

3.5AcuteTreatment……………………………………………………………… P.10

3.5.1EndovascularTherapy……………………………………… P.10

3.5.2MedicalTherapy………………………………………………. P.11

3.6ClinicalandImagingEvaluations……………………………………… P.11

3.6.1Studyassessmentsandfollow‐upvisits…………….. P.12

3.7SiteApprovalandMonitoringPlan……………………………………. P.12

3.8Samplesize,AdaptiveDesignandStatisticalAnalysis………… P.14

3.9RiskAnalysis……………………………………………...……………………. P.18

3.10DeviceDescription………………………………………………………… P.21

3.11MonitoringProceduresandAdverseEventReporting.……. P.22

4.Investigator’sAgreement&CurrentInvestigators……..……………………..….. P.25

5.ExecutiveCommittee/Keyinvestigators…………………………………………...... P.27

6.InstitutionalReviewBoard………………………………………………………………….. P.27

7.Costs…………………………………………………………………………………………………… P.27

8.References…………………………………………………………………………………………... P.28

9.Appendix:PatientInformedConsentForm...…………………………………………. P.32


2

1. SUMMARYOFTRIALDEFUSE3isaprospective,randomized,multi‐center,PhaseIII,adaptive,blindedendpoint,controlledtrial.Amaximumof476patientswillberandomizedandtreatedbetween6and16hoursofsymptomonset.Subjectswillberandomized1:1toendovasculartherapyplusmedicalmanagementormedicalmanagementalone.OnlythedeviceslistedinthisprotocolareapprovedforuseinDEFUSE3.Thechoiceofdeviceordevicesemployedisatthediscretionoftheclinicalinvestigator.

2. SCIENTIFICBACKGROUNDAlthoughstrokeisthenumberonecauseofadultdisabilityintheUnitedStates1,treatmentoptionsforstrokearelimited.TheonlyFDAapprovedtreatmentforstrokeisadministrationofintravenous(iv)tissueplasminogenactivator(tPA)within3hoursaftersymptomonset.Nationwide,onlyabout4percentofstrokepatientsreceivethistherapy.2Themainreasonforthislowtreatmentrateisthatmostpatientspresenttothehospitaloutsidethetime‐windowfortPA.3,4Evenwhenadministered,tPAisoftennoteffectivebecauseiteitherfailstorecanalizetheoccludedartery5‐7orbecausethebrainisalreadyirreversiblyinjured.8Asaresult,itisestimatedthatonly12‐25%oftreatedpatientsbenefitfromtPA.9Thus,inordertoimproveoutcomesfromstroke,weneedbettertreatmentsthatareavailabletoagreaterproportionofstrokepatients.

2.1StateofthescienceonendovascularstroketherapyforacutestrokeEndovascularstroketherapy,theremovalofbloodclotswithmechanicaldevicesorthrombolyticdrugsadministeredintra‐arterially,isthemostpromisingnewtreatmentforpatientswho“fail”treatmentwithivtPAorarenoteligibleforivtPA.Themainadvantageofendovasculartherapyisthatithasahighrateofrecanalization.10,11Bloodflowcanberestoredwithasuccessrateofupto82%withmodernthrombectomydevicesand66%forintra‐arterialthrombolysis.10,12,13ThisisapproximatelytwiceaseffectiveasivtPAwhichhasarecanalizationrateof10‐50%dependingonthelocationofthebloodclot.6,7Despitehigherratesofrecanalizationwithendovasculartherapy,tworecentrandomizedcontrolledtrialsofendovasculartherapy,IMSIIIandMRRESCUE,havefailedtodemonstrateaclinicalbenefit.Patientandtreatmentrelatedfactorslikelybothcontributedtotheneutralresultsofthesetrials(seebelowfordetails).

Patient‐relatedfactors:Acentralconsiderationintheoptimizationofpatientselectionforacutestroketherapiesistheconceptoftheischemicpenumbra.Ischemicpenumbraisdefinedasischemictissuethatispotentiallysalvageableandisdistinguishedfromtheischemiccorethathasalreadysustainedirreversibleinjury.Clearly,thetargetofacutestroketherapiesissalvageoftheischemicpenumbra,preventinginfarctgrowthand,mostimportantly,improvedfunctionaloutcome.Acutestroketrialsshouldthereforeideallybelimitedtopatientswithanischemicpenumbra.MRI‐basedstudies,suchasDEFUSE1and2,indicatethatMRIcanbeusedtoidentifythesepatients.8,14,1516,17

Treatment‐relatedfactors:Recentstudieshaveemphasizedtheimportanceofrecanalizationrates,demonstratingtheinfluenceonpatientoutcomeofhighlyeffectiveendovascularproceduresthatleadtocompleteornear‐completereperfusion.18,19Currently,themostcommonmetricforratingthequalityofreperfusionisthemodifiedThrombolysis


3

InCerebralInfarction(mTICI)scale,andaclearrelationshipexistsbetweenthedegreeofreperfusiononthemTICIscaleandpatientoutcome.20‐22Patientswith>50%reperfusion(mTICI2B‐3)aremuchmorelikelytohaveagoodoutcomethanpatientswith<50%reperfusion.

IMSIII,thelargestendovasculartrialtodate,didnotuseadvancedimagingcriteriatoselectpatients.23Insteaditusedrelativelystricttime‐criteria,anticipatingthatthiswouldyieldahighproportionofpatientswithasubstantialpenumbra.However,severalcategoriesofpatientswholikelydidnothavesubstantialpenumbrawereenrolled.Nearlyonethirdofthepatientsdidnothaveavesselocclusionatangiographyand23%haddistalMCAocclusions;bothofthesesubgroupsareunlikelytohavesubstantialpenumbraltissue.Inaddition,42%ofpatientshadsomeevidenceofirreversibletissueinjury(ASPECTS<8)ontheirbaselineCTand14%hadevidenceofextensiveirreversibleinjury(ASPECTS<5).24Finally,theendovasculardevicesthatwereavailableduringIMSIIIhadrelativelylowratesofearlyreperfusion;mTICI2B‐3wasonly40%inIMSIII.MRRESCUE,astroketrialthataimedtodemonstratebenefitofendovasculartherapyinpatientswithapenumbrabasedonMRI,hadneutralresults.25Severalfactorslikelycontributedtothis.First,therateofendovascularreperfusionwasextremelylow.Only8patients(24%)intheMRRESCUEpenumbralgroupachievedTICI2B‐3reperfusionduringendovasculartherapy.Second,patientsinthepenumbralgroupinMRRESCUEhadlargerbaselineinfarctcorelesions(medianvolume36ml;IQR24–51ml)thantheTargetMismatchpatientsinDEFUSE2(medianvolume13ml;IQR5–26ml).Thecombinationoflowratesofendovascularreperfusionandrelativelylargecorelesions,bothstrongpredictorsofpoorclinicaloutcome,likelyexplainsthelackofatreatmenteffectin“penumbralpatients”inMRRESCUE.Moreover,withonly8endovascularpatientsinthepenumbralcohortachievingTICI2B‐3reperfusion,MRRESCUEwassubstantiallyunderpowered.NewGenerationTrials:Recently,aseriesofpositiverandomizedstudiesofendovasculartherapywithtreatmentinitiatedwithin6hoursofstrokeonsetinthevastmajorityofpatientswerereported.Thishaspromptednewguidelinesendorsingendovasculartherapyupto6hoursaftersymptomonset.TheAmericanHeartAssociationisnowcallingforlatewindowstudiesusingadvancedimagingforpatientselection:“"Furtherrandomized,controlledtrialsshouldbedonetodeterminewhetheradvancedimagingparadigmsusingCTperfusionandMRIperfusion,CTA,anddiffusionimaging,includingmeasuresofinfarctcore,collateralflowstatus,andpenumbra,arebeneficialforselectingpatientsforacutereperfusiontherapywhoarebeyond6hoursfromsymptomonset.(Newrecommendation,2015AHAGuidelines).DEFUSE3willaddressthisnewmandate.Enrollmentislimitedtopatientswithsalvageabletissue(TargetMismatchpatients)whoarelikelytorespondfavorablytoendovascularreperfusioninthe6‐16hourwindow.Useofthelatestgenerationthrombectomydevices,coupledwithstrictqualificationandoversightcriteriafortheneuro‐interventionalists,willresultinhighratesofreperfusion.BasedonthecompellingpreliminarydatafromDEFUSE2,thetrialisadequatelypoweredtodemonstrateacleartreatmenteffect.

2.2.Priorstudiesandrationalefordevelopment


4

Thisprotocolaimstoshifttheselectionofpatientsforreperfusiontherapyfromarelativelyarbitrarydecisionbasedonpoorlyvalidatedclinicalcharacteristicstoanobjectivedecisionbasedonscientificevidence.Manyfactorsaffecttheevolutionoftheischemicpenumbraintotheischemiccore,andtherateofprogressionofirreversibleinjuryishighlyvariablebetweenindividuals.Thisvariabilityislikelymediatedbytheadequacyofcollateralbloodflowaswellasthemetabolicmilieuofindividualstrokepatients.Theindividualityofpenumbralevolutionamongstrokepatientsimpliesthatidentifyingtheextentoftheischemiccoreandpenumbraisusefulformakingtreatmentdecisions.Currentlydiffusion‐weightedimaging(DWI)/perfusion‐weightedimaging(PWI)magneticresonanceimaging(MRI)affordsthebestopportunityforapproximatingtheischemiccoreandpenumbrainrealtimeclinicalpractice.27

TheDWIlesionprovidesadependableestimationoftheischemiccoreandonlyveryrarelyshowspermanentreversalfollowingearlyreperfusion.28,29PWIidentifieshypoperfusedischemictissue.RegionsdefinedasabnormalonPWIthatdonotdemonstrateaDWIabnormality,oftenreferredtoastheDWI/PWImismatch,canestimatetheischemicpenumbra.30ItiscriticalthatPWIutilizesanappropriatelyvalidatedthresholdparameterthatexcludesischemictissuewithmodestbloodflowreduction(i.e.benignoligemia),becausethistissueisunlikelytoinfarctevenifreperfusiondoesnotoccur.WhichPWIparameterisoptimal,aswellaswhatthresholdtousetodefinecriticalhypoperfusion,hasbeenthefocusofmultipleresearchefforts.31PriorworkfromourgroupandotherssupportstheuseofTmax,thresholdedat>6seconds,astheoptimalPWIparametertoidentifyischemictissuedestinedtobecomeinfarctediftimelyreperfusiondoesnotoccur.32ThisTmaxthresholdcorrelateswellwiththepenumbralrangeofcerebralbloodflowdeclineasdeterminedbybothpositronemissiontomographyandXenonCT.33,34

UsingadifferencebetweenthevolumeofthebaselinePWITmaxlesionandtheDWIvolumetoidentifymismatch,theDEFUSEandEPITHETstudiesfoundthatmostpatientswithaPWI/DWImismatchrespondedfavorablyifreperfusionoccurredfollowingivtPAtreatmentinthe3‐to6‐hourtimewindow.However,despitehavingamismatch,patientswithverylargebaselineDWIlesions(largecoreinfarctvolumes)hadhighlyunfavorableoutcomesfollowingreperfusion.PatientswiththisMRIpattern,referredtoastheMalignantprofile,hadasignificantlyhigherrateofbothparenchymalhemorrhageandseveredisability/deathifreperfusionoccurred.35MismatchpatientswhodonothavetheMalignantprofilehavebeendesignatedashavingaTargetMismatch,andthesepatentsrespondextremelyfavorablytoreperfusionfollowingivtPAtherapy.InapooledanalysisofDEFUSEandEPITHET,TargetMismatchprofilepatientswhoexperiencedreperfusionhada5‐foldincreaseinfavorableclinicalresponseat90daysandsignificantlylessinfarctgrowthwhencomparedtothosewhodidnotreperfuse.36Noassociationbetweenreperfusionandfavorableoutcomes,orareductionininfarctgrowth,wasapparentforpatientswithoutthemismatchprofile.

TheDEFUSE2studyutilizedanautomatedmismatchanalysisprogram(RAPID)toprospectivelyestablishMRIprofilesinaconsecutivecohortofpatientswhothenunderwentendovasculartherapy.DEFUSE2confirmedtheconceptsdemonstratedinDEFUSEandEPITHET;TargetMismatchpatientswhoachieveearlyreperfusiontherapyhavelessinfarctgrowthandmorefavorableclinicaloutcomes(8·8,95%CI2·7–29·0).37Noassociationbetweenreperfusionandfavorableoutcomesorinfarctgrowthwaspresentinpatients


5

withoutTargetMismatch.Furthermore,thepositiveassociationbetweenreperfusion,favorableclinicalresponse,andattenuationofinfarctgrowthdidnotdiminishinDEFUSE2patientswithTargetMismatchwhoweretreatedrelativelylate(6‐12hoursaftersymptomonset,seeFigure1).

Figure1:Theeffectoftimeontheassociationbetweenreperfusionandgoodfunctionaloutcome(graphontheleft)andtheeffectoftimetotreatmentontheprobabilityofgoodfunctionaloutcomepatientswithreperfusion(graphontheright)inTargetmismatchpatients.95%CIsareindicatedbydashedlines.Estimatesarebasedonmultivariatelogisticregression,adjustedforageandbaselineDWIvolume.Thereisnosignificanteffectoftimeineithermodel.38

Thisfindingcontrastssharplywithpriorstudiesthatdidnotusepenumbralimagingtoselectpatientsandsuggeststhatimagingfindingsmaybeofequal,orpotentiallyevengreater,importancethantimefromsymptomonsetforidentificationofoptimalpatientswhomightbenefitfromreperfusiontherapy.

HowcouldTargetMismatchpatientswhoaretreatedlatehaveoutcomesthatareasfavorableasthoseofearliertreatedpatients?Atlatertimepoints,theTargetMismatchprofileidentifiespatientsinwhomtheinfarctisevolvingatarelativelyslowrate;aDWIlesionthatisstillconsiderablysmallerthanthePWIlesionreflectsgoodcollateralcirculation.Thesecollateralstypicallyallowprolonged,butnotpermanent,survivalofthehypoperfusedmismatchregion.Evidencethatthemismatchregionisstillatconsiderableriskforinfarctexpansion,evenatlatertimepoints,wasprovidedbytheDEFUSE2findingthatTargetMismatchpatientsimagedbetween6and12hoursfromsymptomonsetconsistentlydemonstratedsubstantialinfarctgrowthifreperfusionwasnotachieved.38Patientswithslowlyevolvinginfarctcoresareidealcandidatesforlatertimewindowreperfusiontherapy,particularlyendovasculartherapies.Oneofthedrawbacksoftheendovascularapproachisthatthetimebetweenhospitalarrivalandachievementofendovascularreperfusionistypicallyatleast90to120minutes.Forpatientswithrapidlygrowinginfarctcores(suchaspatientswiththeMalignantprofile),substantialgrowthoftheinfarctcorehasbeenreporteddespiteendovascularreperfusion.Therefore,removingthispopulation,whichrepresentsabout20%ofeligiblepatients,fromarandomizedendovasculartrial,hasimportantadvantages.

DEFUSE2confirmedthatearlyDWIlesionsareanexcellentsurrogatefortheischemiccore.Despiteendovasculartherapy,only2patientshadafinalinfarctthatwassmallerthanthebaselineDWIlesionandthesizeandlocationoftheearlyDWIlesionwasareliablepredictorofthefinalinfarctvolumeinpatientswithcompletereperfusion.39,40InDEFUSE2,youngerageandsmallerDWIvolumeweresignificantindependentpredictorsoffavorableoutcome.SubgroupanalysisofEPITHETidentifiedDWIlesionsize≤25mlasastrongpredictorofa


6

favorableresponsetoreperfusion.37,41Thesefindingssuggestthatcertainsubgroups,inparticularindividualswithTargetmismatchandsmallDWIlesionsaremostlikelytobenefitfromreperfusion.TheadaptivedesignofDEFUSE3(seebelow)hasthepotentialtofocuspatientenrollmentonasubgroupofpatients(e.g.thosewithsmallerDWIlesionvolumesand/orshortertimesfromsymptomonsettorandomization)whorespondmostfavorablytoendovasculartherapy.Thiswillallowthestudytoidentifythelargestpopulationthathasastatisticallyreliablebenefitofendovasculartherapy.

Newdata(presentedattheInternationalStrokeConference,February2015)suggestthatCTPerfusionstudies,processedwiththesamesoftware(RAPID)usedinthestudiesdescribedabove,canidentifytheischemiccorewithaccuracysimilartoMRI(Cereda,etalISC2015)andselectpatientswhorespondtoendovascularreperfusiontherapyinearlytimewindows(Campbell,etalNEJM2015,EXTEND‐IAstudy,Saver,etalNEJM2015,SWIFTPRIMEstudy).Therefore,DEFUSE3willallowpatientselectionwithbothMRIandCTPerfusion.

3. INVESTIGATIONALPLAN3.1.Purpose

DEFUSE3isaprospectiverandomizedPhaseIIImulticentercontrolledtrialofpatientswithacuteischemicanteriorcirculationstrokesduetolargearteryocclusiontreatedbetween6‐16hoursofstrokeonsetwithendovascularthrombectomytherapyvs.control.Theprimaryendpoint,themodifiedRankinScale,willbeassessedat3months.Thepatients’participationinthestudyconcludesatthattime(3monthsfromstrokeonset).Thestudywillrandomizeupto476patientsover4years.ThepurposeofDEFUSE3istoassessthesafetyandefficacyofthrombectomyincarefullyselectedpatientsinanextendedtimewindow.Onlythedeviceslistedinthisprotocolwillbeused.Selectionofthespecificdevice(ordevices)isdeterminedbytheindividualendovasculartherapist.

3.2.ProtocolDesignDEFUSE3isaprospectiverandomizedPhaseIIImulticentercontrolledtrialofpatientswithacuteischemicanteriorcirculationstrokesduetolargearteryocclusiontreatedbetween6‐16hoursofstrokeonset.PatientswhomeettheinclusioncriteriawillundergoeitherCTPerfusion/CTAorMRDWI/PWI/MRAstudiespriortorandomization.PatientswhohaveevidenceofanICAorMCAM1occlusionandaTargetMismatchProfilewillberandomizedina1:1ratiototreatmentwithendovasculartherapy(usingoneormoreDEFUSE3approvedthrombectomydevices)plusstandardmedicaltherapyversusstandardmedicaltherapyalone.Patientswhoareconsented,butnotrandomized,willreceivestandardtherapyaccordingtolocalguidelines.Baselinedata,andinformationaboutearlystroketherapies,willbecapturedforthisgroupofpatients.Randomizationofamaximumof476patientsisplanned.Atthefirstinterimanalysiswhen200subjectscompletefollow‐up,iftheoverallanalysiscrossesthefutilityboundary,anoveladaptivedesignwillidentify,ifitexists,asubgroupwiththebestprospectforshowingbenefitfromendovasculartreatment,basedonbaselineischemiccorelesionvolumesandthetimetotreatment.Thesecondinterimanalyseswillbeconductedat340patientsatwhichtimethestudymaystopforefficacy/futility,ortheinclusioncriteriamaybeadjustedinthecaseoffutility.


7

Approximately45siteswillbechosen.Individualsiteselectionwillbebasedonanumberoffactorsincludingendovascularvolume,MRIand/orCTperfusionaccess,numberofcompetingtrials,clinicaltrialexperience,andthediversityoftheirpatientpopulation.Ifasitedoesnotconsentapatientwithin4monthsofactivation,itwillbeplacedonprobation.Ifnopatientconsentoccursinthenext2months,thesitewillbereplacedwitha“back‐up”site.

3.3.EnrollmentCriteria3.3.1.ClinicalInclusionCriteria:1. Signsandsymptomsconsistentwiththediagnosisofanacuteanteriorcirculation ischemicstroke2. Age18‐90years3. BaselineNIHSSSis≥6andremains≥6immediatelypriortorandomization4. Endovasculartreatmentcanbeinitiated(femoralpuncture)between6and16 hoursofstrokeonset.Strokeonsetisdefinedasthetimethepatientwaslastknownto beattheirneurologicbaseline(wake‐upstrokesareeligibleiftheymeettheabovetime limits).5. modifiedRankinScalelessthanorequalto2priortoqualifyingstroke(functionally independentforallADLs)6. Patient/LegallyAuthorizedRepresentativehassignedtheInformedConsentform.3.3.2.ClinicalExclusionCriteria:1. Otherserious,advanced,orterminalillness(investigatorjudgment)orlifeexpectancy islessthan6months.2. Pre‐existingmedical,neurologicalorpsychiatricdiseasethatwouldconfoundthe neurologicalorfunctionalevaluations3. Pregnant4. UnabletoundergoacontrastbrainperfusionscanwitheitherMRIorCT5. Knownallergytoiodinethatprecludesanendovascularprocedure6. TreatedwithtPA>4.5hoursaftertimelastknownwell7. TreatedwithtPA3‐4.5hoursafterlastknownwellANDanyofthefollowing:age >80,currentanticoagulantuse,historyofdiabetesANDpriorstroke,NIHSS>258. Knownhereditaryoracquiredhemorrhagicdiathesis,coagulationfactordeficiency; recentoralanticoagulanttherapywithINR>3(recentuseofoneoftheneworal anticoagulantsisnotanexclusionifestimatedGFR>30ml/min).9. SeizuresatstrokeonsetifitprecludesobtaininganaccuratebaselineNIHSS10. Baselinebloodglucoseof<50mg/dL(2.78mmol)or>400mg/dL(22.20mmol)11. Baselineplateletcount<50,000/uL12. Severe,sustainedhypertension(SystolicBloodPressure>185mmHgorDiastolic BloodPressure>110mmHg)13. Currentparticipationinanotherinvestigationaldrugordevicestudy14. Presumedsepticembolus;suspicionofbacterialendocarditis15. Clotretrievalattemptedusinganeurothrombectomydevicepriorto6hoursfrom

symptomonset16. Anyotherconditionthat,intheopinionoftheinvestigator,precludesanendovascular procedureorposesasignificanthazardtothesubjectifanendovascularprocedurewas performed.


8

3.3.3.NeuroimagingInclusionCriteria:1. ICAorMCA‐M1occlusion(carotidocclusionscanbecervicalorintracranial;withor

withouttandemMCAlesions)byMRAorCTAAND

2. TargetMismatchProfileonCTperfusionorMRI(ischemiccorevolumeis<70ml,mismatchratiois>1.8andmismatchvolume*is>15ml)Notes: The mismatch volume is determined by the RAPID software in real time based on the difference between the ischemic core lesion volume and the Tmax>6s lesion volume. If both a CT perfusion and a multimodal MRI scan are performed prior to enrollment, the later of the 2 scans is assessed to determine eligibility. Only an intracranial MRA is required for patients screened with MRA; cervical MRA is not required. Cervical and intracranial CTA are typically obtained simultaneously in patients screened with CTA, but only the intracranial CTA is required for enrollment.

Alternativeneuroimaginginclusioncriteria(ifperfusionimagingorCTA/MRAistechnicallyinadequate):A)IfCTA(orMRA)istechnicallyinadequate:

Tmax>6sperfusiondeficitconsistentwithanICAorMCA‐M1occlusionAND

TargetMismatchProfile(ischemiccorevolumeis<70ml,mismatchratiois>1.8andmismatchvolumeis>15mlasdeterminedbyRAPIDsoftware)

B)IfMRPistechnicallyinadequate:

ICAorMCA‐M1occlusion(carotidocclusionscanbecervicalorintracranial;withorwithouttandemMCAlesions)byMRA(orCTA,ifMRAistechnicallyinadequateandaCTAwasperformedwithin60minutespriortotheMRI)

ANDDWIlesionvolume<25ml

C)IfCTPistechnicallyinadequate:

PatientcanbescreenedwithMRIandrandomizedifneuroimagingcriteriaaremet.

3.3.4.NeuroimagingExclusionCriteria:1. ASPECTscore<6onnon‐contrastCT(ifpatientisenrolledbasedonCTperfusion

criteria)2. Evidenceofintracranialtumor(exceptsmallmeningioma)acuteintracranial

hemorrhage,neoplasm,orarteriovenousmalformation3. Significantmasseffectwithmidlineshift4. Evidenceofinternalcarotidarterydissectionthatisflowlimitingoraorticdissection5. Intracranialstentimplantedinthesamevascularterritorythatprecludesthesafe

deployment/removaloftheneurothrombectomydevice6. Acutesymptomaticarterialocclusionsinmorethanonevascularterritoryconfirmedon

CTA/MRA(e.g.,bilateralMCAocclusions,oranMCAandabasilararteryocclusion).

3.4.EnrollmentandRandomization

3.4.1Enrollment:Allpatientswhomeettheclinicalcriterialistedaboveareeligiblefor


9

enrollment.Thisincludesbothpatientswhoaredirectlyadmittedtothestudysiteandpatientswhoaretransferredfromanoutsidehospital.Thetimeofenrollmentisthetimewhentheinformedconsentissigned.AfterobtainingconsenttheRAPIDoutputfromaCTperfusionormultimodalMRIscanwillbeassessed.Ifthepatientisconfirmedtomeettheneuroimagingeligibilitycriterialistedaboveandnothaveanyoftheneuroimagingexclusioncriteria,thenthepatientwillberandomized.Ingeneral,patientswillbeconsentedpriortoobtainingtheRAPIDoutputmaps.InsomesituationstheCTperfusion/multimodalMRImayhavebeenperformedaspartofstandardcarepriortothepatientbeingassessedforstudyeligibility.Patientswhoareconsentedbutdonotmeettheimagingcriteriawillnotberandomized.

DeterminationofTargetMismatchandLargeArteryOcclusion:AttheconclusionoftheMRIorCTPerfusionscan,thetechnologistsendsthesequencesfromtheconsoletoRAPIDwithasinglemouseclickforautomatedprocessing.TheRAPIDsoftwarewasdevelopedbasedondatafromDEFUSE1andwasprospectivelyvalidatedinDEFUSE2.Thesystemprovidesfullyautomatedprocessingofbrainimages.TheRAPIDoutputmaps,whichidentifythevolumeandlocationofischemiccoreandperfusionlesions,areemailedtoinvestigators(protectedhealthinformationisautomaticallyremoved)andauto‐senttoPACSaswellastoasecureemailsystemforviewingwithin5minutesofcompletionofthescan.Immediatelyaftertheimagesareavailable,theinvestigatorwillreviewtheresultsoftheRAPIDmismatchmap(Figure2)andtheMRA/CTAtodetermineifthepatientmeetstheimagingcriteria(listedabove).Ifapatienthasundergonemultipleimagingevaluations(bothMRIandCTormultipleCTsorMRIs),themostrecentimagingstudywillbeusedtodetermineifthepatientmeetstheimagingcriteria.Theaccuracyofthesoftwareforidentifyingthesizeandlocationofperfusionanddiffusionlesionshasbeenestablishedbyextensivevalidationandtestingonbloodflowphantoms;thesoftwarereceivedFDA510Kclearanceforclinicalusein2013.TheagreementbetweenlocalinvestigatorsandtheImagingCoreLabforidentificationofthemismatchprofileinDEFUSE2was97%,κ0·92;95%CI0·83–1·

Baselinedatawillbecapturedforallconsentedpatients,includingforpatientswhodonotmeettheimagingcriteria.Patientswhodonotmeettheimagingeligibilitycriteriawillreceivestandardcareperlocalhospitalpractice.3.4.2RandomizationusingaDynamicStratificationAlgorithm:OnceaconsentedpatientisdeterminedtomeetallNeuroimagingcriteria,thepatientwillimmediatelybe

Figure2.TheRAPIDmismatchsummaryallowsinvestigatorstoquickly,accuratelyandeasilydetermineifthepatientmeetstheimagingcriteriaforenrollment.ThepatientabovemeetstheTargetMismatchcriteria:corevolumeis<70ml,mismatchratiois>1.8andmismatchvolumeis>15ml.


10

randomizedontheWebDCUTMwebsite.Adynamicstratificationsystemwillensurewell‐balancedsubgroups.Therandomizationalgorithm,whichwillbeprogrammedintothedatacapturesystem,willemploybiased‐coinminimizationandthevariancemethodwithstratificationweights.44Thestrategyistobalancetreatmentassignmentalongthemarginaldistributionofeachstratificationfactor.Thestratificationfactorsusedwillbe:1)age,2)corelesionvolume,3)timefromsymptomonsettoenrollment,4)baselineNIHSS,and5)studysite.Whenanewpatientisenrolled,thesitewillenterthestratificationfactorvaluesintotheeCRF(electroniccasereportform)onWebDCUTM.Thedynamicrandomizationalgorithmwilldetermineanimbalancemeasureforeachtreatmentgroup.Thetreatmentgroupassociatedwiththesmallestimbalancemeasurewillreceivethelargestprobabilityofassignmentinthebiased‐coinrandomization.Thebiased‐coinacceptanceregionandstratificationweightsarespecifiedintheRandomizationPlan.Thesuperiorbalancingcharacteristicsofdynamicrandomizationoverblockedrandomizationhavebeenwellestablished.

Patientswillbeassignedtoeitherendovasculartherapyplusmedicaltherapyortomedicaltherapyalone(1:1randomization).Crossoverfrommedicaltoendovasculartherapyisstrictlyprohibited;endovasculartomedicaltherapycrossoverisdefinedasapatientwhoisassignedtoendovasculartherapybutdoesnothaveaconventionalangiogramperformed.Endovasculartomedicaltherapycrossoverisonlyallowedifanendovascularcontraindicationarisesafterrandomization.Siteswillbecloselymonitoredforcrossovers(seesitemonitoringplanbelow).

3.5 AcuteTreatment

3.5.1EndovascularTherapy:Inpatientsrandomizedtoendovasculartherapy,thegoalforfemoralarterypuncturewillbewithin45minutesofrandomization;femoralarterypuncturemustoccurwithin90minutesofthecompletionofthequalifyingimaging.Patientswillbetreatedwiththrombectomydevices(stent‐retrievers)and/orsuctionthrombectomysystemscurrentlyclearedbytheFDAforthrombusremovalinpatientsexperiencinganacutestrokewithin8hoursofsymptomonsetfollowingthepublishedinstructionsforuseforthesedevices.Thesedeviceswillbeusedupto16hoursfollowingsymptomonsetinDEFUSE3basedonanFDAIDE.ThedeviceswhichwillbeusedaretheTrevoRetriever,theSolitaireRevascularizationDevice,CovidienMindFrameCaptureRevascularizationDeviceandthePenumbrathrombectomysystem.

Standardmedicaltherapy,basedoncurrentAHAguidelines,willalsobeprovidedforallpatients.IndividualinvestigatorsmayuseanyofthesedevicesoranycombinationofthesedevicestoremovethrombusfromtheICA,MCAM1segmentor,ifneeded,fromM2segmentsoftheintracranialcirculation.Theseareallapprovedanatomiclocationsforthesedevices.Theuseofthrombectomydevicesshouldbeperformedinaccordancewiththeindicationsforuse.Ifthereisaseverestenosisofthecommoncarotidarteryortheproximalinternalcarotidartery,investigatorsmayalsouseotherFDAdevicesapprovedforangioplastyorFDAdevicesapprovedforstentingofthecarotidarteryasdeemedappropriate.Theuseofadjuvantintra‐arterial(IA)thrombolyticmedicationisnotcurrentlyapprovedbytheFDAforstroketreatmentandcannotbeusedinDEFUSE3.

Siteswilluselocalprotocolsforfemoralaccess,sedation,heparininfusion,monitoring,etc.Siteswillperformacervicalinjectionoftheinvolvedcarotidcirculationasabaseline


11

angiogram.Attheconclusionoftheprocedure,apost‐treatmentangiogramasacervicalinjectionoftheinvolvedcarotidcirculationwillalsobeobtained.ImagingwillcoverthefullregionofthenormalcirculationinAPandlateralprojectionsat2‐3filmspersecondthroughtheentirevenousphase.Allbrainimagingfromstrokeonsetthroughhospitaldischarge,includingthebaselineMRIandCT,aswellasangiographicimagesobtainedforthediagnosticandtherapeuticportionsoftheprocedure,willbetransmittedtothecorelab.

3.5.2MedicalTherapy:PatientsrandomizedtomedicaltherapywillreceivestandardmedicaltherapybasedoncurrentAHAguidelines.BasedonthetimewindowforDEFUSE3,itisanticipatedthatveryfewofthepatientsenrolledinDEFUSE3willhavereceivedivtPApriortorandomization(“tPAfailures”).Forthesepatients,thesites’post‐tPAprotocolwillbefollowed.Non‐tPAtreatedpatientsrandomizedtomedicaltherapywillbetreatedwithaspirin,325mgonDay1,and81‐325mg/day(investigator’spreference)onday2‐5,unlessanindicationforearlyanticoagulationispresent(asdeterminedbythepatient’sattendingphysician).AllpatientswillreceivestandardDVTpreventiontherapy.Intravenousanticoagulantsareprohibited(unlessaclearindicationforearlyanticoagulationisdocumented);dualantiplatelettherapyisprohibitedunlesscarotidstentingwasperformedduringtheendovascularprocedureoraclearindicationfordualantiplatelettherapyisdocumented.Subsequentantithrombotictherapywillbedeterminedbythepatient’sattendingphysician.

3.6ClinicalandImagingEvaluations

Follow‐up(imagingandclinical):Randomizedpatientswillbefollowedclinicallyfor90daysandwillhaveanMRI/MRA/MRperfusionat24hours(range18‐30hours)toassessinfarctvolume,recanalization,hemorrhageandreperfusion(Table1below).

Table1ScheduleofEvents

Screening Enrollment Baseline /

Randomization

EndovascularProcedure

24 hours (+/‐6 hrs)

Hospital Discharge

Day 30 (+/‐ 7 days)

Day 90 (+/‐ 14 days)

Screen Failure Log X

Informed Consent X

Subject Enrollment X

Inclusion and Exclusion Criteria X

MRI or CTP X X++

Randomization X

Medical History X

Vital Signs X

NIH Stroke Scale X X X X X

Modified Rankin Scale X+ X X X

Baseline ASPECTS Score X

Baseline Labs* X

EndovascularTherapy X

24 Hour Labs X

Hospital Discharge X

Adverse Event Assessment X X X X X

NeuroQOL X

*LaboratoryEvaluationincludesCBCwithPlatelets,Creatinine,Glucose,INR,activatedPTT,andPregnancytest(ifapplicable).At24hourfollow‐uponlycreatinineisrequired.+HistoricalmRSatbaseline,mRS/NIHSStobeperformedbyanNIHSS/mRScertifiedmemberoftheresearchteamwhoisblindedtotreatmentallocationat30and90days.++PatientswillpreferablyundergoanMRIwithMRAANDMRPerfusionat24hours;ifanMRcannotbeperformed,aCTwithCTAANDCTPcanbesubstituted.Forpatientswhoareconsentedbutnotrandomized,thescheduleofeventsislimitedtoasummaryofstroketherapiesreceivedwithin24‐hrsofstrokeonset.


12

3.6.1Assessmentsandfollow‐upvisits

Baselinevisit:AllitemsinTable1abovelistedunder“baseline”aretobeperformedpriortorandomization.TheMRIorCTscanshouldbeperformedwiththeDEFUSE3(baseline)protocol,whichwillbeinstalledatallstudysites.Inaddition,theinclusion/exclusionpageofthecasereportformmustbecompletedtodetermineifthepatientmeetstheeligibilityrequirementsforthestudy.Ifthepatientiseligibleandtheconsentformissignedbythepatientorauthorizedrepresentative,thentherandomizationprocedureshouldoccurimmediately.

24hourvisit(+/‐6hours):TheitemslistedforthisvisitinTable1shouldbeperformedbetween18and30hoursfromthetimeofrandomization.Theonlylaboratoryvaluerequiredatthe24hourvisitisaserumcreatinine.Ifpossible,the24hourfollow‐upimagingstudyshouldbeperformedwithmultimodalMRI,ratherthanCTperfusion.TheMRIorCTscanperformedatthistimeshouldbeperformedwiththeDEFUSE3protocol.

Dischargevisit:TheitemslistedforthisvisitinTable1shouldbeperformedonthedayofhospitaldischarge

30and90dayvisits:TheitemslistedforthisvisitinTable1shouldbeperformedonDay30(+/‐7)daysandDay90(+/‐14)days.ThemRSscoremustbeperformedbyanmRScertifiedinvestigatorwhoisblindedtotreatmentallocationatboththe30and90dayvisits.Ifaninpersonvisitisnotpossible,thenthemRSshouldbeperformedbyphonebyanmRScertifiedinvestigatorwhoisblindedtotreatmentallocation.Ifaninpersonvisitisnotpossible,thentheNIHSSscorewillbemarked“notavailable”inthecasereportform.

Neurologicalworsening:Ifclinicalworsening(definedasa>4pointincreaseontheNIHSSscore)occurspriortodischarge,aCTscanorMRIshouldbeobtainedassoonaspossible.Neurologicalworseningisareportableadverseevent.3.6.2SourcesofMaterialsInformationontheclinicalstatusofpatientswillbeobtainedfromthepatient’smedicalrecord.StudycoordinatorsatthesitewillcompletetheDEFUSE3casereportformstocollectbasicdemographicandmedicalinformationaboutthepatients.DatawillsubsequentlybeenteredintotheStrokeNet’sWebDCUelectronicdatacapturesystem.ImagingdatawillbeelectronicallytransmittedtothecoordinatingcenteratStanfordviaRAPID(allpatientidentifiersareremovedbythesoftwarepriortoexportingthedataoutsideofthesite’sfirewall).Allstudysiteswillcompleteastrokescreeninglogthatdocumentsallpatientstreatedinthecathlabbeyond6hoursattheircenter,andreasonforexclusionofpatientsnotenrolled,intheStrokeNet’sWebDCUelectronicdatacapturesystem.Seriousadverseevents(SAEs)willbereportedwithin24hoursoftheeventintheStrokeNet’sWebDCUelectronicdatacapturesystem.Thedatacollectionprocesswillincludepatientdemographics,medicalhistory,vitalsigns,laboratoryassessments,NIHSSandmRSscores,andresultsofdiagnosticstudiesperformedtoclarifystrokeetiology.

3.7.SiteApprovalandMonitoringPlanSiteapproval:Individualsitesapprovedforparticipationinthestudywillbehigh‐volumesites.SelectedsiteswillhaveaccesstoemergentCTperfusionand/orMRimaging24/7.Priortoactivatingasite,wewillverifythatRAPIDisfunctionalatthesite.Togetherwiththe


13

site’sCTand/orMRtechnologists,wewillinstalltheDEFUSE3scanprotocolonthelocalscannersandperformadummy‐runtoassessimagequalityandtrainthetechnologistsinsoftwarehandlinganddatasending.AsitewillbeactivatedforenrollmentaftertestcasesprocessedwithRAPIDhaveensuredgoodqualitymaps.

Monitoringforimagingquality:TheImagingCoreLabwillmonitorimagequalitythroughoutthestudy.Ifsignificantinadequaciesorprotocolerrorsarenotedatasite,enrollmentwillbehalted.Enrollmentwillresumeafterallimagingproblemshavebeenresolvedandrepeatdummyrunshavebeenobtainedthatdemonstrateadequateimagequality.Table 2. Example Imaging Sequences for DEFUSE 3 scans

Sequence Scan Parameters (3T) Time

MRI 6 min

Localizer 128X256; 28 FOV;5/5mm, GRE 24 sec

Calibration 5 sec DWI 128x128, 24 FOV, 5/0mm, 30 slices, 1 NEX, R=2; b=0 and 1000 s/mm2

over 3 axes, TE/TR=min/7000ms. 25 sec

GRE 256x192; 24 FOV; 5/0 mm, 30 slices, TE/TR= 25/800ms, flip 20, interleaved EPI, 16 shots

27 sec

MRA intracranial

256x192, 1 mm; 4 slabs, 26 phase-encodes; 6 overlap, 22 FOV, 0.8 rFOV, fractional echo, ZIPx2, ZIPx512, minTE, flowcomp, TR=18ms, flip=18, inferior->superior ramppulse, R=2; 19 MIPS

143 sec

PWI 128x128; 24 FOV; 5/0 mm, 17 slices, TE/TR=35ms/1800ms, R=2 using 0.1mmol/kg Gadolinium @ 4ml/sec.

108 sec

CT (example below for GE VCT; comparable protocols will be used for other scanner models)

5-6 min

Non-con head 2.5 – 5mm, 40 slices, 120-140kV, 265-290mA 120-180 sec

CTA 0.625mm, 0.984:1/39.37cm, 120kV, 550mA , inject and observe for 15 sec until contrast concentration in ascending aorta reaches 80HU (smart prep) then the CT gantry moves along with the bolus of the contrast material from the aortic arch up to the apex of the brain in 5sec.

90 sec

CTP 22 FOV, 40mm, 8x5mm, 1.8sec time interval, 45 cycles, 80kV, 125mA; 2 runs

90 sec

Monitoringforbias:Adetailedsite‐monitoringplanhasbeendevelopedtodetectbias.Thisplanwillprotectthestudyfromenrollment,randomization,andtreatmentbias.Thefirstcomponentinvolvesmonitoringthepercentageofeachsite’sendovascularvolume(within6‐16hrs)thatisenrolledinDEFUSE3.Siteswillreporttheirvolumeofendovascularstrokeprocedures(within6‐16hrs)eachmonthonascreeninglog.IfaDEFUSE3eligiblepatientistreatedwithendovasculartherapyoutsidetheDEFUSE3study,anexplanationwillberequireddetailingwhythepatientwasnotenrolled.Thesecondcomponentoftheplaninvolvestrackingofpatientswhoareconsentedbutnotrandomized.ThesepatientswillrequireanentryintheWebDCUwithanexplanationwhythepatientwasnotrandomizedaswellasdocumentationwhetherendovasculartherapywasperformedoutsideofthestudy.Athirdcomponentinvolvesmonitoringofcrossoverafterrandomization.TheExecutiveCommitteewillreviewthedatadescribedaboveforeachsiteevery6months.Ifevidenceofenrollmentbiasissuspected,itwillbeinvestigated.Ifconfirmed,thesitewillbeplacedonprobation.Ifadditionalincidentsof


14

suspectedbiasareconfirmed,thesitewillbewithdrawn.RoutinemonitoringoftheclinicalsitesforsourcetodatabaseverificationwillbeperformedbytheStrokeNetDataManagementCenter.

3.8.SampleSize,AdaptiveDesignandStatisticalAnalysisDEFUSE3willfeatureanoveladaptivetrialdesignthatwillallowthestudytofocusonasubpopulationifinterimorfinalanalysesindicatefutilityintheoverallpopulation.59TheadaptivedesignwasdevelopedspecificallyforDEFUSE3.ItisbasedonclosedtestingtheoryandthegroupsequentialmethodsfortheGeneralizedLikelihoodRatio(GLR)statisticdevelopedbyLaiandShih.60Theadaptivedesignwaschosenbecausethereisstrongpreliminarydatathatsuggeststhattheeffectofendovasculartreatmentismodifiedbytwobaselinevariables:corelesionsizeandtime‐to‐treatment.Thewaytheadaptivedesigntakesadvantageofthesebiologicalassumptions(whentheyaretrue)isbyreallocationoffutureaccrualtothesubgroupwiththebestprospectsforshowingefficacy.Specifically,ifasubgroupischosenataninterimanalysis,subsequentenrollmentislimitedtopatientsinthatsubgroup.Asaresult,thissubgroupwillbecomelargerthanitwouldhavebeenintheabsenceoftheadaptivedesign.Thecriterionfordecidingwhichsubgrouphasthebestchanceofshowingabenefitfromendovasculartherapycombinesboththeestimatedsizeoftheeffectinthesubgroupandthesamplesizeofthesubgroup.TheGLRstatistic(Kullback‐Leiblercriterion)isusedtoidentifythissubgroupbecauseitoptimallybalancesthosetwocriteria.Itselectsthesubgroupthathasthebestchanceofshowinganeffectbecauseithasanapparentlylargeeffectandisalsoofsubstantialsize(notethereare5subgroupsofincreasinglylargersize,figure3).Theadaptivedesignemploystwobiologically‐basedassumptionstolimittheinflationofsamplesize;amonotonicity/contiguityassumptionandanapriorassumptionthattheeffectislargestinthepatientswiththesmallestDWIlesionsandtheshortesttimetorandomization(cellC11infigure3).Theboundariesofthecategories(cells)willbedeterminedjustpriortothe1stinterimanalysisbasedonthedistributionofpatientsacrossthesetwodimensions(lesionvolumeandtime‐to‐treatment).

Primaryanalysis:TheprimaryendpointisthedistributionofscoresonthemodifiedRankinScale(mRS)atday90.WewilltestthenullhypothesisattheinterimandfinalanalysisusinganormalapproximationoftheWilcoxon‐Mann‐Whitneytest(thegeneralizedlikelihoodratio[GLR]test).Theprimaryanalysiswillbeintentiontotreat,adjustedfordesignandnotadjustedforcovariates.

Foreachanalysis,anefficacyboundwillbesettocontroltheoverall(one‐sided)TypeIerrorrateat2.5%.Ateachinterimanalysisafutilityboundwillbesettodecideifthestudyshouldcontinuerecruitmentintheoverallgroup,shiftaccrualandtestingtoasubgroup,orstopinitsentirety.Thefutilityboundaryadaptswhenasubgroupisselectedtothefactthatthemaximumanalyzedsamplesizeisarandomvariablethatisnolargerthanthefixedmaximumnumberofpatientsrandomized(n=476).Becausesubgroupselectionreducesthe

Figure 3. The cohort is stratified according to core lesion volume and time to randomization. Exact boundaries of the stratification will be determined based on the distribution of patients just prior to the first interim analysis. Depending on the results of the 1st interim analysis, subsequent enrollment will continue in all 6 cells or will be limited to one of 5 sub-groups (C11, C11+21, C11+21+12, C11+21+12+22, or C11+21+12+22+13).


15

maximumnumberofpatientsavailableforanalysisatcompletionofthestudy,thismethodeffectivelyallowsaneasierfutilitystopaftersubgroupselection.Thissetupreplacesconditionalpoweranalyseswithanautomaticandmorepowerfuladjustmentofboundaries.

Firstinterimanalysis(n=200randomizedandcompleted90dayfollow‐up):Thenullhypothesisistestedintheentirepatientpopulation:1. Ifneitherefficacynorfutilityboundiscrossed,thetrialcontinuesenrollmenttothe2nd

interimanalysis.2. Iftheefficacyboundiscrossed,thetrialstopsandefficacyisdeclaredintheoverall

population.3. Ifthefutilityboundiscrossed,theoptimalsubgroupisselectedbasedontheKullback‐

Leiblercriterionandthenullistestedinthatsubgroup.Thefutilityboundisrelaxedasdescribedabove,basedontheexpectedmaximumnumberofpatientsinthetrialatcompletion(ie476minusthenumberofpatientsalreadyenrolledincellsthatwillnolongerbeopenforenrollment).3.1. Ifneitherboundiscrossed,thetrialwillcontinuewithenrollmentlimitedtothe

selectedsubgroup3.2. Iftheefficacyboundiscrossed,thetrialstopsandefficacyisdeclaredintheselected

subgroup3.3. Ifthefutilityboundiscrossed,thetrialstopsforfutility.

Secondinterimanalysis(n=340randomizedandcompleted90dayfollow‐up):If,afterthefirstinterimanalysis,thestudyproceedswithenrollmentintheoverallpopulation(option1above),thetestingatthe2ndinterimanalysisisidenticaltothefirstinterim.Ifenrollmentislimitedtoaselectedsubgroup(option3.1),thenullistestedinthatsubgroup:1. Ifneitherboundiscrossed,thetrialcontinuestothefinalanalysiswithenrollmentof136

additionalpatientslimitedtotheselectedsubgroup2. Iftheefficacyboundiscrossed,thetrialstopsandefficacyisdeclaredintheselected

subgroup3. Ifthefutilityboundiscrossed,thetrialstopsforfutility.

Finalanalysis(n=476randomizedandcompleted90dayfollow‐up):If,afterthesecondinterimanalysis,thestudyproceedswithenrollmentintheoverallpopulation,thenullistestedintheoverallpopulation:1. Iftheefficacyboundiscrossed,endovasculartherapyisdeclaredefficaciousintheoverall

population.2. Iftheefficacyboundisnotcrossed,theoptimalsubgroupisselectedandthenullistestedin

thatgroup:2.1. Iftheefficacyboundiscrossed,endovasculartherapyisdeclaredefficaciousinthat

subgroup2.2. Iftheefficacyboundisnotcrossed,endovasculartherapywillbedeclaredofnobenefit.


16

Ifenrollmentafteroneoftheinterimanalysesislimitedtoaselectedsubgroup,thenullwillbetestedinthatsubgrouponlyandefficacyorlackthereofwillbedeclaredasperoptions2.1and2.2above.

Powerandsamplesizeconsiderations:Theprojectedoveralleffectofendovasculartherapyisbasedon1)theobserved90‐daymodifiedRankinScaleoutcomesinDEFUSE2oftargetmismatchpatientstreated>6hrsaftersymptomonsetand2)theassumptionthatearlyreperfusionwillbeachievedin75%oftheendovasculararmvs.20%ofthemedicaltherapyarm.20,21,61Usingthesedata,weprojectedthedistributionsonthemRSat90daysintheendovascularandcontrolarmsofDEFUSE3:

mRS at day 90

0 1 2 3 4 5 6 Total

Endovascular group 18.0% 11.5% 19.6% 11.5% 16.4% 11.5% 11.5% 100%

Medical group 9.7% 7.9% 15.0% 17.7% 14.4% 17.7% 17.7% 100%

Thisdistributioncorrespondstoastandardizedeffectof0.36fortheprimaryanalysis.Basedonthesedata,thefixedsamplesizeforanon‐adaptivedesignrequiresatotalof376patients(188/arm)tohave90%poweratanalphaof5%(Wilcoxon‐Mann‐Whitneytest);100additionalpatientsareaddedfortheadaptivedesigntoreachamaximumsamplesizeof476forDEFUSE3.

Weransimulations(n=5000)tocomparetheperformanceofatraditionalfixedsample‐sizedesign(fixedn=476)totheadaptivedesign(maxn=476)undervariousscenarios(seeTable3,below).Forthesimulationstheeffectsizeisexpressedasastandardizedeffect,whereastandardizedeffectof0.3correspondstoaconservativeprojectedeffectofendovasculartherapy(anticipatedeffect0.36;seeabove).

Imagingoutcomes:Wehypothesizethatendovasculartreatmentimprovesradiologicaloutcomesinstrokepatientswithfavorableclinicalandimagingcharacteristics.DEFUSE2demonstratedasubstantialreductionininfarctgrowthamongTargetmismatchpatientstreatedinthe6‐12hourtime‐windowwhoachievedearlyreperfusion:mediangrowth0.5ml(IQR:‐2–10)withreperfusion(n=23)vs.39ml(IQR:18‐121)withoutreperfusion(n=13),p<0.001.ThesedatahavebeenextrapolatedtoDEFUSE3usingthesameassumptionsdescribedabove;anticipatedanearlyreperfusionrateof75%intheendovasculararmvs.20%inthemedicalarm.Thisyieldsasamplesizeof42pergroupfor90%power.Therefore,DEFUSE3ishighlypoweredtodemonstratedifferencesinlesiongrowth.Infarctvolumes,ischemiclesiongrowth,andreperfusionratesat24hourswillbecomparedbetweengroupswiththeMann‐WhitneyUtest.The24hourendpointisbasedondatademonstratingthatassessmentofinfarctvolumeat24hourscapturestheeffectofreperfusiontherapiesoninfarctgrowthandpredictsoutcomessimilarlytoday90infarctvolumes.29,62RAPID‐assessedischemiccorevolumeatbaselinewillbecorrelatedwith24hinfarctvolume(DWIvolume)insubjectswhoachievereperfusionwithoutPH1orPH2intracranialhemorrhage.Pearson’scorrelationcoefficientwillbecalculatedandthemedianabsoluteerror(ml)willbereported.Similarly,correlationofthebaselineTmax>6volumeandthe24hinfarctvolumeinpatientswithoutPH1orPH2intracranialhemorrhagewhohave<10%reperfusionwillbeperformed.CorrelationofRAPIDpredictedinfarctvolume(coregisteredbaselineischemiccoreand24hTmax>6volume)withtheactual24hinfarctvolumewillalsobeperformed.


17

Scenario Standardized effect in cells

C11, C12, C21, C22, C31, C32

Average standard. effect

Adaptive Design Fixed Design

Average No. randomized

Power Number randomized

Power

#0 0 0 0 0 0 0 0 361 2.2% 476 2.5%

#1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 354 80% 476 89%

#2 0.5 0.4 0.3 0 0 0 0.2 400 86% 476 55%

#3 0.5 0.5 0 0 0 0 0.17 403 87% 476 41%

Table 3. Under the null (Scenario #0), the adaptive design controls the total Type 1 error below 2.5%, stops early for futility 63% of the time, and the average number of randomizations is 361. If the effect is uniform across cells (scenario #1), the fixed-sample design is optimal, but the adaptive design results in only a small loss of power (from 89 to 80%). The adaptive design performs much better (higher power and smaller expected sample size) than the fixed sample, conventional trial when the effect size distribution across the subgroups is in accord with the biological assumptions (scenarios #2 and 3). If the effect is concentrated in two cells with small core volumes (scenario #3), the adaptive design maintains power (87%) while the conventional design collapses (41% power). The adaptive design also performs well compared to a non-adaptive, fixed sample that includes efficient multiple comparisons-adjusted testing for effect in subgroups at the end of the study. (see Lai et al59)

Secondaryanalysis:OursecondaryendpointistheproportionofpatientswithmRS0‐2atday90(indicatingfunctionalindependence).ThedifferenceintheproportionsofpatientswithmRS0‐2betweentreatmentarmswillbeassessedusinglogisticregression.

Subgroupanalyses:Subgroupanalysesoftheeffectofendovasculartherapyontheprimaryandsecondaryendpointswillbeperformed.Subgroupswillbedefinedbasedonthestratificationvariables,keydemographicfactors(suchasraceandethnicity),tPAvs.notPA,CTPvs.MRIselection,andwitnessedvs.unwitnessedsymptomonset,wake‐upvsnon‐wake‐upstroke,andTICI0‐2avs.TICI2b/3resultsincathlab.

Missingdata/losttofollow‐up(LTFU):

Alleffortisputforthtoensurenearcompletefollow‐up,inparticularwiththeassessmentoftheprimaryoutcome(mRSat90days),death(mRS=6),andstrokerecurrence.Iftheprimaryoutcome(mRSat90days)cannotbeassessedintheclinic,itwillinsteadbeobtainedbyphoneusingastructuredinterview.Ifthesubject’smRScannotbeobtainedinclinicorbyphonewithinthewindowof60to120daysfromrandomization,thenforprimaryanalysestheday30mRSscorewillbeusedastheprimaryoutcome(ieday30mRScarriedforward).Ifneitherthe30‐daynorthe90‐daymRSisavailable,thenthemRSwillbeimputed.WedonotexpectthistoalterthemainstudyresultsgivenourestimatedverylowLTFUrate(<2%).

DEFUSE3Timetable

Year1 Year2 Year3 Year4 Year5 InstallRAPIDatallsites

Beginenrollment(anticipatedtobeginmid‐year)

Ptenrollmentcontinuesforatotalof4yrs

1stinterimanalysisPotentialmodificationofenrollmentcriteriabasedonadaptivedesign

2ndinterimanalysisPotentialmodificationofenrollmentcriteriabasedonadaptivedesign

Finishenrollment

DataAnalyses Publicationofresults


18

3.9.RiskanalysisDescriptionandanalysisofallincreasedriskstotheresearchsubjects:PotentialcomplicationsofMRIscanincludelocalizedtwitchingsensationduetothemagneticfieldchangesduringthescan,anxietyduetoclaustrophobiaandallergicreactiontothecontrastagent.Theallergicreactionmayincludeheadache,nausea,rash,hives,nasalcongestion,sneezing,itchingorswelling.Ifaseverereactionoccurs,swellingofthethroat,chesttightness,oramarkeddropinbloodpressuremayoccur.Inaddition,pain,bleeding,bruising,coldnessorinflammationattheinjectionsitemayoccur.Precautionswillbetakenforearlydetectionandrapidtreatmentifsuchreactionsoccur.PotentialcomplicationsofCTscanincluderadiationexposureandallergicreactiontoCTcontrastagents.Radiationdoses:Combinedscanningwithcomprehensivestrokeimaging,whichincludesanoncontrastheadCTscan,perfusionimaging,andCTangiographyofthecervicocranialvesselsstartingattheaorticarchresultsinadoseofapproximately7‐10mSv.(AJNR201031:1003‐1009).AccordingtotheNationalCouncilonRadiationProtectionandMeasurement,theaverageannualradiationdoseperpersonintheU.S.is620millirem(6.2milliSieverts).Reactionstocontrastagents:

MildNausea,Vomiting,Headache,Cough,Nasalstuffiness,Alteredtaste,Flushing,Itching,Rash,Hives,Sweats,Swellingofeyesorface

ModerateMildhypotension,TachycardiaorBradycardia,Bronchospasm,Wheezing,Dyspnea,Laryngealedema,Generalizedordiffuseerythema

SevereCardiopulmonaryarrest,Clinicallymanifestedarrhythmias,Profoundhypotension,Convulsions,Unresponsiveness,Respiratoryfailure,LaryngealedemaTherateofmajorreactions(e.g.,anaphylaxis,death)isverylow,estimatedatonein170,000administrations.

Potentialcomplicationsofendovasculartherapyincludestroke;newclotinanartery;totalblockageofanartery;infectionandpainintheregionofinsertionsite;lackofbloodflowtothebrain;ruptureorpunctureofanartery;significanttearingofthevesselwall;bleedingrequiringbloodtransfusion;allergicreactiontocontrastdye;abnormallowbloodpressurerequiringtreatment;temporaryclosingoftheartery(vesselspasm);formationofordislodgmentsofclotswhichblockthearteries(embolism).Inrarecircumstances,theprocedurecouldresultindeath.Atthepuncturesiteinthegroin,abloodclotorotherbloodvesselinjurymayoccurandrequirebloodtransfusionorsurgicalrepair.Infectionmayoccuratthepuncturesite;thiscouldcausepainandrequireadditionalmedications.Thereissomechanceofanallergicreactiontothex‐raycontrast(dye)usedduringtheangiogramprocedure.Minorallergicreactionsmayincludearashorhives.Thereisalsothepossibilityofaseriousallergicreactionthatcouldincludeshortnessofbreathandswelling,dropin


19

bloodpressure,andevendeath.Patientswillbecloselymonitoredforthesereactionsandreceiveprompttreatmenttoreverseanyallergicreactions.Safetyofendovasculartherapybeyond8hoursMechanicalthrombectomydeviceshavebeenusedbeyond8hoursofstokeonsetinanumberofclinicaltrialsandregistries.Nosafetyconcernshavebeenassociatedwithlatewindowtherapy.InDEFUSE237,patientsweretreatedupto12hoursaftersymptomonset,andnosafetyconcernswereidentifiedinanytimewindow.Basedonbothfavorablesafetydataandencouragingefficacydata,StrykerNeurovascularhasinitiatedtheDAWNStudywitha24hourtreatmentwindow.NosafelyissueshavebeenidentifiedtodateinDAWN.TheDAWNstudyisbeingrununderanFDAIDE.MethodstomitigateriskstosubjectsinthetrialMethodstomitigateriskstosubjectsinthetrialincludeexclusionofsubjectswithbleedingdisordersandselectionofsubjectsvianeuroimaging(infarctcorelesionslessthan70ml)tominimizetheriskofsymptomaticintracranialhemorrhage.ComputedTomography(CT)scanswillbeperformedforneurologicaldeterioration(≥4pointincreaseinNationalInstitutesofHealthStrokeScale(NIHSS)score)toidentifynewstrokes,hemorrhage,oredema.Hospitalswillfollowtheirlocalstandardofcaresafetyproceduresinordertoreducetheriskofkidneydysfunctioncausedbycontrastagents.Onlyinvestigatorswhoaretrainedandexperiencedwithuseofthedevicesallowedwithinthetrialareeligibletoparticipate(seeSiteApprovalandMonitoringPlan)above.Theadaptivedesignwilleliminatesubgroupswithanunfavorabletherapeuticresponse.PatientswillbecarefullyscreenedforCT,MRIandendovasculartreatmentcontraindicationsaccordingtotheinclusion/exclusioncriteriaandexcludedfromenrollmentifanyarepresent.Radiationexposure:Radiationexposureduringalltestswillbeminimizedbyoptimizingtheimagingprotocolsandbylimitingfluoroscopy‐timeduringtheendovascularprocedure.AllCTsequences,includingtheCTPsequence,meetallFDAguidelinesforradiationexposure.StoppingrulesorsafetytriggersforthestudySymptomaticICHordeathratesthatexceedpre‐specifiedthresholdswilltriggerameetingoftheDSMBtodiscusstheeventsandmakeadeterminationonthecontinuationofthetrial.Belowarethepre‐specifiedtriggers:TheDEFUSE3hasestablishedthefollowingautomaticstoppingrules,basedonidentifyingwith95%probability:1) therateofsymptomaticICH(NIHSSworseningof4ormorepointsassociatedwithICH)in

theendovasculargroupisexceeding10%2) therateof90daymortalityintheendovasculargroupisexceeding20%.Ifeitherthresholdiscrossed,thestudywillbeautomaticallyplacedonholduntiltheinvestigatorsandtheDSMBcanconductareviewofevents.


20

Adverseevents(AEs)willbecollected,recorded,andanalyzedinaccordancewithSection3.11below.SafetyoversightforthisstudywillbeprovidedbyboththeDSMBandanindependentMedicalSafetyMonitor.PleaseseeSection3.11fordetails.PatientPopulation:Fourhundredandseventysixacutestrokepatientsmeetingthepre‐definedinclusioncriteriawillbeenrolledinthetrial.Themeanageisanticipatedtobe69yearsofage.Ourtargetedplannedenrollmentbreakdownisasfollows:

RacialCategories

EthnicCategories

TotalNothispanicorLatino HispanicorLatino

Female Male Female Male

American Indian/Alaska Native 2 2 1 1 6

Asian 13 13 1 1 28

Native Hawaiian or Other Pacific Islander 2 3 0 0 5

Black or African American 32 32 1 1 66

White 175 174 11 11 371

Racial Categories: Total of All Subjects 224 224 14 14 476

Imagingcorelab:TheStanfordimagingcorelabhas15yearsofexperiencewithMRimagestorageandprocessing.Itwillperformtheorganization,archivingandblindedanalysisofallimagingdatacollectedinDEFUSE3.TheywillberesponsibleforMRIandCTPerfusionimageprocessingandartifactremovalandwillgeneratefinallesionvolumesforallMRIscansperformedinthestudy.

DSAAngiogramssenttothecorelabwillincludeabaseline(pre‐treatment)andafinalangiogramfortheterritoryoftreatment.Inaddition,angiographicimagesfromMRAorCTAwillbesenttothecorelab.TheMRAorCTAwillbeusedtoassignaprimaryarteriallesion(AOL)fromnon‐invasiveimaging.ThebaselineDSAangiogramwillalsobeusedtoassignaprimaryarterialocclusivelesion(AOL)andapre‐treatmentmTICIscore.Thefinalangiogramwillhaveapost‐treatmentmTICIscoreassigned.22,52ThisscoringsystemdefinesTICI2Aaspartialperfusionof<50%ofthevasculardistributionoftheoccludedarteryand2Baspartialperfusionof>50%ofthevasculardistribution.Thescoringsystemwillusethesepreviouslydescribeddefinitions22:

Grade0Noperfusion

Grade1Antegradereperfusionpasttheinitialocclusion,butlimiteddistalbranchfillingwithlittleorslowdistalreperfusion

Grade2aAntegradereperfusionoflessthanhalfoftheoccludedtargetarterypreviouslyischemicterritory(eg,in1majordivisionoftheMCAanditsterritory)

Grade2bAntegradereperfusionofmorethanhalfofthepreviouslyoccludedtargetarteryischemicterritory(eg,in2majordivisionsoftheMCAandtheirterritories)


21

Grade3Completeantegradereperfusionofthepreviouslyoccludedtargetarteryischemicterritory,withabsenceofvisualizedocclusioninalldistalbranches

Twoseniorneurointerventionalistswillperformtheangiographicanalysis,blindedtotheclinicaldata,MRimagingandCTPerfusionresults,andtheanalysisoftheothercorelabreader.AnydisagreementsintheAOLinterpretationorthemTICIscoreswillbeadjudicatedbycommonreviewofthosecasesandaconsensusreadingwillbeapplied.

NationalDataManagementCenter:DatamanagementandsitemonitoringwillbeperformedbytheStrokeNetNationalDataManagementCenter(NDMC)atMedicalUniversityofSouthCarolina(Director, ,PhD,seeletterofsupport).TheNDMCwillcreatethedatabaseandsetuptheinterfaceonthewebsite(WebDCU™)whereclinicalsitepersonnelwillenterthedataintotheelectronicCRF.Dataqualityassuranceprocessesinclude:(1)logicandrulechecksbuiltintothedatabase;(2)monitoringbytheDataManagerattheNDMC;(3)centralmonitoringbythestatisticalprogrammerattheNDMC;and(4)risk‐basedsourceverificationmonitoringbytheClinicalResearchAssociates.DEFUSE3data,includingneuroimaging,willbesharedinaccordancewiththeStrokeNetdatasharingpolicies.AnonymizedneuroimagingwillbestoredonsecureserversattheStanfordStrokeCenterwithnightlyback‐up.NINDSCommonDataElementswillbeusedforbothclinicalandimagingdata.

3.10.DescriptionofdevicesThefollowingFDAapprovedthrombectomydeviceswillbeincluded:

1) TrevoRetriever2) Solitaire™RevascularizationDevice3) CovidienMindFrameCaptureRevascularizationDevice4) Penumbrathrombectomysystemincludingthefollowingdevicesandpumps:

Penumbra Aspiration Pump (1115V) Penumbra System [026, 032, 041]

Penumbra System 054 Penumbra System Separator Flex [026, 032, 041, 054]

Penumbra System MAX Penumbra Pump MAX

Penumbra System 110 Aspiration Tubing Penumbra System Reperfusion Catheter ACE64 & ACE68

3.11.Monitoringprocedures

ThecoordinationoftheDEFUSE3Trialoperationswillbecentralizedthroughthefollowing:NIHStrokeNetNationalCoordinatingCenter(NCC)/PI: ,MDUniversityofCincinnati

Cincinnati,Ohio LeadingtheNCCteamwillbetheProjectManager,whowillbeassignedtocoordinatethefollowingstudyoversight:trialcommunicationrequiredtrainingactivities,siteassessmentand/orinitiationvisits,collectionoftrialrelatedregulatorydocuments,recruitmentperformancetracking,sitemonitoring,andperformanceanalysis.StudyoversightwillbehandledaccordingtotheDataMonitoringStandardOperatingProcedure(SOPNumberADM19).


22

DEFUSE3willhaveanindependentDataandSafetyMonitoringBoard(DSMB)appointedbytheNIHtooverseestudysafety.Patientsinbothstudyarmswillbeassessedfortheincidenceofstroke‐relatedmortalityat90days,theincidenceofsymptomaticintracranialhemorrhageat36hoursfromsymptomonset(definedasa≥4pointworseningofimmediatepre‐deteriorationNIHSSneurologicalstatusvs.postdeteriorationandassociatedwithbrainhemorrhage),andtheincidenceofsignificantneurologicdeteriorationpriortodischarge(definedas≥4pointworseningoftheimmediatepre‐deteriorationNIHSSneurologicalstatusvs.postdeteriorationandnotattributedtosedation).Intheendovasculararmpatientswillbeassessedforintra‐proceduralcomplicationsincluding:intra‐proceduralmortality,vesselperforation,arterialdissection,accesssitecomplicationrequiringsurgicalrepairorbloodtransfusion,embolizationanddevicefailure.SAEswillbereportedwithin24hoursofawarenessoftheevent.TheDSMBwillmeetinpersonorbyteleconference,onasemi‐annualbasis,tomonitorthecumulativesafetydataduringparticipantfollow‐up.Innoinstancewillmorethan12monthselapsebetweenDSMBreviewsofcumulativesafetydataafterthefirstparticipanthasbeenrandomized.TheDSMBwillmonitorthestudyaccordingtotheguidelinesspecifiedinthestudyprotocolandtheoperatingproceduresestablishedattheinitialmeeting,unlesstheDSMBdeterminesduringthecourseofthetrialthatmodificationoftheguidelinesisinthebestinterestofthestudyanditsparticipants.IndependentMedicalSafetyMonitor:InadditiontotheDSMB,Dr. hasbeenappointedastheindependentMedicalSafetyMonitor(MSM)forDEFUSE3.Dr. isnotinvolvedinthestudyandhasnoconflictofinterest.HewillberesponsibleforongoingmonitoringofreportsofSAEssubmittedbytheclinicalcentersinrealtimetoensuregoodclinicalpracticeandtoidentifysafetyconcernsquickly.Dr. maysuggestprotocolmodificationstopreventtheoccurrenceofparticularAEs,e.g.,modifyingtheprotocoltorequirefrequentmeasurementoflaboratoryvaluespredictiveoftheeventortoimproveexpeditiousidentificationofSAEs.IntheeventofunexpectedSAEsoranundulyhighrateofSAEs,Dr. willpromptlycontacttheDSMBLiaisonwhowillnotifytheDSMBChair.Intheeventthatheisunavailableforanextendedperiodoftime(i.e.,extendedvacation,sabbatical,illness,etc.),aback‐upMSMwillbenominatedbythestudyPIandapprovedbytheDSMB.AdverseEventReportingConsiderationofadverseeventswillhereafterconsistofadverseevents,seriousadverseevents,andadversedeviceeffects,includinganticipatedadversedeviceeffectsandunanticipatedadversedeviceeffects.

• Adverse event (AE)is definedas anyuntoward/undesirableclinicaloccurrence inaclinicalinvestigationofasubjectwhichdoesnotnecessarilyhaveacausalrelationshipwiththetreatmentunderinvestigation.AnAdverseEventcanthereforebeanyunfavorableand/orunintendedsign, symptom, or disease temporally associatedwith the use of a device product, whether or notconsideredrelated to thedeviceproduct.Onlyabnormal laboratoryvalues thataredeemedclinicallysignificantbytheinvestigatorwillbeclassifiedasadverseevents.


23

• Seriousadverseevent(SAE)isdefinedasanyuntoward/undesirableadverseexperiencethatresultsinanyof the followingoutcomes:1)death;2)a life‐threateningadverseexperience;3) inpatienthospitalization or prolongation ofexisting hospitalization; 4) a permanent/persistent or significantdisability/incapacityora congenitalanomaly/birthdefect;5) importantmedicalevents thatmaynotresultindeath,belife‐threatening,orrequirehospitalizationmaybeconsideredaseriousadverseeventwhen, based upon appropriatemedicaljudgment, theymay jeopardize the subject andmay requiremedicalorsurgicalinterventiontopreventoneoftheoutcomeslistedinthisdefinition.Thiscategoryincludestheuseofintra‐arterialthrombolyticsand/orintracranialstents.

• Anticipated adverse device effect (AADE) is defined asany adverse effectrelated to

the device orprocedure,whichisidentifiedintheprotocolortheIFUforthedevice.

• UnanticipatedAdverseDevice Effects (UADEs) is defined as any serious adverseeffect onhealthorsafetyorany life‐threateningproblemordeathcausedby,orassociatedwithadevice, if that effect,problem, ordeathwasnotpreviouslyidentifiedinnature,severity, ordegreeof incidenceintheinvestigationalplanorapplication,oranyotherunanticipatedseriousproblemassociatedwithadevicethatrelatesto therights,safety,orwelfareofsubjects.

ReportsofUADEswillbemadetotheFDAwithin10daysofreceivingnotificationoftheUADE(asrequiredin21CFR81218p(b)(1)).SafetyMonitoring

TheMSMwillmonitorallAEreportstoidentifyandtrendalleventsthatwouldrequiretemporarydiscontinuationofstudyenrollment,tofullycharacterizedevicesafety,tomodifythestudyprotocol,ortoterminatethestudy.

ReportingProceduresforAllAdverseEventsAll Adverse Events,whetherornotattributedtothestudyand/orthedevices,observedbytheinvestigatororreportedbythesubject, will be recordedfromthetimeofrandomizationthroughDay5ordischarge,whicheverisearlier.AllSAEswillberecordedthroughDay90.

The following attributes w i l l beassignedbythereportinginvestigator:

1. Descriptionofevent2. Dateofonset3. Dateofresolution(ifapplicable)4. Seriousness5. Relationshiptothestudydeviceand/orprocedure(s)6. Severity7. Action(s)taken8. Outcome(s)


24

Severityisdefinedasameasureoftheintensityofareaction,effectorexperience.Themeasurement(s)aredescribedasmild,moderate,severe,lifethreateningordeath.Theeventitself,however,maybeofrelativeminormedicalsignificance.TheseverityofAdverseEventsisassessedaccordingtothefollowingindexscale:

Mildasymptomaticormildsymptoms;clinicalordiagnosticobservationsonly;interventionnotindicated

Moderateminimal,localornoninvasiveinterventionindicated;limitingage‐appropriateinstrumentalActivitiesofDailyLiving.

Severemedicallysignificantbutnotimmediatelylife‐threatening;hospitalizationorprolongationofhospitalizationindicated;disabling;limitingselfcareActivitiesofDailyLiving.

Life‐threateningconsequences;urgentinterventionindicated

DeathrelatedtoAE

TherelationshipofanAEtothestudydeviceorprocedurewillbegradedasfollows: Unrelated Unlikely Reasonablepossibility Definitely

SeriousAdverseEventsAllSeriousAdverseEventsincludingdeathswillbereportedtotheMSM, theCentra l InstitutionalReviewBoard (IRB)andtheFDA,as required.


25

4. INVESTIGATOR’SAGREEMENT

AsthisstudywillbecarriedoutbytheNIHStrokeNet,thenamesofthespecificsitesandinvestigatorsarenotyetavailable.Investigator’sagreementsforkeyinvestigatorsatthecoordinatingsiteareincluded.Allinvestigatorswillberequiredtosignthefollowingagreement: INVESTIGATOR AGREEMENT FOR THE CLINICAL INVESTIGATION OF THE DEFUSE 3 TRIAL I___________________________________________________ agree to participate as an Investigator on the DEFUSE 3 trial. I have been provided a copy of the following Food and Drug Administration (FDA) regulations: 21 CFR Part 812, Investigational Device Exemptions; 21 CFR Part 50, Protection of Human Subjects; and 21 CFR Part 54, Financial Disclosure by Clinical Investigators. I agree and/or certify that:

1. I will conduct the clinical investigation in accordance with this agreement, all requirements of the investigational plan, IDE regulations, other applicable regulations of the FDA, and any conditions of approval imposed by my reviewing Institutional Review Board (IRB) or FDA. I agree to abide by all of the responsibilities of Investigators addressed under 21 CFR Part 812, Subpart E and Subpart G, including but not limited to the following:

2. I will obtain written approval from the authorized IRB for the institution at which this

investigation will be conducted.

3. I will ensure that Informed Consent is obtained from each subject participating in this clinical investigation in accordance with the informed consent regulation found in 21 CFR Part 50, and that a signed copy of the informed consent is available to the sponsor (sponsor-investigator) and the sponsor’s (sponsor- investigator’s) designated monitor.

4. I will ensure the accurate completion of protocol case report forms and, if I am not also the

sponsor- investigator of the corresponding IDE application, I will submit completed protocol case report forms to the sponsor (sponsor-investigator) at the time frames specified in the Protocol and/or FDA regulations.

5. I have the appropriate, relevant qualification to conduct and to oversee the conduct of the

clinical investigation as documented by the following: (initial applicable statement) ____ My relevant qualifications, including dates, location, extent and type of experience are listed in my most recent curriculum vitae (CV), which is attached to the Agreement and which will be maintained by the sponsor (sponsor-investigator) of the corresponding IDE application. ____ My curriculum vitae (CV) does not reflect my relevant qualifications, therefore attached to this Agreement is a statement of my relevant experience (including dates,


26

location(s), extent and type of experience) which will be maintained by the sponsor (sponsor-investigator) of the corresponding IDE application.

6. There are no reasons to question my ability to oversee the appropriate conduct of this

clinical investigation. (Initial applicable statement) ____ I have never participated in an investigation or other research activity which was terminated (disqualified) by the FDA, IRB (or equivalent), or sponsor of a study due to non-compliance issue. ____ I have participated in an investigation or other research activity which was terminated (disqualified) by the FDA, IRB (or equivalent), or sponsor of a study due to non-compliance issue. The specific circumstances leading to this termination and my role in the respective problems or issues and the resolution of these problems or issues are summarized in an attachment to this Agreement.

I further certify that I have not been debarred under the Generic Drug Enforcement Act of 1992, 21 USC §§ 335a and 335b. In the event that I become debarred or receive notice of an action or threat of action with respect to my debarment during the term of this Agreement, I agree to immediately notify the sponsor (sponsor-investigator) and the authorized IRB for my study site. If I am the sponsor-investigator of the corresponding IDE application I will notify the authorized IRB and the FDA. As required by 21 CFR Part 54, Financial Disclosure by Clinical Investigator, I will disclose sufficient and accurate financial information to the sponsor (sponsor-investigator) by completing the Certification of Financial Interest Form (attached) and if applicable, the Disclosure of Financial Interest Form (attached). I will also notify the sponsor (sponsor-investigator) if my disclosed financial information changes at any time during the clinical investigation or up to one year following the closure of the study. Site Name and Address: __________________________________________ __________________________________________ __________________________________________ ________________________________________ __________ Investigator Signature Date


27

5. EXECUTIVECOMMITTEE/KEYPARTICIPATINGINVESTIGATORS

TheExecutivecommittee,composedofexpertsinvascularneurology,endovasculartherapyandneuroimaging,willprovidetheoverallscientificguidanceforthestudy.Thecommitteewilltypicallymeetmonthlybyphone(1hour/month)forthefulldurationofthestudy.Responsibilitiesincludeoversightoftheoverallconductofthestudywithregardtoprotocolcomplianceandmodifications/amendments,studyprogress,andproblem‐solving.Dr.Alberswillchairtheexecutivecommittee.KeyParticipatingInvestigatorsatCoordinatingSiteGregoryW.Albers,MDPrincipalInvestigatorStanfordStrokeCenter780WelchRd.Suite350PaloAlto,CA94305650‐723‐[email protected]

MichaelMarks,MDCo‐PrincipalInvestigatorStanfordUniversityMedicalCenter300PasteurDr.Stanford,CA94305‐5105650‐723‐[email protected]

MaartenLansberg,MD,PhDProtocolDirectorStanfordStrokeCenter780WelchRd.Suite350PaloAlto,CA94305650‐723‐[email protected]

6. INSTITUTIONALREVIEWBOARDTheUniversityofCincinnatiInstitutionalReviewBoardwillserveastheNationalCentralInstitutionalReviewBoardforallparticipatingsites.TheCentralInstitutionalReviewBoard(CIRB)formulticenterprotocolsisthesingleIRBofrecord.IthasregulatoryresponsibilityforassuringtheprotectionoftherightsandwelfareofresearchparticipantsinaccordancewithStandardOperatingProcedureADM12;CentralInstitutionalReviewBoardReporting.TheNationalInstituteofNeurologicalDisordersandStroke(NINDS)selectedtheUniversityofCincinnatiInstitutionalReviewBoard(IRB)toserveastheCIRBfortheNIHStrokeNet(StrokeNet).UniversityofCincinnatiIRBRegistration#00000180FWA#:00003152ExpirationDate:6/27/2016

,PhD,CIPChairmanCIRB

7. COSTS

Alloftheeligibledevicesthatwillbeusedinthisstudyarecurrentlyonthemarket.Therewillbenochargesbeyondthetypicalstandardofcareforuseoftheseapproveddevices.Thesedeviceswillbeusedandbilledaccordingtothestandardofcareforeachinstitution.


28

8. REFERENCES1. LopezAD,MathersCD,EzzatiM,JamisonDT,MurrayCJL.Globalandregionalburdenof

diseaseandriskfactors,2001:Systematicanalysisofpopulationhealthdata.TheLancet.2006;367:1747‐1757

2. KleindorferD,LindsellCJ,BrassL,KoroshetzW,BroderickJP.Nationalusestimatesofrecombinanttissueplasminogenactivatoruse:Icd‐9codessubstantiallyunderestimate.Stroke.2008;39:924‐928

3. KatzanIL,HammerMD,HixsonED,FurlanAJ,Abou‐CheblA,NadzamDM.Utilizationofintravenoustissueplasminogenactivatorforacuteischemicstroke.ArchNeurol.2004;61:346‐350

4. QureshiAI,KirmaniJF,SayedMA,SafdarA,AhmedS,FergusonR,etal.Timetohospitalarrival,useofthrombolytics,andin‐hospitaloutcomesinischemicstroke.Neurology.2005;64:2115‐2120

5. SaqqurM,UchinoK,DemchukAM,MolinaCA,GaramiZ,CallejaS,etal.Siteofarterialocclusionidentifiedbytranscranialdopplerpredictstheresponsetointravenousthrombolysisforstroke.Stroke.2007;38:948‐954

6. MoriE,YonedaY,TabuchiM,YoshidaT,OhkawaS,OhsumiY,etal.Intravenousrecombinanttissueplasminogenactivatorinacutecarotidarteryterritorystroke.Neurology.1992;42:976‐982

7. delZoppoGJ,PoeckK,PessinMS,WolpertSM,FurlanAJ,FerbertA,etal.Recombinanttissueplasminogenactivatorinacutethromboticandembolicstroke.AnnNeurol.1992;32:78‐86

8. AlbersGW,ThijsVN,WechslerL,KempS,SchlaugG,SkalabrinE,etal.Magneticresonanceimagingprofilespredictclinicalresponsetoearlyreperfusion:Thediffusionandperfusionimagingevaluationforunderstandingstrokeevolution(defuse)study.AnnalsofNeurology.2006;60:508‐517

9. LansbergMG,SchrootenM,BluhmkiE,ThijsVN,SaverJL.Treatmenttime‐specificnumberneededtotreatestimatesfortissueplasminogenactivatortherapyinacutestrokebasedonshiftsovertheentirerangeofthemodifiedrankinscale.Stroke.2009;40:2079‐2084

10. FieldsJD,LindsayK,LiuKennethC,NesbitGM,LutsepHL.Mechanicalthrombectomyforthetreatmentofacuteischemicstroke.ExpertReviewofCardiovascularTherapy.2010;8:581‐592

11. LeeM,HongK‐S,SaverJL.Efficacyofintra‐arterialfibrinolysisforacuteischemicstroke:Meta‐analysisofrandomizedcontrolledtrials.Stroke.2010;41:932‐937

12. SmithWS,SungG,SaverJ,BudzikR,DuckwilerG,LiebeskindDS,etal.Mechanicalthrombectomyforacuteischemicstroke:Finalresultsofthemultimercitrial.Stroke.2008;39:1205‐1212

13. FurlanA,HigashidaR,WechslerL,GentM,RowleyH,KaseC,etal.Intra‐arterialprourokinaseforacuteischemicstroke.Theproactiistudy:Arandomizedcontrolledtrial.Prolyseinacutecerebralthromboembolism.Jama.1999;282:2003‐2011

14. ThomallaG,SchwarkC,SobeskyJ,BluhmkiE,FiebachJB,FiehlerJ,etal.Outcomeandsymptomaticbleedingcomplicationsofintravenousthrombolysiswithin6hoursinmri‐selectedstrokepatients:Comparisonofagermanmulticenterstudywiththepooleddataofatlantis,ecass,andnindstpatrials.Stroke.2006;37:852‐858

15. HsiaA,KidwellC.Developmentsinneuroimagingforacuteischemicstroke:Diagnosticandclinicaltrialapplications.CurrentAtherosclerosisReports.2008;10:339‐346


29

16. KakudaW,LansbergMG,ThijsVN,KempSM,BammerR,WechslerLR,etal.Optimaldefinitionforpwi/dwimismatchinacuteischemicstrokepatients.JCerebBloodFlowMetab.2008;28:887‐891

17. LansbergMG,ThijsV,BammerR,KakudaW,HamiltonS,WechslerL,etal.Clinicalandmri‐basedriskfactorsforsymptomaticintracerebralhemorrhagefollowingtreatmentwithtissueplasminogenactivator.Stroke.2006;37:654

18. SaverJL,JahanR,LevyEI,JovinTG,BaxterB,NogueiraRG,etal.Solitaireflowrestorationdeviceversusthemerciretrieverinpatientswithacuteischaemicstroke(swift):Arandomised,parallel‐group,non‐inferioritytrial.Lancet.2012;380:1241‐1249

19. NogueiraRG,LutsepHL,GuptaR,JovinTG,AlbersGW,WalkerGA,etal.Trevoversusmerciretrieversforthrombectomyrevascularisationoflargevesselocclusionsinacuteischaemicstroke(trevo2):Arandomisedtrial.Lancet.2012;380:1231‐1240

20. MarksMP,LansbergMG,MlynashM,KempS,McTaggartRA,ZaharchukG,etal.Angiographicoutcomeofendovascularstroketherapycorrelatedwithmrfindings,infarctgrowth,andclinicaloutcomeinthedefuse2trial.IntJStroke.2014

21. YooAJ,SimonsenCZ,PrabhakaranS,ChaudhryZA,IssaMA,FugateJE,etal.Refiningangiographicbiomarkersofrevascularization:Improvingoutcomepredictionafterintra‐arterialtherapy.Stroke.2013;44:2509‐2512

22. ZaidatOO,YooAJ,KhatriP,TomsickTA,vonKummerR,SaverJL,etal.Recommendationsonangiographicrevascularizationgradingstandardsforacuteischemicstroke:Aconsensusstatement.Stroke.2013;44:2650‐2663

23. BroderickJP,PaleschYY,DemchukAM,YeattsSD,KhatriP,HillMD,etal.Endovasculartherapyafterintravenoust‐paversust‐paaloneforstroke.NEnglJMed.2013;368:893‐903

24. HillMD,DemchukAM,GoyalM,JovinTG,FosterLD,TomsickTA,etal.Albertastrokeprogramearlycomputedtomographyscoretoselectpatientsforendovasculartreatment:Interventionalmanagementofstroke(ims)‐iiitrial.Stroke.2014;45:444‐449

25. KidwellCS,JahanR,GornbeinJ,AlgerJR,NenovV,AjaniZ,etal.Atrialofimagingselectionandendovasculartreatmentforischemicstroke.NEnglJMed.2013;368:914‐923

26. O.A.Berkhemer,P.S.S.Fransen,D.Beumer,etal.ARandomizedTrialofIntraarterialTreatmentforAcuteIschemicStroke.NEnglJMed2015;372:11‐20.

27. MuirKW,BuchanA,vonKummerR,RotherJ,BaronJ‐C.Imagingofacutestroke.TheLancetNeurology.2006;5:755‐768

28. CampbellB,PurushothamA,ChristensenS,DesmondP,NagakaneY,ParsonsM,etal.Theinfarctcoreiswellrepresentedbytheacutediffusionlesion:Sustainedreversalisinfrequent.Journalofcerebralbloodflowandmetabolism:officialjournaloftheInternationalSocietyofCerebralBloodFlowandMetabolism.2012;32:50‐56

29. ChemmanamT,CampbellBC,ChristensenS,NagakaneY,DesmondPM,BladinCF,etal.Ischemicdiffusionlesionreversalisuncommonandrarelyaltersperfusion‐diffusionmismatch.Neurology.2010;75:1040‐1047

30. BairdAE,WarachS.Magneticresonanceimagingofacutestroke.JCerebBloodFlowMetab.1998;18:583‐609

31. DaniKA,ThomasRG,ChappellFM,ShulerK,MacLeodMJ,MuirKW,etal.Computedtomographyandmagneticresonanceperfusionimaginginischemicstroke:Definitionsandthresholds.AnnNeurol.2011;70:384‐401

32. OlivotJM,MlynashM,ThijsVN,KempS,LansbergMG,WechslerL,etal.Optimaltmaxthresholdforpredictingpenumbraltissueinacutestroke.Stroke.2009;40:469‐475


30

33. Zaro‐WeberO,Moeller‐HartmannW,HeissWD,SobeskyJ.Mapsoftimetomaximumandtimetopeakformismatchdefinitioninclinicalstrokestudiesvalidatedwithpositronemissiontomography.Stroke.2010;41:2817‐2821

34. OlivotJM,MlynashM,ZaharchukG,StrakaM,BammerR,SchwartzN,etal.Perfusionmri(tmaxandmtt)correlationwithxenonctcerebralbloodflowinstrokepatients.Neurology.2009;72:1140‐1145

35. MlynashM,LansbergMG,DeSilvaDA,LeeJ,ChristensenS,StrakaM,etal.Refiningthedefinitionofthemalignantprofile:Insightsfromthedefuse‐epithetpooleddataset.Stroke;ajournalofcerebralcirculation.2011;42:1270‐1275

36. LansbergMG,LeeJ,ChristensenS,StrakaM,DeSilvaDA,MlynashM,etal.Rapidautomatedpatientselectionforreperfusiontherapy:Apooledanalysisoftheechoplanarimagingthrombolyticevaluationtrial(epithet)andthediffusionandperfusionimagingevaluationforunderstandingstrokeevolution(defuse)study.Stroke.2011;42:1608‐1614

37. LansbergMG,StrakaM,KempS,MlynashM,WechslerLR,JovinTG,etal.Mriprofileandresponsetoendovascularreperfusionafterstroke(defuse2):Aprospectivecohortstudy.LancetNeurology.2012;11:860‐867

38. LansbergMG,CeredaCW,MlynashM,MishraNK,InoueM,KempS,etal.Responsetoreperfusionnottime‐dependentinpatientswithsalvageabletissue.Neurology,inpress2015

39. InoueM,MlynashM,ChristensenS,WheelerHM,StrakaM,TipirneniA,etal.Earlydiffusion‐weightedimagingreversalafterendovascularreperfusionistypicallytransientinpatientsimaged3to6hoursafteronset.Stroke.2014;45:1024‐1028

40. WheelerHM,MlynashM,InoueM,TipirneniA,LigginsJ,ZaharchukG,etal.Earlydiffusion‐weightedimagingandperfusion‐weightedimaginglesionvolumesforecastfinalinfarctsizeindefuse2.Stroke.2013;44:681‐685

41. ParsonsMW,ChristensenS,McElduffP,LeviCR,ButcherKS,DeSilvaDA,etal.Pretreatmentdiffusion‐andperfusion‐mrlesionvolumeshaveacrucialinfluenceonclinicalresponsetostrokethrombolysis.JCerebBloodFlowMetab.2010;30:1214‐1225

42. MarksMP,OlivotJM,KempS,LansbergMG,BammerR,WechslerLR,etal.Patientswithacutestroketreatedwithintravenoustpa3‐6hoursafterstrokeonset:Correlationsbetweenmrangiographyfindingsandperfusion‐anddiffusion‐weightedimaginginthedefusestudy.Radiology.2008;249:614‐623

43. WeisstannerC,GratzPP,SchrothG,VermaRK,KochlA,JungS,etal.Thrombusimaginginacutestroke:Correlationofthrombuslengthonsusceptibility‐weightedimagingwithendovascularreperfusionsuccess.EurRadiol.2014

44. PocockSJ,SimonR.Sequentialtreatmentassignmentwithbalancingforprognosticfactorsinthecontrolledclinicaltrial.Biometrics.1975;31:103‐115

45. ChimowitzMI,LynnMJ,TuranTN,FiorellaD,LaneBF,JanisS,etal.Designofthestentingandaggressivemedicalmanagementforpreventingrecurrentstrokeinintracranialstenosistrial.JStrokeCerebrovascDis.2011;20:357‐368

46. PereiraVM,GrallaJ,DavalosA,BonafeA,CastanoC,ChapotR,etal.Prospective,multicenter,single‐armstudyofmechanicalthrombectomyusingsolitaireflowrestorationinacuteischemicstroke.Stroke.2013;44:2802‐2807

47. MolinaCA,MontanerJ,AbilleiraS,IbarraB,RomeroF,ArenillasJF,etal.Timingofspontaneousrecanalizationandriskofhemorrhagictransformationinacutecardioembolicstroke.Stroke.2001;32:1079‐1084


31

48. WolpertSM,BruckmannH,GreenleeR,WechslerL,PessinMS,delZoppoGJ.Neuroradiologicevaluationofpatientswithacutestroketreatedwithrecombinanttissueplasminogenactivator.Thert‐paacutestrokestudygroup.AJNRAmJNeuroradiol.1993;14:3‐13

49. AlexandrovAV,MolinaCA,GrottaJC,GaramiZ,FordSR,Alvarez‐SabinJ,etal.Ultrasound‐enhancedsystemicthrombolysisforacuteischemicstroke.NEnglJMed.2004;351:2170‐2178

50. RubinD.Multipleimputationfornonresponseinsurveys.Wiley&Sons;1987.51. CampbellBC,TuHT,ChristensenS,DesmondPM,LeviCR,BladinCF,etal.Assessing

responsetostrokethrombolysis:Validationof24‐hourmultimodalmagneticresonanceimaging.ArchNeurol.2012;69:46‐50

52. TomsickT,BroderickJ,CarrozellaJ,KhatriP,HillM,PaleschY,etal.Revascularizationresultsintheinterventionalmanagementofstrokeiitrial.AJNRAmJNeuroradiol.2008;29:582‐587

53. ChristouI,AlexandrovAV,BurginWS,WojnerAW,FelbergRA,MalkoffM,etal.Timingofrecanalizationaftertissueplasminogenactivatortherapydeterminedbytranscranialdopplercorrelateswithclinicalrecoveryfromischemicstroke.Stroke.2000;31:1812‐1816

54. NguyenTN,MalischT,CastonguayAC,GuptaR,SunCH,MartinCO,etal.Balloonguidecatheterimprovesrevascularizationandclinicaloutcomeswiththesolitairedevice:Analysisofthenorthamericansolitaireacutestrokeregistry.Stroke.2014;45:141‐145

55. SanRomanL,ObachV,BlascoJ,MachoJ,LopezA,UrraX,etal.Single‐centerexperienceofcerebralarterythrombectomyusingthetrevodevicein60patientswithacuteischemicstroke.Stroke.2012;43:1657‐1659

56. TurkAS,FreiD,FiorellaD,MoccoJ,BaxterB,SiddiquiA,etal.Adaptfaststudy:Adirectaspirationfirstpasstechniqueforacutestrokethrombectomy.JNeurointervSurg.2014;6:260‐264

57. BangOY,SaverJL,LeeKH,KimGM,ChungCS,KimSJ,etal.Characteristicsofpatientswithtargetmagneticresonancemismatchprofile:Datafromtwogeographicallyandraciallydistinctpopulations.CerebrovascDis.2010;29:87‐94

58. HackeW,FurlanAJ,Al‐RawiY,DavalosA,FiebachJB,GruberF,etal.Intravenousdesmoteplaseinpatientswithacuteischaemicstrokeselectedbymriperfusion‐diffusionweightedimagingorperfusionct(dias‐2):Aprospective,randomised,double‐blind,placebo‐controlledstudy.Lancet.Neurology.2009;8:141‐150

59. LansbergMG,ThijsVN,BammerR,OlivotJM,MarksMP,WechslerLR,etal.Themra‐dwimismatchidentifiespatientswithstrokewhoarelikelytobenefitfromreperfusion.Stroke.2008;39:2491‐2496

60. MishraNK,AlbersGW,ChristensenS,MarksM,HamiltonS,StrakaM,etal.Comparisonofmagneticresonanceimagingmismatchcriteriatoselectpatientsforendovascularstroketherapy.Stroke.2014;45:1369‐1374


32

APPENDIX

I. DEFUSE3PatientInformedConsentForm

Documents

Endovascular Therapy Following Imaging Evaluation for ... · Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3 Protocol Version/Version Date Version 2.4 April