UPICT CT Lung Nodule V1.3A

QIBA Volumetric CT Technical Committee Proffered Protocol to UPICT (continued)

QIBAv-CTChestV1.4

2010.09.09

X.ImagingProtocol–LungTumorVolumesastheBasisforResponseEvaluationCriteriaInSolidTumors(RECIST)oftheChest

InstructionstoClinicalTrialistswhoareadaptingthisimagingprotocolforinclusionintheirClinicalTrialProtocolareshowninitalics.Allitalictextshouldgenerallyberemovedaspartofpreparingthefinalprotocoltext.

0.ExecutiveSummary

Thisprotocoldescribesimageacquisition,processing,analysis,changemeasurementsandinterpretationforquantitativelyevaluatingtheprogression/regressionofmeasurabletargetlesions,thatis,lungtumorsgreaterthan1cminLongestDiameter(LD)atbaselinethatareselectedforfollowupduringacourseoftreatment.Itisintendedtoprovide“twicethesensitivityofusingsumoflongestdiameters(SLD)asthebasisforRECIST1.1”.

TheprotocolusesthesumofwholetumorvolumesinsteadofSLDsasthebasisforRECIST.Otherwise,mostallofthedefinitionsandrulesdescribedbyRECIST1.1stillapply.Theonlyexceptionsarethatthereneednotbealimitonthenumberofmeasurabletargetlesionsselected,asthelong-termgoalistoquantifythewholetumorburdensufferedbyagivenpatientateachtime-pointduringtreatment.

1.ContextoftheImagingProtocolwithintheClinicalTrial

1.1.UtilitiesandEndpointsoftheImagingProtocol

Thisimageacquisitionandprocessingprotocolisappropriateforquantifyingthevolumesofsolidtumorsofthelung,andlongitudinalchangesinvolumeswithinsubjects.

Followingthisprotocolisexpectedtoprovideintra-ratertest-retestprecisionofmeasurementforwholetumorvolumeofnolessthan18%.Thisvalueshouldbe“twiceassensitiveasRECIST”,basedontheideathatforuniformlyexpandingcubesandsolidspheres,a20%increaseintheRECISTdefineduni-dimensionalLongestDiameterofaMeasurableLesioncorrespondstoanincreaseinvolumeofabout73%,andtodiagnoseProgressiveDiseaseatachangeofaboutonehalfthatvolume,36%,thenoiseneedstobelessthanabout18%.

Thisprotocolisotherwiseagnostictotheclinicalsettingsinwhichthemeasurementsaremadeandthewaythemeasurementswillbeusedtomakedecisionsaboutindividualpatientswithcancerornewtreatmentsforpatientswithcancer.Typicalusesmightincludeassessingresponsetotreatment,establishingthepresenceofcertainprogressioneventsfordeterminingTTP,PFS,etc.,distinguishingbetweenarmsofclinicaltrials,ordeterminingeligibilityofpotentialsubjectsinclinicaltrials.

1.2.TimingofImagingwithintheClinicalTrialCalendar

Thisprotocoldoesnotpresumeaspecifictiming.

Generally,perRECIST1.1,"allbaselineevaluationsshouldbeperformedascloseaspossibletothetreatmentstart".Inclinicaltrials,thereisanexpectationthatfollowupscanswillbeacquiredatregular,protocol-specificintervals.

1.3.ManagementofPre-enrollmentImaging

Toquantifyvolumesandvolumechangeswiththeprecisionclaimedinthisprotocol,thepre-treatmentimageacquisitionandprocessingmustmeetorexceedtheminimumspecificationsdescribedinthisprotocol.Scansthatmeetcriteriacanserveasthe“baseline”scanonwhichchangemeasurementsarebased.

Managementofpre-enrollmentimaging,includingdecisionsaboutwhethertoacceptlowerprecisionortorequireanew,protocol-specificbaselinescan,arelefttoeachspecificclinicaltrialprotocolauthor.

1.4.ManagementofOn-ProtocolImagingPerformedOff-Schedule

Thisprotocoldoesnotpresumeauniversal,orevenaspecificimagingschedule.Itisintendedtomeasuretumorvolumechangebetweentwoarbitrarytimepoints,includingscansthatareacquiredoutsideoftheprotocol-specifiedtime-window(OOWscans).

Managementoftheclinicaltrialcalendar,deviationsfromtheprotocolspecifiedtimewindow,andpotentialimpactsofdeviationsornon-uniformityofintervaltimingonderivedoutcomessuchareTime-To-Progression(TTP)orProgression-Free-Survival(PFS)arelefttoeachclinicaltrialprotocolauthor.

1.5.ManagementofOn-ProtocolImagingPerformedOff-Specification

Deviationfromtheimageresolutionspecificationsdefinedinthisprotocolwilllikelydegradethequalityofmeasurements.

Managementofoff-specificationimaging,includingdecisionsaboutwhethertoacceptlowerprecisionscansortorequirerepeatscans,arelefttotheclinicaltrialprotocolauthor.

1.6.ManagementofUnscheduled,Off-ProtocolImaging

Unscheduledimagingexaminationsthatarenotpartoftheprotocolspecifiedproceduresformeasuringtumorvolumesmaybeusedasindicatorsofprogressiononly.Forexample,inasubjectwithlungcancerwhoisbeingfollowedwithCTscansofthebody,ifanunscheduled,off-protocolMRIscanoftheheadisacquiredinthemiddleofacycletoevaluateanewcomplaintofheadache,thenitmaybereadeitherasconfirmingprogressionorbeingnegativeforprogressiondependingonwhetherornotnewbrainmetastasesarediscovered.

1.7.SubjectSelectionCriteriaRelatedtoImaging

1.7.1.RelativeContraindicationsandMitigations

Thisprotocolinvolvesionizingradiation.RiskandSafetyconsiderations,e.g.,foryoungchildrenorpregnantwomen,arereferencedinsection13.1.Localstandardsforgoodclinicalpractice(cGCP)andtheALARAPrinciple(AsLowAsReasonablyAchievableradiationexposure)shouldbefollowed.

Thisprotocolinvolvestheuseofintravenouscontrast.RiskandSafetyconsiderations,e.g.,forsubjectswithchronicrenalfailure,arereferencedinsection13.2.Localstandardsforgoodclinicalpractice(cGCP)shouldbefollowed.Theuseofcontrastinsection5assumestherearenoknowncontra-indicationsinaparticularsubject.

1.7.2.AbsoluteContraindicationsandAlternatives

Therearefew,ifany,absolutecontra-indicationstotheCTimageacquisitionandprocessingproceduresdescribedinthisprotocol.Localstandardsforgoodclinicalpractice(cGCP)shouldbefollowed.

Magneticresonanceimaging(MRI)maybeusedwhenclinicalindicated(e.g.,toevaluatemetastasestotheliver).However,themeasurementoftumorvolumewithnon-CTbasedimagingtechnologiesisoutsidethescopeofthisprotocol.

2.SiteSelection,QualificationandTraining

2.1.PersonnelQualifications

ThisprotocoldoesnotpresumespecificpersonnelorqualificationsbeyondthosenormallyrequiredfortheperformanceandinterpretationofCTexamswithcontrast. Localrulesandregulationsforthecertificationofpersonnelprovidingpatientcareshouldbefollowed.Responsibilitiesforthequalificationandmaintenanceofcertificationofimageanalystsinclinicaltrialsislefttoeachclinicaltrialsponsor.

2.2.ImagingEquipment

ThisprotocolrequiresaCTscannerwiththefollowingcharacteristics:

• whilemulti-sliceisnotrequired,itwillproducebetterresults.
  Acceptable:Singleslice,Target:16detectorsorgreater,Ideal:64orgreater
• seesection7forrequiredacquisitioncapabilities
• conformstotheMedicalDeviceDirectiveQualitySystemandtheEssentialRequirementsoftheMedicalDeviceDirective
• designedandtestedforsafetyinaccordancewithIEC601-1,aswellasforElectroMagneticCompatibility(EMC)inaccordancewiththeEuropeanUnion’sEMCDirective,89/336/EEC
• Labelledfortheserequirements,aswellasISO9001andClassIILaserProduct,atappropriatelocationsontheproductandinitsliterature
• CSAcompliant

MeasurementSoftware: Seesection9forrequiredcapabilities

Participatingsitesmayberequiredtoqualifyfor,andconsistentlyperformat,aspecificlevelofcompliance.(SeediscussionofBulls-eyeComplianceLevelsinAppendixC.) DocumentationofAcceptable/Target/IdealLevelsofCompliancewillappearinrelevantsectionsthroughoutthisdocument.

2.3.Infrastructure

Noparticularinfrastructureorphysicalenvironmentisspecified. Itisassumedthatimagingprocedureswillbeperformedinlocationsthatareincompliancewithlocalregulationsforoperatingmedicalimagingfacilities.

2.4.QualityControl

2.4.1.Procedures

See12.1.1forproceduresthesitemustdocument/implement.

2.4.2.BaselineMetricsSubmittedPriortoSubjectAccrual

See12.1.2formetricsubmissionrequirements.

2.4.3.MetricsSubmittedPeriodicallyDuringtheTrial

See12.1.3formetricsubmissionrequirements.

Additionaltask-specificQualityControlisdescribedinsectionsbelow.

2.5.Protocol-specificTraining

NoUPICTprotocol-specifictrainingisspecifiedbeyondfamiliaritywiththerelevantsectionsofthisdocument.

3.SubjectScheduling

3.1.TimingRelativetoIndexInterventionActivity

3.2.TimingRelativetoconfoundingActivities(tominimize“impact”)

Thisprotocoldoesnotpresumeanytimingrelativetootheractivities.

FastingpriortoacontemporaneousFDGPETscanortheadministrationoforalcontrastforabdominalCTarenotexpectedtohaveanyadverseimpactonthislungtumorprotocol.

3.3.SchedulingAncillaryTesting

Thisprotocoldoesnotdependonanyancillarytesting.

4.SubjectPreparation

4.1.PriortoArrival

NopreparationisspecifiedbeyondthelocalstandardofcareforCTwithcontrast.

4.2.UponArrival

4.2.1.Confirmationofsubjectcompliancewithinstructions

NopreparationisspecifiedbeyondthelocalstandardofcareforCTwithcontrast.

4.2.2.AncillaryTesting

NoancillarytestingisspecifiedbeyondthelocalstandardofcareforCTwithcontrast.

4.2.3.PreparationforExam

NoexampreparationisspecifiedbeyondthelocalstandardofcareforCTwithcontrast.

5.Imaging-relatedSubstancePreparationandAdministration

5.1.SubstanceDescriptionandPurpose

Theuseofcontrastisnotanabsoluterequirementforthisprotocol.However,theuseofintravenouscontrastmaterialisoftenmedicallyindicatedforthediagnosisandstagingoflungcancerinmanyclinicalsettings.

Contrastcharacteristicsinfluencetheappearanceandquantificationofthetumors;therefore,agivensubjectmustbescannedwiththesamecontrastagentandadministrationproceduresforeachscan,evenifthatmeansnocontrastisgivenduetoitnotbeinggiveninpreviousexamsofthissubjectinthistrial.

Asubjectshouldbescannedwiththesamebrandofcontrastagentforeachscan(Target).Anotherbrandorswitchofcontrastagenttypemaybeusedifmedicallyindicated,e.g.,aswitchfromionictonon-ioniccontrastmedia(Acceptable).

5.2.DoseCalculationand/orSchedule

Foragivensubject,thesamecontrastdoseshouldbeusedforeachscan(Target).Ifadifferentbrandortypeofcontrastisused,thedosemaybeadjustedtoensurecomparabilityasindicatedandasdocumentedbypeer-reviewedliteratureand/orthecontrastmanufacturers’packageinserts(Acceptable).

Site-specificslidingscalesthathavebeenapprovedbylocalmedicalstaffsandregulatoryauthoritiesshouldbeusedforpatientswithimpairedrenalfunction(e.g.,contrastdoseRreductionbasedoncreatinineclearance).

5.3.Timing,SubjectActivityLevel,andFactorsRelevanttoInitiationofImageDataAcquisition

Foragivensubject,imageacquisitionshouldstartatthesametimeaftercontrastadministrationforeachscan(Target).

Scandelayaftercontrastadministrationisdependentupontheboththedoseandrateofadministration,aswellasthetypeofscannerbeingused.Contrastadministrationshouldbetailoredforboththevasculartreeaswellasoptimizationoflesionconspicuityinthesolidorgans.(Theseguidelinesdonotrefertoperfusionimagingofsingletumors.)Generally,sincetherearemultipleconcentrationsofcontrastaswellasadministrationratesandscanningspeeds,itisdifficulttomandatespecificvalues.Generally,institutionalguidelinesshouldbefollowedsoastooptimizereproducibilityofthescantechnique.

5.4.AdministrationRoute

Intravenous.

5.5.Rate,DelayandRelatedParameters/Apparatus

Contrastmaybeadministeredmanually(Acceptable),preferablyatthesamerateforeachscan(Target),whichismosteasilyachievedbyusingapowerinjector(Ideal).

Ifadifferentbrandortypeofcontrastisused,theratemaybeadjustedtoensurecomparabilityifappropriateandasdocumentedbypeer-reviewedliteratureand/orthecontrastmanufacturers’packageinserts(Acceptable).

5.6.RequiredVisualization/Monitoring,ifany

NoparticularvisualizationormonitoringisspecifiedbeyondthelocalstandardofcareforCTwithcontrast.

5.7.QualityControl

See12.2.

6.IndividualSubjectImaging-relatedQualityControl

See12.3.

7.ImagingProcedure

7.1.RequiredCharacteristicsofResultingData

Thissectiondescribescharacteristicsoftheacquiredimagesthatareimportanttothisprotocol.Characteristicsnotcoveredherearelefttothediscretionoftheparticipatingsite.

Additionaldetailsaboutthemethodforacquiringtheseimagesareprovidedinsection7.2.

7.1.1.DataContent

Theseparametersdescribewhattheacquiredimagesshouldcontain/cover.

Parameter / Compliance
Level*
AnatomicCoverage / Acceptable / entirelungfields,bilaterally
(lungapicesthroughbases)
Target / entirelungfields,bilaterally
(lungapicesthroughadrenalglands)
FieldofView:PixelSize / Acceptable / completethorax:0.55to1.0mm
Target / rib-to-rib:0.55to0.8mm

*SeeAppendixCforadiscussionofBulls-eyeComplianceLevels

FieldofViewaffectspixelsizeduetothefixedimagematrixsizeusedbymostCTscanners.Ifitisclinicallynecessarytoexpandthefieldofviewtoencompassmoreanatomy,theresultinglargerpixelsareacceptable.

7.1.2.DataStructure

Theseparametersdescribehowthedatashouldbeorganized/sampled.

Parameter / Compliance
Level*
CollimationWidth / Acceptable / 5to160mm
Target / 10to80mm
Ideal / 20to40mm
SliceInterval / Acceptable / contiguousorupto50%overlap
SliceWidth / Acceptable / <=53.0mm
Target / 1.0to2.5mm
Ideal / <=1.0mm
PixelSize / see7.1.1
IsotropicVoxels / Acceptable / (5:1)slicewidth<=5xpixelsize
Target / (1:1)slicewidth=pixelSize
ScanPlane / Acceptable / sameforeachscanofsubject
Target / 0azimuth
RotationSpeed / Acceptable / manufacturer’sdefault

*SeeAppendixCforadiscussionofComplianceLevel

CollimationWidth(definedasthetotalnominalbeamwidth)isoftennotdirectlyvisibleinthescannerinterface.Widercollimationwidthscanincreasecoverageandshortenacquisition,butcanintroduceconebeamartifactswhichmaydegradeimagequality.

Sliceintervals(a.k.a."reconstructionintervals"thatresultindiscontiguousdataareunacceptableastheymay“truncate”thespatialextentofthetumor,degradetheidentificationoftumorboundaries,confoundtheprecisionofmeasurementfortotaltumorvolumes,etc.

Pitchimpactsdosesincetheareaofoverlapresultsinadditionaldosetothetissueinthatarea.Overlapsofgreaterthan20%haveinsufficientbenefittojustifytheincreasedexposure.

SliceWidthdirectlyaffectsvoxelsizealongthesubjectz-axis.Smallervoxelsarepreferabletoreducepartialvolumeeffectsand(likely)providehigherprecisionduetohigherspatialresolution.

PixelSizedirectlyaffectsvoxelsizealongthesubjectx-axisandy-axis.Smallervoxelsarepreferabletoreducepartialvolumeeffectsand(likely)providehighermeasurementprecision.

IsotropicVoxelsareexpectedtoimprovethereproducibilityoftumorvolumemeasurements,sincetheimpactoftumororientation(whichisdifficulttocontrol)isreducedbymoreisotropicvoxels.

ScanPlanemaydifferforsomesubjectsduetotheneedtopositionforphysicaldeformitiesorexternalhardware,butshouldbeconstantforeachscanofagivensubject.

FasterRotationSpeedreducesthebreathholdrequirementsandreducesthelikelihoodofmotionartifacts.

7.1.3.DataQuality

Theseparametersdescribethequalityoftheimages.

Parameter / Compliance
Level*
MotionArtifact / Acceptable / minimal(seebelow)
Target / noartifact
NoiseMetric / Acceptable / std.dev.in20cmwaterphantom40HU
Target
Ideal
SpatialResolutionMetric / Acceptable / >=6lp/cm
Target / >=7lp/cm
Ideal / >=8lp/cm

*SeeAppendixCforadiscussionofBulls-eyeComplianceLevels

MotionArtifactsmayproducefalsetargetsanddistortthesizeofexistingtargets.“Minimal”artifactsaresuchthatmotiondoesnotdegradetheabilityofimageanalyststodetecttheboundariesoftargetlesions.

NoiseMetricsquantifythelevelofnoiseintheimagepixelvalues.NoisemetricsarenotdefinedbythisUPICTprotocol.Theyarelefttothelocalstandardofcareforbalancingimagequalitywiththerisksofincreasingradiationexposures.

SpatialResolutionMetricquantifiestheabilitytoresolvespatialdetails.Itisstatedintermsofthenumberofline-pairspercmthatcanberesolvedinascanofresolutionphantom(suchasthesyntheticmodelprovidedbytheAmericanCollegeofRadiologyandotherprofessionalorganizations).Lowerspatialresolutioncanmakeitdifficulttoaccuratelydeterminethebordersoftumors,andasaconsequence,decreasestheprecisionofvolumemeasurements.

Spatialresolutionismostlydeterminedbythescannergeometry(whichisnotusuallyunderusercontrol)andthereconstructionalgorithm(whichisunderusercontrol).

ProceduresformeasuringandoptimizingspatialresolutionarenotspecificallydefinedbythisUPICTprotocol,otherthantonotethathigherresolutiontendstoincreasetheprecisionofmeasurement,andasaconsequence,enhancebothpatientcareandscientificmeritinclinicaltrials.

7.2.ImagingDataAcquisition

7.2.1.SubjectPositioning

Foragivensubject,theymaybeplacedinadifferentpositionifmedicallyunavoidableduetoachangeinclinicalstatus(Acceptable),butotherwisethesamepositioningshouldbeusedforeachscan(Target)andifpossible,thatshouldbeSupine/ArmsUp/FeetFirst(Ideal).

Ifthepreviouspositioningisunknown,thesubjectshouldbepositionedSupine/ArmsUp/FeetFirstifpossible.Thishastheadvantageofpromotingconsistency,andreducingcaseswhereintravenouslines,whichcouldintroduceartifacts,gothroughgantry.

Subjectpositioningshallberecorded,manuallybythestaff(Acceptable)orintheimagedatasetheader(Target).

Consistentpositioningisrequiredtoavoidunnecessaryvarianceinattenuation,changesingravityinducedshape,orchangesinanatomicalshapeduetoposture,contortion,etc.Carefulattentionshouldbepaidtodetailssuchasthepositionoftheirupperextremities,theanterior-to-posteriorcurvatureoftheirspinesasdeterminedbypillowsundertheirbacksorknees,thelateralstraightnessoftheirspines,and,ifprone,thedirectiontheheadisturned.

Factorsthatadverselyinfluencepatientpositioningorlimittheirabilitytocooperate(breathhold,remainingmotionless,etc.)shouldberecordedinthecorrespondingDICOMtagsandcasereportforms,e.g.,agitationinpatientswithdecreasedlevelsofconsciousness,patientswithchronicpainsyndromes,etc.

7.2.2.InstructionstoSubjectDuringAcquisition

BreathHold

Subjectsshouldbeinstructedtoholdasinglebreathatfullinspiration(Target)oratleastnearthehighendinspiration(Acceptable)forthedurationoftheacquisition.

Breathholdingreducesmotionwhichmightdegradetheimage.Fullinspirationinflatesthelungswhichisnecessarytoseparatestructuresandmakelesionsmoreconspicuous.

7.2.3.Timing/Triggers

Foreachsubject,thetime-intervalbetweentheadministrationofintravenouscontrastandthestartoftheimageacquisitionshouldbedeterminedinadvance,andthenmaintainedaspreciselyaspossibleduringallsubsequentexaminations.

Forlungmasses,imageacquisitionshouldbetimedtocoincidewithvisualizationofthethoracicarteries.Forsub-diaphragmaticacquisitions,timingshouldcoincidewithopacificationoftheportal-venousbloodvessels.

Acceptable:useastandardtime;Target:evaluate“manually”.

7.2.4.Model-SpecificParameters

AppendixG.1listsacquisitionparametervaluesforspecificmodels/versionsthatcanbeexpectedtoproducedatameetingtherequirementsofSection7.1.

7.2.5.ArchivalRequirementsforPrimarySourceImagingData

See11.3.

7.3.ImagingDataReconstruction

Theseparametersdescribegeneralcharacteristicsofthereconstruction.

Parameter / Compliance
Level*
ReconstructionKernelCharacteristics / Acceptable / softtooverenhancing
Target / standardtoenhancing
Ideal / slightlyenhancing
ReconstructionInterval / Acceptable / <=5mm
Target / <=3mm
Ideal / <=1mm
ReconstructionOverlap / Acceptable / contiguous(e.g.,5mmthickslices,spaced5mmapartor1.25mmspaced1.25mmapart)
Target / 20%overlap(e.g.5mmthickslices,spaced4mmapartor1.25mmspaced1mmapart)

*SeeAppendixCforadiscussionofBulls-eyeComplianceLevels

ReconstructionKernelCharacteristicsshouldbethesameforeachscanofagivensubject.Asofterkernelcanreducenoiseattheexpenseofspatialresolution.Anenhancingkernelcanimproveresolvingpowerattheexpenseofincreasednoise.Moderationonbothfrontsisrecommendedwithaslightbiastowardsenhancement.

ReconstructionIntervalshouldbethesameforeachscanofagivensubject.

ReconstructionOverlapshouldbethesameforeachscanofagivensubject.

•Decisionsaboutoverlapshouldconsiderthetechnicalrequirementsoftheclinicaltrial,includingeffectsonmeasurement,throughput,imageanalysistime,andstoragerequirements.

•Reconstructingdatasetswithoverlapwillincreasethenumberofimagesandmayslowdownthroughput,increasereadingtimeandincreasestoragerequirements.

ItshouldbenotedthatformultidetectorrowCT(MDCT)scanners,creatingoverlappingimagedatasetshasNOeffectonradiationexposure;thisistruebecausemultiplereconstructionshavingdifferentkernel,slicethicknessandintervalscanbereconstructedfromthesameacquisition(rawprojectiondata)andthereforenoadditionalradiationexposureisneeded.

Asaconsequence,MDCTscannersaretheTargetscannersforthisUPICTprotocol,andthemorerowsofdetectors,theclosertheacquisitioncomestoIdealspecifications.

7.3.1.Model-SpecificParameters

AppendixG.2listsreconstructionparametervaluesforspecificmodels/versionsthatcanbeexpectedtoproducedatameetingtherequirementsofSection7.1.

7.3.2.ArchivalRequirementsforReconstructedImagingData

See11.4.

7.3.3.QualityControl

See12.4.

8.ImagePost-processing

Nopost-processingshallbeperformedonthereconstructedimagessentforimageanalysis.

Suchprocessing,ifperformed,hasthepotentialtodisrupttheconsistencyoftheresults.

9.ImageAnalysis

Eachlunglesionshallbecharacterizedasdescribedinthissection.

Lesionsofinterestinclude:

a)smallpulmonarynodulessurroundedbyair;

b)smalltomediumpulmonarynodulessurroundedbyairand/orwithadjacentnormalandabnormal(non-neoplastic)anatomicstructures;

c)largepulmonarymassessurroundedbyairand/orwithadjacentnormalandabnormal(non-neoplastic)anatomicstructuresand/orconfluentwithmediastinum,chestwall,anddiaphragm.Thus,thecriteriaforselectinglargemassesastargetlesionsisdependentonthecontrastbetweenneoplasticandnon-neoplastictissue.

Fluid,blood,necroticdebris,andthelikeshouldnotbeincludedinthemeasurementoftumorvolume.ProceduresforsegmentingtissuetypeswithinamassarenotdescribedbythisUPICTprotocolbutshouldbeimplementedwhentechnicallyfeasible.

9.1.InputDatatoBeUsed

Thereconstructedimagesmaybeuseddirectlysincenopost-processingisspecified.

NootherdataisrequiredforthisAnalysisstep.

9.2.MethodstoBeUsed

Eachlesionshallbecharacterizedbydeterminingtheboundaryofthelesion(referredtoassegmentation)andtakingcertainmeasurementsofthesegmentedlesion.

Segmentationmaybeperformedautomaticallybyasoftwarealgorithm,manuallybyahumanobserver,orsemi-automaticallybyanalgorithmworkingwithhumanguidance/intervention.

Measurementsmaybeperformedautomaticallybyasoftwarealgorithm,manuallybyahumanobserverwith“e-calipers”,orsemi-automaticallybyanalgorithmworkingwithhumanguidance/intervention.

Itisexpectedthatautomatedboundarydetectionalgorithmswillplacesegmentationedgeswithgreaterprecision,accuracyandspeedthananoperatorcandrawbyhandwithapointingdevice.ItisalsoexpectedthatautomatedalgorithmsforfindingtheLongestDiameter(LD)andLongestPerpendicular(LP)withineachROIwillhavegreaterspeedandprecisionofmeasurementthananoperatorusingelectroniccalipers.Theperformanceofthealgorithmswill,however,dependonthecharacteristicsofthelesionsmaybechallengedbycomplexlungtumors.

Foreachmethodofsegmentationandmeasurementasitechoosestouse,thebaselineintra-andinter-raterreliabilityforsegmentationandforlinearmeasurementshallbemeasuredusingthemethodsdescribedinsection9.6.

MethodsforadjudicatingdiscordantresultsarenotdescribedinthisUPICTprotocol.Varioussystemsofadjudicationaretobeselectedbyattendingphysiciansandclinicaltrialsponsorsortheirdesignees.

Theintra-raterreliabilityoffullyautomatedtumorsegmentationshallbegreaterthan80%(Acceptable),preferablygreaterthan90%(Target),andcanbegreaterthan95%(Ideal).Operatorassistedsemi-automaticsegmentationshouldproducethesamelevelofintra-andinter-raterreliabilityforthevolumemeasurementsofeachtargetlesion.

9.3.RequiredCharacteristicsofResultingData

Whileallmeasurementmetricsareproxiesfortumorburden,itisstilluncertainwhichmeasurementmetricisoptimalforassessingchangesinhealthstatus.Accordingly,multipleoverlappingmeasurementsarespecifiedhere.Themetadatashallinclude:

Foreachtargetlesion,thelesionvolume,inmm3ormL(volumetricmetric)whichisdefinedasthesumofallthevoxelvolumeswithintheboundariesofadiscretetumormassonallthetomographicslicesonwhichitisvisible.

SumofTargetLesionVolumes:Avaluecomputedbyaddingupallofthetargetlesionvolumes.

longestdiameter(LD),inmm(uni-dimensionalmetric)whichisdefinedasthelongestcontinuous,in-planeline-lengththatcanbeplacedwithinanon-nodaltumormassonatransaxialimage.TheLDshouldcorrespondtothegreatestdistancebetweentwoin-planevoxelsanywhereinthestackofimagesonwhichthemasscanbevisualized.ItisexpectedthattheaxiallevelonwhichtheLDwillbederivedwillvaryfromtime-pointtotime-point.

ShortAxis:Theshortaxisisdefinedasthelongestlinethatisperpendiculartothelongest,in-planelinelength(seenearbydiagram).Incontrasttoextra-nodalmasses,thelengthoftheshort-axisistheRECIST1.1outcomemeasureforlymphnodes.

Diagram:Illustrationshowingthelongaxisandtheshortaxisofalymphnode.

thelongestperpendicular(LP),inmm(bi-dimensionalmetric),thatis,thelongest,in-planelinethatcanbeplacedata90degreeangletotheLDontheoneslicecontainingtheLDforatumor

SumoftheDiameters(SOD):Avaluecomputedbyaddingupallofthelongestdiameters(LDs)ofallofthenon-nodalTargetLesionsandtheShortAxisofeachTargetlymphnode.

Thefollowingdefinitionsapplytotheresultingdata:

Baseline:Themeasurementsbasedonthepre-treatmentscansetacquiredmostcloselytothestartoftreatment.

Nadir:Thelowestvalueforthesumofthelongestdiametersorsumofthevolumesofalltargetlesions.Ineffect,thenadiristhenew"baselineequivalent"valueforassessingprogression.ProgressiveDisease(PD)isdefinedasanincreaseinthesumofthelongestdiametersby20%fromnadir,ifbutonlyif,thenewsumofthediameters(SOD)exceeds5mm.Forvolumes,PDisdefinedasanincreaseofmorethan20%ortwotimestheStandardDeviationofmeasurement,whicheverisgreater.

Non-MeasurableLesions:Neoplasticmassesthat,intheirlongestuninterrupteddiameteratbaseline,aretoosmalltomeasurebecausethegreatestdistancebetweenanytwoin-planepixelsislessthantwotimestheaxialslicethickness.

Non-TargetLesions(NTL):Additionalneoplasticmassesthatmeetthecriteriaoftargetlesionsbutarenotselectedforquantitativeassessment,neoplasticmassesthateitherdonotmeettheminimumsizecriteriaorarenotsuitableforrepeatmeasurement,andanytrulynon-measurablelesions,suchasbonemetastases,leptomeningealmetastases,malignantascites,pleural/pericardialeffusion,inflammatorybreastdisease,lymphangitiscutis/pulmonis,cysticlesions,illdefinedabdominalmasses,etc.Non-targetlesionsmustbefollowedqualitatively.

9.4.Platform-specificInstructions

AppendixG.4listsparametervaluesand/orinstructionsforspecificmodels/versionsthatcanbeexpectedtoproducedatameetingtherequirementsofSection9.3.

9.5.ArchivalandDistributionRequirements

See11.6.

9.6.QualityControl

See12.6.

Forallmeasurements,thecoefficientsofvariationshouldbecharacterized,andthe95%confidenceintervalsurroundingthemshouldbecalculated.SpecificqualityassuranceproceduresforestimatingvariancearenotspecifiedinthisUPICTprotocol.

10.ImageInterpretation

WhileAnalysisisprimarilyaboutcomputation;Interpretationisprimarilyaboutjudgment.Interpretationmaybeperformedatboththelesional/targetlevelandintheaggregateatthesubjectlevel(e.g.,inanoncologystudyeachindexlesionmaybemeasuredinlongestdiameterduringtheanalysisphase,butinthisphaseajudgmentmaybemadeastowhetherthereisanew“non-index”lesion;theaggregationofthemeasuredlesionswithcomparisontopreviousstudiescoupledwiththejudgmentastothepresenceorabsenceofanewlesionwillresultintheRECISTclassificationatthesubjectlevel).

10.1.InputDatatoBeUsed

Describerequiredinputdataandanynecessaryvalidationoradjustmentswhichshouldbeperformedonit.Mayalsospecifydatawhichshouldnotbeuseduntilaftertheclinicaltrialinterpretationisrecorded.

(e.g.particularimageseriesorviews;beforeandafterprocessingversionsofimagestoevaluate/validatetheeffectsofprocessing;analysisresults)

10.2.MethodstoBeUsed

Describehowtheinterpretationshouldbeperformed.(For example, definitionofkeyanatomicalpointsorpathologyboundaries;scoringscalesandcriteriasuchasBIRADS,interpretationschemasuchasRECIST,relatedannotations)

10.3.RequiredCharacteristicsofResultingData

10.4.Platform-specificInstructions

AppendixG.5providesinstructionsforspecificmodels/versionsthatcanbeexpectedtoproducedatameetingtherequirementsofSection10.3.

10.5.ReaderTraining

10.6.ArchivalRequirements

See11.7.

10.7.QualityControl

See12.7.

11.ArchivalandDistributionofData

Describetherequireddataformats,transmissionmethods,acceptablemedia,retentionperiods,…

(e.g.Isthesiterequiredtokeeplocalcopiesinadditiontotransmittingtothetrialrepository?Mustallintermediatedatabearchived,orjustfinalresults?Atwhatpointmayvariousdatabediscarded?)

11.1.CentralManagementofImagingData

Ideal:electronictransmissionofencrypteddataoverasecurenetwork

Target:electronictransmissionwithasecurefiletransferprotocol

Acceptable:couriershipmentofphysicalmediacontainingelectroniccopiesofthedata

Note:Thesubmissionoffilmsfordigitizationisrarelyacceptable.Whendigitizedfilmsaresubmitted,theymustcontainarulerorquantificationwillnotbepossible.

Imagingdataforanalysisatcentrallaboratoriesshouldbede-identifiedaccordingto

11.2.De-identification/AnonymizationSchema(s)toBeUsed

Thede-identificationsoftwareshouldbecertifiedasfit-for-purposebyregulatoryauthoritiesatboththesiteoforiginandsiteofreceipt.

Allpersonalpatientinformationthatisnotneededforachievingthespecificaimsofthetrialshouldberemoved.

Pre-specifieddata,suchasheight,weight,andinsomecases,sex,race,orage,mayberetainedifitisessentialforachievingthespecificaimsofthestudyandassuchhasbeenapprovedforusebyregulatoryauthorities.Qualityassuranceproceduresmustbeperformedbytherecipienttoverifythattheimagesthatwillbesubmittedforanalysishavebeenproperlyde-identified.

Acceptable:Datashouldbetransferredtoa"quarantinearea"ofa"safeharbor"forcleaningandcertificationofde-identificationbyprofessionalresearchorganizationsortrainedoperatorsusingproceduresthathavebeencertifiedbyregulatoryauthoritiesatthesiteofreceipt.Qualityassuranceproceduresperformedbytherecipientshouldverifythattheimagesthatwillbesubmittedforanalysishavebeenproperlyde-identified.Imagesthatwerenotproperlyde-identifiedpriortoreceiptbythecentralarchivingfacilityshouldbeobliteratedafterassuringthatcopiesconformtoqualitystandardsforpatientprivacy.

11.3.PrimarySourceImagingData

Thisprotocolpresumesnoarchivingthepre-reconstructionimagedata.

11.4.ReconstructedImagingData

Reconstructedimagesshallbearchivedlocally,formattedaseitherDICOMCTimageobjectsorDICOMEnhancedCTimageobjects.

RetentionperiodandpolicyislefttotheClinicalTrialProtocolauthor.

11.5.Post-ProcessedData

Nopostprocessingisspecified,howeverifpost-processingisperformed,theimagesshallbearchivedthesameas11.4.

11.6.AnalysisResults

SegmentationresultsmayberecordedasDICOMSegmentationObjects,orSTLModelFiles.

Measurementresultsmayberecordedas…

Thedatadescribedin9.3maybeprovidedinanyofthefollowingformats:

•DICOMSR

•DICOMRTSS

•DICOMsecondarycapture

•XLS,CSV,XML

11.7.InterpretationResults

12.QualityControl

12.1.QCAssociatedwiththeSite

12.1.1.QualityControlProcedures

DescriberequiredproceduresanddocumentationforroutineandperiodicQCforthesiteandvariouspiecesofequipment.

12.1.2.BaselineMetricsSubmittedPriortoSubjectAccrual

Listrequiredbaselinemetricsandsubmissiondetails.

12.1.3.MetricsSubmittedPeriodicallyDuringtheTrial

Listrequiredperiodicmetricsandsubmissiondetails.

12.2.QCAssociatedwithImaging-relatedSubstancePreparationandAdministration

12.3.QCAssociatedwithIndividualSubjectImaging

AcquisitionSystemCalibration

Ideal:AprotocolspecificcalibrationandQAprogramshallbedesignedconsistentwiththegoalsoftheclinicaltrial.

Thisprogramshallinclude(a)elementstoverifythatsitesareperformingthespecifiedprotocolcorrectly,and(b)elementstoverifythatsites’CTscanner(s)is(are)performingwithinspecifiedcalibrationvalues.Thesemayinvolveadditionalphantomtestingthataddressissuesrelatingtobothradiationdoseandimagequality(whichmayincludeissuesrelatingtowatercalibration,uniformity,noise,spatialresolution-intheaxialplane-,reconstructedslicethicknessz-axisresolution,contrastscale,CTnumbercalibrationandothers).ThisphantomtestingmaybedoneinadditionaltotheQAprogramdefinedbythedevicemanufacturerasitevaluatesperformancethatisspecifictothegoalsoftheclinicaltrial.

Target:AprotocolspecificcalibrationandQAprogramshallbedesignedconsistentwiththegoalsoftheclinicaltrial.

Thisprogrammayinclude(a)elementstoverifythatsitesareperformingthespecifiedprotocolcorrectly,and(b)elementstoverifythatsites’CTscanner(s)is(are)performingwithinspecifiedcalibrationvalues.Thesemayinvolveadditionalphantomtestingthataddressalimitedsetofissuesprimarilyrelatingdoseandimagequality(suchaswatercalibrationanduniformity).ThisphantomtestingmaybedoneinadditionaltotheQAprogramdefinedbythedevicemanufacturerasitevaluatesperformancethatisspecifictothegoalsoftheclinicaltrial.

Acceptable:SitestaffshallconformtotheQAprogramdefinedbythedevicemanufacturer.

12.3.1.PhantomImagingand/orCalibration

[Documenttheprocedureforacquiringimagesandmeasuringtheimagequalitymetricsintheacquisitionprotocoldescription,e.g.uniformity,noise,effectiveresolution]

12.3.2.QualityControloftheSubjectImageandImageData

12.4.QCAssociatedwithImageReconstruction

12.5.QCAssociatedwithImageProcessing

12.6.QCAssociatedwithImageAnalysis

12.7.QCAssociatedwithInterpretation

13.Imaging-associatedRisksandRiskManagement

13.1.RadiationDoseandSafetyConsiderations

ItisrecognizedthatX-rayCTusesionizingradiationandthisposessomesmall,butnon-zerorisktothepatientsinanyclinicaltrial.Theradiationdosetothesubjectsinanytrialshouldconsidertheageanddiseasestatus(e.g.knowndiseaseorscreeningpopulations)ofthesesubjectsaswellasthegoalsoftheclinicaltrial.Theseshouldinformthetradeoffsbetweendesiredimagequalityandradiationdosenecessarytoachievethegoalsoftheclinicaltrial.

13.2.ImagingAgentDoseandSafetyConsiderations

13.3.ImagingHardware-specificSafetyConsiderations

13.4.ManagementandReportingofAdverseEventsAssociatedwithImagingAgentandEnhancerAdministration

13.5.ManagementandReportingofAdverseEventsAssociatedwithImageDataAcquisition

AppendixA:AcknowledgementsandAttributions

ThisimagingprotocolisprofferedbytheRadiologicalSocietyofNorthAmerica(RSNA)QuantitativeImagingBiomarkerAlliance(QIBA)VolumetricComputedTomography(v-CT)TechnicalCommittee.

Thev-CTtechnicalcommitteeiscomposedofscientistsrepresentingtheimagingdevicemanufacturers,imageanalysissoftwaredevelopers,imageanalysislaboratories,biopharmaceuticalindustry,academia,governmentresearchorganizations,professionalsocieties,andregulatoryagencies,amongothers.Allworkisclassifiedaspre-competitive.Amoredetaileddescriptionofthev-CTgroupanditsworkcanbefoundatthefollowingweblink:

TheVolumetricCTTechnicalCommittee(inalphabeticalorder):

• Avila,R Kitware,Inc.

• Buckler,A (Chair)BucklerBiomedicalLLC

• Dorfman,G (UPICTliaison)Cornell

• Fenimore,C (WG1Cleader)NatInstStandardsTechnology

• Ford,R RadPharm,Inc.

• Gottlieb,R RoswellParkCancerCenter

• Hayes,W BristolMyersSquibb

• Hillman,B Metrix,Inc.

• McNitt-Gray,MUniversityCaliforniaLosAngeles

• Mozley,PD (pharmaindustryco-chair)MerckCoInc/PhRMA

• Mulshine,JL Rush

• Nicholson,D Definiens,Inc.

• O'Donnell,K (IHEliaison)Toshiba

• Petrick,N (WG1Aleader)USFoodandDrugAdministration

• Schwartz,LH (academicco-chair)

• Sullivan,DC (RSNAExecutiveSponsor)DukeUniversity

• Zhao,B., MemorialSloanKetteringCancerCenter

Thev-CTCommitteeisdeeplygratefulfortheremarkablesupportandtechnicalassistanceprovidedbythestaffoftheRadiologicalSocietyofNorthAmerica,includingSusanAnderson,LindaBresolin,JosephKoudelik,andFionaMiller.

AppendixB:BackgroundInformation

Thelong-termgoalofthev-CTcommitteeistoqualifythequantificationofanatomicalstructureswithx-raycomputedtomography(CT)asbiomarkers.Thev-CTgroupselectedsolidtumorsofthechestinpatientswithlungcancerasitsfirstcase-in-point.Therationaleforselectinglungcancerasaprototypeisthatthesystemsengineeringanalysis,thegroundwork,profileclaimsdocuments,androadmapsforbiomarkerqualificationinthisspecificsettingcanserveasageneralparadigmforeventuallyquantifyingvolumesinotherstructuresandotherdiseases.

Thespecificaimofthisimageacquisitionandprocessingprotocolistodescribeproceduresthatseemsufficientforquantifyingthevolumesofneoplasticmassesinthechestthathaverelativelysimplegeometricshapesandareadequatelydemarcatedfromsurroundingnon-neoplastictissues.Thisparticularimageacquisitionandprocessingprotocolislimitedtomassesthathavemeasurablediametersof10mmormore.Theprofileclaimsdocumentonwhichthisprotocolisbasedassertsthatfollowingtheseimageacquisitionandprocessingprocedureswillproducevolumemeasureswithlessthan18%test-retestvariability.

Theprotocoldescribes,inpredominantlychronologicalorder,proceduresthatarerequiredtoachievethislevelofprecision.

Theprotocoldescribesproceduresthatshouldbeuniversallyfollowedinthissetting,regardlessoftheinstrumentthatisusedtoacquirethedata.Italsoprovideslinkstotablesthatlistspecificsettingsonvariousmakes-and-modelsofCTscanners.

Thisprotocolshouldbeconsideredforuseinthecareofindividualpatientsinconventionalmedicalsettings,aswellasinclinicaltrialsofnewtherapiesforlungcancer.Table1summarizeshowstagingrelatestolungcancerdrugtherapyapproaches,theimagingapproachesusedinthosestagesandissuesrelativetotheimagerequirements.

Table1:SummaryofImageProcessingIssuesRelativetoStageofLungCancer

Stage / %ofCases / 5-yearSurvival% / ImagingFocus/TherapyFocus / ImagingTool / Issues / ThoracicSegment. / Hi-Res
I / 16% / 49% / Primarytumor/NeoandadjuvantRX / sCT / Smallcancerssurroundedbyair / Canbestraightforward / Need
II/III / 35% / 15.2% / Primary,hilarandmediastinallymphnodes/Combinedmodality / sCT,PET / Largertumorsandnodesabutotherstructures / Oftenchallenging / Opt.
IV / 41% / 3% / Primary/regionalnodesandmetastaticsites/Chemotherapy / sCT,PET,Bone,BrainScan / Tumorresponseoftendeterminedoutsidethechest / Oftenchallenging / Opt.

AppendixC:ConventionsandDefinitions

Bulls-eyeComplianceLevels

Acquisitionparametervaluesandsomeotherrequirementsinthisprotocolarespecifiedusinga“bullseye”approach.Threeringsareconsideredfromwidesttonarrowestwiththefollowingsemantics:

ACCEPTABLE:failingtomeetthisspecificationwillresultindatathatislikelyunacceptablefortheintendeduseofthisprotocol.

TARGET:meetingthisspecificationisconsideredtobeachievablewithreasonableeffortandequipmentandisexpectedtoprovidebetterresultsthanmeetingtheACCEPTABLEspecification.

IDEAL:meetingthisspecificationmayrequireunusualeffortorequipment,butisexpectedtoprovidebetterresultsthanmeetingtheTARGET.

AnACCEPTABLEvaluewillalwaysbeprovidedforaspecifiedparameter.Whenthereisnoreasontoexpectbetterresults(e.g.intermsofhigherimagequality,greaterconsistency,lowerdose,etc.),TARGETandIDEALvaluesarenotprovided.

Someprotocolsmayneedsitesthatperformathighercompliancelevelsdosoconsistently,sositesmayberequestedtodeclaretheir“levelofcompliance”.IfasitedeclarestheywilloperateattheTARGETlevel,theymustachievetheTARGETspecificationwheneveritisprovidedandtheACCEPTABLEspecificationwhenaTARGETspecificationisnotprovided.Similarly,iftheydeclareIDEAL,theymustachievetheIDEALspecificationwheneveritisprovided,theTARGETspecification

Acquisitionvs.Analysisvs.Interpretation

Thisdocumentorganizesacquisition,reconstruction,post-processing,analysisandinterpretationasstepsinapipelinethattransformsdatatoinformationtoknowledge.

Acquisition,reconstructionandpost-processingareconsideredtoaddressthecollectionandstructuringofnewdatafromthesubject.Analysisisprimarilyconsideredtobecomputationalstepsthattransformthedataintoinformation,extractingimportantvalues.Interpretationisprimarilyconsideredtobejudgmentthattransformstheinformationintoknowledge.

(Thetransformationofknowledgeintowisdomisbeyondthescopeofthisdocument.)

OtherDefinitions

Anonymization / Anonymizationistheprocessofde-identificationandfurtherremovaloramiguationofinformationtoreducetheprobabilityofre-identificationoftheimagedespiteaccesstootherinformationsources
AdjudicationorAdjudicationRead / Adjudicationistheprocessesofdecisionmakingthatinvolvesanindependentpartywiththeauthoritytodetermineabindingresolutionthroughaprespecifiedprocess.Inthestandardreaddesignonceaprimaryanalysisofalltimepointdataforeachpatienthasbeencompleted,thefindingsarecomparedinordertoidentifyanydiscrepanciesinoverallassessmentsofBestOverallResponse,DateofProgression,DateofResponseandDateofResponseConfirmation.Soastoascertainthefinaldeterminationfordiscrepantcases,athirdReviewerwillreviewthepatientdataandchoosetoagreewiththefindingsofoneofthetwoprioranalyses.ThethirdReviewerevaluatesthesamesetofimagesusedbyReviewers1and2andwillnothaveseentheimagesbefore.
AvariationofthisreaddesignistohaveathirdReviewerwhoisblindedtothepreviousfindingsandreviewsthecasesinexactlythesamefashionastheinitialReviewers.Ifagreementisnotreachedinthreeseparateanalyses,aconsensusofReviewersisrequiredinordertoverifythefinaldeterminationforthatpatient.ReaddesignswillbeoutlinedindetailintheIndependentReviewCharterinadvanceofeCRFdesign,or
Adjudicationisameansofdecidinganoutcomewheretwodifferentopinionsareformed,or
Generallyreferringtoablindedindependentreaddesignedtoresolvediscrepanciesbetweentwoprimaryreaders.
AdjudicationRate / TheAdjudicationRateisthenumberofcasesthatareadjudicateddividedbythetotalnumberofcasesevaluated.
BaselineFollowedbyRandomizedTemporalImagePresentation / BaselineFollowedbyRandomizedTemporalImagePresentationisthesequenceofimagepresentationsuchthatthebaseline(earliest)timepointisshowntothereviewerforthepurposeofidentifyingregionsofinterest,suchasselectingneoplasticmassesastargetlesions.Subsequenttimepointsarepresentedinarandomorderwithrespecttothedate.
Blinding / Blindingisaprocedureinwhichoneormorepartiestothetrialarekeptunawareofthetreatmentassignmentsandotherinformationthatmightintroducebias.Singleblindingusuallyreferstothesubjectsbeingunaware,anddouble-blindingusuallyreferstothesubjects,investigators,monitor,and,insomecases,dataanalystsbeingunawareofthetreatmentassignments,or
Blindingistheoutcomeofallprocessesandproceduresthataredeployedtopreventimageanalysisoperatorsorreviewersfrombecomingawareoforinformedaboutthecircumstancessurroundingacase,suchastheinformationinthefollowingabbreviatedlist
investigational(testdiagnosticortesttherapeutic)arm(oranydatathatmightrevealtheinvestigationalarm)
subjectdemographics
siteassessments(includingsitechoiceoflesions)
situationalspecificdescriptionsofthescans(suchas“confirmation”or“endoftreatmentscans”)
resultsorassessmentsofotherreviewersparticipatinginthereadingprocess(exceptduringsomeadjudicationscenarios)
clinicaldataotherthanthatwhichhasbeenpre-specifieddescribedintheimagingcharter.
BlindedRead / ABlindedReadistheanalysisofimagestodetermineresultsofthetestinginwhichtheradiologistisunawareofanysubjectorsiteinformation.
Burned-inInformation / Burned-inInformationisinformationthatispartoftheactualpixeldataasopposedtopresentintheimageheader.
Comment / Acommentinthisinstanceisgenerallyreferringtoatextfieldthatcancaptureadditionalreaderinsightintothereadprocessorreaderthoughtprocesses.CommentsaregenerallyrequiredwhenthereaderindicatesanimageinNotEvaluableortheiropiniondiffersfromthederivedresponse.
ClinicalReadorSiteread / Generallyreferringtoanindependentreadthatcombinestheresultoftheimagingreviewwithpre-definedclinicalinformationtocometoafinaloutcomeassociatedwiththeefficacyendpoint.Readersaregenerallyblindtotreatmentgroupsbutmaybeprovidedavarietyofclinicalinformationappropriatetothediseaseassessment.Or
Imageinterpretationdoneattheinvestigationalsite
ComputerGeneratedQuantitativeImageAnalysis / ComputerGeneratedQuantitativeImageAnalysisisananalysisperformedautomaticallybyacomputerwithlittleornohumaninteractionusingsignalprocessingalgorithmstoquantifyanimagingoutcomemeasure.Thistypeofanalysisshouldbedeterministic(alwaysproduceidenticaloutputfromthesameinput)orhavelowvariability.
ClinicalData
ConfirmationRead / AConfirmationReadisgenerallyreferringtoacentralreadthatoccursbasedonan“on-site”event.Confirmationreadsareassociatedwitheligibilitycriteria,diseaseprogressionorothereventsthatmaybenefitfromathirdpartyconfirmation.
DataLock / TheDataLockisthepointandmethodwhentheresultsofanassessmentorimagingoutcomevariableareconsideredfinalandareprotected.Thismustbepre-definedintheanalysis.Lockingmustnotbeconstruedtomeananassessmentcannotbeoverturnedasindicatedbyemergingdataaslongas:(1)theprocessispre-definedintheImagingCharter;(2)theprocessisdrivenbydatathat,bydesign,emergesaftertheinitialassessment;and(3)thereareadequateaudittrailsthatcansubstantiatethechanges.
De-identification / De-identificationistheprocessofremovingrealpatientidentifiersortheremovalofallsubjectdemographicsfromimagingdataforanonymization
De-personalization / De-personalizationistheprocessofcompletelyremovinganysubject-relatedinformationfromanimage,includingclinicaltrialidentifiers.
DerivedResponse / ADerivedResponseisanoutcomemeasurealgorithmicallyderivedbasedoninformationfromtheblindedreaderassessments.
EndofReviewDataLock / IntheEndofReviewDataLock,thedataarelockedwhenthereviewsofallthetimepointsforthesubjecthavebeencompleted.
Evaluable / Generallyreferringtoimagequality.Basedonpresenceorabsenceofnecessaryimagingandtheassociateimagequality.Evaluableistheresponsegeneratedwhenanimageand/ortimepointcanbeinterpreted.
ExamLevelDataLock / IntheExamLevelDataLock,thedataare"locked"in"finalform"aftereachExam(medicalimagingprocedure)isassessed.ThepurposeoftheExamLockistoassessthedifferentialcontributionofeachExamtotheoverallassessment.
HumanInterfacedImageAnalysis / HumanInterfacedImageAnalysisisimageanalysisthatisdrivenprimarilybyahumanreviewerwhomayusecomputergeneratedanalysistoolstoquantifyanimagingoutcomemeasure.
HybridRandomizedImagePresentation / HybridRandomizedImagePresentation.Inthisparadigm,thefirststageoftheassessmentisfullyrandomizedorthepost-baselinescansarerandomized.Oncetheresultshavebeenlockedforeachtimepoint,theimagesarere-presentedinknownchronologicalorderforreconsideration.Changesinanyoftherandomizedassessmentsaretrackedandhighlightedinthefinalassessment.
Inwithin-patient-controltrials(e.g.comparativeimaging)imagesobtainedbeforeandafterthetestagentshouldbepresentedinrandomizedun-pairedfashion.Theminimumsizeoftherandomizationblocknecessarytominimizerecallshouldbeconsidered.
ImageAnalysis,ImageReview,and/orRead / Proceduresandprocessesthatculminateinthegenerationofimagingoutcomemeasures,suchasbrainvolume,cardiacoutput,ortumorresponsecriteria.Reviewscanbeperformedforeligibility,safetyorefficacy.Thereviewparadigmmaybecontextspecificanddependentonthespecificaimsofatrial,theimagingtechnologiesinplay,andthestageofdrugdevelopment,amongotherparameters.
ImagingData / ImagingDataarevariablesderivedfromtheimagereviewor,ImagingDataarequantitativeorqualitativevariablesresultingfromimagereview.Suchvariablesmaybeusedtoassesseligibilityforstudyandtreatmentresponse,or
ImagingDataisinformationthatresultsfromorisproducedbytheimageanalysisorreviewprocesses(suchaslesionselectionandtheirassociatedspatialmeasurements),orfromalgorithmicallyderivedassessmentsspecifiedintheprotocol.Inthiscontext,thetermalsorefersto"marks"placedonimages,suchasregionsofinterestboundaries,annotationssuchas"TargetLesion4",etc.
ImagingEndpoint / Endpointbasedonobjectiveimagefeatureschosentoevaluatetheactivityofastudytreatment(e.g.retardationofjointdestructioninpatientswithrheumatoidarthritis)
ImagingExaminationorImagingExamorExam / AnExaminationorExamisasinglesetofintimatelyrelatedimagesacquiredcontemporaneouslywithasingletechnology,suchasaCTscanofthechest,awholebodybonescintigram,oranechocardiogram.
ImageHeader / TheImageHeaderisthatpartofthefileordatasetcontainingtheimageotherthanthepixeldataitself
ImagingCaseReportForms: / ImagingCaseReportFormsareIRC-specificformsdesignedtocaptureelementsofimageacquisition,and/orimageinterpretationand/orderivedresponsesforenrollmentand/oreligibilityreviewand/orconfirmationreadsand/orefficacyassessment.
ImagingPhantoms / Devicesusedforperiodictestingandstandardizationofimageacquisition.Thistestingmustbesitespecificandequipmentspecificandconductedpriortothebeginningofatrial(baseline),periodicallyduringthetrialandattheendofthetrial.,
ImageReviewPlanor
RadiologyTechnicalManual / TheImageReviewPlanorRadiologyTechnicalManualisathatdocumentthatsummarizestheplanfortheacquisitionofimagingdata.
ImagingSurrogateEndpoint / Imagingendpointthatiscorrelatedwithaclinicaloutcomebutisnotsufficienttoshowclinicalbenefit
IndependentReview / UsedefinitionsinGFIDevelopingMedImagingDrugPart32004
IndependentReviewCharter(IRC): / TheImageReviewCharterisadocumentsubmittedtoaregulatoryagencytodocumentandsupporttheuseofimagingtosupportanIND,NDAorBLA.Thedocumentidentifiesandlistsimagingresources,imagingsurrogatecriteria,processesforreceipt,handling,preparationandarchiveofimages,theprocessstepsforreviewandassessmentofimagesandthevariousmethodologiesforqualityassuranceandqualitycontrol,or
TheImageReviewCharterisadetailedtechnicaldocumentgoverningtheacquisition,processing,display,interpretation,transfer,andintegrityofimagingdatainefficacytrialsthatuseimagingforassessmentofefficacyoutcomes.IRCsareanintegralcomponentoftheclinicaltrialprotocolandpromotequalityandverifiabilityofimagingdata.TheIRCisdesignedforusebytheclinicalinvestigators,bythecentralimagelaboratoryandbyregulatoryagencies.
TheIRCmightbesubmittedtoaregulatoryagencyforreviewandcommentandtoreachagreementontheuseofimagingintrialsintendedtosupportanNDAorBLA.TheIRCcontainsasummaryoftheclinicalprotocolincludingstudydesign,studypopulation,efficacyendpointsandprimaryefficacyanalysis.TheIRCdescribesimagingresources,processesforreceipt,handling,preparationandarchiveofimages,processstepsforreviewandassessmentofimagesandvariousmethodologiesforqualityassuranceandqualitycontrol.TheIRCneedstocrossreferencerelevantportionsoftheclinicalprotocolincludingenrollmentcriteria,outcomemeasures,andstatisticalanalysisplan(includingprimaryefficacyanalysis,proceduresforhandlingmissingoruninterpretabledataetc.)ForeaseofregulatoryreviewtheIRCshouldincludeaprotocolsynopsisthatissufficientlydetailedtopermitverificationofconsistencyoftheIRCwiththeclinicalprotocolandstatisticalanalysisplan.
IndividuallyIdentifiableInformation / IndividuallyIdentifiableInformationisdatathataloneorincombinationmaybeusedtoidentifyanindividual.
Interpretable / Generallyreferringtoimagequalityassessedbytheblindedreader.Basedonpresenceorabsenceofnecessaryimagingandtheassociateimagequality.Groundsfortheassessmentarecommonlycaptured.Forexample,notOptimalbutEvaluableistheresponsegeneratedwhenanimageand/ortimepointisofquestionablequalitybutcanbeinterpreted.
Intra-ObserverVariabilityorIntra-ReaderVariability / Intra-ObserverVariabilityorIntra-ReaderVariabilityisthevariabilityintheinterpretationofasetofimagesbythesamereaderafteranadequateperiodoftimeinsertedtoreducerecallbias.
Inter-ObserverVariabilityorInter-ReaderVariability / Inter-ObserverVariabilityorInter-ReaderVariabilityisthevariabilityintheinterpretationofasetofimagesbythedifferentreaders.
NotEvaluable / Generallyreferringtoimagequality. Basedonpresenceorabsenceofnecessaryimagingandtheassociateimagequality.NotEvaluableistheresponsegeneratedwhenanimageand/ortimepointcannotbeinterpreted.Maybeassessedinrealtimebyablindedthirdpartyqualityassessorindependentlyofimagereader.Provisionforreimaging(wherefeasible)shouldbeprespecified.ListingofcriteriaisprovidedandresponsesarecapturedintheCRF.
“N”TimePointDataLock / Inthe“N”TimePointDataLock,avariablenumberoftimepoints“N”canbecombinedandshowntogetherataparticularstageofthereviewprocess.Forexample,thebaseline/screeningandthefirstsubsequenttimepointafterbaseline/screeningmaybereviewedtogethertoestablishthebaselineextentofdisease.
Off-ProtocolImaging / Off-ProtocolimagingisimagingthatmayhavebeenperformedduringatrialandshouldnotbereviewedbytheIRCor
Imagingwhichisdoneduringatrialbutnotrequiredbytheprotocol.
On-ProtocolImaging / On-Protocolimagingisimagingthatisperformedduringatrialasrequiredbyanddefinedintheprotocol,or
On-ProtocolimagingisimagingthatisperformedduringatrialasrequiredbyanddefinedintheprotocolthatshouldbereviewedbytheIRC.
OrderofImagePresentation / TheOrderofImagePresentationisthesequencethatimagesarepresentedtoreviewersforformalreviewandgenerationoftheimagingoutcomemeasures.Sufficientlywellestablisheddesignsinclude:
Personalinformation / Personalinformationisdatarelatedtopersonidentification-seeEUguidance(e.g.,Age)
PrimaryReadorPrimaryReview / APrimaryReadistheblindedreviewofimagingdatainwhichoneormoreindependentreviewersreviewimagestogeneratetheoutcomeinformationassociatedwiththeefficacyendpoint,or
APrimaryReviewistheblindedreviewofdatainwhichoneormoreindependentreviewersreviewthedatatogeneratetheoutcomeinformationassociatedwiththeefficacyendpoint.
Pseudonymization / Pseudonymizationistheprocessofde-identificationandreplacementofidentifierswithapseudonymthatisuniquetotheindividualandknownwithinthecontextofatrialbutnotlinkedtotheindividualintheexternalworld.
RandomizedIndependentTemporalImagePresentation / RandomizedIndependentTemporalImagePresentationisthesequenceofimagepresentationthateachtimepointispresentedalone,inarandomorderwithrespecttothedateofacquisition,andreviewedindependentlywithoutaccesstoothertimepoints.
ReaderIndependence / Academic,financial,trialconduct
ScheduledimagingorScheduledImagingAssessment / Scheduledimagingisimagingthatisperformedduringatrialatoneormoreofthetimepoints(orwindowassignedtoatimepoint)designatedforimagingassessmentintheprotocol.Appliestoeitheron-protocoloroff-protocolimaging,or
Examsthatarescheduledasroutineassessments.
SecondaryRead / ASecondaryReadisablindedreviewofimagingdatainwhichoneormoreindependentreviewersreviewimagestogenerateoutcomedatathatisnotpartoftheefficacyendpoints.AnexamplewouldbeareadthatispartofIntra-readeranalysis.
SensitivePersonalInformation / SensitivePersonalInformationisdatarelatedtopersonalpreferencesanddisposition.-seeEUguidance(e.g.,Ethnicity).
SequentialChronologicImagePresentation / SequentialChronologicImagePresentationisthesequenceofimagepresentationthatimagesareshowntothereviewerintheorderinwhichtheywereactuallyacquired.Inthisformat,thereviewershouldnotknowthetotalnumberoftimepointstobeassessedunlessthatinformationhasbeenpre-specifiedintheimagingcharter.(Forexample,pre-specificationisusualandcustomaryinimagingstudiesofneurodegenerativedisorders,arthritis,osteoporosis,andcongestiveheartfailure,amongothers.)
SimultaneousChronologicalImagePresentation / SimultaneousChronologicalImagePresentationisthesequenceofimagepresentationthatallimagesassociatedwithasubjectareshowntotherevieweratthesametimewithoutblindingthedateorsequenceortotalnumber.
SimultaneousRandomizedTemporalImagePresentation / SimultaneousRandomizedTemporalImagePresentationisthesequenceofimagepresentationthatallimagesassociatedwithasubjectareshowntotherevieweratthesametimeinarandomorderwithrespecttothedatebutwithoutblindingtototalnumber.
SequentialUnblinding / SequentialUnblindingSequentialinterpretationofimageswithandwithoutclinicalinformation(e.g.demography,clinicalassessments).
StatisticalAnalysisPlanforMedicalImaging / Analysisplanfocusedonprimaryefficacyanalysisandincludingstatementofnullhypothesis,studypower,statisticaltest,efficacypopulation,andhandlingofmissingoruninterpretableimages,sensitivityanalyses
TimePoint / ATimePointisadiscreteperiodduringthecourseofaclinicaltrialwhengroupsofimagingexamsorclinicalexamsarescheduledasdefinedinthestudyprotocol.
TimePointDataLock / IntheTimePointDataLock,thedataarelockedafterallofthepre-specifiedinformationassociatedwitheachtimepointisassessed.Insomeparadigms,thetimepointsareknowntobepresentedinchronologicalorder;inothers,thetimepointsmayberandomizedduringtheearlystagesoftheimageanalysisprocess(vidainfra).
TruthStandard / Singleormultipleimagemodalities
UniqueIdentifiers(UIDs): / UniqueIdentifiers(UIDs)aregloballyuniqueidentifierusedtoidentifiersimages,setsofimages,orcomponentswithinanimage.
Uninterpretable / Generallyreferringtoimagequalityassessedbytheblindedreader.
UnscheduledImaging / Unscheduledimagingisimagingthatisperformedduringatrialatatime/dateoutsidethewindowassignedtoatimepointdesignatedforimagingassessmentintheprotocol. Itmaybeadhocimagingperformedtoevaluateanunscheduledclinicalcircumstance.Itmaybeon-protocoloroff-protocolimaging.

AppendixD:Documentsincludedintheimagingprotocol(e.g.,CRFs)

Attached.

AppendixE:AssociatedDocuments(derivedfromtheimagingprotocolorsupportiveoftheimagingprotocol)

Attached.

AppendixF:TBD

AppendixG:Model-specificInstructionsandParameters

Thefollowingsectionsprovideinstructionsforvariousequipmentmodels/versionsthatareexpectedtoproducedatameetingtherequirementsoftherelevantactivity.

Thepresenceofspecificproductmodels/versionsinthefollowingtablesshouldnotbetakentoimplythatthoseproductsarefullycompliantwiththeQIBAProfile.Compliancewithaprofileinvolvesmeetingavarietyofrequirementsofwhichoperatingbytheseparametersisjustone.Todetermineifaproduct(andaspecificmodel/versionofthatproduct)iscompliant,pleaserefertotheQIBAConformanceDocumentforthatproduct.

G.1.ImageAcquisitionParameters

Thefollowingtechniquetableslistacquisitionparametervaluesforspecificmodels/versionsthatcanbeexpectedtoproducedatameetingtherequirementsofSection7.1.

Thesetechniquetablesmayhavebeenpreparedbythesubmitterofthisimagingprotocoldocument,theclinicaltrialorganizer,thevendoroftheequipment,and/orsomeothersource.(Consequently,agivenmodel/versionmayappearinmorethanonetable.)Thesourceislistedatthetopofeachtable.

Sitesusingmodelslistedhereareencouragedtoconsiderusingtheseparametersforbothsimplicityandconsistency.SitesusingmodelsnotlistedheremaybeabletodevisetheirownacquisitionparametersthatresultindatameetingtherequirementsofSection7.1andconformtotheconsiderationsinSection13.

Insomecases,parametersetsmaybeavailableasanelectronicfilefordirectimplementationontheimagingplatform.

TableG.1a

Generic:ThisrepresentsparametersforagenericCT.Source:QIBAv-CTCmte Date:2009-mm-dd

Parameter / ComplianceLevel* / Generic
kVp / Acceptable / 110to140
Target / 110to130
mAs
(mediumpatient) / Acceptable / 40to350
Target / 80to160
ScanDuration / Acceptable / 30sec.
Target / 15sec.
Ideal / 10sec.
TableSpeed / Acceptable
Target

*SeeAppendixCforadiscussionoftheLevelsofCompliance

kVpandmAsshouldbeadjustedasnecessary,dependingonthebodyhabitusofindividualpatients.Thevaluesshouldbeconsistentforallscansofthesamepatient.

ScanDurationvaluesareintendedtoallowcompletionofthescaninasinglebreathholdformost/amajority/nearlyallsubjectsrespectively.

TableSpeedvaluesareintendedtoyieldanIECPitchValueofapproximately1whileachievingthecorrespondingScanDuration.

TableG.1b:"Target"CompliantProtocolsforSpecificSystems

Thefollowingtableprovidessampleparameterssetsthatmeetthe“Target”LevelofComplianceforspecificmodels.

SeeAppendixCforadiscussionoftheLevelsofCompliance.

Parameters / vCT1A(Philips) / GE / ACRIN
MxIDT8000
(Thin) / MxIDT8000
(Thick) / Ultra / VCT-64 / 6678
DataContent
AnatomicCoverage
FieldofView:PixelSize / Rib-to-rib:0.55-.75mm
DataStructure
CollimationWidth / 16x0.75mm / 16x1.5mm / (TBA)
SliceInterval
SliceWidth / 0.8mm / 5.0mm / 1.0mm
PixelSize / 0.55mm
IsotropicVoxels / (2:1)
ScanPlane
RotationSpeed / 0.5sec
DataQuality
MotionArtifact
NoiseMetric
SpatialResolutionMetric
Acquisition
TubeVoltage / 120kVp / 120kVp / 120kVp
Exposure / 100mAs / 100mAs / 100mAs
Pitch / 1.2 / 1.2
Reconstruction
Recon.Kernel / Detailedfilter / Detailedfilter / Standard
Recon.Interval
Recon.Overlap / 50% / 50% / 20%

G.2.ImageReconstructionParameters

Seeabove.

G.3.Post-ProcessingInstructions

Noneprovided.

G.4.AnalysisInstructions

Noneprovided.

G.5.InterpretationInstructions

Noneprovided.

Page 1 of 36Version 1.4 of 9 September 2010