Incorporating liquid biopsies into treatment decision-making: obstacles and possibilities

feature

Incorporating

liquid

biopsies

into

treatment

decision-making:

obstacles

and

possibilities

NickBeijez,n.beije@erasmusmc.nl,JohnW.M.MartenszandStefanSleijferz

Circulating

tumor

cells

(CTCs)

and

cell-free

DNA

(cfDNA)

together

with

newer

emerging

liquid

biopsies

have

a

unique

potential

to

deal

with

key

issues

in

oncology.

For

example,

they

can

be

used

to

assess

prognosis,

direct

treatment

with

certain

kinds

of

drug,

or

provide

information

about

response

to

treatment.

However,

despite

an

overflow

of

literature

on

the

subject,

clinical

implementation

of

these

liquid

biopsies

has

been

scarce.

This

is

mainly

because

there

is

a

lack

of

preanalytical

standardization,

multiple

different

techniques

or

platforms

are

being

used,

and

a

lack

of

prospective

studies

investigating

a

meaningful

clinical

question

are

performed.

Here,

we

provide

an

overview

of

the

current

state

of

liquid

biopsies

and

make

suggestions

for

how

liquid

biopsies

can

reach

the

tipping

point.

Introduction

Intheeraofprecisionmedicine,liquidbiopsies haveattractedsignificantinterestforpersonalizing thetreatmentofpatientswithcancer.Thepremise ofliquidbiopsiesisthat,byobtainingasimple bloodsample,allsortsofcancer-related charac-teristicscanbedeterminedinrealtime,andcanbe usedtopersonalizecancertreatment.However, theclinicalutilityofliquidbiopsiesisyettobe confirmed.Here,wediscussthecurrentstandingof liquidbiopsiesinoncology,theirhighlightsuntil now,aswellastheirpitfallsandcaveats. Liquidbiopsies:lotstochoosefrom Thecurrentlandscapeinvolvesmultipletypesof liquidbiopsy.Mostoftheresearchhasbeen

doneonCTCsandcfDNA.CTCsareintacttumor

cellsthathavedetachedfromasolidtumor,

whereascfDNAisfragmentedDNAcomprising

germlineDNAandpotentiallycirculatingtumor

DNA(ctDNA)thatoriginatesmainlyfrom

apo-ptotictumorcells(Fig.1).GiventhatcfDNAonly offersthepossibilitytoanalyzeDNA,CTCs theoreticallyoffertheanalysisofallmolecular materials,includingRNAandproteinsinthe samecell.Therearealsoemergingliquid bi-opsies,suchascell-freeRNA,extracellular vesi-cles,suchasexosomesandmicrovesicles, circulatingproteins,andplateletsexhibiting

tumor-specificRNAprofiles.Promisingdataon

theuseoftheseemergingliquidbiopsiesin cancerscreeninghavebeenpresented[1–4], but,giventherelativescarcityofdataonthese

newerliquidbiopsies,wefocushereonCTCs

andcfDNA.Forboth,therearesomeclinicaldata underliningtheirpotentialrelevancefor per-sonalizingcancertreatments,butrobust evi-denceshowingtheirclinicalutilityiswarranted.

Circulatingtumorcells

TheCellSearchsystemiscurrentlytheonly

systemapprovedbytheUSFoodandDrug

Administration(FDA)forenumeratingCTCsin

patientswithmetastaticbreast,colorectal,or prostatecancer.Thesystemenrichesforallcells

expressingEpCAMbyusinganti-EpCAM

mag-neticbeads.Subsequently,cellsarestainedfor anti-cytokeratin(CK)8/18/19(positiveonCTCs),

DAPI(anucleusmarker),andanti-CD45(to

excludecontaminatingleukocytes)toidentify

CTCs.Themethodhasshowntobehighly

specific,giventhatCTCsarerareinhealthy donors[5].TheprognosticvalueofCTCs

countedwithCellSearchhasbeenshownnot

onlyformanydifferentmetastaticepithelial malignancies,butalsoinpatientswith non-metastaticcancer.Forexample,inmetastatic breastcancer(MBC),aCTCcountoffiveormore

Features

PERSPECTIVE

CTCsisstronglyassociatedwithpoorprognosis; similarly,patientswithprimarybreastcancer

withoneormoreCTCshavedecreasedoverall

survival[6,7].Inaddition,changesinCTCshave

beenassociatedwithresponse.However,

al-thoughenumeratingCTCundoubtedlyhas

clinicalvalidity,studiesinvestigatingclinical utilityhavebeenrareand,thus,thetestis seldomusedbyclinicians.

Onestudyrandomizingpatientsbasedon

CTCchangesduringtreatmentforMBC

dem-onstratednosurvivalbenefitforswitchingto anotherlineofsystemictreatmentbasedonCTC countsratherthanontraditionalmeans[8]. Similarly,astudyinwhichpatientswith

HER2-negativeprimarybreastcancerwithdetectable

CTCsaftersurgeryandstandard(neo)adjuvant

therapyreceivedadditionaltrastuzumab

showednoadditionalbenefit[9].Theonlystudy

demonstratingapossiblebenefitofcounting

CTCsfortreatmentdecision-makingwas

re-centlypresentedanddemonstratedthata

baselineCTCcountcanbeusedsafelytodirect patientswithMBCtoreceiveeitherfirst-line

chemotherapy(iffiveormoreCTCs)or

endo-crinetreatment(iflessthanfiveCTCs)[10]. Interestingly,approximatelyhalfofthepatients

whothetreatingphysicianintendedtogive

chemotherapybasedonclinicalgroundscould

besafelyde-escalatedtoreceiveendocrine

therapy.However,giventhatpatientswithMBC

nowincreasinglyreceivecombinedtreatment

withCDK4/6inhibitorsandendocrinetherapy,it isunlikelythatthesefindingswillresultin

widespreadchangesinthetreatmentofsuch

patients.Thelackoftrueclinicalutilityof

countingCTCsmeansthatbaselineCTCcounts

andchangesinCTCcountsarenowmainlyused

inclinicaltrialsonlyasaprognosticmarkeroras

earlyresponsemarker.

OneofthemainconcernswithCellSearchis

itsdependencyonEpCAM,withthepotentialof

missingoutonEpCAM-negativeCTCs[11].

Therefore,aplethoraofalternativeassayshave becomeavailablethathavetriedtoaddressthis issue:forexample,byusingalternativemarkers todetectCTCsorusingsize-basedpropertiesof

CTCstoenrichforthem.However,head-to-head

comparisonsoftheseveralavailableassaysfor

CTCdetectionhavebeenanecdotal.Oneofthe

mainproblemsincomparingassaysisthe

ab-senceofagroundtruth,meaningthat,forsome CTCdetectionassays,increasedsensitivity

comparedwithCellSearchhasbeendescribed,

althoughitisunclearwhetherthiswasatthe expenseofspecificity.Itisunlikelythat,for

countingCTCs,anyotherassaythanCellSearch

willundergosuchvigorousvalidationofits

prognosticvalueinsomanytumortypesand

settings.

Beyond

counting

CTCs

CountingCTCsisarather1DuseofCTCs.Most

CTCdetectionassayshavedescribedmethods

tocharacterizeCTCsattheRNA,DNA,and

proteinlevel.Initially,thesetypesofanalysis werelimitedtothedetectionofgenesor

mutationsonamixedpoolofCTC-enriched

materialandleukocytes.Althoughitispossible

toidentifysomaticmutationsandRNAprofiles

onthesematerials,itislaborintensiveand, moreover,specificityissuesarisewhen

attemptingtomeasureanytumor-specific

sig-nalinabackgroundofleukocyte-derived

ma-terial.Oneofthemostclinicallyrelevantmarkers

thathavecomefromCTCsistheandrogen

receptorsplicevariantV7(AR-V7)inmetastatic prostatecancer(mPC),whichisaRNAvariant thatpredictsforresistancetoantiandrogen therapiesinmPC[12,13].Prospectiveclinical

studiesinvestigatingwhethertheCTCAR-V7

statusisusefultoguidetreatment

decision-makinginmPCareongoing.

Givenitsgreatpromise,manyeffortsare

underwaytofurtherimprovemethodsto

mo-lecularlycharacterizeCTCs.Forexample,efforts tocharacterizeCTCsatthesinglecelllevelby

DNAorRNAsequencinghavetakenoffwiththe

availabilityofseveralmethodstoisolatesingle cells[14],buthavesofarbeenlimitedto proof-of-conceptstudies.AlthoughsingleCTC char-acterizationwillprobablyoffermore informa-tiononheterogeneity,itisuncleartowhat extentthiswillreflecttheentirelandscapeof tumorheterogeneity,especiallyifperformedin

limitednumbersofCTCs.

AnothermethodtoimproveCTC

characteri-zationisbyincreasingthebloodvolumethatis

analyzed,sometimesevenbyusing

leukapher-esis[15].Althoughusingleukapheresisdoes yieldmoreCTCstoanalyze[16],itsomewhat

compromisestheminimallyinvasiveand

easy-Tumor tissue

(primary tumor or metastasis)

Vascular endothelial cells

Blood stream

Circulating tumor cell (CTC)

intact cell from tumor tissue

Circulating tumor DNA (ctDNA)

cell-free DNA mainly from apoptotic or necrotic tumor cells

Drug Discovery Today

FIGURE1

Currentleadingliquidbiopsies:circulatingtumorcells(CTCs)andcirculatingtumorDNA(ctDNA).Fromthesolidtumor,whichcaneitherbetheprimarytumoror ametastaticsite,tumorcellscanextravasateandappearinthebloodasCTCsorgointoapoptosisandshedtheirDNA.CTCsareintactcellsandcanformdistant metastasesorgointoapoptosis.CtDNAiscell-freeDNA(cfDNA)thatismainlyfromapoptoticornecrotictumorcellsandhasunknownbiologicalrelevance. BothCTCsandcfDNAcanbeisolatedfromwholebloodusingsophisticatedtechniquesandthenbequantifiedorsubjectedtomultiplemolecularanalyses.

Features



to-collectnatureofliquidbiopsies,whilestill

requiringpurificationofCTCsamong

contami-natingbloodleukocytes.

Anotherstrategyforobtaininghigher

num-bersofCTCsfordownstreamanalysisisculturing ofCTCs,forexamplewiththeintentto subse-quentlyrundrug-sensitivityanalysesonthem

[17].Althoughmultiplegroupshavedescribed successfullong-termculturesofCTCs,the

chanceofsuccessislowandseeminglylimited

topatientswithveryhighCTCcountsof>300 CTCs/7.5ml[18].Giventhislowsuccessrateand

becauseittakesseveralweekstomonthsto

culturetheseCTCs,broadimplementation

appearsunlikely,otherthanperhapsforgroups ofpatientswithhigherlevelsofCTCs,suchas mostpatientswithsmall-celllungcancer[19]. Cell-freeDNA

EvenmoreasisthecaseforCTCs,manydifferent assaystodetectaberrationsincfDNAarenow

commerciallyavailable.ComparedwithCTC

de-tection,cfDNAismoreeasilyanalyzedwithtools

suchasnext-generationsequencing(NGS)

machinesandPCRmachinesalreadyavailablein

mostresearchlaboratories.ctDNAisthefraction ofcfDNAthatistumorderived.Thepercentageof ctDNAisoften<1%,meaningthatclassical

techniques,suchasSangersequencingand

quantitativePCR,oftenlackthesensitivityto detecttumor-specificmutations.Initially,mostly

digitalPCR(dPCR)-basedapproacheswereused.

However,usingdPCR,onlyalimitedamountof

mutationscanbedetectedinonerun.Therefore, NGSprotocolsusinguniquemolecularidentifiers arenowmoreoftenused,increasingsensitivity andallowingthedetectionofctDNAmutationsin

multiplegenesofinterest.Newermethodseven

allowforthedetectionofcopynumbervariations

usinglow-coveragewhole-genomesequencing

[20]ormethylomeprofiles[21].

Preanalytical

and

postanalytical

processing

Regardlessofthetechnique,preanalyticalsteps,

suchassamplecollectionandprocessing,have

tooccurandcangreatlyaffectthefinalresults. Therearesignificantdifferencesbetweenthe variousprotocols,asdescribedintheliterature.

Itisknownthatbloodsampleshavetobe

processedwithin24horinspecialtubeswitha stabilizingagent(e.g.,CellSearchtubesorBCT tubes)topreventlysisofleukocytesincreasing

theamountofnontumor-derivedcfDNA,

therebyloweringthesensitivity[22]. Addition-ally,somecfDNAisolationkitsselectforsmaller orlongercfDNAfragmentsthanotherkits[23], ofwhichthesignificanceisunknown.Besides

thesefactors,thereisalsolimitedinsightinto theeffectoffreezingandthawingofsamples

andtowhatextentvariouscomedications,

un-derlyingcomorbidities,or,forexample,circadian rhythmsinfluencecfDNAconcentrations[24].

Aftersampleprocessing,thepostanalytical

phaseinwhichthecfDNAsomaticmutations

havetobeidentifiedisalsocrucial.Although commerciallyavailablecfDNAassaysusually

comewithassociatedsoftware,thesubsequent

analysisandcallingofsomaticmutationsis oftenlefttotheusers.

Anissuerelatedtothesedownstream

anal-ysesishowsomaticvariantsshouldbereported.

Somaticvariantsaremostcommonlyexpressed

asvariantallelefrequency(VAF)orthenumber ofmutantcopies/ml.Withtheavailabilityof highlysensitivityassays,wedonotknow whetherdetectionofavariantisclinically

rele-vantatalowVAForlowmutantcopynumber.A

complicatingfactorisalsothattheVAFin par-ticularisinfluencedbythebackground.Hence, anincreaseordecreaseinbackground(e.g.,

increasedapoptosisofleukocytes)canchange

theVAF,whichmightespeciallybeimportantin

monitoringofVAFduringtreatment.The

num-berofmutantcopynumbersappearstobemore

constant[22];however,itisunclearwhatchange

inVAForcopynumberisbothbiologicallyand

clinicallyrelevant.

GivenallthestepsthatmightinfluencecfDNA analysis,TorgaandPienta[25]recentlysent

bloodsamplesfrompatientstakenatthesame

momenttotwoCLIA-licensedcommercial

lab-oratoriesforcfDNANGSsequencingfor

head-to-headcomparison.Theresultswereworrisome,

becauseresultsfrom64%ofthetestedsamples

wereincongruentbetweenthetwolaboratories.

Atthispoint,itisunclearhowthedifferences

betweenbothassayscanbeexplained.

Potential

clinical

use

of

ctDNA

Althoughitisclearthatthereisstillsomewayto gobeforeoptimalanalyticalvalidityisreached, therearealsoexitingdatapublishedthatopen doorstoallkindsofkeyclinicalproblemsin

oncology.ThisincludestheuseofctDNAasa

predictivemarkerforcertaintreatments,asa markertodetectdiseaserelapse,asamarkerto detectemergingresistancetoa(targeted)

treatment,orasameanstoscreenpeoplefor

thepresenceofcancer.

Nowadays,cfDNAassaysaremostcommonly

usedaspredictivemarkersinthesettingof

EGFR-mutatedmetastaticnon-smallcelllung

cancer(mNSCLC),becausethepresenceof

ac-tivatingEGFRmutationsintreatment-naive

mNSCLCisaprerequisitefortreatmentwith

EGFRtyrosine-kinaseinhibitors(TKIs).Theonly cfDNAtestthatiscurrentlyapprovedbytheFDA isthecobastest,whichisapprovedtoidentify

EGFRexon19deletions,L858R,andT790M

mutations.Overallconcordancebetweentissue

andcfDNAusingthistesthasbeenhighat89– 91%[26,27]andpatientswithmNSCLC har-boringEGFRmutationsincfDNAidentifiedusing thistestbenefittedfromtreatmentwithEGFR TKIscomparedwithplacebo[26].Therefore, cfDNAisincreasinglyusedtoscreenpatientsfor first-lineEGFRTKItreatmentandmostclinicians useitiftissuespecimensareunavailable[27]. FortheT790MmutationintheEGFR,conferring

resistancetoTKIsinpatientswithmNSCLCand

renderingthesepatientssuitablefortreatment

withosimertinib,somegroupsevenadvocate

usingcfDNAinitiallytodetectEGFRT790M

followedbyatumorbiopsyincaseofnegative

results[27].However,theFDAapprovedthe

T790MEGFRcfDNAtestonlywhenitisnot

possibletoobtainatissuebiopsyatthetimeof progressiononaTKI.Theircautiousnessis

probablybecauseofpoorerconcordancerates

betweentissueandcfDNAforEGFRT790M

mutationsat70%[28].However,concordance

studiesbetweentissueandcfDNAwillnever

showperfectsimilarity.Fordrivermutations,

suchasEGFRexon19deletionsandL858R

mutations,whicharepresentinmosttumor

cells,concordancewillprobablybegood.By

contrast,forsubclonalorresistancemutations,

suchasEGFRT790M,concordancerateswill

likelybelower.Thesediscrepanciesbetween

tissueandcfDNAmightoccurbecauseof

ana-lyticalissues,lownumbersofctDNA,or bio-logicalfactors,suchasheterogeneity.

Anotherimportantfactorforcautiousness

withusingcfDNAwithouttissue-based

confir-mationisthatthereisnoformalproofthat

cfDNA-mutant-positive,tissue-mutant-negative

patientshavesimilarresponsestotherapyas cfDNA-mutant-positive,tissue-mutant-positive patients.Forexample,patientswithEGFR

T790M-positivecfDNAhadpoorer

progression-freesurvival(PFS)onosimertinibiftheyhada

T790M-negativetumorthaniftheyhada

T790M-positivetumor(PFS4.2versus9.3

months,respectively;P=0.0002)[28].

BesidestheuseofcfDNAinEGFR-mutated

mNSCLC,cfDNAisnotyetroutinelyusedasa

predictivemarkertoselectpatientswithcancer forcertaintargetedtherapies.Althougharecent reportunderlineditspotentialusetodirect patientswithanactionablemutationto ap-propriatePhaseItrialswithinatimeframe ac-ceptableforclinicaldecision-making[29], furthertrialsareeagerlyawaited.

Features

WithrespecttotheuseofctDNAtodetect

earlyrelapse,therehavebeenexamplesin

multiplenon-metastasizedtumorsinwhich

somaticcfDNAmutationsweredetectedafter

curativeoperativetreatmentsandwere

associ-atedwithdecreasedrelapse-freesurvival

[30,31].Thismightmeanthatpatientsatriskfor relapsecanbeidentifiedearlyandtreated ac-cordingly,althoughtheextenttowhichthere mightbesometypeoflead-timebiasremainsto beseen.

MutationsincfDNAcanalsobeusedasa

surrogateofdrugresponseinlongitudinal

monitoring.Forexample,interestingdatawere

describedforKRASmutationsthatcause

resis-tancetomonoclonalantibodiestargetingthe

EGFR(EGFR-MoAbs).Thereissomeevidence

that,upondiscontinuationofEGFR-MoAbs,

KRASmutationsdecayanddrugsensitivityis

regained,whichmightmeanthatthesepatients

canberechallengedwithEGFR-MoAbs[32],

whichisnowbeingtestedinaproof-of-concept

study(CHRONOStrial,NCT03227926).

Lastly,it hasbeensuggestedthatcfDNA

couldbeusedasacancer-screeningtool.Inan

analysisof1005patientswithnon-metastatic

cancers,atestfordetectingcancerbasedon

ctDNAmutationsandproteintumormarkers

hadamediansensitivityof70%andspecificity of>99%[33].Althoughthishighlightedthe potentialofcfDNAanalysestodetectcancerat anearlystage,itsapplicationasascreening toolinahealthypopulationstillhasalong

waytogo,especiallybecausethepositive

predictivevalueofthetestdeclined

signifi-cantlywhentestedinapopulationinwhich

theprevalenceofcancerwaslow,whichis

typicallythecasewhenscreeningthegeneral

population[34].

Reachingthetippingpoint

Researchonliquidbiopsieshasskyrocketed

overthepastdecadeandhasledtonovel

insightsintocancerbiology.However, incor-porationofliquidbiopsiesintoclinical

work-flowsremainsamajorchallenge.Giventhe

plethoraofarticlesonliquidbiopsiesthathave beenpublishedtodate,itisdisappointingthat approvalonlyexistsfortheCellSearchCTC

assayandcertaincfDNAassaysabletodetect

mutationsintheEGFRreceptor.Withrespectto

cfDNA,arecentASCOandCollegeofAmerican

Pathologistjointreviewevenconcludedthat

thereis‘littleevidenceofclinicalvalidityand clinicalutilitytosupportwidespreaduseof

ctDNAassaysinmostpatientswithadvanced

cancer’[35].

Oneofthefactorscontributingtothelackof clinicalapplicabilityisthecurrent overabun-danceofliquidbiopsyassays.Thisisespecially thecaseinthecfDNAfield,inwhichdozensof

assaysarenowcommerciallyavailableandalso

animportantsubsetofassaysdeveloped

in-housebyacademia.Alloftheseassayshavetheir ownlimitofdetection,sensitivity,and specific-ity,meaningthatfindingsononeparticular liquidbiopsyplatformarenotnecessarily ap-plicabletoallotherplatforms,hampering clin-icalapplicabilityandunderliningtheneedfor externalqualityassessmentstudiesbefore im-plementationinroutinediagnostics(general issuesincfDNAresearcharelistedinTable1).

Anotherkeyissueisthattrialsthatexaminea truechangeinclinicaldecision-makingbased

onliquidbiopsiesarerare.From642trials registeredatclinicaltrials.govthatareevaluating liquidbiopsiesinsomeway,only21trials(3.3%) areinvestigatingaparticularinterventionbased

onliquidbiopsies(supplementalinformation

online).Inalmosthalfoftheseinterventional studies,patientsreceivedaparticular

interven-tion(suchasatargetedtreatment)basedon

baselineCTCorcfDNAstatuswithoutthe

in-clusionofacontrolgroupthatdoesnotreceive thatparticularintervention,meaningthat,inthe caseofapositivestudyresult,thesestudiesare unlikelytoresultinimmediatechangesin clinicalpractice.Althoughobservationalstudies arealsoofkeyimportanceinincreasingour

understandingofcancerandmightbe

hy-pothesisgenerating,ourcurrentunderstanding

ofliquidbiopsiesdoesallowformoredirect interventionalresearch,butatthispointthat seemstoberare.

Concludingremarks

Inconclusion,forliquidbiopsiestoeventually reachthetippingpoint,inadditiontoasmuch harmonizationofpreanalyticalandanalytical conditionsaspossible,theneedformoreclinical trialsthatinvestigateameaningfulclinical questionisespeciallyhigh.Whetherthis

re-searchisdonewithCTCs,cfDNA,oroneofthe

emergingliquidbiopsiesisnotthatimportant: theonethatprovestobeofclinicalutility,in combinationwithanalyticalvalidity,will

even-tuallybeadaptedbythecommunity.

AppendixA. Supplementarydata Supplementarymaterialrelatedtothisarticle canbefound,intheonlineversion,atdoi:

https://doi.org/10.1016/j.drudis.2019.05.028.

References

1 Sozzi,G.etal.(2014)Clinicalutilityofaplasma-based

miRNAsignatureclassifierwithincomputed

tomographylungcancerscreening:acorrelativeMILD

trialstudy.J.Clin.Oncol.32,768–773

2 Melo,S.A.etal.(2015)Glypican-1identifiescancer

exosomesanddetectsearlypancreaticcancer.Nature

523,177–182

3 Best,M.G.etal.(2018)Tumor-educatedplateletsasa

noninvasivebiomarkersourceforcancerdetectionand

progressionmonitoring.CancerRes.78,3407–3412

4 IntegrativeAnalysisofLungCancerEtiologyandRisk

(INTEGRAL)ConsortiumforEarlyDetectionofLung

Canceretal.(2018)Assessmentoflungcancerriskon

thebasisofabiomarkerpanelofcirculatingproteins.

JAMAOncol.4,e182078

5 Allard,W.J.etal.(2004)Tumorcellscirculateinthe

peripheralbloodofallmajorcarcinomasbutnotin

healthysubjectsorpatientswithnonmalignant

diseases.Clin.CancerRes.10,6897–6904

6 Bidard,F.C.etal.(2014)Clinicalvalidityofcirculating

tumourcellsinpatientswithmetastaticbreastcancer:

TABLE1

FactorsinfluencingtheclinicalvalidityandutilityofcfDNA

Problem Solution

Preanalytical

LargevarietyinuseofbloodtubesandcfDNA isolationassays

Establishingevidence-basedstandardizedprotocol forbloodcollectionandprocessing

Unknowninfluenceofcircumstantialbiological factors(e.g.,comedications,comorbidity)

Researchinginfluenceofthesecircumstancesinlarge cohortstudies

Analytical

Multitudeofassaysusedtodetectand characterizectDNA

Vigorousanalyticalvalidationofeachassayand performcross-assaycomparisons

Postanalytical

Somaticvariantsreportedindifferentvariables Useofmutantcopynumber/mlinsteadofVAF Clinicalutility

Unknownrelevanceoffindingasomatic mutationat(extremely)lowfrequency

Performingclinicaltrialsinvestigatingatwhichcut-off asomaticmutationisrelatedtoresponsetoatargeted treatmentorclinicalresidualdisease

Unknownwhatchangeinsomaticmutation overtimeisclinicallyrelevant

Establishbiologicalandanalyticalvariationforagiven somaticmutationandsubsequentlyperformclinical trialsthattakethisvariationintoaccount

Features



apooledanalysisofindividualpatientdata.Lancet

Oncol.15,406–414

7 Janni,W.J.etal.(2016)Pooledanalysisofthe

prognosticrelevanceofcirculatingtumorcellsin

primarybreastcancer.Clin.CancerRes.22,2583–2593

8 Smerage,J.B.etal.(2014)Circulatingtumorcellsand

responsetochemotherapyinmetastaticbreastcancer:

SWOGS0500.J.Clin.Oncol.32,3483–3489

9 Ignatiadis,M.etal.(2018)Trastuzumabversus

observationforHER2nonamplifiedearlybreastcancer

withcirculatingtumorcells(EORTC90091-10093,BIG

1-12,TreatCTC):arandomizedphaseIItrial.Ann.Oncol.

29,1777–1783

10Bidard,F.C.etal.(2018)Clinicalutilityofcirculating

tumorcellcountasatooltochosebetweenfirstline

hormonetherapyandchemotherapyforER+

HER2-metastaticbreastcancer:resultsofthephaseIIISTIC

CTCtrial.SanAntonioBreastCancerSymp.2018

AbstractGS3–07

11Sieuwerts,A.M.etal.(2009)Anti-epithelialcell

adhesionmoleculeantibodiesandthedetectionof

circulatingnormal-likebreasttumorcells.J.Natl.

CancerInst.101,61–66

12Antonarakis,E.S.etal.(2014)AR-V7andresistanceto

enzalutamideandabirateroneinprostatecancer.N.

Engl.J.Med.371,1028–1038

13Onstenk,W.etal.(2015)Efficacyofcabazitaxelin

castration-resistantprostatecancerIsindependentof

thepresenceofAR-V7incirculatingtumorcells.Eur.

Urol.68,939–945

14Alberter,B.etal.(2016)Single-cellanalysisofCTCswith

diagnosticprecision:opportunitiesandchallengesfor

personalizedmedicine.ExpertRev.Mol.Diagn.16,25–

38

15Stoecklein,N.H.etal.(2016)ChallengesforCTC-based

liquidbiopsies:lowCTCfrequencyanddiagnostic

leukapheresisasapotentialsolution.ExpertRev.Mol.

Diagn.16,147–164

16Fischer,J.C.etal.(2013)Diagnosticleukapheresis

enablesreliabledetectionofcirculatingtumorcellsof

nonmetastaticcancerpatients.Proc.Natl.Acad.Sci.U.

S.A.110,16580–16585

17 Yu,M.etal.(2014)Cancertherapy.exvivocultureof

circulatingbreasttumorcellsforindividualizedtesting

ofdrugsusceptibility.Science345,216–220

18 Cayrefourcq,L.etal.(2015)Establishmentand

characterizationofacelllinefromhumancirculating

coloncancercells.CancerRes.75,892–901

19 Blackhall,F.etal.(2018)Willliquidbiopsiesimprove

outcomesforpatientswithsmall-celllungcancer?

LancetOncol.19,e470–e481

20 Stover,D.G.etal.(2018)Associationofcell-freeDNA

tumorfractionandsomaticcopynumberalterations

withsurvivalinmetastatictriple-negativebreast

cancer.J.Clin.Oncol.36,543–553

21 Moss,J.etal.(2018)Comprehensivehumancell-type

methylationatlasrevealsoriginsofcirculatingcell–free

DNAinhealthanddisease.Nat.Commun.9,5068

22 vanDessel,L.F.etal.(2017)Applicationofcirculating

tumorDNAinprospectiveclinicaloncologytrials–

standardizationofpreanalyticalconditions.Mol.Oncol.

11,295–304

23 vanDessel,L.F.etal.(2018)High-throughputisolation

ofcirculatingtumorDNA:acomparisonofautomated

platforms.Mol.Oncol.13,392–402

24 vanGinkel,J.H.etal.(2017)Preanalyticalbloodsample

workupforcell-freeDNAanalysisusingDropletDigital

PCRforfuturemolecularcancerdiagnostics.Cancer

Med.6,2297–2307

25 Torga,G.andPienta,K.J.(2018)Patient-pairedsample

congruencebetween2commercialliquidbiopsytests.

JAMAOncol.4 8688–8670

26 Mok,T.etal.(2015)Detectionanddynamicchangesof

EGFRmutationsfromcirculatingtumorDNAasa

predictorofsurvivaloutcomesinNSCLCpatients

treatedwithfirst-lineintercalatederlotiniband

chemotherapy.Clin.CancerRes.21,3196–3203

27 Rolfo,C.etal.(2018)Liquidbiopsyforadvanced

non-smallcelllungcancer(NSCLC):astatementpaperfrom

theIASLC.J.Thorac.Oncol.13,1248–1268

28 Oxnard,G.R.etal.(2016)Associationbetweenplasma

genotypingandoutcomesoftreatmentwith

osimertinib(AZD9291)inadvancednon-small-celllung

cancer.J.Clin.Oncol.34,3375–3382

29 Rothwell,D.G.etal.(2019)UtilityofctDNAtosupport

patientselectionforearlyphaseclinicaltrials:the

TARGETstudy.Nat.Med.25,738–743

30 Garcia-Murillas,I.etal.(2015)Mutationtrackingin

circulatingtumorDNApredictsrelapseinearlybreast

cancer.Sci.Transl.Med.7 302ra133

31 Tie,J.etal.(2016)CirculatingtumorDNAanalysis

detectsminimalresidualdiseaseandpredicts

recurrenceinpatientswithstageIIcoloncancer.Sci.

Transl.Med.8 346ra92

32 Siravegna,G.etal.(2015)Clonalevolutionand

resistancetoEGFRblockadeinthebloodofcolorectal

cancerpatients.Nat.Med.21,827

33 Cohen,J.D.etal.(2018)Detectionandlocalizationof

surgicallyresectablecancerswithamulti-analyte

bloodtest.Science359,926–930

34 Diamandis,E.P.andLi,M.(2016)Thesideeffectsof

translationalomics:overtesting,overdiagnosis,

overtreatment.Clin.Chem.Lab.Med.54,389–396

35 Merker,J.D.etal.(2018)CirculatingtumorDNAanalysis

inpatientswithcancer:AmericanSocietyofClinical

OncologyandCollegeofAmericanPathologistsJoint

Review.J.Clin.Oncol.36,1631–1641

NickBeijez,*

JohnW.M.Martensz

StefanSleijferz

ErasmusMCCancerInstitute,Departmentof

Med-icalOncologyandCancerGenomicsNetherlands,

Erasmus University Medical Center, Rotterdam,

TheNetherlands

*Correspondingauthor.

z_{Currentaddress:ErasmusMCCancerInstitute,}

Wytemaweg 80, 3015 CN, Rotterdam, The

Netherlands.

Features