Challenges in molecular testing in non-small-cell lung cancer patientswith advanced disease

Hiley, C. T., Le Quesne, J., Santis, G., Sharpe, R., de Castro, D. G., Middleton, G., & Swanton, C. (2016).Challenges in molecular testing in non-small-cell lung cancer patients with advanced disease. Lancet,388(10048), 1002-11. https://doi.org/10.1016/S0140-6736(16)31340-X

Published in:Lancet

Document Version:Peer reviewed version

Queen's University Belfast - Research Portal:Link to publication record in Queen's University Belfast Research Portal

Publisher rights© Elsevier Ltd. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ which permits distribution and reproduction for non-commercial purposes, provided the author and source are cited.

General rightsCopyright for the publications made accessible via the Queen's University Belfast Research Portal is retained by the author(s) and / or othercopyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associatedwith these rights.

Take down policyThe Research Portal is Queen's institutional repository that provides access to Queen's research output. Every effort has been made toensure that content in the Research Portal does not infringe any person's rights, or applicable UK laws. If you discover content in theResearch Portal that you believe breaches copyright or violates any law, please contact openaccess@qub.ac.uk.

Download date:26. Jul. 2020

1

Challengesinmoleculartestinginnon-smallcelllungcancerpatientswithadvanced1

disease2

3

CrispinTHiley1,2,JohnLeQuesne3,GeorgeSantis4,RowenaSharpe5,DavidGonzalezde4

Castro6,7,GaryMiddleton8,9andCharlesSwanton1,10*.5

6

1. TranslationalCancerTherapeuticsLaboratory,TheFrancisCrickInstitute,London,UK.72. DivisionofCancerStudies,King'sCollegeLondon,London,UK.8

3. DepartmentofCancerStudies,UniversityofLeicester,Leicester,UK.9

4. DepartmentofRespiratoryMedicineandAllergy,KingsCollegeLondon,UK.10

5. CancerResearchUK,London,UK.11

6. CentreforMolecularPathology,RoyalMarsdenHospital,Sutton,UK.12

7. SchoolofMedicine,DentistryandBiomedicalSciences,QueensUniversityBelfast,UK.13

8. InstituteofImmunologyandImmunotherapy,UniversityofBirmingham,Birmingham,14

UK.15

9. UniversityHospitalBirminghamNHSFoundationTrust,QueenElizabethHospital,16

Birmingham,UK.17

10. CRUKLungCancerCentreofExcellence,UCLCancerInstitute,London,UK.18

19

*Correspondenceto:ProfessorCharlesSwanton.TheFrancisCrickInstitute,44LincolnsInn20Field.WC2A3LY.charles.swanton@crick.ac.uk21

22

Summary:23

Lungcancerdiagnosticshaveevolvedsignificantlyinthepreviousdecade.Thechallengeof24

moleculartestingtoidentifyanincreasingnumberofpotentiallyclinicallyactionablegenetic25

variants, using smaller samples obtained via minimally invasive techniques, is significant.26

Tumour heterogeneity and cancer evolution in response to therapy means that repeat27

biopsiesorcirculatingbiomarkersarelikelytobeincreasinglyusefultoadapttreatmentas28

resistancedevelops.Wehighlight someof the current challenges faced in clinical practice29

formolecular testing of EGFR, ALK and new biomarkers such as PDL1. Implementation of30

nextgenerationsequencing(NGS)platformsformoleculardiagnosticsinnon-smallcelllung31

cancer is increasingly common allowing testing of multiple genetic variants from a single32

sample.TheuseofNGSto recruit tomolecularlystratifiedclinical trials isdiscussed in the33contextoftheUKStratifiedMedicineProgrammeandTheUKNationalLungMatrixTrial.34

2

35

Historicaloverviewoflungcancerdiagnostics36

37

Lungcancer is themostcommoncauseofmortality in theUK,accounting for1 in5ofall38

cancer deaths.(1) With the estimated global incidence in 2012 of 1.83 million cases it is39

important to reflect that a century ago, lung cancer diagnosis was a rare event. In40

comparisontothecurrentepidemic, in1912IsaacAdler’scollectionof374casereports in41hispublicationPrimaryMalignantGrowthsoftheLungsandBronchirepresentedtheentire42

known global incidence at the time. A century later theWHOhistological classification of43

malignantepithelialtumoursofthelungrecognizesdifferenthistologieswithmanyvariants44

foreachsubtypeandanalysesfromnextgenerationsequencing(NGS)studieshavedivided45

thisdisease intomolecularsubtypesdefinedbydistinctsomaticalterations.(2-4)Thisreview46

will focusonkey challenges faced in current clinical practice formolecular testing innon-47

small cell lung cancer (NSCLC). In broad terms the challenges are technical, logistical and48

relatedtotumourbiologyandsomeofthepertinentissuesarehighlighted(Figure1).49

50

Identificationoftumourhistology5152Historically the treatment focus for those with advanced NSCLC was selection of an53

appropriate cytotoxic chemotherapy regimen irrespective of histological subtype. Several54

large studies were published that showed the efficacy of various platinum doublet55

combinations were comparable but with differing drug specific toxicities.(5-7) However56accurate classification of NSCLC subtype has become fundamental in themanagement of57

advanced NSCLC following the results of phase III clinical trials showing improved58

progression free survival in EGFR mutation positive adenocarcinoma treated with EGFR59

tyrosinekinaseinhibitors(TKI),(8,9)andimprovedoverallsurvivalwithpemetrexedinthefirst60

lineandmaintenancesettingforthosewithnon-squamoushistology.(10)(11)61

62(12,13)The number of tumours that cannot be given an accurate histological diagnosis (i.e.63

adenocarcinomaversussquamouscellcarcinoma)hasreducedsignificantlywiththeuseof64

immunohistochemicalmarkers.Theuseofmarkersforp63,p40andcytokeratinCK5/6help65

to identify squamouscell carcinomas,while thyroid transcription factor1 (TTF1),NapsinA66

and CK7, as well as mucin stains, are indicative of adenocarcinomas.(14) However67

interpretationofimmunohistochemistrypanelsstillrequirestheexpertiseofanexperienced68

histopathologist, as markers are not reliable in isolation.(15) TTF1 for example, a marker69

3

synonymouswithadenocarcinoma, is expressed inonly80-90%of cases and is commonly70

expressedinneuroendocrinetumours.(14,16)Immunohistochemistrycanonlybemeaningfully71

interpretedinadetailedmorphologicalcontext.72

73

SamplingchallengesinadvancedNSCLC7475Theanalysisof lungcancer tissue isparticularlychallengingasprimary lungtumoursoften76

show much lower tumour cellularity than other tumour types. Even with macroscopic77selection of areas of frank carcinoma the tumour purity (the fraction of a given region78

containingtumourcells)canoftenbe<20%becauseofthehighproportionofstromalcells,79

lymphocytic infiltrationandnecrosis(unpublishedobservationsfromtheUKLungTRACERx80

longitudinalcohortstudy).(17).81

82

This challenge is compounded by the nature of the specimen types routinely received by83

histopathologyandmoleculardiagnosticslaboratories.Presentationwithmetastaticdisease84

is common and only a small proportion of patients with NSCLC undergo curative surgical85

resection.(18) The large tissue samples obtained via open thoracotomy (wedge resection,86

lobectomy,pneumonectomy)areusuallyofsufficientquantityandquality foranumberof87

histologicalandmolecularassaysifhandledappropriately.Howeverpatientswithadvanced88

diseasearepredominatelydiagnosedwithCTguidedpercutaneousorUSguidedendoscopic89

biopsywith18gaugeneedlesorwithfineneedleaspiration.Thesepatientsarethecohort90

wheremoleculardiagnosticsaremost important fordetermining thestandardofcareand91enablingparticipationinclinicaltrialsyetthesamplequalityandquantityfromsuchneedle92

biopsiesisthemostlimitingforhistologicalandmoleculartesting.93

94

Obtainingadequatetissuefordiagnosis,tissuesub-typing,molecularprofilingandtreatment95

planningarethereforekeytopatientmanagement.Thetargettumour isnotalwayseasily96

accessible in patients presenting with a probable lung cancer. The development of97

endobronchial ultrasound transbronchial needle aspiration (EBUS-TBNA) is proving98

increasingly important in the investigationandmanagementof thoracicmalignanciesas it99

offersaminimally invasiveapproach to samplingofmediastinal lymphnodesandmasses.100

EBUS-TBNAisnowincreasinglyembeddedinroutineclinicalpracticewithwiderusebeyond101

high volume tertiary centres in theUK andUSA. It is now generally accepted that EBUS-102

TBNAaloneorincombinationwithendoscopicultrasoundfineneedleaspiration(EUS-FNA)103

canpotentiallysparesurgicalmediastinoscopyorthoracotomyinthestagingofNSCLC.(19,20)104

4

Importantly,EBUS-TBNAalsooffers thepossibilityof combiningdiagnosisandstagingasa105

single procedure in patients with suspected lung cancer. In contrast to tissue biopsies or106

surgicalsamplesthatallowsub-typing(adenocarcinomaversussquamouscellcarcinoma)on107

morphological criteria alone in themajority of cases, evaluation of cytological specimens108

obtainedbyEBUS-TBNAposesadditionalchallengesthatcanbepartlyovercomewithwider109

useofimmunohistochemistry.(21)110

111Identifying driver mutations, such as EGFR and ALK, in these small samples is central to112

managementofpatientswithadvanceddisease.Whethermolecularanalysisissuccessfully113

performeddependsontheabsolutenumberoftumourcells,theproportionoftumourcells114

compared to total nucleated cells present and themethodused formolecular analysis. In115

caseofEBUS-derivedsamples, there isevidencetoconcludethatsimplemutationanalysis116

(EGFR,KRAS,ALK)canbesuccessfullyperformedinmostcases.(22-24)Theuseofmulti-gene117

targeted NGS panels, using only nanograms of DNA, to sequence fine needle aspiration118

samplesisachievableandisbecomingmorecommonlyusedinclinicalpractice.(25,26)Whole119

exome sequencing (WES) and whole genome sequencing (WGS) analysis which require120

greater amounts of DNA,micrograms in the case ofWGS, will bemore challenging from121

EBUS-TBNAsamples.122

123

124CurrentchallengesinmoleculardiagnosticsforEGFRmutationanalysisinclinicalpractice125

126

The initial randomised phase II studies of gefitinib demonstrated clinical activity,(27,28) and127

phaseIIIstudiesalthoughnegativefortheprimaryoutcomemeasure,suggestedabenefitin128

patients with adenocarcinoma, those of Asian origin and never smokers.(29,30) During this129

periodanumberofseminalcaseseriesidentifiedEGFRmutationsasamarkerofsensitivity130

to EGFR TKIs,(31-33) and analysis of samples from these early trials supported this.(34)131

Subsequent phase III trials have incorporated EGFR mutation status and showed higher132

responseratesandprogressionfreesurvival (PFS) inpatientswithEGFRactivatingsomatic133

mutationstreatedwithEGFRTKIscomparedtowhentreatedwithchemotherapy.(8,9,35)134

135

Developmentofdiversemechanismsofresistanceandselectionofresistantclonesin136

responsetotreatment137ThecommonEGFRmutationsarelocatedinthetyrosinekinasedomain(exon18-21)ofthe138

5

EGFRgene,withdetectionof L858Randdeletions inexon19 the clinicalpriority as these139

determinesensitivitytofirstandsecondgenerationTKIs.(36)TheT790Mmutationinexon20140

results in resistance to these therapies.(37) Sensitive assays suggest that tumour clones141

harbouringtheT790Mmutationareoftendetectablepriortoinitiationofafirstgeneration142

TKIbutcanalsooccurbygeneticevolutioninT790Mmutationnegativedrugtolerantcellsin143

response to treatment.(38-40) Identification of this resistance mutation is more critical144

followingthedevelopmentofthethirdgenerationEGFRTKIsactiveagainstT790Mmutation145positive NSCLCs.(41,42) But whether these T790M resistant clones pre-exist or evolve in146

response to treatment may have clinical implications with differing sensitivities to third147

generationTKIs.(40) The capacity for tumours toevolve in response to first generationTKIs148

results in anadditionaldiversearrayofmechanismsof resistance suchas amplificationof149

MET, selection for PIK3CA or BRAF mutations and transformation to a small cell150

phenotype.(43)Clearly,cancerevolutionandselectionofresistantsubclonesisnotrestricted151

to first generation TKIs. This is highlighted by recent reports of the emergence of T790M152

mutation negative disease and the development of novel secondary EGFR resistance153

mutations (C797S) after treatment of T790M mutation positive patients with third154

generationTKIs.(44,45)155

156

EGFRmutationtesting157

The nature of EGFR sensitizingmutations, being single nucleotide variants (SNV) or short158deletions, lends themselves to molecular analysis of formalin fixed small samples which159

contain fragmented DNA.(46) There are a variety of methods to detect EGFR mutations160

including conventional Sanger sequencing, amplification refractory mutations systems161

(ARMS), restriction fragment lengthpolymorphismsandmore recently aspartof targeted162

NGS panels.(47,48) Reporting the limitations of an assay along with the result is critical for163

clinicalinterpretation.Bi-directionalSangersequencingwithoutamutationenrichmentstep164

hasalowerlimitofdetectionof10-25%oftotalDNAmeaningthattheuseofsampleswith165

low tumour cellularity can result in falsenegativemutation calls. Consequently theuseof166

methods that can detect mutations in low tumour cellularity samples (<10%) is167

recommended.Polymerasechainreaction(PCR)basedultrasensitiveandNGSmethodscan168

generateartefactmutations leadingtofalsepositiveresults. Howevertechniques,suchas169

duplexsequencing,arebeingdevelopedtoovercometheinherenterrorrateinsequencing170

technologis(49) Formalin fixed samples are particularly prone to DNA damage and display171disproportionate levelsofC>T/G>Achanges inthe1-10%allelefrequencyrangewhichcan172

6

result in falsepositivemutationcalls.(50)Publicationof clinical trials resultson response to173

EGFR TKI in patients with real but less common EGFR mutations can help guide clinical174

decision-making.(51)Detectionof EGFRmutations aspart of aWESorWGSanalysis allows175

multiple driver mutations to be queried simultaneously but the performance of176

bioinformaticstoolstocallmutationsfromNGSdatavaries. Suchcomplexitiesneedtobe177

considered as these technologies are increasingly adopted into mainstream clinical178

practice.(52-54)179180

EGFRmutations,resistanceandtumourheterogeneity181

ThereareveryfewreportsofdiscordanceofEGFRmutationstatusbetweenprimarydisease182

andmetastaticsitesandthesemaybeduetotechnicallimitationsoftheassaysused.(55)Loss183

oftheEGFRmutationwasnotamechanismseeninseminalstudies.(43)Studieslookingatthe184

extent of intra tumour heterogeneity (ITH) in early lung cancer have shown EGFR to be185

exclusivelyaclonaleventprevalentthroughoutthetumour.(56,57)AsresistancetoEGFRTKIis186

usually due to acquisition of secondarymutations in EGFR or other driver genes the key187

challenge at the time of disease progression is to obtain a contemporaneous sample to188

informselectionofsecondlinetherapy.Ingeneralthemosteasilyaccessiblelesionisused189

butinpatientswithapoorperformancestatusthismaynotbeatrivialtask.Duetotumour190

heterogeneityitispossiblethatasinglesamplemaybeinsufficienttoaccuratelyrepresent191

all the resistance mechanisms present or the breadth of subclonal driver events present192acrossmultiplediseasesitesfollowingprogressionontherapy.193

194

CurrentchallengesinmoleculardiagnosticsforALKtestinginclinicalpractice195

196

Thediscoveryofanoncogenicanaplasticlymphomakinasefusiongene(EML4-ALK)in2007197

identified another importantmolecular cohort inNSCLC.(58) Present in 2-7%ofNSCLCALK198

fusiongenesarerestrictedtoadenocarcinomasubtypesandaremorecommoninyounger199

patients and never-smokers.(59-61) Identification of this cohort is critical given the high200

response rates (57-74%) to ALK inhibition both as a first line therapy and after platinum-201

based chemotherapy.(59,62,63) Subsequently other rare fusion genes have been identified202

involving ROS1, with similar exquisite sensitivity to kinase inhibition,(64) but also RET and203

NTRKwhereobjectiveresponseratesarelower.(65,66)204

205ALKfusiongenedetection206

7

TestingforALKfusiongenesbrings itsownparticularsetofchallenges.ALK isactivatedby207

genomic rearrangement, leading to the expression of a chimeric protein containing the208

effectorpartoftheALKtyrosinekinasefusedtotheproximalportionofanotherprotein.In209

NSCLCcancerthisistypicallyabalancedtranslocationwiththeubiquitouslyhighlyexpressed210

EML4gene,(58)althoughrarelyotherpartnergenesmaybe involved.(67,68)Expressionofthe211

chimericproteinleadstoupregulationofmitogenicsignallingthroughtheRAS/RAFpathway212

and interruption of this pathway by ALK inhibitors causes cancer cell death and tumour213regression.(62) ALK-mutated tumours often show unusual features on conventional214

microscopy,suchascribriformgrowthpatternsand‘signetring’cellswithlargevacuoles,(69)215

butthisisnotsufficientlysensitiveorspecifictoguidetreatment.216

217

The first widely adopted test for ALK-driven tumours was FISH (fluorescence in situ218

hybridisation),approvedbytheFDA(USFoodandDrugAdministration)in2011.(70)FISHisa219

technically demanding method, requiring specialised equipment and experienced220

practitioners.Tissuesectionsorcytologyspecimensaresubjectedtoaprotocol that labels221

eithersideoftheALKbreakpointlocuswithredandgreenfluorescentDNAprobes.Innon-222

transformedcellnucleithecoloureddotsoverlapandlookyellow,whileintranslocatedcells223

isolated red and/or green signals are seen. For a reliable FISH assay the tissue must be224

adequate inquantity andquality. This canbemore challengingwith small biopsy samples225

whichmaycontainfewcellsorwhichshowcrushingartefactsthatcanimpairinterpretation.226227

In 2015, an immunohistochemistry method was approved by the FDA. This approach is228

simpler in principle, using an antibody stain to detect abnormal ALK antigen expression.229

However, currently available antibodies do not give a strong signal and so an additional230

signalamplificationstepsneedstobeemployed.Thisplacesthetestbeyondthecapacityof231

many small labs.Nonetheless, themodified test is cheaper than FISH, easier to interpret,232

and has the theoretical advantage of additionally detecting ALK expression following rare233

atypical rearrangements. After much investigation, recent studies suggest234

immunohistochemistrycanbeanadequatestand-alonediagnostic,showingextremelyhigh235

concordancewith FISH.(71) UK guidelines do not dictatewhich test should be applied, and236

practices vary regionally, though FISH is still often regarded as the ‘gold standard’ and is237

considered thedefinitive test in theUS.(48)Asourunderstandingof tumour taxonomyand238

genotypesadvances, it seems inevitable that some formofNGSplatformwillbecome the239clinical standard for gene fusiondetection.(72) Thesemethodshave thepotential todetect240

8

ALK(andother)rearrangementsineitheratargetedpaneloraWESorWGSapproach.(73,74)241(72)242

243

ALKfusiongenes,resistanceandtumourheterogeneity244

ALK fusion genes are considered to be clonal events with minimal discordance between245

primaryandmetastaticlesions.(75)TheywereconsideredtobemutuallyexclusivewithEGFR246

mutations however recent reports suggests a smallminority of tumours can contain both247ALK and EGFR positive clones.(76-78) The mechanisms of resistance seen following ALK248

inhibitortherapyagaindemonstratetumourevolutionwithsecondarymutationsinALK,ALK249

copynumbergain,secondarydrivermutationsinothergenesandoutgrowthofALKfusion250

gene negative clones reported.(79-82) Consequently contemporaneous sampling of251

progressivedisease,byneedlebiopsyoranalysisofcfDNA,mayallowrealtimeanalysisof252

tumourevolutionandguidetherapy.253

254

Integrationofmulti-geneNGStestinginclinicalpractice255

Routinemolecularprofilingcanbeperformedatscaleonanationallevel.Largecooperative256

efforts in France and the USA used combinations of mutation specific PCRs, Sanger257

sequencingandFISHanalysistoassay6-10oncogenicdriversinthousandsofpatientswith258

NSCLC and survival was improved for those treated with gene directed targeted259

therapies.(83,84) The use of next generation sequencing to simultaneously assay multiple260oncogenicdriversisattractivebecauselessDNAisrequiredcomparedtomultipleindividual261

assays, there is a reduction in hands-on laboratory time, and complex FISH analysis for262

detectionoffusiongenesmaybeavoided.ArecentNGSapproachusedanampliconbased263

approach to assay 14 genes used only 50ng of DNA from FFPE samples.(85) This study264

providedacomprehensiveassessmentofthespectrumofmutations,andco-occurrenceof265

mutations, in adenocarcinoma and squamous cell carcinomaswith detection turn around266

timesoflessthantwoweeks.ThesestudiesandthoseofTheCancerGenomeAtlashighlight267

theinter-patientmolecularheterogeneityofNSCLC(Figure2).Evenwithinthesemolecular268

cohorts intra-tumour heterogeneity could have significant effects on outcome as269

exemplified by a recent study showing that the clonality of FGFR amplification is an270

importantpredictorofresponsetoFGFRinhibition.(86)Adeeperunderstandingoftheclonal271

or subclonal nature of driver events in NSCLC from sufficiently powered studies, is still272

awaited. Recruitment of patients with rare mutations to molecularly stratified trials is273challenging,(87) and some advocate that modifications to existing paradigms in drug274

9

development are required in the eraof genomic studies andprecisionmedicine.(88)Multi-275

geneorWESNGSassaysarelikelytobecomestandardpracticeintheyearstocomeandthe276

ultimatelyautomatedprovisionofreadable,applicablereportsofcomplexgenomicdata is277

anotherimportantchallenge.278

279

CurrentchallengesinmoleculardiagnosticsforPDL1testinginclinicalpractice280

281Activation of inhibitory T cell checkpoint interactions in established tumours has been282

demonstratedinanumberofsolidtumours,includingNSCLC,andthissuppressestheanti-283

tumour immune response.(89,90) The aim of immunotherapy using antagonists of these284

inhibitoryTcellcheckpointinteractionsistoreactivateanti-tumourimmunity.PDL1(B7-H1)285

isa ligandpresentonantigenpresentingcells (APCs), including tumourcells that interacts286

withitsreceptor(PD-1)onTcellsandinhibitsTcelleffectorfunctions.PD-1andCD8positive287

effectorTcellpopulationarethoughttobethetumourreactivesubsetresponsibleforanti-288

tumourimmunity.(91)Thereislimitedknowledgeofthespatialorfunctionalheterogeneityof289

tumour infiltrating lymphocyte (TIL) populations and the T cell checkpoint ligand-receptor290

interactionswithinsolidtumours.291

292

Recent randomised trials have shown activity of PD1 and PDL1 targeting antibodies in293

squamous and non-squamous NSCLC.(92-96) In most instances these agents have shown294greateractivity inpatientswhosetumourexpressesPDL1whentestedusing IHC.However295

durable responsesareseen inpatientswithoutPDL1expression.This isunsurprisinggiven296

thetechnicalandspatialheterogeneityofPDL1expressioninNSCLC,whichhampersitsuse297

asapredictivebiomarker.(97-99)StudiesoftheexpressionofPDL1onAPCsinNSCLCarealso298

contradictorywith respect to any correlationwith tumour infiltration of the effector CD8299

positiveTcells.(89,97,98)300

301

Regulation of PDL1 expression is complex and controlled by both cell intrinsic and cell302

extrinsic factors.(100) This means that oncogene driven expression of PDL1 can result in303

increased expression in the absence of significant underlying immunogenicity.(101) This304

underlying immunogenicity is thought to be a result of non-synonymous SNVs which305

generate neoantigens, mutated proteins, recognised by the TIL population.(102,103) The306

number of neoantigens harboured by a tumour could act as a potential biomarker for307immunotherapy although there are technical challenges inherent with such complex308

10

assessments.Recentdataalsosuggestthatneo-antigenintratumourheterogeneitymayalso309

beassociatedwithalteredcheckpointinhibitorresponse,whichmayfurthercomplicatethe310

useofsuchassaysinaclinicalsetting.(104)311

312

The advent of immunotherapy presents additional challenges formolecular diagnostics in313

NSCLC.Although IHC for PDL1 canbeperformedon the small samples oftenused in lung314

cancerdiagnosticsthereistheriskofsignificantsamplingbiasbecauseofITH.Thedynamic315nature of PDL1 gene expression,(105,106) means a contemporaneous sample obtained by316

repeatbiopsymaybe themostaccurateaddingadditionalburdenandexpense tocurrent317

clinicalpathways.Characterisationofneo-antigensasapotentialbiomarkerwould require318

sufficient tumourDNA forWESandcarries significantexpensebutgiven thecostof these319

therapieswouldbe justified if theassayweresufficientlypredictive.Howeverneo-antigen320

prediction algorithms are still in their infancy and evidence suggests that there are a321

proportionofpatientswhoderivenoclinicalbenefitfromcheckpointinhibitortherapy,yet322

have tumourswithaneo-antigenburdenabove thresholds associatedwith sensitivity and323

converselypatientswithlowneo-antigenburdenwhobenefit.324

325

Moleculardiagnosticsinpractice:TheUnitedKingdomNationalLungMatrixTrial326

327

The Cancer Research UK Stratified Medicine Programme 2 (SMP2) screens samples from328advanced NSCLC patients using NGS for known drivers that are considered clinically329

actionable.TheaimofSMP2istoestablishhigh-throughputandqualitygenomicscreening330

atanational level intheUK.Basedontheseresults,patientsarerecruitedtoTheNational331

Lung Matrix Trial (NLMT) (NCT02664935) a phase II umbrella study with both targeted332

therapyandimmunotherapyarmsforpatientswhohaveprogressedonfirstlinetherapy.(107)333

In comparison, the Lung-MAP (NCT02154490) and SAFIR02 Lung trial (NCT02117167) are334

umbrellastudies,outsideoftheUK,forpatientswithNSCLCwhererecruitmentispreceded335

bymolecularstratification(Table1).336

337

SMP2 molecular pathology workflow utilises DNA from excess diagnostic biopsy tissue.338

Sections are sent from the referring clinical site and extracted by one of three central339

technologyhubs.SampleswithsufficientamountsofDNA(>50ng)arethenanalysedusinga340

custom 28-gene targeted NGS panel. Having successfully screened over 1000 patients,341patterns of mutation and prevalence are emerging across the genomic and clinical data.342

11

Preliminary analysis indicates prevalence and distribution of SNVs consistent to published343

reports, including 31.6% KRAS (of which 19.7% show concomitant STK11 mutation) and344

15.1%EGFRmutations inpatientswithadenocarcinoma.Overthepastyear,SMP2has led345

to the recruitment of over 60 patients to the NLMT. A number of detailed audits have346

identified areas of improvement along the SMP2 pathway; from patient recruitment, to347

samplepreparationandresultanalysis.348

349Whilst utilising excess DNA from the FFPE diagnostic biopsy has significant advantage for350

patients and clinical workload (as repeat biopsies are not required), only 70% of samples351

senthavesufficientDNAtoenterthesequencingpipeline.ThisisinpartduetoFFPEblocks352

being exhausted during the diagnostic process and a general reduction in the size of353

diagnosticcoresovertime.Consequentlytheminimumnumberofsectionshassincebeen354

increasedtoensureenoughDNAisobtainedupfront.SomerecruitingcentresquantifyDNA355

upfront,which allows a faster feedback loop if insufficientDNA is present. Sites can then356

obtain additional sample from the diagnostic block or through re-biopsy, if appropriate.357

However, differences in quantification methodology between local clinical centres and358

central technologyhubshave led to samplesbeing sentwith less than the required50ng,359

resulting some of these samples failing quality control metrics prior to sequencing. As a360

result changes in extraction methods and a standardized DNA concentration have been361

introduced.362363

Unique toNLMT is theneed todeterminewild type statusof somegenes foreligibility to364

certainarms.PatientsrecruitedtotheCDK4/6inhibitorpalbociclibarmmusthavewildtype365

retinoblastoma 1 protein (RB1) in addition to deficiencies in cell cycle regulation. The366

determinationofwildtypestatusrequiresapre-sequencingassessmentoftumourcellularity367

todetermineappropriatesequencingdepth.Howevertherecanbesignificantdiscordance368

between pathologist assessments of this.(108) Clearer guidance and online training should369

ensuremoreconcordanceforvisualassessment,whilstdigitalsolutionsmayprovideauseful370

alternative.Anumberofcomputationalmethodsexisttoassesstumourpurityandcontrol371

forbothstromalcelladmixtureandcancercellploidyinDNAsamplesfromnextgeneration372

sequencingdata.(109,110)373

374

Extremes of GC nucleotide content in certain genes (RB1 and FGFR3) can result in an375increasednumberof sequencing failures.Additionalprobe coverage in the targetedpanel376

12

and correction for GC content in the data processing stagewill improve results for these377

difficulttosequenceregions.Followingtheseincrementalimprovementsateachstepofthe378

molecularpathologyworkflowwehave shown that thenumberof successfully sequenced379

samples thatwould allow recruitment to theNLMT increased and therehas alsobeen an380

increaseinidentificationofpotentiallyactionablemutationsthatwouldpermitrecruitment381

totrialsotherthantheNLMT.382

383Futuresolutions384

385

Understandingtumourheterogeneityandcancerevolution386

Atpresentthetechnicallimitationofthesmall,andpotentiallylowtumourcellularityNSCLC387

samples, obtained from bronchoscopic and EBUS-TBNA samples means that the main388

challenge facing clinicians and pathologists is the need for ever greater amounts of389

informationfromdiminishingamountsoftissue.Itisthereforeimperativethatthequalityof390

diagnostic samples in the advanced NSCLC setting is of the highest order. How best to391

achievethisrepresentsachallengeforhealthserviceprovidersthathasreceivedverylittle392

attention thus far.However the spectreof ITHand cancer evolutionmeans that sampling393

biasandthepresenceofsubclonaldrivermutations,causingresistancetotherapy,arelikely394

to hinder clinical benefit of targeted therapeutics.(111,112) The UK Lung TRAcking Cancer395

EvolutionthroughTherapytrial(NCT01888601)iscurrentlycharacterizingtheextentofITH396in early surgically resected NSCLC and with longitudinal follow-up aims to determine the397

originsoftumoursubclonescontributingtorelapse.(17)Thereisevidencefromothertumour398

typesofparallelevolution,acquisitionofmutationsinthesamegeneorsignallingpathway399

in distinct subclones, that may highlight an ‘evolutionary bottle neck’ that could be an400

Achilles heel for subsequent cancer therapy.(113-115) Clonal analyses of a drug target and401

putativeresistanceevents,whethertheyarepresentinalltumorcellsoronlyaproportion,402

mayaffect the response rateandprogression free survival timeson targeted therapyand403

this is being addressed in clinical trials including the DARWIN studies (NCT02314481,404

NCT02183883). Ultimately it may be that only through ‘warm’ autopsy studies, where405

subclonal phylogenetic structures can be determined through sampling multiple sites of406

disease,thateffectivestrategiestoforestallcancerevolutioncanbeelucidated.(116)407

408

Circulatingbiomarkers409The use of minimally invasive methods to detect mutations in circulating cell-free DNA410

13

(cfDNA)or‘liquidbiopsies’offersthepotentialtoobtainamutationcallinapatientwhere411

an invasive biopsymay not be feasible. As tumour DNA from all sites of disease has the412

potential to enter the blood stream it may also be a better reflection of tumour413

heterogeneitythanasinglebiopsy.(117,118)TheuseofcfDNAtodetectresistancemechanisms414

in patients treated with EGFR TKIs, often prior to radiographic progression, has been415

demonstrated.(45,119,120)Thishasresultedinthedevelopmentandapprovalofacommercially416

available assay of cfDNA in plasma that can detect a spectrum of EGFR mutations in417includingtheT790MmutationamenabletotargetingwiththirdgenerationTKIs.418

419

Circulating tumour cells (CTCs) are tumour cells that can be isolated from the peripheral420

bloodandareacomplementarycirculatingbiomarker tocfDNA. CTCsareaversatile tool,421

cellenumerationcanbeprognostic,immunohistochemistrypermitsfurthercharacterization,422

singlecellDNAorRNAsequencing ispossibleand tumourxenografts canbegenerated to423

assess drug response.(121-125) However at present the complexity of separation from other424

cellsintheperipheralcirculationandtheneedtoprocesssamplespromptlyforfunctionalor425

genomic studies results in greater expense in comparison to cfDNA analysis. Circulating426

biomarkers will make a significant impact on cancermanagement in the near future and427

readers are directed to more extensive reviews focusing on CTCs, cfDNA and other428

circulatingnucleotides.(121,126-129)429

430Conclusion431

ThechallengesformoleculardiagnosticsinNSCLCarelargelyparalleledacrossothertumour432

types.Resolvingtheseissueswillrequiretechnologyimprovementsinadditiontoagreater433

understanding of tumour biology. The logistical challenges of implementing the next434

generationofmoleculardiagnostics intoclinicalpracticeareequallyaschallenging.Clinical435

governance, information technology infrastructure, data storage, pathways in sample436

processing and training and professional developments in histopathology, respiratory437

medicineandoncologywillneedinvestment.Withthesegreatchallengescomessignificant438

opportunity to improvethesuccess rateandefficiencyofdrugdevelopment inNSCLCand439

ultimatelypatientoutcomes.440

441

Conflictsofinterest:None442

14

Figure1:Summaryofthekeytechnical,logisticalandbiologicalchallengesformoleculartestinginNSCLC.Figure2:Piechartsshowingtheapproximatedistributionofclinicallyrelevantdrivermutationsidentified to date in individuals with NSCLC. The genomic variants shown are potentiallyclinically actionable variants.(130) Theproportionspresentedarebasedonestimates from thereferenced studies and data sources, including the Stratified Medicine Programme 2(unpublisheddata).(2,3,85,131)ThesestudiesexamineSNVs,copynumbervariantsandgenefusionproductsusingdifferentsequencingtechnologiesandsequencingdepthresultingininter-studyvariation and therefore the data is presented in aggregate form and represents anapproximation.CCGA(Cellcyclegenomicaberration);lossofCyclin-DependentKinaseInhibitor2A or amplification of Cyclin-Dependent Kinase 4 or Cyclin D1 in the presence of wildtypeRB1.(107) EGFR, Epidermal growth factor receptor; LKB1, Liver Kinase B1; ALK, anaplasticlymphoma kinase; MET, MET Proto-Oncogene, Receptor Tyrosine Kinase; FGFR, fibroblasticgrowth factor receptor; NRAS, neuroblastoma RAS viral (v-ras) oncogene homolog; DDR2Discoidin Domain Receptor Tyrosine Kinase 2; AKT1, v-akt murine thymoma viral oncogenehomolog 1; PTEN, Phosphatase And Tensin Homolog; PIK3CA, phosphoinositide-3-kinase,catalytic, α polypeptide; BRAF, v-raf murine sarcoma viral oncogene homolog B1; ERBB2,humanepidermalgrowthfactorreceptor2;KRAS,v-Ki-ras2Kirstenratsarcomaviraloncogenehomolog;TP53,TumorProteinP53.Table1:ComparisonofmolecularlystratifiedumbrellastudiesinNSCLC.PFS=progressionfreesurvival, ORR = objective response rate, R = randomised, NR = non-randomised, SCC =squamouscellcarcinoma,CCGA=cellcyclegenomicaberration.

15

REFERENCES

1. LungCancerKeyStatistics.UnitedKingdom:CancerResearchUK,2015.

2. CancerGenomeAtlasResearchN.Comprehensivegenomiccharacterizationofsquamouscelllungcancers.Nature.2012;489(7417):519-25.

3. CancerGenomeAtlasResearchN.Comprehensivemolecularprofilingoflungadenocarcinoma.Nature.2014;511(7511):543-50.

4. SwantonC,GovindanR.ClinicalImplicationsofGenomicDiscoveriesinLungCancer.TheNewEnglandjournalofmedicine.2016;374(19):1864-73.

5. SchillerJH,HarringtonD,BelaniCP,etal.Comparisonoffourchemotherapyregimensforadvancednon-small-celllungcancer.TheNewEnglandjournalofmedicine.2002;346(2):92-8.

6. ScagliottiGV,DeMarinisF,RinaldiM,etal.PhaseIIIrandomizedtrialcomparingthreeplatinum-baseddoubletsinadvancednon-small-celllungcancer.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2002;20(21):4285-91.

7. KellyK,CrowleyJ,BunnPA,Jr.,etal.RandomizedphaseIIItrialofpaclitaxelpluscarboplatinversusvinorelbinepluscisplatininthetreatmentofpatientswithadvancednon--small-celllungcancer:aSouthwestOncologyGrouptrial.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2001;19(13):3210-8.

8. MokTS,WuYL,ThongprasertS,etal.Gefitiniborcarboplatin-paclitaxelinpulmonaryadenocarcinoma.TheNewEnglandjournalofmedicine.2009;361(10):947-57.

9. RosellR,CarcerenyE,GervaisR,etal.Erlotinibversusstandardchemotherapyasfirst-linetreatmentforEuropeanpatientswithadvancedEGFRmutation-positivenon-small-celllungcancer(EURTAC):amulticentre,open-label,randomisedphase3trial.Thelancetoncology.2012;13(3):239-46.

10. CiuleanuT,BrodowiczT,ZielinskiC,etal.Maintenancepemetrexedplusbestsupportivecareversusplaceboplusbestsupportivecarefornon-small-celllungcancer:arandomised,double-blind,phase3study.Lancet.2009;374(9699):1432-40.

11. ScagliottiGV,ParikhP,vonPawelJ,etal.PhaseIIIstudycomparingcisplatinplusgemcitabinewithcisplatinpluspemetrexedinchemotherapy-naivepatientswithadvanced-stagenon-small-celllungcancer.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2008;26(21):3543-51.

12. ReckM,PopatS,ReinmuthN,etal.Metastaticnon-small-celllungcancer(NSCLC):ESMOClinicalPracticeGuidelinesfordiagnosis,treatmentandfollow-up.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2014;25Suppl3:iii27-39.

13. MastersGA,TeminS,AzzoliCG,etal.SystemicTherapyforStageIVNon-Small-CellLungCancer:AmericanSocietyofClinicalOncologyClinicalPracticeGuidelineUpdate.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2015;33(30):3488-515.

16

14. KerrKM,BubendorfL,EdelmanMJ,etal.SecondESMOconsensusconferenceonlungcancer:pathologyandmolecularbiomarkersfornon-small-celllungcancer.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2014;25(9):1681-90.

15. TerryJ,LeungS,LaskinJ,LeslieKO,GownAM,IonescuDN.Optimalimmunohistochemicalmarkersfordistinguishinglungadenocarcinomasfromsquamouscellcarcinomasinsmalltumorsamples.AmJSurgPathol.2010;34(12):1805-11.

16. ZhangC,SchmidtLA,HatanakaK,ThomasD,LagsteinA,MyersJL.EvaluationofnapsinA,TTF-1,p63,p40,andCK5/6immunohistochemicalstainsinpulmonaryneuroendocrinetumors.AmJClinPathol.2014;142(3):320-4.

17. Jamal-HanjaniM,HackshawA,NgaiY,etal.Trackinggenomiccancerevolutionforprecisionmedicine:thelungTRACERxstudy.PLoSbiology.2014;12(7):e1001906.

18. AuditNLC.NLCAAnnualReport.London:RoyalCollegeofPhysicians,2015.

19. RiveraMP,MehtaAC,WahidiMM.Establishingthediagnosisoflungcancer:Diagnosisandmanagementoflungcancer,3rded:AmericanCollegeofChestPhysiciansevidence-basedclinicalpracticeguidelines.Chest.2013;143(5Suppl):e142S-65S.

20. KinseyCM,ArenbergDA.Endobronchialultrasound-guidedtransbronchialneedleaspirationfornon-smallcelllungcancerstaging.AmJRespirCritCareMed.2014;189(6):640-9.

21. NavaniN,BrownJM,NankivellM,etal.Suitabilityofendobronchialultrasound-guidedtransbronchialneedleaspirationspecimensforsubtypingandgenotypingofnon-smallcelllungcancer:amulticenterstudyof774patients.AmJRespirCritCareMed.2012;185(12):1316-22.

22. SantisG,AngellR,NicklessG,etal.ScreeningforEGFRandKRASmutationsinendobronchialultrasoundderivedtransbronchialneedleaspiratesinnon-smallcelllungcancerusingCOLD-PCR.PLoSOne.2011;6(9):e25191.

23. NakajimaT,YasufukuK,NakagawaraA,KimuraH,YoshinoI.Multigenemutationanalysisofmetastaticlymphnodesinnon-smallcelllungcancerdiagnosedbyendobronchialultrasound-guidedtransbronchialneedleaspiration.Chest.2011;140(5):1319-24.

24. vanEijkR,LichtJ,SchrumpfM,etal.RapidKRAS,EGFR,BRAFandPIK3CAmutationanalysisoffineneedleaspiratesfromnon-small-celllungcancerusingallele-specificqPCR.PLoSOne.2011;6(3):e17791.

25. Kanagal-ShamannaR,PortierBP,SinghRR,etal.Next-generationsequencing-basedmulti-genemutationprofilingofsolidtumorsusingfineneedleaspirationsamples:promisesandchallengesforroutineclinicaldiagnostics.ModPathol.2014;27(2):314-27.

26. QiuT,GuoH,ZhaoH,WangL,ZhangZ.Next-generationsequencingformoleculardiagnosisoflungadenocarcinomaspecimensobtainedbyfineneedleaspirationcytology.Scientificreports.2015;5:11317.

27. KrisMG,NataleRB,HerbstRS,etal.Efficacyofgefitinib,aninhibitoroftheepidermalgrowthfactorreceptortyrosinekinase,insymptomaticpatientswithnon-smallcelllungcancer:arandomizedtrial.JAMA.2003;290(16):2149-58.

17

28. FukuokaM,YanoS,GiacconeG,etal.Multi-institutionalrandomizedphaseIItrialofgefitinibforpreviouslytreatedpatientswithadvancednon-small-celllungcancer(TheIDEAL1Trial)[corrected].Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2003;21(12):2237-46.

29. ThatcherN,ChangA,ParikhP,etal.Gefitinibplusbestsupportivecareinpreviouslytreatedpatientswithrefractoryadvancednon-small-celllungcancer:resultsfromarandomised,placebo-controlled,multicentrestudy(IressaSurvivalEvaluationinLungCancer).Lancet.2005;366(9496):1527-37.

30. GiacconeG,HerbstRS,ManegoldC,etal.Gefitinibincombinationwithgemcitabineandcisplatininadvancednon-small-celllungcancer:aphaseIIItrial--INTACT1.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2004;22(5):777-84.

31. LynchTJ,BellDW,SordellaR,etal.Activatingmutationsintheepidermalgrowthfactorreceptorunderlyingresponsivenessofnon-small-celllungcancertogefitinib.TheNewEnglandjournalofmedicine.2004;350(21):2129-39.

32. PaezJG,JannePA,LeeJC,etal.EGFRmutationsinlungcancer:correlationwithclinicalresponsetogefitinibtherapy.Science(NewYork,NY.2004;304(5676):1497-500.

33. PaoW,MillerV,ZakowskiM,etal.EGFreceptorgenemutationsarecommoninlungcancersfrom"neversmokers"andareassociatedwithsensitivityoftumorstogefitinibanderlotinib.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica.2004;101(36):13306-11.

34. BellDW,LynchTJ,HaserlatSM,etal.Epidermalgrowthfactorreceptormutationsandgeneamplificationinnon-small-celllungcancer:molecularanalysisoftheIDEAL/INTACTgefitinibtrials.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2005;23(31):8081-92.

35. YangJC,WuYL,SchulerM,etal.Afatinibversuscisplatin-basedchemotherapyforEGFRmutation-positivelungadenocarcinoma(LUX-Lung3andLUX-Lung6):analysisofoverallsurvivaldatafromtworandomised,phase3trials.Thelancetoncology.2015;16(2):141-51.

36. SharmaSV,BellDW,SettlemanJ,HaberDA.Epidermalgrowthfactorreceptormutationsinlungcancer.NaturereviewsCancer.2007;7(3):169-81.

37. PaoW,MillerVA,PolitiKA,etal.AcquiredresistanceoflungadenocarcinomastogefitiniborerlotinibisassociatedwithasecondmutationintheEGFRkinasedomain.PLoSMed.2005;2(3):e73.

38. InukaiM,ToyookaS,ItoS,etal.PresenceofepidermalgrowthfactorreceptorgeneT790Mmutationasaminorcloneinnon-smallcelllungcancer.CancerRes.2006;66(16):7854-8.

39. SuKY,ChenHY,LiKC,etal.Pretreatmentepidermalgrowthfactorreceptor(EGFR)T790MmutationpredictsshorterEGFRtyrosinekinaseinhibitorresponsedurationinpatientswithnon-small-celllungcancer.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2012;30(4):433-40.

40. HataAN,NiederstMJ,ArchibaldHL,etal.Tumorcellscanfollowdistinctevolutionarypathstobecomeresistanttoepidermalgrowthfactorreceptorinhibition.Naturemedicine.2016;22(3):262-9.

18

41. JannePA,YangJC,KimDW,etal.AZD9291inEGFRinhibitor-resistantnon-small-celllungcancer.TheNewEnglandjournalofmedicine.2015;372(18):1689-99.

42. SequistLV,SoriaJC,GoldmanJW,etal.RociletinibinEGFR-mutatednon-small-celllungcancer.TheNewEnglandjournalofmedicine.2015;372(18):1700-9.

43. SequistLV,WaltmanBA,Dias-SantagataD,etal.GenotypicandhistologicalevolutionoflungcancersacquiringresistancetoEGFRinhibitors.SciTranslMed.2011;3(75):75ra26.

44. PiotrowskaZ,NiederstMJ,KarlovichCA,etal.HeterogeneityUnderliestheEmergenceofEGFRT790Wild-TypeClonesFollowingTreatmentofT790M-PositiveCancerswithaThird-GenerationEGFRInhibitor.Cancerdiscovery.2015;5(7):713-22.

45. ThressKS,PaweletzCP,FelipE,etal.AcquiredEGFRC797SmutationmediatesresistancetoAZD9291innon-smallcelllungcancerharboringEGFRT790M.Naturemedicine.2015;21(6):560-2.

46. WilliamsC,PontenF,MobergC,etal.Ahighfrequencyofsequencealterationsisduetoformalinfixationofarchivalspecimens.TheAmericanjournalofpathology.1999;155(5):1467-71.

47. PirkerR,HerthFJ,KerrKM,etal.ConsensusforEGFRmutationtestinginnon-smallcelllungcancer:resultsfromaEuropeanworkshop.Journalofthoraciconcology:officialpublicationoftheInternationalAssociationfortheStudyofLungCancer.2010;5(10):1706-13.

48. LindemanNI,CaglePT,BeasleyMB,etal.MoleculartestingguidelineforselectionoflungcancerpatientsforEGFRandALKtyrosinekinaseinhibitors:guidelinefromtheCollegeofAmericanPathologists,InternationalAssociationfortheStudyofLungCancer,andAssociationforMolecularPathology.Journalofthoraciconcology:officialpublicationoftheInternationalAssociationfortheStudyofLungCancer.2013;8(7):823-59.

49. SchmittMW,KennedySR,SalkJJ,FoxEJ,HiattJB,LoebLA.Detectionofultra-raremutationsbynext-generationsequencing.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica.2012;109(36):14508-13.

50. WongSQ,LiJ,TanAY,etal.Sequenceartefactsinaprospectiveseriesofformalin-fixedtumourstestedformutationsinhotspotregionsbymassivelyparallelsequencing.BMCmedicalgenomics.2014;7:23.

51. YangJC,SequistLV,GeaterSL,etal.Clinicalactivityofafatinibinpatientswithadvancednon-small-celllungcancerharbouringuncommonEGFRmutations:acombinedpost-hocanalysisofLUX-Lung2,LUX-Lung3,andLUX-Lung6.Thelancetoncology.2015;16(7):830-8.

52. LaiZ,MarkovetsA,AhdesmakiM,etal.VarDict:anovelandversatilevariantcallerfornext-generationsequencingincancerresearch.NucleicAcidsRes.2016.

53. O'RaweJ,JiangT,SunG,etal.Lowconcordanceofmultiplevariant-callingpipelines:practicalimplicationsforexomeandgenomesequencing.Genomemedicine.2013;5(3):28.

54. RehmHL,BaleSJ,Bayrak-ToydemirP,etal.ACMGclinicallaboratorystandardsfornext-generationsequencing.GenetMed.2013;15(9):733-47.

55. ParkS,Holmes-TischAJ,ChoEY,etal.Discordanceofmolecularbiomarkersassociatedwithepidermalgrowthfactorreceptorpathwaybetweenprimarytumorsandlymphnodemetastasisinnon-

19

smallcelllungcancer.Journalofthoraciconcology:officialpublicationoftheInternationalAssociationfortheStudyofLungCancer.2009;4(7):809-15.

56. ZhangJ,FujimotoJ,ZhangJ,etal.Intratumorheterogeneityinlocalizedlungadenocarcinomasdelineatedbymultiregionsequencing.Science(NewYork,NY.2014;346(6206):256-9.

57. deBruinEC,McGranahanN,MitterR,etal.Spatialandtemporaldiversityingenomicinstabilityprocessesdefineslungcancerevolution.Science(NewYork,NY.2014;346(6206):251-6.

58. SodaM,ChoiYL,EnomotoM,etal.IdentificationofthetransformingEML4-ALKfusiongeneinnon-small-celllungcancer.Nature.2007;448(7153):561-6.

59. KwakEL,BangYJ,CamidgeDR,etal.Anaplasticlymphomakinaseinhibitioninnon-small-celllungcancer.TheNewEnglandjournalofmedicine.2010;363(18):1693-703.

60. KoivunenJP,MermelC,ZejnullahuK,etal.EML4-ALKfusiongeneandefficacyofanALKkinaseinhibitorinlungcancer.ClinCancerRes.2008;14(13):4275-83.

61. ZhaoW,ChoiYL,SongJY,etal.ALK,ROS1andRETrearrangementsinlungsquamouscellcarcinomaareveryrare.LungCancer.2016;94:22-7.

62. ShawAT,KimDW,NakagawaK,etal.CrizotinibversuschemotherapyinadvancedALK-positivelungcancer.TheNewEnglandjournalofmedicine.2013;368(25):2385-94.

63. SolomonBJ,MokT,KimDW,etal.First-linecrizotinibversuschemotherapyinALK-positivelungcancer.TheNewEnglandjournalofmedicine.2014;371(23):2167-77.

64. ShawAT,OuSH,BangYJ,etal.CrizotinibinROS1-rearrangednon-small-celllungcancer.TheNewEnglandjournalofmedicine.2014;371(21):1963-71.

65. TakeuchiK,SodaM,TogashiY,etal.RET,ROS1andALKfusionsinlungcancer.Naturemedicine.2012;18(3):378-81.

66. VaishnaviA,CapellettiM,LeAT,etal.Oncogenicanddrug-sensitiveNTRK1rearrangementsinlungcancer.Naturemedicine.2013;19(11):1469-72.

67. IyevlevaAG,RaskinGA,TiurinVI,etal.NovelALKfusionpartnersinlungcancer.CancerLett.2015;362(1):116-21.

68. TakeuchiK,ChoiYL,TogashiY,etal.KIF5B-ALK,anovelfusiononcokinaseidentifiedbyanimmunohistochemistry-baseddiagnosticsystemforALK-positivelungcancer.ClinCancerRes.2009;15(9):3143-9.

69. YoshidaA,TsutaK,NakamuraH,etal.ComprehensivehistologicanalysisofALK-rearrangedlungcarcinomas.AmJSurgPathol.2011;35(8):1226-34.

70. IASLCatlasofalktestinginlungcancer:InternationalAssociationfortheStudyofLungCancerPress;2013.

71. TakeuchiK,TogashiY,KamiharaY,etal.ProspectiveandclinicalvalidationofALKimmunohistochemistry:resultsfromthephaseI/IIstudyofalectinibforALK-positivelungcancer(AF-001JPstudy).AnnOncol.2016;27(1):185-92.

20

72. DavareMA,TognonCE.Detectingandtargettingoncogenicfusionproteinsinthegenomicera.BiolCell.2015;107(5):111-29.

73. DrilonA,WangL,ArcilaME,etal.Broad,HybridCapture-BasedNext-GenerationSequencingIdentifiesActionableGenomicAlterationsinLungAdenocarcinomasOtherwiseNegativeforSuchAlterationsbyOtherGenomicTestingApproaches.Clinicalcancerresearch:anofficialjournaloftheAmericanAssociationforCancerResearch.2015;21(16):3631-9.

74. JangJS,WangX,VedellPT,etal.CustomGeneCaptureandNextGenerationSequencingtoResolveDiscordantALKStatusbyFISHandIHCinLungAdenocarcinoma.Journalofthoraciconcology:officialpublicationoftheInternationalAssociationfortheStudyofLungCancer.2016.

75. HouLWC.ComparativestudyofALKrearrangementbetweenprimarytumorandpairedlymphaticmetastasisinNSCLCpatients.ASCO20152015.

76. CaiW,LinD,WuC,etal.IntratumoralHeterogeneityofALK-RearrangedandALK/EGFRCoalteredLungAdenocarcinoma.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2015;33(32):3701-9.

77. BirkbakNJ,HileyCT,SwantonC.EvolutionaryPrecisionMedicine:ARoleforRepeatEpidermalGrowthFactorReceptorAnalysisinALK-RearrangedLungAdenocarcinoma?Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2015;33(32):3681-3.

78. SasakiT,KoivunenJ,OginoA,etal.AnovelALKsecondarymutationandEGFRsignalingcauseresistancetoALKkinaseinhibitors.CancerRes.2011;71(18):6051-60.

79. ChoiYL,SodaM,YamashitaY,etal.EML4-ALKmutationsinlungcancerthatconferresistancetoALKinhibitors.TheNewEnglandjournalofmedicine.2010;363(18):1734-9.

80. DoebeleRC,PillingAB,AisnerDL,etal.MechanismsofresistancetocrizotinibinpatientswithALKgenerearrangednon-smallcelllungcancer.Clinicalcancerresearch:anofficialjournaloftheAmericanAssociationforCancerResearch.2012;18(5):1472-82.

81. KatayamaR,ShawAT,KhanTM,etal.MechanismsofacquiredcrizotinibresistanceinALK-rearrangedlungCancers.SciTranslMed.2012;4(120):120ra17.

82. ShawAT,FribouletL,LeshchinerI,etal.ResensitizationtoCrizotinibbytheLorlatinibALKResistanceMutationL1198F.TheNewEnglandjournalofmedicine.2016;374(1):54-61.

83. BarlesiF,MazieresJ,MerlioJP,etal.Routinemolecularprofilingofpatientswithadvancednon-small-celllungcancer:resultsofa1-yearnationwideprogrammeoftheFrenchCooperativeThoracicIntergroup(IFCT).Lancet.2016;387(10026):1415-26.

84. KrisMG,JohnsonBE,BerryLD,etal.Usingmultiplexedassaysofoncogenicdriversinlungcancerstoselecttargeteddrugs.JAMA.2014;311(19):1998-2006.

85. KonigK,PeiferM,FassunkeJ,etal.ImplementationofAmpliconParallelSequencingLeadstoImprovementofDiagnosisandTherapyofLungCancerPatients.Journalofthoraciconcology:officialpublicationoftheInternationalAssociationfortheStudyofLungCancer.2015;10(7):1049-57.

86. PearsonA,SmythE,BabinaIS,etal.High-LevelClonalFGFRAmplificationandResponsetoFGFRInhibitioninaTranslationalClinicalTrial.Cancerdiscovery.2016.

21

87. Lopez-ChavezA,ThomasA,RajanA,etal.Molecularprofilingandtargetedtherapyforadvancedthoracicmalignancies:abiomarker-derived,multiarm,multihistologyphaseIIbaskettrial.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2015;33(9):1000-7.

88. BiankinAV,PiantadosiS,HollingsworthSJ.Patient-centrictrialsfortherapeuticdevelopmentinprecisiononcology.Nature.2015;526(7573):361-70.

89. KonishiJ,YamazakiK,AzumaM,KinoshitaI,Dosaka-AkitaH,NishimuraM.B7-H1expressiononnon-smallcelllungcancercellsanditsrelationshipwithtumor-infiltratinglymphocytesandtheirPD-1expression.Clinicalcancerresearch:anofficialjournaloftheAmericanAssociationforCancerResearch.2004;10(15):5094-100.

90. TaubeJM,AndersRA,YoungGD,etal.ColocalizationofinflammatoryresponsewithB7-h1expressioninhumanmelanocyticlesionssupportsanadaptiveresistancemechanismofimmuneescape.SciTranslMed.2012;4(127):127ra37.

91. GrosA,RobbinsPF,YaoX,etal.PD-1identifiesthepatient-specificCD8(+)tumor-reactiverepertoireinfiltratinghumantumors.TheJournalofclinicalinvestigation.2014;124(5):2246-59.

92. GaronEB,RizviNA,HuiR,etal.Pembrolizumabforthetreatmentofnon-small-celllungcancer.TheNewEnglandjournalofmedicine.2015;372(21):2018-28.

93. HerbstRS,BaasP,KimDW,etal.Pembrolizumabversusdocetaxelforpreviouslytreated,PD-L1-positive,advancednon-small-celllungcancer(KEYNOTE-010):arandomisedcontrolledtrial.Lancet.2016;387(10027):1540-50.

94. BorghaeiH,Paz-AresL,HornL,etal.NivolumabversusDocetaxelinAdvancedNonsquamousNon-Small-CellLungCancer.TheNewEnglandjournalofmedicine.2015;373(17):1627-39.

95. BrahmerJ,ReckampKL,BaasP,etal.NivolumabversusDocetaxelinAdvancedSquamous-CellNon-Small-CellLungCancer.TheNewEnglandjournalofmedicine.2015;373(2):123-35.

96. FehrenbacherL,SpiraA,BallingerM,etal.Atezolizumabversusdocetaxelforpatientswithpreviouslytreatednon-small-celllungcancer(POPLAR):amulticentre,open-label,phase2randomisedcontrolledtrial.Lancet.2016.

97. TaubeJM,KleinA,BrahmerJR,etal.AssociationofPD-1,PD-1Ligands,andOtherFeaturesoftheTumorImmuneMicroenvironmentwithResponsetoAnti-PD-1Therapy.Clinicalcancerresearch:anofficialjournaloftheAmericanAssociationforCancerResearch.2014.

98. VelchetiV,SchalperKA,CarvajalDE,etal.Programmeddeathligand-1expressioninnon-smallcelllungcancer.Laboratoryinvestigation;ajournaloftechnicalmethodsandpathology.2014;94(1):107-16.

99. McLaughlinJ,HanG,SchalperKA,etal.QuantitativeAssessmentoftheHeterogeneityofPD-L1ExpressioninNon-Small-CellLungCancer.JAMAoncology.2016;2(1):46-54.

100. PardollDM.Theblockadeofimmunecheckpointsincancerimmunotherapy.NaturereviewsCancer.2012;12(4):252-64.

101. AkbayEA,KoyamaS,CarreteroJ,etal.ActivationofthePD-1pathwaycontributestoimmuneescapeinEGFR-drivenlungtumors.Cancerdiscovery.2013;3(12):1355-63.

22

102. RizviNA,HellmannMD,SnyderA,etal.Cancerimmunology.MutationallandscapedeterminessensitivitytoPD-1blockadeinnon-smallcelllungcancer.Science(NewYork,NY.2015;348(6230):124-8.

103. BrownSD,WarrenRL,GibbEA,etal.Neo-antigenspredictedbytumorgenomemeta-analysiscorrelatewithincreasedpatientsurvival.Genomeresearch.2014;24(5):743-50.

104. McGranahanN,FurnessAJ,RosenthalR,etal.ClonalneoantigenselicitTcellimmunoreactivityandsensitivitytoimmunecheckpointblockade.Science(NewYork,NY.2016;351(6280):1463-9.

105. PatelSP,KurzrockR.PD-L1ExpressionasaPredictiveBiomarkerinCancerImmunotherapy.MolCancerTher.2015;14(4):847-56.

106. HerbstRS,SoriaJC,KowanetzM,etal.Predictivecorrelatesofresponsetotheanti-PD-L1antibodyMPDL3280Aincancerpatients.Nature.2014;515(7528):563-7.

107. MiddletonG,CrackLR,PopatS,etal.TheNationalLungMatrixTrial:translatingthebiologyofstratificationinadvancednon-small-celllungcancer.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2015;26(12):2464-9.

108. ShawEC,HanbyAM,WheelerK,etal.ObserveragreementcomparingtheuseofvirtualslideswithglassslidesinthepathologyreviewcomponentofthePOSHbreastcancercohortstudy.JClinPathol.2012;65(5):403-8.

109. FaveroF,JoshiT,MarquardAM,etal.Sequenza:allele-specificcopynumberandmutationprofilesfromtumorsequencingdata.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2015;26(1):64-70.

110. VanLooP,NordgardSH,LingjaerdeOC,etal.Allele-specificcopynumberanalysisoftumors.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica.2010;107(39):16910-5.

111. YatesLR,CampbellPJ.Evolutionofthecancergenome.NatRevGenet.2012;13(11):795-806.

112. HileyC,deBruinE,McGranahanN,SwantonC.Decipheringintratumorheterogeneityandtemporalacquisitionofdrivereventstorefineprecisionmedicine.Genomebiology.2014;15(8):453.

113. MelchorL,BrioliA,WardellCP,etal.Single-cellgeneticanalysisrevealsthecompositionofinitiatingclonesandphylogeneticpatternsofbranchingandparallelevolutioninmyeloma.Leukemia.2014.

114. FisherR,HorswellS,RowanA,etal.DevelopmentofsynchronousVHLsyndrometumorsrevealscontingenciesandconstraintstotumorevolution.Genomebiology.2014;15(8):433.

115. JuricD,CastelP,GriffithM,etal.ConvergentlossofPTENleadstoclinicalresistancetoaPI(3)Kalphainhibitor.Nature.2015;518(7538):240-4.

116. GundemG,VanLooP,KremeyerB,etal.Theevolutionaryhistoryoflethalmetastaticprostatecancer.Nature.2015;520(7547):353-7.

117. Jamal-HanjaniM,WilsonGA,HorswellS,etal.Detectionofubiquitousandheterogeneousmutationsincell-freeDNAfrompatientswithearly-stagenon-small-celllungcancer.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2016;27(5):862-7.

23

118. MurtazaM,DawsonSJ,TsuiDW,etal.Non-invasiveanalysisofacquiredresistancetocancertherapybysequencingofplasmaDNA.Nature.2013;497(7447):108-12.

119. OxnardGR,PaweletzCP,KuangY,etal.NoninvasivedetectionofresponseandresistanceinEGFR-mutantlungcancerusingquantitativenext-generationgenotypingofcell-freeplasmaDNA.Clinicalcancerresearch:anofficialjournaloftheAmericanAssociationforCancerResearch.2014;20(6):1698-705.

120. ThressKS,BrantR,CarrTH,etal.EGFRmutationdetectioninctDNAfromNSCLCpatientplasma:Across-platformcomparisonofleadingtechnologiestosupporttheclinicaldevelopmentofAZD9291.LungCancer.2015;90(3):509-15.

121. HouJM,KrebsM,WardT,etal.Circulatingtumorcellsasawindowonmetastasisbiologyinlungcancer.TheAmericanjournalofpathology.2011;178(3):989-96.

122. KrebsMG,SloaneR,PriestL,etal.Evaluationandprognosticsignificanceofcirculatingtumorcellsinpatientswithnon-small-celllungcancer.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2011;29(12):1556-63.

123. MaheswaranS,SequistLV,NagrathS,etal.DetectionofmutationsinEGFRincirculatinglung-cancercells.TheNewEnglandjournalofmedicine.2008;359(4):366-77.

124. LohrJG,AdalsteinssonVA,CibulskisK,etal.Whole-exomesequencingofcirculatingtumorcellsprovidesawindowintometastaticprostatecancer.Naturebiotechnology.2014;32(5):479-84.

125. HodgkinsonCL,MorrowCJ,LiY,etal.Tumorigenicityandgeneticprofilingofcirculatingtumorcellsinsmall-celllungcancer.Naturemedicine.2014;20(8):897-903.

126. DiazLA,Jr.,BardelliA.Liquidbiopsies:genotypingcirculatingtumorDNA.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2014;32(6):579-86.

127. Alix-PanabieresC,PantelK.Challengesincirculatingtumourcellresearch.NaturereviewsCancer.2014;14(9):623-31.

128. Alix-PanabieresC,PantelK.ClinicalApplicationsofCirculatingTumorCellsandCirculatingTumorDNAasLiquidBiopsy.Cancerdiscovery.2016;6(5):479-91.

129. SchwarzenbachH,HoonDS,PantelK.Cell-freenucleicacidsasbiomarkersincancerpatients.NaturereviewsCancer.2011;11(6):426-37.

130. DienstmannR,JangIS,BotB,FriendS,GuinneyJ.Databaseofgenomicbiomarkersforcancerdrugsandclinicaltargetabilityinsolidtumors.Cancerdiscovery.2015;5(2):118-23.

131. CeramiE,GaoJ,DogrusozU,etal.ThecBiocancergenomicsportal:anopenplatformforexploringmultidimensionalcancergenomicsdata.Cancerdiscovery.2012;2(5):401-4.

Technical• Novelbiopsytechniques(egEBUS)genera=ngsmallersampleswithdiminishedtumourcellularity

• Mul=pletestswiththepoten=alfordiscordantresults(egIHCvsFISHforALKmuta=on)

• Technology-specificfailuresduetodifferencesinsensi=vity/knownartefacts(egsequencingthroughrepeats/highGCareas)

• Qualityassuranceofgenomicmedicinedespiteacrossmul=pleplaQormsanddataanalysisalgorithms

Logis=cal• Turnaround=meofassaysinaclinicallyrelevant=meframe

• Desirabilityofcentralisedvsdistributed/localtes=ngapproaches

• Educa=onandtrainingoflaboratoryandclinicalstaffnnewtechnologies

• Dis=lla=onofhighvolumedataintousefulstandardisedreportsusablebyclinicicans

• Computa=onalanddatastoragecapacityforNGSwithinahealthcaresystem

TumourBiology• DiversityofmolecularsubgroupswithinNSCLC•  inter-pa=entheterogeneity

•  Intra-tumourheterogeneity• samplingbias• differen=alresponses

• Cancerevolu=onandresistanceinresponsetotreatments• needforlongitudinalsampling

• Evolvingtreatmentparadigms•  immuno-oncolog&newbiomarkers(egPDL-1,neoan=genload)

•  Increasingcomplexityofdetectablegenomicchangesincancer• egepigene=cchanges&non-codingvariants

Adenocarcinoma

35%

KRAS

No variant

EGFRLKB1

ALK

Cell cycle variant

MET

DDR2Other

BRAFFGFR

ERBB2

75%

mutant wildtype

TP53 status

Squamous Cell Carcinoma

No variant

Cell cycle variant

FGFR

PIK3CA

PTEN

KRAS

METDDR2

Other

57% 43%

mutant wildtype

Study Line Phase POM Molecularsubgroups LocationNationalLungMatrixTrial

2ndorlater NRPhaseII PFSORR

AKT,PIK3CA/PTEN,TSC,LKB1,KRAS,NRAS,NF1,MET,ROS1,EGFR(T790M),CCGA,immunotherapy

UK

SAFIR_02LungStudy

1stlinemaintenance

RPhaseII PFS mTOR,AKT,FGFR,HER2,EGFR,MEK,immunotherapy

France

LungMAP 2ndorlater(SCC)

NRPhaseII PFSORR

PIK3CA,FGFR,CCGA,immunotherapy

USA

DarwinI/II 1storlater NRPhaseII PFS EGFR,HER2,ALK,RET,BRAF,immunotherapy

UK