Bias-Corrected Targeted Next-Generation Sequencing...

Personalized Medicine and Imaging

Bias-Corrected Targeted Next-GenerationSequencing for Rapid, Multiplexed Detectionof Actionable Alterations in Cell-Free DNA fromAdvanced Lung Cancer PatientsCloud P. Paweletz1, Adrian G. Sacher2, Chris K. Raymond3, Ryan S. Alden2,Allison O'Connell1, Stacy L. Mach2, Yanan Kuang1, Leena Gandhi2, Paul Kirschmeier1,Jessie M. English1, Lee P. Lim3, Pasi A. J€anne1,2, and Geoffrey R. Oxnard2

Abstract

Purpose: Tumor genotyping is a powerful tool for guidingnon–small cell lung cancer (NSCLC) care; however, comprehen-sive tumor genotyping can be logistically cumbersome. To facil-itate genotyping, we developed a next-generation sequencing(NGS) assay using a desktop sequencer to detect actionablemutations and rearrangements in cell-free plasma DNA (cfDNA).

Experimental Design: An NGS panel was developed targeting11 driver oncogenes found in NSCLC. Targeted NGS was per-formed using a novel methodology that maximizes on-targetreads, and minimizes artifact, and was validated on DNA dilu-tions derived from cell lines. Plasma NGS was then blindlyperformed on 48 patients with advanced, progressive NSCLC anda known tumor genotype, and explored in two patients withincomplete tumor genotyping.

Results: NGS could identify mutations present in DNA dilu-tions at�0.4% allelic frequency with 100% sensitivity/specificity.Plasma NGS detected a broad range of driver and resistance

mutations, including ALK, ROS1, and RET rearrangements,HER2insertions, and MET amplification, with 100% specificity. Sensi-tivity was 77% across 62 known driver and resistance mutationsfrom the 48 cases; in 29 cases with common EGFR and KRASmutations, sensitivity was similar to droplet digital PCR. In twocases with incomplete tumor genotyping, plasma NGS rapidlyidentified a novel EGFR exon 19 deletion and a missed case ofMET amplification.

Conclusions: Blinded to tumor genotype, this plasma NGSapproach detected a broad range of targetable genomic alterationsin NSCLC with no false positives including complex mutationslike rearrangements and unexpected resistance mutations such asEGFR C797S. Through use of widely available vacutainers and adesktop sequencing platform, this assay has the potential to beimplemented broadly for patient care and translational research.Clin Cancer Res; 22(4); 915–22. �2015 AACR.

See related commentary by Tsui and Berger, p. 790

IntroductionGenotype-directed targeted therapies are revolutionizing can-

cer care. Genomic alterations in genes such as EGFR, ALK, KRAS,and BRAF have been validated as powerful predictive biomarkersin the management of non–small cell lung cancer (NSCLC),colorectal cancer, and melanoma; it is now standard to test forthese mutations to personalize treatment decisions (1–7). Devel-opment of new genotype-directed therapies is widespread in solid

tumor oncology, leading to increasing application of next-gener-ation sequencing (NGS) panels that can test tumor biopsies for awide range of potentially targetable mutations (8, 9). However,routine use of NGS for tumor genotyping presents practicalchallenges including the availability of adequate biopsy speci-mens, slow turnaround time, and the need for repeat biopsiesafter development of drug resistance (9).Given these challenges, itis clear that there is an unmet need for noninvasive assays that canbroadly detect actionable genomic alterations.

Many groups, including our own, have investigated nonin-vasive tumor genotyping of cell-free plasma DNA (cfDNA) asan alternative to tissue genotyping (10–15). Rather than study-ing circulating cells, these technologies study the free floatingDNA contained in the plasma; in advanced cancer patients, aportion of this cfDNA may be derived from the tumor. Plasmagenotyping has the potential to be less invasive and faster thantumor genotyping, while also allowing serial assessment ofgenotype during development of treatment resistance. Werecently reported on a highly specific and rapid droplet digitalPCR (ddPCR) assay for quantifying the concentration of EGFRand KRAS mutations in cfDNA of advanced NSCLC patients(16, 17). Such PCR-based plasma assays test for mutations at asingle site in a gene, but are limited by their inability to detectmore complex genomic alterations such as chromosomal

1Belfer Center for Applied Cancer Science, Dana-Farber Cancer Insti-tute, Boston, Massachusetts. 2Lowe Center for Thoracic Oncology,Dana-Farber Cancer Institute, Boston, Massachusetts. 3ResolutionBioscience, Bellevue,Washington.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Prior presentation: A portion of this data was presented previously as an oralpresentation at the AACR Annual Meeting 2015.

Corresponding Author: Geoffrey R. Oxnard, Dana-Farber Cancer Institute, 450Brookline Avenue, Boston, MA 02115. Phone: 617-632-6049; Fax: 617-632-5786;E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-15-1627-T

�2015 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 915

on May 24, 2018. © 2016 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 12, 2015; DOI: 10.1158/1078-0432.CCR-15-1627-T

http://clincancerres.aacrjournals.org/

rearrangements and their inability to multiplex across severalgenes. Others have studied NGS of cfDNA using PCR ampliconsor tagged DNA baits to enrich for target DNA sequences;however, many such assays are unable to detect rearrange-ments, whereas other assays rely on massive sequencing andcomputational processing resulting in unacceptable costs andslow turnaround time (11, 18). Although detection of mutantcfDNA present at low concentration is possible with theseapproaches despite the more abundant wild-type (germline)DNA, a universal challenge with these highly sensitive geno-typing assays is the risk of false positives due to PCR artifact.

In this study, we piloted a novel-targeted NGS approach forthe detection of driver mutations and rearrangements incfDNA from advanced NSCLC patients. Taking cues fromtraditional hybrid capture approaches that isolate genomicsubsets by pull down with probes to genes of interest, ourmethodology improves on key steps during library generationto reduce sequencing demands and turnaround times. First, tomaximize on-target reads to �90%, a two-step pull-downprocess was used that includes both a thermodynamicallycontrolled hybridization step and a kinetically controlledextension step under conditions that neutralize GC bias.Then, to improve signal-to-noise ratio, tags were connectedto each captured DNA fragment, so that every read is anchoredto its clonal family and to its pull-down probe of origin,facilitating identification of low-frequency mutant alleles andquantification of subtle changes in gene copy numbers (Fig. 1and Supplementary Methods). We hypothesized that thisapproach would allow for accurate detection of a broad rangeof targetable genotypes, including insertions/deletions andrearrangements, in cfDNA from advanced NSCLC patients.Our goal was to leverage a desktop sequencing platform toenable a rapid turnaround time and facilitate widespreadclinical adoption.

Materials and MethodsPlasma NGS

Targeted NGS of cell-line DNA and plasma cfDNA was per-formed at Resolution Bioscience as described in SupplementaryMethods. Chimeric gene fusions were detected using tiled probesthat allow sequencing-based discovery of de novo rearrangements(Supplementary Fig. S1 and Supplementary Table S1).

Plasma ddPCRFor comparison to plasmaNGS, plasma ddPCRwas performed

using an established andvalidated assaywhichhasbeendescribedpreviously (16). Briefly, this assay emulsifies extracted plasmacfDNA into thousands of droplets which subsequently undergoindividual PCR with custom fluorescently labeled probesdesigned to detect EGFR L858R, EGFR exon 19 deletions, or KRASG12X (16). Individual droplets are then read in a flow cytometerand the number of positive droplets are quantified (Bio-Rad).Each sample is analyzed in triplicate.

Cell line validationThe targeted NGS panel was validated using genomic DNA

from 14 independent, genetically-annotated cell lines harboringfour gene fusions, 19 point mutations, and two insertions/dele-tions (Supplemental Table S2). Cell lineswere combined into twoseparate DNA pools, each containing the genomes of seven celllines, and systematically blended with normal, wild-type DNA toproduce admixtures at 2.5%, 1.0%, 0.4%, and 0.1% dilutions.Prior to NGS, DNA pools were acoustically fragmented to anaverage size distribution centered around 165 bp and purified bytwo-sided SPRI to give fragment profiles of 150 to 200 bp thatclosely approximate cfDNA.Cell lines for the analytical validationexperiment were obtained from ATCC (A549, H2228, SK-MEL-2,H1666, SW48), RIKEN (Lc-2/ad), the Broad institute (H1781,SW480, HCT116, H2347, HCC78), and the NCI (H3122). PC-9and H1975 cells were obtained from the laboratory of Dr. PasiJ€anne. All cell lines were validated to be correct by short tandemrepeat (STR) analyses.

Patient populationPatients were identified during their routine lung cancer care at

Dana-Farber Cancer Institute. Patients were deemed eligible ifthey had advanced NSCLC with a known tumor genotype, eitheruntreated or progressive on therapy. Tumor genotyping wasperformed as part of routine care, either using conventionalgenotyping assays (PCR, FISH) or a targeted NGS panel whenavailable (5, 19). All patients consented to plasma collection andanalysis on an IRB-approved prospective plasma collection pro-tocol or associated correlative science protocols. Following clin-ical validation, plasma NGS was explored in two patients withhigh suspicion of a targetable genotype missed on tumorgenotyping.

Plasma collectionPlasmawas collected prior to initiation of therapy for untreated

or progressive advanced NSCLC. Whole blood was collected into10 mL EDTA containing "purple-top" vacutainer tubes, centri-fuged for 10minutes at 1,200� g and the plasma supernatant wasfurther cleared by centrifugation for 10 minutes at 3,000 � g.Cleared plasma was stored in cryostat tubes at �80�C until use.cfDNAwas isolated using theQIAmpCirculating Nucleic Acid Kit(Qiagen) according to the manufacturer's protocol. DNA waseluted in AVE buffer (100 mL) and stored at �80�C until use.

Blinding of specimensTo ensure data integrity in these experiments, the samples were

identified only by sample key. Only the clinical team (GRO, AGS,RSA) that identified patients for study had access to the tissuegenotype results. The teams involved in ddPCR and NGS analysis

Translational Relevance

Noninvasive genotyping of cell-free plasma DNA (cfDNA)is a potentially powerful tool for advancing cancer care andtranslational research, but most established assays are PCR-based and limited to detection of hotspotmutations. Here, wedescribe the development of a novel rapid targeted next-generation sequencing (NGS) assay for study of cfDNA. Study-ing 48 cases using a desktop sequencer, this assay was able todetect targetable oncogenic genomic alterations and resistancemechanisms in advanced non–small cell lung cancer withoutany false positives. The comprehensive coverage afforded bythis assay while utilizing a widely available NGS platform hasgreat potential for broad uptake as a tool for noninvasivetumor genotyping.

Paweletz et al.

Clin Cancer Res; 22(4) February 15, 2016 Clinical Cancer Research916




were blinded during data acquisition (plasma isolation, cfDNAextraction, library generation, NGS, and ddPCR analysis) andunblinding was done after NGS and ddPCR results had beenreported to the clinical team.

ResultsA probe set was developed that covers portions of 11 genes

known to be targetable oncogenic drivers in NSCLC. Selectedcoding regions of eight genes were sequenced (KRAS, EGFR, ALK,HER2, BRAF, NRAS, PIK3CA, MET, and MEK1). In addition,intronic probes were designed to detect genome-level rearrange-

ments that create chimeric gene fusions in ALK, ROS1, and RET.The coding regions of the tumor suppressor TP53 were alsoincluded as a control because this gene is commonly mutated inNSCLC. Analysis of the performance of this probe set on plasmaDNA showed >80% on-target percentage, which compares favor-ably with the less than 50% on-target percentage seen in previ-ously published data using standard hybridization selection onplasma DNA (Supplementary Fig. S2 and ref. 18).

The targeted NGS panel was initially validated with dilutionsof cell line DNA. Four different dilutions of two DNA poolswere sequenced, each derived from seven cell lines harboring

BB B B

B B BB

A

B

C

D

E

Standard hybrid capture

120 bp probe

cfDNA (~135–480 bp)

Figure 1.Key differences between standard hybrid capture (left) and bias-correctedNGS (right). A,mono-, di-, and trimeric nucleosome cfDNA fragments ranging from 130 to480 basepairs are isolated. B, in standard hybrid capture, cfDNA fragments are end-repaired and ligated with single primers. In contrast, bias-correctedNGS uses multifunctional adaptors that include sequences for single-primer amplification (red), tags for sample identification (green), and sequence identificationtags (blue) that, in conjunction with the fragmentation site (blue dot) identify unique sequence clones. C, in standard hybridization cfDNA fragments arecaptured with large capture probes (up to 120 bp) that span the genetic region of interest andmay result in off-target fragments being isolated (e.g., daisy-chainingoff-target DNA). Bias-corrected NGS uses small capture probes (�40 bp) that are designed to be adjacent to the region of interest. Primer extension offragments copies genomic and adaptor sequences. Finally, amplification with tailed PCR primers creates sequencing ready clones. D, although both approachesallow sequencing of gene rearrangements, large capture probes designed to target one gene will inefficiently target fragments containing a large amountof fusion partner gene sequence, resulting in poor sensitivity. In bias-corrected NGS, gene junction and partner gene sequence is replicated during primer extension.E, in standard hybrid capture all pulled-down cfDNA (specific and nonspecific) is amplified and sequenced without knowing the exact read or probe whichcaptured the fragment. In bias-correctedNGS, READ_1 identifies the sample ID and the unique sequence identifiers, whereas READ_2 identifies the probe that pulleddown each clone, facilitating read analysis, and probe optimization.

Targeted NGS of Cell-Free DNA from Advanced NSCLC

www.aacrjournals.org Clin Cancer Res; 22(4) February 15, 2016 917




previously characterized mutations (Supplemental Table S2).Dilutions of 2.5% to 0.1% resulted in calculated allelic frequen-cies ranging from4.5% to 0.01%(Supplementary Fig. S3). Variantcalling algorithms (Supplementary Methods) were able to iden-tifymutations that were present at 0.1% or greater with sensitivityand specificity of 88% and 100%. Diagnostic performanceimproved to 100% sensitivity and specificity for mutations pres-ent at an allelic frequency 0.4% or greater (Supplementary Fig. S3and Table S2). This preliminary analysis of the cell lines allowedthe setting of thresholds that ensured high specificity in thesubsequent analysis of patient samples.

After validating the NGS platform with DNA dilutions, plasmasamples from 48 patients with advanced NSCLC were studied,blinded to the tumor genotyping results (Supplemental Table S3).The median age of these patients was 57, 61% were female and92%had extra-thoracicmetastatic disease.Mean reads per sample

was 8.9 million, with a mean coverage per base of 983 uniquereads.

The sensitivity of theNGSplatformwasfirst studied in 29 of the48 patients known by tumor genotyping to harbor EGFR andKRAS driver mutations readily assayed with ddPCR (Fig. 2A).Using the tumor genotyping as the gold standard, ddPCR ofplasma had a sensitivity of 86%, whereas NGS had a sensitivityof 79%, not significantly different (P¼ 0.43). Bothmethods weremore sensitive when more cfDNA was available. The allelicfraction of the mutant allele in cfDNA, calculated as the numberof mutant reads over wild-type reads, was closely correlated forplasma NGS and plasma ddPCR (Pearson correlation¼ 0.93, P <0.001; Fig. 2B).

Detection of rare mutations and rearrangements in cfDNA wasnext studied in a blinded fashion in 20 of the 48 patients withNSCLC known to harbor a rare mutation or rearrangement on

2.0

4.3

6.1

11

2.0

1.9

15

9.7

17

0.8

1.11.2

2.9

9.2

11

21

1.9

44

5.3

19

33

0.2

1.1

3.7

3.1

6.59.7 15733

3.95.1

3619

3021

2071

1672

1510

1256

1087

1035

1001

746

605

463

382

382

300

89

289

100

17

4680

14 1480

0.4

30

0.1

7.9

4084 1202

3358 11852

7.14.1 8636

2734 3206

1119 5617

1018 4852

7.13.4 1013

312

4562

4429

3990

2071

1253

858

475

420

265

0.5

4.0

0.4

3.9

6.0

2.1

4.0

1818.0

69614

1800.5

118105

773440

5832

82625

85921

636513

1719

P < 0.0001

B

C

A

genotype

Alle

le fr

eque

ncy

(%)

(NG

S)

Allele frequency (%) (ddPCR)

Figure 2.Plasma NGS compared with knowntumor genotype across a range ofgenomic equivalents (GE) in thesequencing library. The mutant allelefrequency is provided when detectedby the plasma genotyping assay(green circle) but not if undetected(red circle). In patients with commonEGFR and KRAS mutations (A),plasma NGS has similar sensitivity toplasma ddPCR. Quantification ofallelic frequency with plasma NGS andplasma ddPCR are closely correlated(B). In patients with rare genotypes(C), plasma NGS is able to detect awide range of genomic alterations. Inboth groups of patients, the rate ofdetection by plasma NGS increases asthe number of GE increases (A, C).

Paweletz et al.





tumor genotyping (Fig. 2C); these included 19 new cases plus onepreviously studied above, harboring both a KRASmutation and aPIK3CA mutation. Plasma NGS was able to blindly detect six ofeight caseswith rearrangements and four caseswith rare targetableHER2 and EGFR mutations. Sensitivity in this cohort was 75%(15/20), similar to the prior analysis (Fig. 2C). For each of the sixrearrangements detected, the exact breakpoints and fusion part-ners could bemapped to the genome (Fig. 3A and SupplementaryFig. S4).

Specificity was studied in these 48 patients, each with tumorgenotyping positive for an oncogenic driver mutation in EGFR,KRAS, ALK, ROS1, RET, BRAF, or HER2. These oncogenes areestablished as nonoverlapping on tumor genotyping and aretherefore ideal gold standards for assessment of false positives(5, 7). Specificity of the plasmaNGS platformwas 100% for theseseven driver genotypes, with a false positive rate of 0% (95%confidence interval, 0%–10%).

Detection of resistance mutations was explored in 15 of the 48patients who had plasma collected after development of acquiredresistance to a tyrosine kinase inhibitor (Table 1). Of 12 patientswith acquired resistance to erlotinib or afatinib, plasma NGSdetected T790M in eight. One of the 12 patients had beenrefractory to erlotinib and afatinib despite harboring an EGFRexon 19 deletion, and tumor NGS had identified high METamplification. Blinded to the tumor genotype, plasma NGSsimilarly detected MET amplification, as evidenced by a signifi-cant increase inMET copies compared with control (Fig. 3B). Noresistancemutations were identified in the remaining three EGFR-mutant cases. Studying two patients who had developed acquiredresistance to crizotinib, plasma NGS identified two point muta-tions in the ALK kinase domain in one case with an ALK rear-rangement, and no ROS1 resistance mutations in the other casewith a ROS1 rearrangement. One patient was studied who hadpreviously developed T790M-positive resistance to erlotinib, and

43,611,340 bp

55,249,070 bp 55,249,080 bp 55,249,090 bp 55,242,460 bp 55,242,470 bp 55,242,480 bp 55,242,490 bp

43,611,360 bp 43,611,380 bp 43,611,400 bp 43,611,420 bp 43,611,440 bp 43,611,460 bp

MET probes

Control probes

A

C D

B

Figure 3.Bias-corrected NGS of cfDNA identifies complex genomic alterations. A, sequencing of the intronic region of RET detects reads extending into KIF5B(inset), predicting a fusion of these two genes. B, in cfDNA from a case with known MET amplification (case 105), MET copy number is significantly increasedcompared with control probes (P < 0.001), which is not seen in cases without MET amplification. C, two mutations encoding for EGFR C797S are detectedin ciswith EGFR T790M after resistance to AZD9291. D, in a case with acquired T790M despite no apparent EGFR sensitizing mutation, plasma NGS detects a noveldouble-deletion in exon 19 of EGFR which would have been missed with many PCR-based genotyping assays.






subsequently developed acquired resistance to the investigationalEGFR kinase inhibitor AZD9291 (20); plasmaNGS identified twodifferent DNA mutations encoding for EGFR C797S (Fig. 3C), amutation recently described as a common mediator of acquiredresistance to AZD9291 (21). Fourteen of 15 cases had resistancebiopsies available for genotyping. Tumor and plasma results wereconcordant in 12 (86%), whereas in two cases plasma NGS didnot detect an acquired T790M mutation detected in tumor.Altogether, sensitivity for the 62 known driver and resistancemutations from the 48 cases was 77% (48/62).

Finally, the plasma NGS assay was explored in two advancedNSCLC patients with high clinical suspicion of possessing a tar-getable genomic alteration which had been missed on prior tissuegenotyping. The initial patient was a 66-year-old female never-smoker who had responded durably to empiric treatment witherlotinib and developed resistance; EGFR genotyping of a rebiopsyusing a commercial PCR assay had identified T790M without anysensitizing mutation evident. Plasma ddPCR was first performedand detected 1508 copies/mL of an exon 19 deletion. PlasmaNGSthen confirmed the presence of a novel double-deletion in exon 19of EGFRwhichwould have beenmissed bymany commercial PCRassays because these often detect only common exon 19 deletionvariants (Fig. 3D). The second patient was a 28-year-old femalenever-smoker who hadprogressed onmultiple lines of therapy, forwhom previous tumor genotyping (including NGS) had revealedno targetable alterations despite four biopsies. Plasma NGSrevealed high-level MET amplification which was subsequentlyconfirmed by fluorescent in situ immunohistochemistry (MET:CEP7 ratio >5), and led to the initiation of crizotinib. For each

case, turnaround time from blood draw to result reporting in thisinitial feasibility study was six business days.

ConclusionIn this blinded study, a targeted NGS of cfDNA from advanced

NSCLC patients was able to accurately detect a broad range oftargetable genomic alterations in NSCLC, including point muta-tions, insertions/deletions, and rearrangements, with no falsepositives. This report is the first, to our knowledge, to describethe accurate detection of ALK, ROS1, and RET rearrangementsusing a single plasma assay without prior knowledge of fusionpartners. This represents a dramatic advance over PCR-basedplasma genotyping assays which are limited to the detection ofhotspotmutations in coding regions (10). This assaywas also ableto detect both canonical and novel resistance mechanisms,including MET amplification and EGFR C797S (22, 23).

Importantly, this approach uses widely available equipmentsuch as standard EDTA-containing vacutainers and a desktopsequencing platform: any accredited molecular pathology labwith a MiSeq could, with the right technical expertise and bioin-formatics support, implement this assay to guide the care ofadvanced NSCLC patients.

Although others have also studied plasma NGS of lung cancer,this is thefirst to describe comprehensive andblinded detectionofa broad range of alterations with one clinic-ready assay. Detectionof hotspot mutations using plasma NGS was described by Cour-aud and using the IonTorrent platform, achieving a sensitivity of58% and 87% specificity but without the ability to detect rear-rangements or amplifications (11). Newman and others recently

Table 1. Detection of resistance mechanisms using plasma NGS compared with tissue genotyping of rebiopsy specimens in patients with acquired resistance tokinase inhibitors

Sample Baseline genotype Therapy received Tissue genotype at resistance Plasma NGS at resistance Allele frequency

11 EGFR exon 19 del Erlotinib EGFR exon 19 del EGFR exon 19 del 1.9%EGFR T790M EGFR T790M 0.5%




48 EGFR L858R Afatinib EGFR L858R EGFR L858R 0.8%EGFR T790M

74 EGFR exon 19 del Erlotinib EGFR exon 19 delEGFR T790M

EGFR exon 19 del 7.9%



22 EGFR G719A Erlotinib EGFR G719A EGFR G719A 2.1%105 EGFR exon 19 del Afatinib EGFR exon 19 del EGFR exon 19 del 6.5%

MET amp MET amp NC522 EGFR exon 19 del AZD9291 EGFR exon del 19 EGFR exon del 19 13.9%

EGFR T790M EGFR T790M EGFR T790M 8.5%EGFR C797S EGFR C797S 7.8%

18 ALK rearrangement Crizotinib (No tissue available) EML4-ALK 4.0%ALK G1128A 2.0%ALK G1156Y 1.6%

127 ROS1 rearrangement Crizotinib EZR-ROS1 EZR-ROS1 0.5%120 EGFR exon 19 del Erlotinib EGFR exon 19 del EGFR exon 19 del 39.7%

EGFR T790M EGFR T790M 8.4%232 EGFR L858R Erlotinib EGFR L858R EGFR L858R 7.1%

EGFR T790M EGFR T790M 6.2%

Paweletz et al.





demonstrated more comprehensive detection of lung cancermutations in cfDNA and tumor tissue by hybrid capture usingbiotinylated oligonucleotide probes on a HiSeq (18); however,this study noted an inefficient capture of fusion rearrangements.Finally, targeted sequencing of cfDNA using PCR amplicons hasbeen successful for detection of single nucleotide variant inmultiple types of cancer (15); however, PCR amplicons are nottrivial to multiplex and are inherently blind to gene rearrange-ments. The NGS platform described here overcomes the weak-nesses of each of these prior approaches, allowing multiplexeddetection of a broad range of alterations with no false positivesusing an efficient platform. Furthermore, turnaround time fromtime of blood draw can be as short as 6 days.

Intuitively, we found that the sensitivity of plasma NGS isimproved in specimens with a higher quantity of cfDNA, althoughin some instances targetable genotypes could be detected in speci-menswitha relatively lownumberofGEsequenced indicatinghighDNA shed from the tumor. In this pilot study, sensitivity of plasmaNGS was 77%, comparable to most plasma genotyping assayswhichhave reported sensitivity in the rangeof 70% (24). SensitivityimproveswithhigherDNAconcentration,highlightinghowimpor-tant it will be to understand why the amount of total cfDNA inplasma varies across such awide range, andwhether this variance isdue to fluctuations in tumor biology or in methods of extraction.For example, we and others have described previously that plasmagenotyping assays are more sensitive in lung cancer patients withextra-thoracic metastases (13, 15, 17, 18). Even with a moderatelyhigh sensitivity, the lack of false positives with this assay results in a100% positive predictive value, meaning that this plasma NGSassay could be used as a screening step before a biopsy sample istaken for genomic analysis: if positive, the results are reliable andcan potentially obviate the need for a biopsy, and if negative, abiopsy for further testing may still have value.

Our detection of various resistancemechanisms in patients withacquired resistance to TKIs, including the detection of two differentEGFR C797S clones in one case, highlights the potential power ofplasma NGS for understanding the heterogeneity of the resistantstate. In addition, the C797S alleles identified with plasma NGSwere clearly in cis with T790M (Fig. 3), a detail that would bedifficult to decipher with a PCR assay and may have treatmentimplications (22). It is increasingly appreciated that resistanttumors are made up of clones with diverse biologies that mayrespond differently to therapy (25, 26). We recently showed thatthree molecular subtypes of acquired resistance to the investiga-tional EGFR inhibitor AZD9291 are apparent by use of serialplasma ddPCR: although all patients started with T790M plus asensitizingmutation, some lose T790Mat resistancewhereas somegainC797S (21).However, this analysis requiredfive separate PCRassays: T790M, 19 deletion, L858R, and assays for two C797Svariants. Alternatively, one plasma NGS assay can detect these fivealterations plus detecting any novel resistance mechanisms thatemerge. Our data further suggest that quantification of allelicfraction is similar using plasmaNGS and ddPCR, suggesting eithercould be used to serially monitor plasma genotype concentration.

The findings described in this report are achieved by applying anovel bias-corrected capture technology that builds upon stan-dard sequencing approaches to maximize the efficiency of on-target versus off-target and redundant sequencing reads. By reduc-ing PCR artifacts, this assay can accurately detectmutation presentin as few as 0.1% of sequencing reads; in contrast, some tumorNGS platforms are unable to accurately call a mutation unless

detected in >10% of sequencing reads. Bias-corrected targetedNGS is an approach that can be applied to any sequencingplatform to improve on-target coverage and reduce noise. Here,we apply bias-corrected targeted NGS to desktop sequencing on aMiSeq platform in order to develop a plasma assay with thepotential to be rapid and clinic-ready, in contrast to more timeintensive assays utilizing HiSeq platforms (18). Yet this NGSapproach could also be applied to clinical sequencing panels ordiscovery efforts to improve sequencing coverage and reduceartifact, particularly when studying clinical specimens from smallbiopsies with scant DNA.

In conclusion, we have developed and successfully piloted aplasma NGS assay that, using a novel capture and analysistechnique, can detect targetable mutations and rearrangementsin plasma from advanced NSCLC patients. This is the first plasmaNGS assay to demonstrate blinded,multiplexed detection of sucha broad range of actionable alterations with no false positiveresults. By using a widely available desktop sequencing platformand standard vacutainers with the potential for a rapid turn-around time, this assay has the potential for broad uptake andapplication. Through reducing the barriers between NSCLCpatients and genotyping, we hope that plasma NGS will be ableto facilitate delivery of targeted therapies and improve outcomesfor patients with advanced NSCLC.

Disclosure of Potential Conflicts of InterestA.G. Sacher reports receiving travel reimbursement from AstraZeneca and

Roche. C.K. Raymond and L.P. Lim have ownership interest in ResolutionBioscience. P.A. J€anne reports receiving a commercial research grant fromAstraZeneca and Astellas; has ownership interest in Gatekeeper Pharmaceu-ticals; and is a consultant/advisory board member for AstraZeneca, ChugaiPharmaceutical, Merrimack Pharmaceuticals, Pfizer, and Roche. G.R. Oxnardis a consultant/advisory board member for Ariad, AstraZeneca, Clovis, andSysmex. No potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: C.P. Paweletz, C.K. Raymond, G.R. OxnardDevelopment of methodology: A.G. Sacher, C.K. Raymond, A. O'Connell,L.P. Lim, G.R. OxnardAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A.G. Sacher, C.K. Raymond, R.S. Alden, S.L. Mach,Y. Kuang, P.A. J€anne, G.R. OxnardAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): C.P. Paweletz, A.G. Sacher, C.K. Raymond,R.S. Alden, A. O'Connell, Y. Kuang, L. Gandhi, L.P. Lim, G.R. OxnardWriting, review, and/or revision of themanuscript:C.P. Paweletz, A.G. Sacher,C.K. Raymond, R.S. Alden, A. O'Connell, S.L. Mach, Y. Kuang, L. Gandhi,P. Kirschmeier, J.M. English, L.P. Lim, P.A. J€anne, G.R. OxnardAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): A. O'Connell, J.M. EnglishStudy supervision: P. Kirschmeier, G.R. Oxnard

Grant SupportThis work was supported in part by the Department of Defense, Conquer

Cancer Foundation of ASCO, Phi Beta Psi Sorority, Stading-Younger CancerResearch Foundation, Expect Miracles Foundation, Harold and Gail KirsteinLung Cancer Research Fund, and US National Institutes of Health grantR01CA135257 (P.A. J€anne).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 15, 2015; revised September 21, 2015; accepted September 29,2015; published OnlineFirst October 12, 2015.






References1. Li T, KungHJ,Mack PC, GandaraDR.Genotyping and genomic profiling of

non-small-cell lung cancer: implications for current and future therapies. JClin Oncol 2013;31:1039–49.

2. Paez JG, J€anne PA, Lee JC, Tracy S, Greulich H, Gabriel S, et al. EGFRmutations in lung cancer: correlation with clinical response to gefitinibtherapy. Science 2004;304:1497–500.

3. Kwak EL, Bang YJ, Camidge DR, Shaw AT, Solomon B, Maki RG, et al.Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. NEngl J Med 2010;363:1693–703.

4. Hauschild A, Grob JJ, Demidov LV, Jouary T, Gutzmer R, MillwardM, et al.Dabrafenib in BRAF-mutated metastatic melanoma: a multicentre, open-label, phase 3 randomised controlled trial. Lancet 2012;380:358–65.

5. Cardarella S,Ortiz TM, Joshi VA, ButaneyM, JackmanDM,KwiatkowskiDJ,et al. The introduction of systematic genomic testing for patients with non-small-cell lung cancer. J Thorac Oncol 2012;7:1767–74.

6. PlanchardD, KimT,Mazi�eres J,Quoix E, RielyG, Barlesi F, et al. Dabrafenibin patients with BRAF V600E-mutant advanced non-small cell lung cancer(NSCLC): a multicenter, open-label, phase II trial (BRF113928). AnnOncol 2014;25 (Supplement 5): v1–v41.

7. Sholl LM, Aisner DL, Varella-Garcia M, Berry LD, Dias-Santagata D,Wistuba II, et al. Multi-institutional oncogenic driver mutation analysisin lung adenocarcinoma: the lung cancermutation consortium experience.J Thorac Oncol 2015.

8. Wagle N, Berger MF, Davis MJ, Blumenstiel B, Defelice M, Pochanard P,et al. High-throughput detection of actionable genomic alterations inclinical tumor samples by targeted, massively parallel sequencing. CancerDiscov 2012;2:82–93.

9. Hagemann IS, Devarakonda S, Lockwood CM, Spencer DH, Guebert K,Bredemeyer AJ, et al. Clinical next-generation sequencing in patients withnon-small cell lung cancer. Cancer 2015;121:631–9.

10. Luke JJ,OxnardGR, Paweletz CP, CamidgeDR,Heymach JV, Solit DB, et al.Realizing the potential of plasma genotyping in an age of genotype-directed therapies. J Natl Cancer Inst 2014;106(8).

11. Couraud S, Vaca-Paniagua F, Villar S, Oliver J, Schuster T, Blanch�e H, et al.Noninvasive diagnosis of actionable mutations by deep sequencing ofcirculating free DNA in lung cancer from never-smokers: a proof-of-concept study from BioCAST/IFCT-1002. Clin Cancer Res 2014;20:4613–24.

12. Goto K, Ichinose Y, Ohe Y, Yamamoto N, Negoro S, Nishio K, et al.Epidermal growth factor receptor mutation status in circulating free DNAin serum: from IPASS, a phase III studyof gefitinibor carboplatin/paclitaxelin non-small cell lung cancer. J Thorac Oncol 2012;7:115–21.

13. Lee YJ, Yoon KA, Han JY, KimHT, Yun T, Lee GK, et al. Circulating cell-freeDNA in plasma of never smokers with advanced lung adenocarcinomareceiving gefitinib or standard chemotherapy as first-line therapy. ClinCancer Res 2011;17:5179–87.

14. Douillard JY, Ostoros G, Cobo M, Ciuleanu T, Cole R, McWalter G, et al.Gefitinib treatment in EGFR mutated caucasian NSCLC: circulating-freetumorDNAas a surrogate for determination of EGFR status. J ThoracOncol2014;9:1345–53.

15. Forshew T, Murtaza M, Parkinson C, Gale D, Tsui DW, Kaper F, et al.Noninvasive identification and monitoring of cancer mutations by tar-geted deep sequencing of plasma DNA. Sci Transl Med 2012;4:136ra68.

16. Oxnard GR, Paweletz CP, Kuang Y, Mach SL, O'Connell A, Messineo MM,et al. Noninvasive detection of response and resistance in EGFR-mutantlung cancer using quantitative next-generation genotyping of cell-freeplasma DNA. Clin Cancer Res 2014;20:1698–705.

17. Sacher A, Oxnard G, Mach S, Messineo M, Jackman D, J€anne P, et al.Prediction of lung cancer genotype noninvasively using droplet digital PCR(ddPCR) analysis of cell-free plasma DNA (cfDNA). American Society ofClinical Oncology Annual Meeting; Chicago, United States 2014.

18. Newman AM, Bratman SV, To J, Wynne JF, Eclov NC, Modlin LA, et al. Anultrasensitive method for quantitating circulating tumor DNA with broadpatient coverage. Nat Med 2014;20:548–54.

19. Oxnard G, Lydon C, Lindeman N, Shivdasani P, Chin G, Kuo F, et al.Implementation of clinical next-generation sequencing (NGS) of non-small cell lung cancer (NSCLC) to identify EGFR amplification as apotentially targetable oncogenic alteration. American Society of ClinicalOncology Annual Meeting; Chicago, USA 2014.

20. J€anne PA, Yang JC, Kim DW, Planchard D, Ohe Y, Ramalingam SS, et al.AZD9291 in EGFR inhibitor-resistant non-small-cell lung cancer. N Engl JMed 2015;372:1689–99.

21. Thress KS, Paweletz CP, Felip E, Cho BC, Stetson D, Dougherty B, et al.Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harboring EGFR T790M. Nat Med 2015.

22. Niederst MJ, Hu H, Mulvey HE, Lockerman EL, Garcia AR, Piotrowska Z,et al. The allelic context of the C797S mutation acquired upon treatmentwith third-generation EGFR inhibitors impacts sensitivity to subsequenttreatment strategies. Clin Cancer Res 2015;21:3924–33.

23. Yu HA, Tian SK, Drilon AE, Borsu L, Riely GJ, Arcila ME, et al. Acquiredresistance of EGFR-mutant lung cancer to a T790M-Specific EGFR inhib-itor: emergence of a third mutation (C797S) in the EGFR tyrosine kinasedomain. JAMA Oncol 2015.

24. Luo J, Shen L, Zheng D. Diagnostic value of circulating free DNA for thedetection of EGFR mutation status in NSCLC: a systematic review andmeta-analysis. Sci Rep 2014;4:6269.

25. de Bruin EC, McGranahan N, Mitter R, Salm M, Wedge DC, Yates L, et al.Spatial and temporal diversity in genomic instability processes defines lungcancer evolution. Science 2014;346:251–6.

26. Zhang J, Fujimoto J,WedgeDC, Song X, Seth S, ChowCW, et al. Intratumorheterogeneity in localized lung adenocarcinomas delineated by multi-region sequencing. Science 2014;346:256–9.


Paweletz et al.




2016;22:915-922. Published OnlineFirst October 12, 2015.Clin Cancer Res Cloud P. Paweletz, Adrian G. Sacher, Chris K. Raymond, et al. from Advanced Lung Cancer PatientsMultiplexed Detection of Actionable Alterations in Cell-Free DNA Bias-Corrected Targeted Next-Generation Sequencing for Rapid,

Updated version

10.1158/1078-0432.CCR-15-1627-Tdoi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2015/10/10/1078-0432.CCR-15-1627-T.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/22/4/915.full#ref-list-1

This article cites 20 articles, 10 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/22/4/915.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/22/4/915To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-15-1627-T

http://clincancerres.aacrjournals.org/content/suppl/2015/10/10/1078-0432.CCR-15-1627-T.DC1

http://clincancerres.aacrjournals.org/content/22/4/915.full#ref-list-1

http://clincancerres.aacrjournals.org/content/22/4/915.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/22/4/915


Bias-Corrected Targeted Next-Generation Sequencing...

Documents

Transcript of Bias-Corrected Targeted Next-Generation Sequencing...