detection of mutant free circulating tumor dna in the ... · monitoring response in gist. however,...

Predictive Biomarkers and Personalized Medicine

Detection of Mutant Free Circulating Tumor DNA in thePlasma of Patients with Gastrointestinal Stromal TumorHarboring Activating Mutations of CKIT or PDGFRA

Jacqueline Maier1, Thoralf Lange1, Irina Kerle3, Katja Specht5, Melanie Bruegel4, Claudia Wickenhauser2,Philipp Jost3, Dietger Niederwieser1, Christian Peschel3, Justus Duyster6, and Nikolas von Bubnoff6

AbstractPurpose: In gastrointestinal stromal tumor (GIST), there is no biomarker available that indicates success

or failure of therapy. We hypothesized that tumor-specific v-kit Hardy-Zuckerman 4 feline sarcoma viral

oncogenehomolog (CKIT)- or platelet-derived growth factor receptor-a (PDGFRA)–mutantDNA fragments

can be detected and quantified in plasma samples of patients with GIST.

Experimental Design: We prospectively collected 291 plasma samples from 38 subjects with GIST

harboring activating mutations of CKIT or PDGFRA detected in tumor tissue, irrespective of current disease

status or treatment. We used allele-specific ligation PCR to detect mutant free circulating DNA (fcDNA).

Results:Wewere able to detect fcDNA harboring the tumor mutation in 15 of 38 patients. Patients with

active disease displayed significantly higher amounts ofmutant fcDNA compared with patients in complete

remission (CR). The amount ofmutant fcDNAcorrelatedwithdisease course.Weobserved repeatedpositive

test results or an increase of mutant fcDNA in five patients with progressive disease or relapse. A decline of

tumor fcDNA or conversion from positive to negative was seen in five patients responding to treatment. A

negative to positive conversion was seen in two patients with relapse and one patient with progression. In

two cases, we aimed to identify additional mutations and found four additional exchanges, including

mutations not known from sequentially conducted tumor biopsies.

Conclusions: Our results indicate that fcDNA harboring tumor-specific mutations in the plasma of

patients withGIST can be used as tumor-specific biomarker. The detection of resistancemutations in plasma

samples might allow earlier treatment changes and obviates the need for repeated tumor biopsies. Clin

Cancer Res; 19(17); 4854–67. �2013 AACR.

IntroductionIn gastrointestinal stromal tumor (GIST), themain objec-

tive of treatment is complete resection. Tyrosine kinase

inhibitors (TKI) such as imatinib or sunitinib are used forneoadjuvant, adjuvant, or palliative treatment. Responsesto imatinib can be achieved. However, treatment is notcurative unless complete resection is possible. In advancedGIST, a partial remission can be attained in 50% of patientstreated with imatinib, and there are single cases of completeresponses. However, most patients experience disease pro-gression receiving imatinib (1–7). Patientswith progressionreceiving imatinib only transiently respond to sunitinib (8).

The response to therapy is evaluated by diagnostic imag-ing, which displays limited sensitivity and specificity. Com-puted tomography (CT) constitutes the gold standard ofimaging in GIST (9, 10). However, sensitivity of CT in GISTis insufficient. A randomized trial investigating the inter-ruption of imatinib in patients with advanced GIST wasterminated prematurely due to a high rate of progressionevents in the interruption arm even in patients displayingCT-morphologic complete remission (CR) at the time ofrandomization (11). Currently, 2[18F]fluoro-2-deoxy-D-glucose positron emission-tomography (FDG-PET) repre-sents the most sensitive imaging technique for staging andmonitoring response in GIST. However, the sensitivity andspecificity of FDG-PET in GIST is limited (12–14). In a

Authors' Affiliations: 1Center for Internal Medicine, Department of Hema-tology/Oncology and Hemostaseology, and 2Institute of Pathology, Uni-versity of Leipzig, Leipzig; 3Medizinische Klinik und Poliklinik; 4Institut f€urDiagnostische und Interventionelle Radiologie, Klinikum Rechts der Isar,Technische Universit€at M€unchen; 5Institut f€ur Allgemeine Pathologie undPathologische Anatomie der Technischen Universit€at M€unchen, M€unchen;and 6Medizinische Universit€atsklinik, Klinik f€ur Innere Medizin 1, Freiburg,Germany

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

J. Maier and T. Lange contributed equally to this work.

All authors had access to the study data and have reviewed and approvedthe final manuscript.

Corresponding Author: Nikolas von Bubnoff, Klinik f€ur Innere Medizin 1,Universit€atsklinik Freiburg, Hugstetter Str. 55, D-79106 Freiburg, Germany.Phone: 49-761-270-33210; Fax: 49-761-270-73570; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-13-0765

�2013 American Association for Cancer Research.

ClinicalCancer

Research

Clin Cancer Res; 19(17) September 1, 20134854

on August 15, 2020. © 2013 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst July 5, 2013; DOI: 10.1158/1078-0432.CCR-13-0765

http://clincancerres.aacrjournals.org/

recent study, the sensitivity and specificity of FDG-PET/CTfor detecting intraoperatively occult metastases in patientswith GIST and R0 resection was 25% and 88%, respectively(15). Even in cases with FDG-PET/CT complete response,lesions contained viable tumor (16).Currently, there is no biomarker available for detecting

success or failure of therapy in GIST. Thus, currently thetrigger to change treatment is clinical or CT-morphologicprogression.Activating mutations of the CKIT or the PDGFRA can be

found in at least 85% of all the cases in GIST and constitutethe transforming event in the pathogenesis of GIST tumors(17–21).Free circulating DNA (fcDNA) can be detected in normal

individuals but, however, is increased in patients withtumor (22–25). Proposed models for tumor-DNA releaseinto the blood include DNA release by tumor cells thatundergo apoptosis or necrosis, and extravasation of tumorcells into the blood where cells undergo lysis and releasetheir DNA content into the blood (26). It has been shownthat levels of long interspersed nucleotide elements 1(LINE1) DNA in plasma samples correlate with tumorprogression in different cancer entities (27–30) but, how-ever, LINE1 sequences are not tumor specific.Recently, it was shown that EGFRmutations can be recov-

ered in fcDNA in serum samples of patients with non–smallcell lung cancer (31), and detection of circulating mutantAPC, p53, or KRAS DNA was shown to correlate with pro-gression-free survival in patients with colorectal cancer (32).A sensitive, noninvasive biomarker for detection and

quantitation of disease activity would be a highly valuabletool for the management of GIST. Genomic tumor DNAfrom the resection specimen can be used to identify the

patient-specific CKIT or PDGFRAmutation present in GISTtumor cells. In this trial, we aimed to evaluate whethertumor DNA carrying mutations for CKIT and PDGFRA canbe detected and quantified in the plasma of patients withGIST, and whether detection can be correlated with theclinical course of disease and results of diagnostic imagingstudies. Using allele-specific ligation PCR (L-PCR), we werefor the first time able to detect fragments of genomic DNAcontaining the tumor-specific mutation in plasma samplesderived from patients with GIST, and our data suggest thatnoninvasive detection of tumor-derived fcDNA measuredin plasma samples correlates with disease activity.

Materials and MethodsTrial design

This study is an open-label, nonrandomized, noninter-ventional, explorative phase IIIb trial that aims to detecttumor-specific mutations of CKIT or PDGFRA in fc plasmaDNA in patients with GIST harboring activating mutationsof CKIT or PDGFRA detected in tumor tissue irrespective ofcurrent disease status and current or planned treatment.Patients were treated at the Klinikum Rechts der Isar,Technische Universit€at M€unchen (M€unchen, Germany).We prospectively collected 291 plasma samples from 38subjects betweenOctober 2007 andMay 2012. All patientsgave written informed consent. After a monocentric pilotphase, the trial was continued as multicenter study. Thestudy was approved by the Institutional Review Board ofthe TechnischeUniversit€atM€unchen. This trial is registeredunder Eudra-CTNo. 2011-002544-27 andClinicalTrials.govNCT01462994. The present report focuses on the mono-centric pilot phase of the trial.

PatientsThe 38 cases included 18 patients with active disease,

which was defined as having at least one GIST lesion thatcould be measured by imaging at any time after inclusion.Of these, 9 displayed active disease before TKI treatmentwasinstituted (first diagnosis, n¼ 2; relapse, n¼ 5; after surgicaldebulking, n¼ 2) and 9 patients had relapse or progressionreceiving TKI treatment, including 3 patients with responseto TKI treatment where no samples were available at pro-gression before TKI. Twenty patients were in CR but withhigh (n ¼ 14), moderate (n ¼ 3), or low (n ¼ 1) risk ofrelapse according to Miettinen criteria (33), or were in CRand Miettinen criteria were not applicable because mitoticindex was not available (n ¼ 1) or treated local peritonealmetastasis (n ¼ 1; see Table 1). Twenty-nine patientsreceived medical treatment with imatinib, sunitinib, nilo-tinib, dasatinib, sorafenib, everolimus, or doxorubicin.Nine patients did not receive medical treatment duringparticipation in the trial. Tumor response was measuredradiographically with CT, MRI, or FDG-PET/CT.

Mutation analysis of DNA from tumor tissueA tumor-specific mutation site of each subject was det-

ermined by Sanger sequencing of genomic DNA extractedfromtumor tissue. Formalin-fixedparaffin-embedded (FFPE)

Translational RelevanceIn gastrointestinal stromal tumor (GIST), response to

treatment is assessed by diagnostic imaging techniques,which is associated with limited sensitivity and specific-ity. There is no blood biomarker available to detectsuccess or failure of therapy. In GIST, activating v-kitHardy-Zuckerman 4 feline sarcoma viral oncogenehomolog (CKIT) or platelet-derived growth factor recep-tor-a (PDGFRA) mutations can be detected by DNAsequencing in tumor specimens in 90% of the cases.Mutated genomic DNA fragments are highly specific forthe tumor. We here show that tumor-specific mutationsin CKIT or PDGFRA can be detected and quantified infree circulatingDNA (fcDNA) in plasma samples derivedfrom patients with GIST by quantitative ligation PCR,and that fcDNA levels correlatewith the clinical course ofthe GIST lesions as measured by diagnostic imaging.Thus, tumor-specific fcDNA in plasma can be used ashighly specific biomarker in patients with GIST, andmight be used to predict response to treatment andrelapse, allowing earlier treatment changes.

Mutant Free Circulating Tumor DNA in GIST

www.aacrjournals.org Clin Cancer Res; 19(17) September 1, 2013 4855




tumor tissue was obtained from each subject. Genomic DNAfrom FFPE sections was extracted with the QIAamp DNAFFPE Tissue Kit (Qiagen) according to the manufacturer’sinstructions. Sequence analysis of CKIT exons 9, 11, 13, 14,17, and 18 and PDGFRA exon18was conducted on a routinebasis. Furthermore, in all cases where tumor specimens wereavailable (30 of 38 cases; 79%), we repeated Sanger sequenc-ing for confirmation of the initial mutation status.

Isolation of tumor DNA from plasmaPeripheral blood samples were collected in EDTA tubes

(Sarstedt). At the same day, tubes were centrifuged at 800�g for 10 minutes. The supernatant was transferred to 10-mL

tubes and centrifuged at 1,600� g for 10minutes to removedebris. Plasma supernatants were transferred in 1.0 mLaliquots to 1.5 mL tubes and stored at �80�C. Total geno-mic DNA of plasma aliquots was extracted from 1 mLplasma using the QIAamp Circulating Nucleic Acid Kit(Qiagen) according to the manufacturer’s instructions.

Detection and quantification of mutant and LINE1-fcDNA from plasma

Analyses of plasma samples were conducted in a blindedfashion with respect to the clinical information of individ-ual patients. The sequences of the primers and probes usedfor each test are listed in the Supplementary Data. fcDNA

Table 1. Overview of test results in relation to clinical setting

Mutant fcDNA

PositiveLow-levelpositive Negative Total

Active disease 5 4 9a 18Postsurgery in CR: riskof relapse (Miettinen)

High 0 5 9 14Moderate 0 1 2 3Low 0 0 1 1N/A 0 0 2 2

Total 5 10 23 38

NOTE: Number of caseswith detection of mutant fcDNA in plasma samples according to the clinical setting. Positive denotes 0.1%ormoremutant allele towt ratio (except for theK642Eassay: 1%ormore due to high cross-reactivity). Low-level positive refers tomutant/wt allele ratios of <0.1% (except for the K642E assay: <1%due to high cross-reactivity). Active disease denotes at least one lesion thatcould be measured and either being pre-TKI treatment or with progression receiving TKI treatment. Patients postsurgery were in CRwithout evidence of residual disease, and the risk of relapse was calculated according to the Miettinen criteria based on tumorlocalization, tumor size, and mitotic index.Abbreviation: N/A, not applicable.aIncludes 3 cases with samples available at the time of response only (see Table 2).

Table 2. Overview of test results in relation to clinical setting for patients with active disease

Mutant fcDNA

Positive Low-level positive Negative Total

Pre TKI First diagnosis 1 1 0 2Relapse 1 0 4 5Postsurgery 0 2 0 2

With response receiving TKIa 0 0 3 3In progression receiving TKI 3 1 2 6Total 5 4 9 18

NOTE: Shown are the numbers of caseswith detection of mutant fcDNA in plasma samples according to the clinical setting in patientswith active disease. Active disease denotes at least one lesion that could be measured and either being pre-TKI treatment or withprogression receiving TKI treatment. Pre-TKI patients were either at first diagnosis, at relapse without treatment, or after debulkingbefore start of TKI treatment.With response receiving TKI refers to patients displaying response to TKI treatmentwhere no samplewasavailable at the time of relapse before TKI was started, and in progression receiving TKI denotes patients displaying progression whileon TKI treatment based on imaging studies. For explanation of positive and low-level positive refer to Table 1. Allocation to patientnumbers corresponding to Fig. 2 is given in Supplementary Table S2.aSamples available at the time of response only.

Maier et al.

Clin Cancer Res; 19(17) September 1, 2013 Clinical Cancer Research4856




purified from 1 mL plasma was used to quantify mutatedand LINE1GIST DNA. The principle used to quantify pointmutations in CKIT (GenBank acc. no. NM_000222.2) orPGFRA (GenBank acc. no. NM_006206.4) gene based onthe MLPA technique as previously described (34, 35). Toassess the load of total fcDNA, LINE1 was selected as targetfor quantitative real-time PCR.Mutant fcDNA. For detection of CKIT- or PDGFRA-

mutant fcDNA, CKIT- or PDGFRA-mutated sequences werefirst amplified in a standard PCR reaction ("accumulationPCR") from total genomic DNA extracted from plasma. Thesecond step involved the stringent (high temperature)hybridization to this PCR product of two probes, A1 andA2, which span the mutation region of interest. Each probeconsists of aCKIT or PDGFRA hybridizing sequence flankedby a PCR primer-binding site. One of these probes (A1)carries the mutation at the 30-end, whereas the 50-end of theother (A2) corresponds to the neighboring base. A high-temperature ligation reaction serves to join these twoprobesonly when there is no mismatch at the mutation site.Finally, the yield of ligated probe is determined by real-time PCR using primers complementary to the primer-binding sites at the nonligated ends of the two probes.Using universal primers, the parallel L-PCR amplificationof mutation-specific CKIT (PDGFRA) and total CKIT(PDGFRA) sequences in the same run ensures comparabil-ity. This allows normalization of the mutation-specificcycle threshold (Ct) values to those of totalCKIT (PDGFRA).The comparative DCt method was used [2E � (DCt) � 100]to calculate the relative percentage. Ct values of mutationand wild-type (wt) reaction were directly comparedto determine mutated allele in the wt background(minus cross-reactivity) in percentage. Positive samplesunderwent a second independent run. The mean of thetwo runs was used to calculate the mutated/wt ratio. Puri-fied DNA harboring the specific mutation was used as apositive control. Because some L-PCR runs do have abackground hybridization of the mutation-specific probesin a nonmutated sample, DNA from healthy subjects wasincluded in each assay as negative control. This cross-reac-tivity determines the sensitivity of the specific test run andthe nonspecific allele level were subtracted from a positiveallele level.LINE1 quantification. LINE1 are one of the most abun-

dant sequences in the human genome and therefore corre-late with the load of fcDNA. To assess the load of totalfcDNA, LINE1 was selected as target for quantitative real-time PCR. In detail, 180 bpwere amplified with the primers50-CAGAGAGCCAAATCATGAGTGAA-30 (forward) and 50-CTTCCTACCCATGAGCATGG-30 (reverse). The final reac-tion mixture for quantitative PCR consisted of 0.2 mmol/Lforward and reverse primers, 1� SYBR Green/polymerasemix (QuantiTect SYBR Green PCR Kit; Qiagen; #2041445)and as template 2 mL plasmaDNA in a total volume of 25 mLPCR reaction. The real-time PCR amplification was con-ducted with an initial activation of 95�C for 15 minutes,followed by 40 cycles at 94�C (15 seconds), 58�C (30seconds), and 72�C (35 seconds) using the 7300 Real-Time

PCR-system (Applied Biosystems). The LINE1 DNA frag-ments in each sample were calculated using a LINE1 plas-mid standard curve.

Statistical analysesStatistical analysis was conducted using Student unpaired

t test with 95% confidence intervals.

ResultsValidation of allele-specific L-PCR application

The application of the L-PCR showed in a total of 25assays an interassay coefficient of variation (SD/mean) of0.2 to 0.5, investigated in five independent runs con-ducted on different days and calculated from five log10dilutions between 104 and 108 CKIT exon 9 insertion 502-503 mutated copies in duplicates. The intraassay coeffi-cients of variations were 0.4 to 0.5 calculated from themeans of eight replicates of plasmid dilutions of CKITexon 9 insertion 502-503 between 104 and 107 copies.Sensitivity has been tested for six cloned key mutations(CKIT exon 9 insertion 502-503; CKIT exon 11 deletion557-558; CKIT exon 11 V559D; PDGFRA exon 18 D842V;CKIT exon 13 V654A; CKIT exon 13 K642E) and rangedfrom 0.01% to 0.1% mutated allele/wt dependent oncross-reactivity of hybridizing probes. This results in adynamic detection range of 4 to 5 log10. One exceptionwas K642E with a higher cross-reactivity of 1% and asubsequent detection range of 3 log10.

Detection of tumor-specific CKIT and PDGFRAmutations in plasma DNA

We developed and validated 25 different allele-specificL-PCR assays covering CKIT and PDGFRA mutations iden-tified in all patients (Supplementary Table S1). Three muta-tions appeared inmore than2patients [CKIT InsAY502-503(patient 5; 11; 25; 38),CKITDelW557-K558 (patient 1; 23;31; 36), and PDGFRAD842V (patient 12; 17; 18; 19; 20; 21;35]. Using our L-PCR assays, we examined 291 plasmasamples from 38 patients (Fig. 2). Eighteen of 38 patientshad active disease, which was defined as having at least onelesion that could be measured at any time after inclusionand being either before institution of TKI treatment (n¼ 9)or with relapse or progression receiving TKI treatment (n¼9; Tables 1 and 2). In 9 of these 18 patients, we detectedfcDNAharboring the tumor-specificmutation, withmutantto wt allele ratios of more than 0.1% (up to 15.6%) in5 cases and less than 0.1% in 4 cases (Table 1).We identifiedmutant fcDNA in 4 of 6 patients with progression or relapsereceiving TKI (patients 5, 7, 11, and 15; Table 2; Fig. 2;Supplementary Table S2). We did not detect mutantfcDNA in 3 patients who were already responding to TKItreatment at the time of the first plasma sample (patients 2,4, and 9). In contrast to patients with active disease, noneof the patients in CR displayed mutant to wt alleleratios above 0.1%, but we were still able to detect low levelsof mutant fcDNA (<0.1%) in 6 of 17 patients with highor intermediate risk of relapse according to Miettinen






(Table 1). When we analyzed "untreated and progression"versus "CR" patients with positive test results, the formercategory displayed significantly higher fcDNA ratios (Fig.1). Thus, in relation to the clinical setting, more patientswith active disease had positive test results compared withCR patients.

Quantification of tumor-derived plasma DNA andcorrelation with radiographic response to treatment

We next examined whether the amount of tumor fcDNAcorrelated with response. The clinical characteristics andtreatments for all 38 patients are shown in Table 3. In 5of 18patients with active disease, we found fcDNA harboring thetumor-specific mutation, with mutant to wt allele ratios ofmore than 0.1% (Table 1 and Fig. 3A–E). We observedrepeated positive test results in patients with progression(patients 5, 11, and15) or relapse (patients 7 and10; see Fig.2). In patients 7 and 10, tumor relapse was paralleled by anegative to positive conversion.

Mutated fcDNA test result of patient 5 shortly aftersurgery on day 0 (0.04%) was low-level positive with thetumor-specific CKIT exon 9 insertion AY502-503 mutation(Fig. 3A). After repeatedly negative samples, a negative topositive conversion of mutated fc CKIT DNA (day 641) to3.87% paralleled the progression of liver metastases (day662, red arrows).Mutant fcDNAdecreased to 0.03% shortlybefore high-frequency thermo-therapy (HFTT). Sunitinibwas stopped because of a stroke. Day 737 imaging showsdisintegration of two known liver lesions after HFTT (greenarrows) but extrahepatic progression (red arrow), andmutant fcDNA levels concomitantly increased to 1.29%and 7.72% at days 737 and 856, respectively. Day 926imaging showed massive progression, paralleled by adecrease of the knownmutant allele, whereas the regressionline of logarithmic LINE1 levels showed a steady increaseover time.

Patient 7 with a CKIT exon 11 deletion Y553-Q556mutation displayed one low-level mutated test result onday 57 (0.005%) 2months after surgery with small residual

Figure 2. Detection of tumor-specific mutations in plasma fcDNA and tumor response. Shown is the detection of tumor-specific mutated fcDNA from 38patients with GIST over time. Patients 1 to 18 with active disease, defined as having at least one lesion (A); patients 19 to 38 in CR (B). Symbols andfillings denote: diamond, active disease; triangle, active with disease progression; inverted triangle, active with response to treatment; circle, CR; square,fcDNA sample without imaging; no filling, negative for mutated tumor DNA; gray filling, low level (0.005%–0.099% mutated tumor DNA); black filling,positive (0.1%–100% mutated tumor DNA).

Figure 1. Correlation of mutant fcDNA and disease setting. Shown are thehighest fcDNA mutant to wt (mut/wt) ratios measured for individualpatients who were test-positive. Patients with untreated GIST orprogressive disease (untreated/progression) versus patients inCR. Therewere no significant differences in mean follow-up (P ¼ 0.8255) andsample frequency (P ¼ 0.8761) between "untreated/progression" (n¼ 6)and "CR" (n ¼ 6) groups.

Maier et al.





peritoneal nodules and converted to repeated negativesamples, whereas peritoneal nodules resolved completelyreceiving adjuvant imatinib (Fig. 3B). Although being inCR, the patient converted to 0.015%mutant fcDNA on theday 716, paralleled by an increase of LINE1DNA fragments.Three months later, this patient presented with peritonealmetastases, and further increased with mutant fcDNA to1.21%, which rapidly returned to 0% after 18 days ofsunitinib treatment. CT conducted 2 months later showedresponse (day 901; see Fig. 3B).Patient 10 displayed two tumor-specific mutations, a

CKIT exon 11 deletion V555-D572 insG known from theprimary tumor and an exon 17 D820Y point mutationidentified in surgically debulked tumor at month 26 (Fig.3C and Table 3). Upon hepatic relapse, mutant fcDNAwiththe CKIT exon 11 deletion/insertion, converted from 0% to0.19% together with an increase of LINE1 DNA fragments.After 2 months of imatinib, CKIT exon 11 mutant fcDNAconverted back to negative with a decrease of LINE1 DNAfragments, andCT scans confirmed tumor regression at days268 and 517. The additional D820Y mutation remaineddetectable until day 367, and thus might have contributedto hepatic relapse.Patient 8 was diagnosed with cardiac GIST and liver

metastasis with a CKIT exon 11 K550-K558 deletion. Weobserved a positive to negative conversion from 14.1% to0% mutant plasma fcDNA after only 14 days of treatmentwith imatinib (Fig. 3D). Of note, CT scans starting fromday94 showed anongoing response of liver lesions (red arrows)with persistently negative tumor fcDNAwith a follow-up of29 months and a corresponding steady decline of LINE1DNA fragments over time.In accordance with the rapid positive to negative conver-

sion seen in patients 7 and 8, we did not detect fcDNA inplasma samples from 3 additional patients who respondedto TKI treatment where pretreatment samples were notavailable (see Table 2, category "With response receivingTKI"; patients 2, 4, and 9 in Table 2).In 2 patients, sequentially conducted molecular analyses

of surgically debulked peritoneal tumor revealed secondaryresistance mutations in addition to the primary mutation(CKIT exon 13 V654A in patient 11, Fig. 3E; CKIT exon 17D820Y in patient 10, Fig. 3C). Both additional mutationsare known to confer imatinib resistance (36). Of note, wewere able to detect these additional mutations also insubsequent plasma samples. In patient 10, earlier plasmasamples were available and interestingly, we were able totrack back the exon 17 D820Y mutation in 3 of 7 earliersamples taken up to 2 years before this mutation was foundin the tumor specimen. Patient 11 had stable metastases inthe liver and rapidly progressingmesentericmetastases (Fig.3E, red arrows). In plasma samples of this patient, fcDNAcontaining the initial mutation (CKIT exon 11 V559D)decreased despite the rapid progression (Fig. 3E). In paral-lel, a secondary CKIT exon 13 V654A mutation identifiedfrom a mesenteric lesion after debulking on day 0 (Fig. 3E,black arrow) emerged in fcDNA during progression on day328 together with a steady increase of LINE1 DNA frag-

ments over time. We therefore suspected polyclonal diseaseand suspected additional secondary mutations. Indeed, wewere able to detect CKIT exon 9 insertion AY502-503 in 4samples (0.01%–0.05%), and exon 13 K642E in 2 samples(0.11%–0.24%; data not shown).

Next, we correlated our findings with the global load offcDNA as measured by LINE1 DNA fragments using quan-titative real-time PCR. Healthy volunteers were comparedwith GIST patients in CR or progression/relapse. Tenhealthy donors showed plasma LINE1 DNA fragments inthe range of 1� 105 to 4� 105 with amedian of 3� 105. Incomparison, 55 consecutive samples of patients with GISTin CR had a highermedian of 9� 105 (range, 1� 105 to 5�107) plasma LINE1 DNA fragments, and 41 patients withprogression or relapse displayed an even higher median of18� 105 (range, 1� 105 to 1� 108; Supplementary Fig. S1)but the difference was not statistically significant. However,the regression curve of serial LINE1 levels generally seemedto follow the clinical course, with a decrease in patientsresponding (patients 8 and 10; Fig. 3C and D) and increasein patients with relapse or progression (patients 5, 7, and11; Fig. 3A, B, and E).

DiscussionMethods for evaluating response in GIST are confined to

imaging. Because the specificity and sensitivity of imagingtechniques are limited (11, 15, 16), there is a need todevelop improved and specific methods for monitoringdisease activity. Most established biomarkers suffer fromlow sensitivity and specificity, and provide clinically valu-able information only in specific settings. In theory, tumor-derived fcDNA might offer advantages. In leukemia, DNA-or RNA-based assays are routinely used for monitoringresponse and minimal residual disease, and for example inchronic myelogenous leukemia (CML) determine progres-sion-free and overall survival (37, 38). Cancer cells releasenucleic acids into the circulation (26, 39). The amount oftumor-derived fcDNA might be determined by tumor size,vascularization, and cell turnover. FcDNA thus might wellreflect disease activity and could be used for monitoringcancer. It has been shown that mutant fcDNA might haveclinical use for monitoring in colorectal cancer (mutationsof APC, KRAS, PIK3CA, and TP53; ref. 32) or melanoma(BRAFV600E; ref. 40).GIST tumors in90%of the cases carrymutations in CKIT or PDGFRA. Thus, mutant DNA in anycompartment indicates viable tumor cells. The individualmutation is known from routine sequencing of the tumorspecimen. Allele-specific PCR allows specific amplificationand quantification of the mutated CKIT and PDGFRA DNAfragments resulting in high specificity and sensitivity.

In this study, we could show that mutant fcDNA specif-ically can be detected and quantified in the plasma ofpatients with GIST. Our data indicate that the amount ofmutant fcDNA correlates with response. We observedrepeatedly positive results and increasing mutant fcDNAin patients with progression, negative to positive conver-sions in patients with relapse, and positive to negativeconversions in patients responding to TKI treatment.






Tab

le3.

Clinical

charac

teris

ticsof

thepatients

Patient

#

Tim

efrom

first

diagno

sis,

mo

Initials

ite

Treatmen

tinitials

ite,

time

Relap

sesite,

time

Treatmen

trelapse

site,

time

Mutation

status

primary

Mutationstatus

follo

w-up(site

andtime;

type)

Ligation

probes

used

Clin

ical

setting

atinclus

ion

Cha

ngeof

clinical

settingto

11

Gas

tric

Res

ectio

n,mo�1

;Im

aad

juva

nt,

mo3–

16

Perito

neal,m

o28

Deb

ulking

perito

neal,m

o29

;Im

a,sinc

emo30

CKIT

delW55

7-K55

8Perito

neum

mo29

;CKIT

delW55

7-K55

8

CKIT

delW55

7-K55

8CR-M

RA,A

-R,C

R

21

Gas

tric

Res

ectio

n;mo�1

Live

r,mo34

Ima,

sinc

emo34

CKIT

insP

573(57

bp)

NE

CKIT

insP

573(57

bp)

CR-M

RA-R

36

Gas

tric,

perito

neal

Res

ectio

n,mo�6

Imaad

juva

nt,

mo�6

to31

Perito

neal,m

o42

Deb

ulking

perito

neal,m

o43

;Im

a,sinc

emo44

CKIT

V56

0DNE

CKIT

V56

0DCR-H

RA,C

R

496

Gas

tric

Res

ectio

n,mo�9

6Live

r,mo�1

Ima,

sinc

emo�1

CKIT

insR

586(43

bp)

NE

CKIT

insR

586(43

bp)

A-R

—

57

Duo

den

umRes

ectio

n,mo�7

;Im

aad

juva

nt,

mo�6

Live

r,mo8;

mes

enteriu

m,

mo32

Sun

,mo9–

24;

HFT

T,mo24

;Das

,mo32

–33

CKIT

insA

502-Y50

3Live

rmo9;

CKIT

insA

502-Y50

3CKIT

insA

502-

Y50

3CR-H

RA,A

-R,

A-P

D

646

Duc

tus

ompha

loen

teric

usRes

ectio

n,mo�4

6Perito

neal,m

o1

Ima,

sinc

emo1;

Deb

ulking

perito

neal,m

o10

NE

Perito

neal

mo1;

CKIT

delY55

3-K55

8CKIT

delY55

3-K55

8A

A,A

-R,

CR

71

Gas

tric,

perito

neal

Res

ectio

n,mo�1

;Im

aad

juva

nt,s

ince

mo1

Perito

neal,m

o27

Sun

,since

mo27

CKIT

delY55

3-Q55

6NE

CKIT

delY55

3-Q55

6CR

A,A

-R

80

Cardia,liver

Ima,

sinc

emo1

——

CKIT

delK55

0-K55

8NE

CKIT

delK55

0-K55

8A

A-R

956

Gas

tric

Res

ectio

n,mo�5

6Live

r,mo�2

Ima,

sinc

emo�2

;Hem

ihep

atec

tomy,

mo3

NE

Live

rmo3;

CKIT

delP55

1-E55

4CKIT

delP55

1-E55

4

A-R

CR

1044

Ileum

,om

entum

Res

ectio

nmo�4

4;Im

aad

juva

ntmo�4

4to

�8

Perito

neal,

liver,m

o8

Ima,

mo8–

26;

deb

ulking

perito

neal,m

o26

;mo40

CKIT

delV55

5-D57

2ins

GPerito

neal,liver;

CKIT

delV55

5-D57

2ins

Gþ

CKIT

D82

0Y(m

o26

)

CKIT

delV55

5-D57

2ins

G;C

KIT

D82

0Y

CR-H

RA,A

-R,

A-P

D

1163

Smallb

owel

Res

ectio

n,mo�6

3Live

r,mes

enteria

l,mo�1

7Im

a,mo�1

7to

1;deb

ulking

mes

enteria

l,mo1;

Sor,m

o1–

3;Sun

,mon

3–5;

Das

,mo

5–6;

Imaþ

Eve

,mo

6 –7;

Dox

,mo9

NE

Mes

enteria

lmo

�12;

CKIT

V55

9D.

Mes

enteria

lmo1;

CKIT

V55

9Dþ

V65

4A.M

esen

teria

lmo11

;CKIT

V55

9D

CKITV55

9D;C

KIT

V65

4A;C

KIT

insA

502-Y50

3;CKIT

delW55

7-K55

8;CKIT

K64

2E;C

KIT

V56

0D

A-P

D—

120

Gas

tric

Res

ectio

n,mo1

——

PDGFR

AD84

2VNA

PDGFR

AD84

2VA

CR-M

R13

13Ileum

Res

ectio

n,mo�1

3Live

r,mes

enteria

l,mo9

Ima,

sinc

emo9

CKIT

delV55

9-N56

6NE

CKIT

delV55

9-N56

6CR

A,A

-R

1445

Gas

tric

Res

ectio

n,mo�4

5Live

r,mo�3

2Im

a,mo�2

3to

11;

Sun

,since

mo11

CKIT

insL

589(45

bp)

NE

CKIT

insL

589(45

bp)

A-R

A-P

D

(Con

tinue

don

thefollo

wingpag

e)

Maier et al.





Tab

le3.

Clinical

charac

teris

ticsof

thepatients

(Con

t'd)

Patient

#

Tim

efrom

first

diagno

sis,

mo

Initials

ite

Treatmen

tinitials

ite,

time

Relap

sesite,

time

Treatmen

trelapse

site,

time

Mutation

status

primary

Mutationstatus

follo

w-up(site

andtime;

type)

Ligation

probes

used

Clin

ical

setting

atinclus

ion

Cha

ngeof

clinical

settingto

1513

2Ileum

Res

ectio

n,mo�1

32;

Ima,

sinc

emo�7

3Mes

enteria

l,mo�2

2CKITV55

5IþL5

76P

Mes

enteria

lmo1;

CKIT

V55

5Iþ

L576

P.M

esen

teria

lmo6;

CKIT

V55

5Iþ

L576

Pþ

PDGFR

AP56

7R

CKIT

L576

PA-P

D—

1685

Smallb

owel,

perito

neal,

liver

Deb

ulking

,mo�8

5;Im

amo�8

1to

�8;

Deb

ulking

,mo�8

;Sun

,mo�8

to�3

;Sor,m

o�3

to2;

Das

,mo2–

5;Deb

ulking

,mo4;

Imaþ

Nil,

mo5–

8

——

NE

Live

rmo4;

CKIT

delE55

4-I571

CKIT

delE55

4-I571

A-P

D—

171

Colon

,perito

neal

Surge

ryPDGFR

AD84

2VNE

PDGFR

AD84

2VA

A-P

D18

1Perito

neal

Deb

ulking

,mo�1

;Im

a,mo1–

2;Das

,sinc

emo2

——

PDGFR

AD84

2VNE

PDGFR

AD84

2VA

A-P

D

1914

Gas

tric

Res

ectio

n,mo�1

4;Im

aad

juva

nt,m

o�1

2to

24

——

PDGFR

AD84

2VNA

PDGFR

AD84

2VCR-H

R—

207

Gas

tric

Res

ectio

n,mo�7

;Im

aad

juva

nt,

mo�6

to30

——

PDGFR

AD84

2VNA

PDGFR

AD84

2VCR-H

R—

215

Gas

tric

Res

ectio

n,mo�5

;Im

aad

juva

nt,

mo�2

to34

PDGFR

AD84

2VNA

PDGFR

AD84

2VCR-H

R—

2253

Rec

tum

Res

ectio

n,mo�5

3Rec

tum,m

o�3

4Res

ectio

n,mo�3

4;Im

atinib,

sinc

emo�3

2

NE

Rec

tum

mo�3

4;CKIT

delY55

3-D57

2CKIT

delY55

3-D57

2CR

—

231

Gas

tric

Res

ectio

n,mo�1

;Im

aad

juva

nt,m

o2–

40

——

CKIT

delW55

7-K55

8NA

CKIT

delW55

7-K55

8CR-M

R—

241

Gas

tric

Res

ectio

n,mo�1

——

CKIT

delP55

1-E55

4NA

CKIT

delP55

1-E55

4CR-M

R—

251

Ileum

Res

ectio

n,mo�1

;Im

aad

juva

ntsinc

emo1

——

CKIT

insA

502-Y50

3NA

CKIT

insA

502-

Y50

3CR-H

R—

2631

Jejunu

mRes

ectio

n,mo�3

1;Im

atinib

adjuva

nt,

mo�3

0to

�18

--

CKIT

K64

2ENA

CKIT

K64

2ECR-H

R—

2743

Eso

phag

usIm

atinib

neoa

djuv

ant,mo

�46to

�43;

rese

ction,

mo�4

3

——

CKIT

V56

0GNA

CKIT

V56

0GCR-M

R—

(Con

tinue

don

thefollo

wingpag

e)






Tab

le3.

Clinical

charac

teris

ticsof

thepatients

(Con

t'd)

Patient

#

Tim

efrom

first

diagno

sis,

mo

Initials

ite

Treatmen

tinitials

ite,

time

Relap

sesite,

time

Treatmen

trelapse

site,

time

Mutation

status

primary

Mutationstatus

follo

w-up(site

andtime;

type)

Ligation

probes

used

Clin

ical

setting

atinclus

ion

Cha

ngeof

clinical

settingto

281

Smallb

owel,

perito

neal

Res

ectio

n,thermoa

blatio

n,mo

�1;Imatinib,s

ince

mo2

——

CKIT

delD57

9NA

CKIT

delD57

9CR

—

291

Jejunu

mRes

ectio

n,mo�1

;Im

aad

juva

nt,s

ince

mo2

——

CKIT

delV56

0-Y57

8NA

CKIT

delV56

0-Y57

8CR-H

R—

3094

Gas

tric

Res

ectio

n,mo�9

4FD

G-P

ETup

take

pan

crea

s,mo�7

2Im

a,sinc

emo�6

9CKIT

delM55

2-V55

9NA

CKIT

delM55

2-V55

9CR

—

3142

Smallb

owel

Res

ectio

n,mo�4

2Live

r,mo�2

4an

dmo�2

0Res

ectio

n,mo�2

4an

dmo

�20;

Ima,

mo�2

0to

�5an

dsinc

emo4

NE

Live

rmo�2

0;CKIT

delW55

7-K55

8CKIT

delW55

7-K55

8CR

—

321

Gas

tric

Res

ectio

n,mo�1

——

CKIT

V56

0DNA

CKIT

V56

0DCR-LR

—

331

Jejunu

mRes

ectio

n,mo�1

;Im

aad

juva

nt,s

ince

mo3

——

CKIT

K55

8N,

delV56

9NA

CKIT

K55

8N,

delV56

9CR-H

R—

341

Gas

tric

Res

ectio

n,mo�1

;Im

aad

juva

nt,s

ince

mo2

——

CKIT

insP

585(42

bp)

NA

CKIT

insP

585

(42b

p)

CR-H

R—

351

Gas

tric

Res

ectio

n,mo�1

——

PDGFR

AD84

2VNA

PDGFR

AD84

2VCR-H

R—

362

Cardia,

jejunu

mRes

ectio

n,mo�2

——

CKIT

delW55

7-K55

8NA

CKIT

delW55

7-K55

8CR-H

R—

371

Gas

tric

Res

ectio

n,mo�1

;Im

aad

juva

nt,s

ince

mo2

——

CKIT

V56

0GNA

CKIT

V56

0GCR-H

R—

3859

Duo

den

umRes

ectio

n,mo�5

9Live

r,mo�4

Res

ectio

n,mo�4

;Im

a,sinc

emo�2

NE

Live

rmo�4

;CKIT

insA

502-Y50

3CKIT

insA

502-

Y50

3CR

—

NOTE

:Sho

wnarethesitesof

dise

ase,co

urse

oftrea

tmen

t,molec

ular

charac

teris

ticsof

thetumor,a

ndtheL-PCRprobes

used

.Patients1to

18displaye

dac

tivedise

ase,defi

nedas

having

atleas

ton

elesion

,atleas

tat

onetim

epoint.P

atients19

to38

werein

CR.

Forp

atientsinCR,the

riskof

relapse

was

calculated

acco

rdingto

theMiettinen

crite

riabas

edon

tumor

loca

lization,

tumor

size

,and

mito

ticindex

,whe

reap

plicab

le.C

R-H

R,C

R-M

R,

andCR-LRden

oteCRwith

high

,mod

erate,

andlow

riskof

relapse

,res

pec

tively.

Abbreviations

:A,a

ctivedisea

se;A

-R,a

ctivedisea

sewith

resp

onse

totrea

tmen

t;A-P

D,a

ctivedisea

sewith

dise

aseprog

ress

ion;

NA,n

otap

plicab

le;N

E,n

otev

alua

ted.

Maier et al.





Figure 3. A–E, LINE1 and mutant fcDNA in correlation with clinical response in individual patients over time. LINE1 (long interspersed nuclear elements; left y-axis) reflects total fcDNA; mutant allele in percentagewt indicates the amount ofmutant fcDNA (right y-axis). Treatments and results of imaging (MRI, CT, andFDG-PET/CT) are shown below the time line (x-axis). Patient numbers correspond to Fig. 2. A, patient 5 (age, 76 years; female),CKIT exon 9 insertion AY502-503 mutation (BSC, best supportive care). B, patient 7 (age, 75 years; female), CKIT exon 11 deletion Y553-Q556 mutation.






Figure 3. (Continued ) C, patient 10 (age, 51 years; female), CKIT exon 11 deletion V555-D572 insG and exon 17 D820Y mutation. D, patient 8 (age, 49 years;female), CKIT exon 11 deletion K550-K558 mutation.

Maier et al.





Our finding of very rapid (within 2 weeks) and durablepositive to negative conversion in 2 patients responding,and a negative to positive conversion preceding radiograph-ic relapse by 3 months suggest that detection of fcDNAmight be used as tumor-specific biomarker to predictresponse early after initiation of treatment, and to predictrelapse. We were able to detect mutant fcDNA in 9 of these18 cases with active disease. There are several possibleexplanations for the negative cases. First, 3 negative casesfulfilled the prespecified criteria for active disease (at leastone lesion that could be measured) but, however, nosamples were available at the time of progression, andsamples at the time of response to TKI were negative (Table2). In addition, biologic properties inherent to the tumorsuch as vascularization, cell turn over, apoptosis rate, activesecretion of DNA and/or RNA, and localization mightimpact the release of nucleic acids into the blood stream(26, 39). Furthermore, 3 of 9 negative patients with activedisease carried large duplications that can hamper detectioncaused by ineffective probe hybridization (43–57 basepairs; patient 2, 4, and 14; Supplementary Table S1). Sen-sitivities of our assays ranged between 0.01% and 0.1%mutation/wt allele. The K642E assay was an exception with0.8% cross-reactivity reducing the sensitivity to 1%, due to ahigh GC content of the hybridization probe. More sensitive

mutation detection assays might reduce false-negativeresults. In addition, polyclonal disease and clonal compe-titionmay hamper the detection of individual mutations inplasma, especially when clones become dominant that donot harbor the initial mutation. Imatinib resistance in GISTlesions is associated with acquisition of secondary CKITor PDGFRA mutations that mediate inhibitor resistance(40–43). Importantly, specific lesions may carry differentmutations, indicating polyclonal resistance (40, 43). How-ever, repeated biopsies of different lesions are not feasible.In 2 patients, wewere able to separately track back and forthdistinct clones harboring individual mutations, includingmutations not known from biopsies, indicating clonalheterogeneity and competition during treatment. In thefuture, alternative technologies such as targeted massiveparallel sequencing or multiplex assays might simulta-neously identify and quantitate secondary disease sub-clones, allowing unbiased detection of sequence changesand thus might mirror clonal heterogeneity. This includessecondary mutations mediating resistance to treatment,allowing earlier treatment changes without repeated tumorbiopsies. We are currently expanding our analysis of sec-ondary resistance mutations in plasma samples of patientswith progressive disease and low or decreasing amounts ofthe initial mutation.

Figure 3. (Continued ) E, patient 11 (age, 65 years; male), CKIT exon 11 V559D and CKIT exon 13 V654A mutations. See text for details.






In 6 of 20 patients in CR, we were able to detect mutantfcDNA in plasma at low levels, indicating that viable tumorremains in a proportion of patients after removal of theprimary tumor. We were not able to establish a correlationof Miettinen risk to fcDNA positivity due to low number ofpatients, and possibly also due to the fact that all high-riskpatients received adjuvant imatinib. However, with longerfollow-up, it will be important to correlate test positivitywith relapse especially in high-risk patients, as it is currentlynot known which subgroup of patients benefit most fromadjuvant treatment, andwhich patients are at risk of relapse.

Together, our results indicate that mutant fcDNA specif-ically can be detected and quantified in the plasma ofpatients with GIST using quantitative L-PCR. The amountof mutant, tumor-derived fcDNA correlated with response,and our approach detected secondary mutations that areknown to confer treatment resistance. Thus, detection oftumor fcDNAmight be used as tumor-specific biomarker topredict both tumor response and relapse in patients withGIST.

Disclosure of Potential Conflicts of InterestT. Lange has commercial research grant from Novartis, honoraria from

Speakers Bureau of Novartis, Bristol-Myers Squibb, Ariad, and Pfizer, and is aconsultant/advisory board member of Novartis, Pfizer, Ariad, and Bristol-Myers Squibb. N. von Bubnoff has honoraria from Speakers Bureau ofNovartis Oncology. No potential conflicts of interest were disclosed by theother authors.

DisclaimerThe submitted material is original research, it has not been previously

published, and has not been presented or submitted for publicationelsewhere.

Authors' ContributionsConception and design: J. Maier, T. Lange, D. Niederwieser, J. Duyster,N. von BubnoffDevelopment of methodology: J. Maier, T. Lange, C. Wickenhauser, J.Duyster, N. von BubnoffAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.):K. Specht,M. Bruegel, C.Wickenhauser, P. Jost, D.Niederwieser, N. von BubnoffAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): J. Maier, T. Lange, I. Kerle, C. Wicken-hauser, D. Niederwieser, N. von BubnoffWriting, review, and/or revision of themanuscript: J. Maier, T. Lange, C.Wickenhauser, D. Niederwieser, C. Peschel, N. von BubnoffAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): T. Lange, M. Bruegel, P. Jost, N. vonBubnoffStudy supervision: T. Lange, D. Niederwieser, N. von Bubnoff

Grant SupportThis workwas supported by a grant to J. Duyster andN. von Bubnoff from

the Sander Stiftung, by a grant to N. von Bubnoff by Novartis Oncology, andby a grant to T. Lange by Novartis Oncology.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received March 19, 2013; revised June 12, 2013; accepted June 17, 2013;published OnlineFirst July 5, 2013.

References1. Verweij J, Casali PG, Zalcberg J, LeCesneA,ReichardtP, Blay JY, et al.

Progression-free survival in gastrointestinal stromal tumours withhigh-dose imatinib: randomised trial. Lancet 2004;364:1127–34.

2. Heinrich MC, Corless CL, Demetri GD, Blanke CD, von Mehren M,Joensuu H, et al. Kinase mutations and imatinib response in patientswith metastatic gastrointestinal stromal tumor. J Clin Oncol 2003;21:4342–9.

3. Debiec-Rychter M, Sciot R, Le Cesne A, Schlemmer M, HohenbergerP, van Oosterom AT, et al. KIT mutations and dose selection forimatinib in patients with advanced gastrointestinal stromal tumours.Eur J Cancer 2006;42:1093–103.

4. Demetri GD, von Mehren M, Blanke CD, Van den Abbeele AD, Eisen-berg B, Roberts PJ, et al. Efficacy and safety of imatinib mesylate inadvanced gastrointestinal stromal tumors. N Engl J Med 2002;347:472–80.

5. Heinrich MC, Owzar K, Corless CL, Hollis D, Borden EC, Fletcher CD,et al. Correlation of kinase genotype and clinical outcome in the NorthAmerican IntergroupPhase III Trial of imatinibmesylate for treatment ofadvanced gastrointestinal stromal tumor: CALGB 150105 Study byCancer and Leukemia Group B andSouthwest OncologyGroup. JClinOncol 2008;26:5360–7.

6. Blanke CD, Rankin C, Demetri GD, RyanCW, vonMehrenM, BenjaminRS, et al. Phase III randomized, intergroup trial assessing imatinibmesylate at two dose levels in patients with unresectable ormetastaticgastrointestinal stromal tumors expressing the kit receptor tyrosinekinase: S0033. J Clin Oncol 2008;26:626–32.

7. Blanke CD, Demetri GD, von Mehren M, Heinrich MC, Eisenberg B,Fletcher JA, et al. Long-term results from a randomized phase II trial ofstandard- versus higher-dose imatinib mesylate for patients withunresectable ormetastatic gastrointestinal stromal tumors expressingKIT. J Clin Oncol 2008;26:620–5.

8. Demetri GD, van Oosterom AT, Garrett CR, Blackstein ME, Shah MH,Verweij J, et al. Efficacy and safety of sunitinib in patients with

advanced gastrointestinal stromal tumour after failure of imatinib: arandomised controlled trial. Lancet 2006;368:1329–38.

9. Hong X, Choi H, Loyer EM, Benjamin RS, Trent JC, Charnsangavej C.Gastrointestinal stromal tumor: role of CT in diagnosis and in responseevaluation and surveillance after treatment with imatinib. Radio-graphics 2006;26:481–95.

10. Choi H, Charnsangavej C, Faria SC, Macapinlac HA, Burgess MA,Patel SR, et al. Correlation of computed tomography and positronemission tomography in patients with metastatic gastrointestinal stro-mal tumor treated at a single institution with imatinib mesylate: pro-posal of new computed tomography response criteria. J Clin Oncol2007;25:1753–9.

11. Blay JY, LeCesneA,Ray-Coquard I,BuiB,DuffaudF,DelbaldoC, et al.Prospective multicentric randomized phase III study of imatinib inpatients with advanced gastrointestinal stromal tumors comparinginterruption versus continuation of treatment beyond 1 year: theFrench Sarcoma Group. J Clin Oncol 2007;25:1107–13.

12. Gambhir SS, Czernin J, Schwimmer J, Silverman DH, Coleman RE,PhelpsME.A tabulated summary of theFDGPET literature. JNuclMed2001;42:1S–93S.

13. AntochG,HerrmannK, Heusner TA, Buck AK. [Imaging procedures forgastrointestinal stromal tumors]. Radiologe 2009;49:1109–16.

14. Antoch G, Kanja J, Bauer S, Kuehl H, Renzing-Koehler K, Schuette J,et al. Comparison of PET, CT, and dual-modality PET/CT imaging formonitoring of imatinib (STI571) therapy in patientswith gastrointestinalstromal tumors. J Nucl Med 2004;45:357–65.

15. Hahn S, Bauer S, Heusner TA, Ebeling P, Hamami ME, Stahl A, et al.Postoperative FDG-PET/CT staging in GIST: is there a benefit follow-ing R0 resection? Eur J Radiol 2011;80:670–4.

16. Bauer S, Hartmann JT, de Wit M, Lang H, Grabellus F, Antoch G, et al.Resection of residual disease in patients with metastatic gastrointes-tinal stromal tumors responding to treatmentwith imatinib. Int JCancer2005;117:316–25.

Maier et al.





17. Corless CL, Heinrich MC. Molecular pathobiology of gastrointestinalstromal sarcomas. Annu Rev Pathol 2008;3:557–86.

18. Hirota S, Ohashi A, Nishida T, Isozaki K, Kinoshita K, Shinomura Y,et al. Gain-of-function mutations of platelet-derived growth factorreceptor alpha gene in gastrointestinal stromal tumors. Gastroenter-ology 2003;125:660–7.

19. Lasota J, Dansonka-Mieszkowska A, Sobin LH, Miettinen M. A greatmajority of GISTs with PDGFRAmutations represent gastric tumors oflow or no malignant potential. Lab Invest 2004;84:874–83.

20. Medeiros F, Corless CL, Duensing A, Hornick JL, Oliveira AM, HeinrichMC, etal. KIT-negativegastrointestinal stromal tumors: proof of conceptand therapeutic implications. Am J Surg Pathol 2004;28:889–94.

21. Corless CL, Schroeder A, Griffith D, Town A, McGreevey L, Harrell P,et al. PDGFRA mutations in gastrointestinal stromal tumors: frequen-cy, spectrum and in vitro sensitivity to imatinib. J Clin Oncol 2005;23:5357–64.

22. Shapiro B, Chakrabarty M, Cohn EM, Leon SA. Determination ofcirculating DNA levels in patients with benign or malignant gastroin-testinal disease. Cancer 1983;51:2116–20.

23. Anker P, Stroun M. Circulating DNA in plasma or serum. Medicina2000;60:699–702.

24. Sidransky D. Emerging molecular markers of cancer. Nat Rev Cancer2002;2:210–9.

25. Sozzi G, Conte D, Leon M, Ciricione R, Roz L, Ratcliffe C, et al.Quantification of free circulating DNA as a diagnostic marker in lungcancer. J Clin Oncol 2003;21:3902–8.

26. Gormally E, Caboux E, Vineis P, Hainaut P. Circulating free DNA inplasma or serum as biomarker of carcinogenesis: practical aspectsand biological significance. Mutat Res 2007;635:105–17.

27. Sunami E, Vu AT, Nguyen SL, Giuliano AE, Hoon DS. Quantification ofLINE1 in circulating DNA as a molecular biomarker of breast cancer.Ann N Y Acad Sci 2008;1137:171–4.

28. Wang BG, Huang HY, Chen YC, Bristow RE, Kassauei K, Cheng CC,et al. Increased plasma DNA integrity in cancer patients. Cancer Res2003;63:3966–8.

29. Anker P, Mulcahy H, Chen XQ, Stroun M. Detection of circulatingtumour DNA in the blood (plasma/serum) of cancer patients. CancerMetastasis Rev 1999;18:65–73.

30. Tomita H, Ichikawa D, Ikoma D, Sai S, Tani N, Ikoma H, et al. Quan-tification of circulating plasma DNA fragments as tumor markers inpatients with esophageal cancer. Anticancer Res 2007;27:2737–41.

31. Kimura H, Kasahara K, Kawaishi M, Kunitoh H, Tamura T, Holloway B,et al. Detection of epidermal growth factor receptormutations in serum

as a predictor of the response to gefitinib in patients with non–small-cell lung cancer. Clin Cancer Res 2006;12:3915–21.

32. Diehl F, Schmidt K, Choti MA, Romans K, Goodman S, Li M, et al.Circulating mutant DNA to assess tumor dynamics. Nat Med 2008;14:985–90.

33. MiettinenM, Lasota J. Gastrointestinal stromal tumors: pathology andprognosis at different sites. Semin Diagn Pathol 2006;23:70–83.

34. Pelz-Ackermann O, Cross M, Pfeifer H, Deininger M, Wang SY, Al-AliHK, et al. Highly sensitive and quantitative detection of BCR-ABLkinase domain mutations by ligation PCR. Leukemia 2008;22:2288–91.

35. Schouten JP, McElgunn CJ, Waaijer R, Zwijnenburg D, Diepvens F,Pals G. Relative quantification of 40 nucleic acid sequences by mul-tiplex ligation-dependent probe amplification. Nucleic Acids Res2002;30:e57.

36. Guo T, Agaram NP, Wong GC, Hom G, D'Adamo D, Maki RG, et al.Sorafenib inhibits the imatinib-resistant KITT670I gatekeepermutationin gastrointestinal stromal tumor. Clin Cancer Res 2007;13:4874–81.

37. Druker BJ, Guilhot F, O'Brien SG, Gathmann I, Kantarjian H, Gatter-mann N, et al. Five-year follow-up of patients receiving imatinib forchronic myeloid leukemia. N Engl J Med 2006;355:2408–17.

38. Marin D, Ibrahim AR, Lucas C, Gerrard G, Wang L, Szydlo RM, et al.Assessment of BCR-ABL1 transcript levels at 3 months is the onlyrequirement for predicting outcome for patients with chronic myeloidleukemia treated with tyrosine kinase inhibitors. J Clin Oncol 2012;30:232–8.

39. Schwarzenbach H, Hoon DS, Pantel K. Cell-free nucleic acids asbiomarkers in cancer patients. Nat Rev Cancer 2011;11:426–37.

40. ShinozakiM,O'DaySJ,KitagoM,Amersi F, KuoC,KimJ, et al. Utility ofcirculating B-RAF DNA mutation in serum for monitoring melanomapatients receiving biochemotherapy. Clin Cancer Res 2007;13:2068–74.

41. HeinrichMC, CorlessCL, BlankeCD, Demetri GD, JoensuuH, RobertsPJ, et al. Molecular correlates of imatinib resistance in gastrointestinalstromal tumors. J Clin Oncol 2006;24:4764–74.

42. Debiec-RychterM,Cools J,DumezH,Sciot R, StulM,MentensN, et al.Mechanisms of resistance to imatinib mesylate in gastrointestinalstromal tumors and activity of the PKC412 inhibitor against imati-nib-resistant mutants. Gastroenterology 2005;128:270–9.

43. Wardelmann E, Merkelbach-Bruse S, Pauls K, Thomas N, SchildhausHU, Heinicke T, et al. Polyclonal evolution of multiple secondary KITmutations in gastrointestinal stromal tumors under treatment withimatinib mesylate. Clin Cancer Res 2006;12:1743–9.






2013;19:4854-4867. Published OnlineFirst July 5, 2013.Clin Cancer Res Jacqueline Maier, Thoralf Lange, Irina Kerle, et al.

PDGFRA or CKITMutations of Patients with Gastrointestinal Stromal Tumor Harboring Activating Detection of Mutant Free Circulating Tumor DNA in the Plasma of

Updated version

10.1158/1078-0432.CCR-13-0765doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2013/07/03/1078-0432.CCR-13-0765.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/19/17/4854.full#ref-list-1

This article cites 43 articles, 18 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/19/17/4854.full#related-urls

This article has been cited by 6 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/19/17/4854To 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-13-0765

http://clincancerres.aacrjournals.org/content/suppl/2013/07/03/1078-0432.CCR-13-0765.DC1

http://clincancerres.aacrjournals.org/content/19/17/4854.full#ref-list-1

http://clincancerres.aacrjournals.org/content/19/17/4854.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/19/17/4854


detection of mutant free circulating tumor dna in the ... · monitoring response in gist. however,...

Documents