You are here

Plasma ctDNA Analysis for Detection of the EGFR T790M Mutation in Patients with Advanced Non–Small Cell Lung Cancer

Journal of Thoracic Oncology

Abstract

Introduction

Tumor biopsies for detecting EGFR mutations in advanced NSCLC are invasive, costly, and not always feasible for patients with late-stage disease. The clinical utility of the cobas EGFR Mutation Test v2 (Roche Molecular Systems, Inc., Pleasanton, CA) with plasma samples from patients with NSCLC at disease progression after previous EGFR tyrosine kinase inhibitor therapy was investigated to determine eligibility for osimertinib treatment.

Methods

Matched tumor tissue and plasma samples from patients screened for the AURA extension and AURA2 phase II studies were tested for EGFR mutations by using tissue- and plasma-based cobas EGFR mutation tests. Plasma test performance was assessed by using the cobas tissue test and a next-generation sequencing method (MiSeq [Illumina Inc., San Diego, CA]) as references. The objective response rate, measured by blinded independent central review, was assessed in patients receiving osimertinib with a plasma T790M mutation–positive status.

Results

During screening, 551 patients provided matched tumor tissue and plasma samples. Pooled analysis of the positive and negative percent agreements between the cobas plasma and tissue tests for detection of T790M mutation were 61% and 79%, respectively. Comparing cobas plasma test with next-generation sequencing demonstrated positive and negative percent agreements of 90% or higher. The objective response rate was 64% (95% confidence interval: 57–70) in T790M mutation–positive patients by both cobas tissue and plasma tests (evaluable for response).

Conclusions

The cobas plasma test detected the T790M mutation in 61% of tumor tissue T790M mutation–positive patients. To mitigate the risk of false-negative plasma results, patients with a negative plasma result should undergo a tissue test where feasible.

Keywords: osimertinib, circulating tumor DNA, Roche cobas EGFR Mutation Test, tumor tissue biopsy, MiSeq next-generation sequencing.

Introduction

The EGFR tyrosine kinase inhibitors (TKIs), including erlotinib, afatinib and gefitinib, are recommended first-line treatments for patients with EGFR mutation–positive, advanced NSCLC.1, 2, 3, 4, 5, 6, 7, 8, and 9 Resistance develops in most patients with EGFR mutation–positive NSCLC treated with a first-line EGFR TKI,1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 and approximately 60% of these patients have tumors that harbor an EGFR T790M resistance mutation.11 and 12 Analysis of tumor tissue biopsy material is considered the conventional standard method for mutation detection. The biopsy tissue sample is typically formalin-fixed and paraffin-embedded before DNA is extracted and analyzed to detect EGFR mutations. Obtaining sufficient tumor tissue for analysis from patients with advanced NSCLC who progressed during a previous therapy is invasive and time- and resource-intensive, and such biopsies are associated with a higher risk for complications to the patient.13 In cases in which it is not possible to obtain a tumor biopsy sample, patients who would otherwise be eligible may not be provided with access to potentially beneficial targeted therapies. Therefore, there is a clear unmet clinical need for an alternative procedure to detect the EGFR T790M mutation in the EGFR TKI resistance setting.

Circulating tumor DNA (ctDNA) has been identified as a specific and sensitive biomarker that can be used for the detection of EGFR mutations.14, 15, 16, and 17 ctDNA comprises small fragments of DNA that are shed from tumor cells during processes such as apoptosis and necrosis. ctDNA can be extracted from plasma and serum and can be tested for tumor-specific molecular markers.18 and 19 The cobas EGFR Mutation Test v2 (Roche Molecular Systems, Inc., Pleasanton, CA) has been developed for detection of EGFR mutations in both tumor tissue and plasma samples.

Osimertinib is a potent oral, irreversible EGFR TKI that is selective for both EGFR TKI–sensitizing mutations and T790M resistance mutations over the wild-type EGFR.20 Osimertinib is recommended for the treatment of patients with EGFR T790M–positive NSCLC after progression during a first-line EGFR TKI treatment.1 and 2

The AURA clinical trial program includes two key phase II studies to assess the efficacy and safety of osimertinib in patients with T790M mutation–positive advanced NSCLC who have progressed after EGFR TKI treatment. The AURA extension (NCT01802632) and AURA2 (NCT02094261) studies (phase II), which were used for conditional approval of osimertinib, investigated the efficacy, tolerability, and safety of orally administered osimertinib in patients with T790M mutation–positive advanced NSCLC.21 and 22 Central detection of the T790M mutation in a tumor tissue sample by using the cobas EGFR Mutation Test v2 (European Conformity research use–only, investigational use–only version) was required for enrollment in these clinical studies. The purpose of this analysis is to assess the clinical utility of the cobas plasma test for detection of the T790M mutation in ctDNA obtained from patient plasma samples collected during screening for the AURA extension and AURA2 studies.

Methods

Study Design and Participants

Full details of the AURA extension and AURA2 studies have been previously published.21 and 22 In brief, the AURA extension and AURA2 studies were phase II, single-arm, open-label, multicenter studies of osimertinib, 80 mg orally once daily. Patients were at least 18 years old (≥21 years in Japan) with advanced NSCLC, and they had disease progression after treatment with an EGFR TKI agent or had received prior therapy with an EGFR TKI and at least one other treatment regimen. Radiological documentation of disease progression during the last treatment administered before enrollment in the studies was required. At baseline all patients were required, on confirmation of disease progression during or after their most recent treatment regimen, to provide a tumor biopsy sample for central T790M analysis. During the screening period, patients were also required to provide plasma samples for retrospective analysis. These studies were designed to assess the efficacy, safety, and tolerability of osimertinib.

The studies were approved by the institutional review board or independent ethics committee associated with each study center. All patients provided written informed consent before any study-specific procedures, sampling, and analyses were performed. Patients who stopped participating in the study for any reason other than disease progression continued with tumor assessments until disease progression. The primary end point of the studies was objective response rate (ORR), defined by blinded independent central review (BICR) using the Response Criteria in Solid Tumors, version 1.1. The efficacy of osimertinib was assessed according to cobas tumor tissue (primary study objective) and plasma test (exploratory objective) results for T790M mutation status.

Tumor Tissue and Plasma Sampling

A mandatory tissue biopsy sample was taken from all patients after confirmation of disease progression during the last treatment regimen. Tumor tissue biopsy and plasma samples were collected from patients being screened for the AURA extension and AURA2 clinical studies. All plasma samples were collected during the initial 28-day baseline period before dosing with osimertinib.

DNA Extraction, Amplification, and Detection

The tumor tissue samples were formalin-fixed and paraffin-embedded, and genomic DNA was extracted from them for mutation testing. The concentration of genomic DNA from the extracted tissue sample was determined with a spectrophotometer and adjusted to 2 ng/μL. The cobas cell-free DNA Sample Preparation Kit (Roche Molecular Systems) was used to extract the ctDNA from the plasma samples. The cobas EGFR Mutation Test (investigational use–only version [Roche Molecular Systems]) was used to detect EGFR mutations in the extracted tissue DNA samples (from here on referred to as the cobas tissue test). The cobas EGFR Mutation Test v2 (European Conformity investigational use–only version [Roche Molecular Systems]) was used to detect EGFR mutations in the extracted plasma ctDNA (from here on referred to as the cobas plasma test). Amplification and detection were attained by using target-specific primer and probe sequences. These tests were performed at three central testing laboratories: Carolinas Pathology Group (Charlotte, NC), Histogenex (Antwerp, Belgium), and Quintiles (Singapore). Assays were performed according to the manufacturer protocols. Detectable EGFR mutations were reported as positive in the analysis output. EGFR mutations below the detectable limit were reported as negative. Any invalid cobas test results were reported as status unknown; these samples were excluded from all corresponding analyses unless stated otherwise. MiSeq sequencing (Illumina, San Diego, CA) was used as the comparator method for the cobas test. MiSeq utilizes deep sequencing of EGFR exons 18, 19, 20, and 21. Deep sequencing, or next-generation sequencing (NGS), is a suitable comparator for the polymerase chain reaction (PCR)-based EGFR mutation detection test because it has very low analytical sensitivity (around 1%–2%).23 Illumina’s MiSeq sequencing can provide a high depth of coverage to identify the prevalence of low-frequency mutations. The MiSeq NGS data were generated by using a two-tube amplicon assay. Testing was performed at Roche Molecular Systems laboratories or SeqWright (Houston, TX).

Assessments

Clinical efficacy was evaluated as the primary end point, with ORR determined by BICR in the evaluable-for-response populations from both the AURA extension and AURA2 trials. The evaluable-for-response patient group included those patients with at least one lesion at baseline that was measureable by BICR.

Statistical Methods

The clinical efficacy data is based on a data cutoff date of November 1, 2015, for both the AURA extension and AURA2 trials. The agreement between the cobas tumor tissue test and cobas plasma test for each EGFR mutation subtype was assessed for sensitivity by calculating the positive percent agreement (PPA), specificity by calculating the negative percent agreement (NPA), and concordance by calculating the overall percent agreement (OPA) with 95% confidence intervals (CIs), with the cobas tumor tissue test used as the reference method. The calculations for these values are outlined in Supplementary Table 1. CIs were calculated by using the Clopper-Pearson exact method for binomial proportions. The ORR data analysis was performed with the evaluable-for-response patients, defined as all patients who received at least one dose of study treatment and had measureable disease at baseline according to the BICR.

Results

Patient EGFR Mutation Status

There were 401 and 472 patients screened for the AURA extension and AURA2 studies, respectively. In the AURA extension study, 210 screened patients provided matched tumor tissue and plasma samples, whereas 341 screened patients provided matched tumor tissue and plasma samples in the AURA2 study (Fig. 1). Thus, a total of 551 out of 710 (78%) patients with a cobas tissue test provided matched tumor tissue and plasma samples. These comprised 416 T790M-positive, 127 T790M-negative, and eight invalid results according to the cobas tissue test. Fewer plasma samples were provided from patients with a negative T790M tissue test result than from those with a positive T790M tissue test result. This was a consequence of sample collection during the AURA extension study, as plasma samples tended to be collected from patients only after tissue test results were known. This resulted in fewer plasma samples from patients with a negative tissue test result who were ineligible for the AURA extension study. During screening for AURA2 study, investigators were asked to send plasma samples from all screened patients (not only those with a positive tissue test result). In the AURA extension and AURA2 studies there were 16 of 117 patients (14%) and 111 of 140 (79%), respectively, with tissue T790M-negative status and a matched plasma test sample. The NPA was not calculated for the AURA extension study because of too few observations.

gr1

Figure 1

Schematics of the AURA extension (A) and AURA2 (B) studies.

 

In patients with a positive T790M tissue test result, those with extrathoracic disease (TNM seventh edition category M1B) had detectable T790M in their plasma more often (72% [pooled data set]) than those with TNM M0-M1A disease (51% [pooled data set]) (Table 1). This is supported by a positive correlation between baseline target lesion size and the percentage of patients with a plasma T790M–positive test result. Among patients evaluable for response and with a valid plasma result, T790M was detected in the plasma of 97% of those (29 of 30) with baseline lesions of 120 mm or larger but in only 53% (69 of 131) with baseline lesions smaller than 40 mm (Supplementary Table 2).

Table 1

Baseline TNM Categories by Baseline Plasma T790M Mutation Status (Pooled AURA Extension and AURA2 Full Analysis Set)

 

Plasma T790M Mutation Status No. (%) of Patients
M0/MX/M1/M1A (n = 243) M1B (n = 154) Missing (n = 14) p Value
Positive 123 (51) 111 (72) 8 (57) <0.0001a
Negative 113 (47) 39 (25) 5 (36)
Invalid 1 (<1) 0 0
No sample 6 (2) 4 (3) 1 (7)

a Chi-square test was performed for positive and negative plasma T790M status by comparing nonmissing baseline metastases categories.

Percent Agreement Between the cobas Plasma Test and cobas Tissue Test

The cobas tissue test was used as a reference method to assess the sensitivity, specificity, and concordance of the cobas plasma test for detection of T790M, L858R, and exon 19 deletion mutations (Table 2). For the screened patient population from the AURA extension study, the PPA for the T790M mutation was 64% (95% CI: 57–71). The NPA value is not reported because the total number of matched patient samples was less than 20. In the screened patient population from the AURA2 study, the PPA for the T790M mutation was 59% (95% CI: 52–65) and the NPA was 80% (95% CI: 72–87). When the AURA extension and AURA2 data were pooled, the PPA and NPA were 61% (95% CI: 57–66) and 79% (95% CI: 70–85), respectively, and the OPA was 65% (95% CI: 61–69) for the T790M mutation. The PPA for T790M detection increased slightly and the NPA decreased slightly when patients with neither T790M nor sensitizing EGFR mutations (G719X, exon 19 deletion, S7681, exon 20 insertion, and L858R) detected in plasma by cobas were excluded from the agreement calculations (n = 101 in the pooled data set): the PPA was 72% (95% CI: 67–77) and the NPA was 69% (95% CI: 59–79) (Supplementary Table 3).

Table 2

Percent Agreement of the cobas Plasma Test with the cobas Tissue Test as a Reference Method for the Detection of EGFR T790M, L858R, and Exon 19 Deletion

 

Percent Agreement (95% CI)
T790M L858R Exon 19 Deletion
AURA Extension (n = 210) AURA2 (n = 341) Pooled AURA Extension and AURA2 (n = 551) AURA Extension (n = 210) AURA2 (n = 341) Pooled AURA Extension and AURA2 (n = 551) AURA Extension (n = 210) AURA2 (n = 341) Pooled AURA Extension and AURA2 (n = 551)
PPA 64 (57–71) 59 (52–65) 61 (57–66) 75 (61–85) 76 (67–84) 76 (69–82) 88 (81–93) 83 (77–88) 85 (81–89)
NPA a 80 (72–87) 79 (70–85) 99 (95–100) 98 (95–99) 98 (96–99) 98 (92–100) 98 (94–100) 98 (95–100)
OPA 65 (58–71) 66 (61–71) 65 (61–69) 92 (88–96) 90 (86–93) 91 (88–93) 91 (86–94) 89 (86–93) 90 (87–92)

a Not calculated because of the low number of samples (total <20).

PPA, positive percent agreement (sensitivity); NPA, negative percent agreement (specificity); OPA, overall percent agreement (concordance).

A pooled data set was used to analyze PPA between plasma and tissue for T790M detection according to prior lines of therapy. The PPA was lower for patients receiving osimertinib as second-line therapy (52% [95% CI: 42–61]) than for those receiving osimertinib as third- or a later-line therapy (65% [95% CI: 59–70]) (Supplementary Table 4). In patients whose line of therapy was unknown, the NPA was 79% (95% CI: 71–86). The NPA by line of therapy is not reported because of the low number of patients with T790M-negative tissue samples for whom line of therapy was recorded (n = 1 for the second line and n = 2 for third or later lines).

From the pooled data set, comparing the cobas plasma test results with the cobas tissue test results produced PPA, NPA, and OPA values for the detection of L858R mutation of 76% (95% CI: 69–82), 98% (95% CI: 96–99), and 91% (95% CI: 88–93), respectively. For detection of the exon 19 deletions, the PPA, NPA and OPA values were 85% (95% CI: 81–89), 98% (95% CI: 95–100), and 90% (95% CI: 87–92), respectively (Table 2).

cobas and MiSeq NGS Concordance for Tumor Tissue and Plasma Samples

The cobas plasma test results were compared with a reference NGS method with plasma samples for detection of T790M in ctDNA (MiSeq). For detection of the T790M mutation, the PPA, NPA, and OPA values were 93% (95% CI: 89–96), 92% (95% CI: 88–95), and 92% (95% CI: 90–94), respectively (Table 3). This is comparable with the PPA, NPA, and OPA values with the cobas tissue test when compared with a reference NGS method with tumor tissue samples (Table 3).24

Table 3

Concordance between the cobas EGFR Mutation Test and MiSeq NGS for Tumor Tissue and Plasma Tests from the AURA Extension and AURA2 Pooled Analysis Data Set

 

Percent Agreement (95% CI) cobas Tissue Test vs. NGS Tissue Analysis (Reference) for the Detection of T790M (n = 673) cobas Plasma Test vs. NGS Plasma Analysis (Reference [n = 562]) for the Detection of
T790M Exon 19 Deletion L858R
PPA 90 (87–93) 93 (89–96) 95 (92–98) 93 (87–97)
NPA 98 (94–99) 92 (88–95) 91 (87–94) 97 (95–98)
OPA 92 (90–94) 92 (90–94) 93 (91–95) 96 (94–97)

NGS, next-generation sequencing; CI, confidence interval; PPA, positive percent agreement (sensitivity); NPA, negative percent agreement (specificity); OPA, overall percent agreement (concordance).

Five patients screened for AURA extension and 22 patients for AURA2 were genotyped as T790M positive with the cobas plasma test and T790M negative with the cobas tissue test. These discordances were assessed further by using NGS results (Table 4). Twenty-three of the 27 discordant plasma samples were confirmed as T790M positive by the NGS plasma test, one plasma sample could not be evaluated with NGS, and three plasma samples had no detectable T790M according to NGS. Interestingly, 11 of the matched tumor tissue samples that were negative by the cobas tissue test had detectable T790M according to the NGS tumor tissue analysis. These results suggest that the true false-positive T790M detection rate of the cobas plasma test is very low.

Table 4

NGS Results for T790M Mutation Detection Using Tissue and Plasma Samples for the AURA Extension and AURA2 Cases in Which T790M was Detected with the Plasma Test but Not Detected with the Tissue Test

 

Study T790M Detected with cobas Plasma Test but Not Detected with cobas Tissue Test NGS Tumor Tissue T790M Status NGS Plasma T790M Status
Positive Negative Positive Negative
AURA extension 5 3a of 5 1 of 5 5 of 5 0 of 5
AURA2 22 8 of 22 14 of 22 18b of 22 3 of 22
Pooled AURA extension and AURA2 27 11 of 27 15 of 27 23 of 27 3 of 27

a One AURA extension tissue sample had invalid NGS test.

b One AURA2 plasma sample not tested by NGS.

NGS, next-generation sequencing.

Objective Response Rate Based on the cobas Tissue and Plasma Tests for T790M Mutation Status

In the AURA extension study patients selected by using the cobas tissue test and meeting the evaluable-for-response criteria (n = 198), the ORR, as previously reported, was 62% (95% CI: 54–68) (Table 5).21 In the subset of those patients who also had a plasma T790M–positive result (n = 126), the ORR was very similar: 59% (95% CI: 50–67). Likewise, the AURA2 study patients selected by using the cobas tissue test and meeting the evaluable-for-response criteria (n = 199) had an ORR, as previously reported, of 70% (95% CI: 64–77),22 and in the subset of those patients with a plasma T790M-positive result (n = 109), the ORR was 70% (95% CI: 60–78). When the AURA extension and AURA2 data were pooled, the ORR in the evaluable-for-response analysis set with a tissue T790M–positive test was 66% (95% CI: 61–71); it was 64% (95% CI: 57–70) in the subset of patients who were also plasma T790M–positive.

Table 5

ORR with Osimertinib (80 mg Once Daily) according to Plasma T790M Status in Patients Identified as T790M Positive by the Tissue Test

 

T790M status n No. (%) of Patients with Response 95% CI
AURA extension
Tumor tissue T790M positive (evaluable for response)a 198 122 (62) 54–68
 Plasma T790M positive 126 74 (59) 50–67
 Plasma T790M negative 68 46 (68) 55–79
 Plasma T790M unknown 0 0
 No plasma sample 4 2 (50) 7–93
AURA2
Tumor tissue T790M positive (evaluable for response)b 199 140 (70) 64–77
 Plasma T790M positive 109 76 (70) 60–78
 Plasma T790M negative 82 59 (72) 61–81
 Plasma T790M unknown 1 1 (100)
 No plasma sample 7 4 (57) 18–90
Pooled AURA extension and AURA2
Tumor tissue T790M positive (evaluable for response) 397 262 (66) 61–71
 Plasma T790M positive 235 150 (64) 57–70
 Plasma T790M negative 150 105 (70) 62–77
 Plasma T790M unknown 1 1 (100)
 No plasma sample 11 6 (55) 23–83

a Data from Yang et al.21

b Data from Goss et al.22

Note: ORR by blinded independent central review (using the Response Evaluation Criteria in Solid Tumors, version 1.1) from the evaluable-for-response populations of the AURA extension and AURA2 studies. Data cutoff date was November 1, 2015.

ORR, objective response rate; CI, confidence interval.

Discussion

The cobas EGFR Mutation Test v2, which is suitable for use with both tumor tissue and plasma samples, has been developed as a diagnostic tool to help identify patients with T790M mutation–positive tumors for treatment with osimertinib. This analysis investigated the sensitivity, specificity, and concordance of the cobas plasma test for the detection of EGFR mutations, with a focus on the T790M mutation when compared with the cobas tissue test. The MiSeq Illumina NGS method was used as a reference test.

The PPA value for detection of T790M mutation in the pooled data set when comparing cobas plasma with cobas tissue as a reference was 61%, meaning that this subgroup of patients could have had their T790M mutation status determined without an invasive biopsy procedure. This is consistent with previous reports of EGFR mutation status investigated with tumor tissue and plasma samples.14 When the cobas plasma test results and the cobas tissue test results are compared, the PPA for the detection of T790M is somewhat lower than the PPA for L858R and exon 19 deletions (76% and 85%, respectively [pooled data set]). This is also consistent with other reports of plasma testing in the EGFR TKI resistance setting25, 26, and 27 and likely due to tumor heterogeneity and a lower abundance of the resistance mutation than the driver mutation in the advanced-disease setting. The likelihood of detection of T790M in the plasma of patients with a T790M-positive tissue test result is greater in the subcategory of patients with extrathoracic disease (TNM seventh edition category M1B) than in those with intrathoracic disease (M0–M1A): 72% and 51%, respectively (p = 0.0001, see Table 1). One possibility is that patients with NSCLC who have extrathoracic disease may have a higher disease burden and are therefore more likely to shed tumor-derived DNA into their bloodstream. This is supported by the positive correlation between the baseline target lesion size and positive detection of T790M by plasma test. Furthermore, in patients who received osimertinib as a second-line therapy, PPA was lower (52%) than in those who received osimertinib as third- or later-line therapy (65%). This suggests that it is more challenging to detect T790M in plasma ctDNA of patients who have just progressed after a previous first-line EGFR TKI. We hypothesize that patients who received more previous lines of therapy may have a greater disease burden and a greater allelic fraction of T790M or that prior chemotherapy in some way promotes shedding of ctDNA from tumor cells. Therefore, further investigation into the reasons for higher detection of T790M in this subset of patients is warranted.

The plasma test did not detect the T790M mutation in plasma ctDNA of approximately 40% of patients with a T790M-positive tissue test result. To mitigate the risk for a false-negative plasma test result, it is advised that, where possible, any plasma T790M–negative test result be explored further with a contemporaneous biopsy and tissue test. This is also suggested in a recent study by Oxnard et al.27 It has been suggested that detection of the original sensitizing mutation in plasma could act as an internal control for plasma testing; accordingly, a plasma sample with no detectable sensitizing mutation can be considered uninformative. Analysis of T790M PPA between the cobas plasma and tissue tests, excluding patients who had no detectable T790M or sensitizing mutation, resulted in a marginal increase in PPA (from 61% to 72%). We believe that caution should be applied when interpreting plasma results and recommend following up a plasma T790M–negative test result with a contemporaneous biopsy and tissue test whenever feasible.

The NPA for detection of T790M mutation in the pooled data set when the cobas plasma test result was compared with the cobas tissue test result as a reference was 79%, which is consistent with that reported by Oxnard et al. when using an alternative plasma testing method.27 When the cobas plasma test result was compared with the cobas tissue test result, the NPA was notably lower for the T790M mutation than for L858R and exon 19 deletions. However, when the cobas plasma test results were also compared with the MiSeq NGS plasma results for detection of the T790M mutation, the PPA, NPA, and OPA values were all higher than 90%, demonstrating strong agreement between the two tests and confirming that the cobas plasma test reliably allows sensitive and specific detection of EGFR mutations in plasma. Furthermore, a low incidence of false-positive results from the cobas plasma test for T790M was demonstrated. In the 27 patients with T790M mutation–negative status by the cobas tissue test and T790M mutation–positive status by the cobas plasma test, 23 (85%) were T790M mutation–positive by the NGS reference method using plasma samples. These data suggest that the lower NPA value observed when comparing tumor tissue and plasma T790M results may be driven by tumor heterogeneity and not by poor specificity of the cobas plasma test.

The T790M mutation is an acquired drug resistance–associated mutation and therefore reflects clonal evolution and increased tumor heterogeneity in the resistance setting. This is expected to contribute to the lower NPA between the plasma and tissue test results because obtaining tumor tissue by biopsy restricts sampling of tumor cells to a single site of disease. Thus, it is unsurprising that some discordance between tissue and plasma tests for the T790M mutation will be observed in the later-line setting.28 and 29

A limitation of this analysis is that patients were selected for the AURA extension and AURA2 clinical trials on the basis of tumor tissue T790M mutation status. Therefore, although these studies demonstrated similar ORRs in patients who were T790M mutation–positive according to the cobas tissue and plasma tests, preselection of patients by using the tissue test limits our assessment of efficacy in patients who would be selected by using a plasma test alone.

In addition to enabling a clinically significant proportion of patients to avoid a biopsy, a well-validated plasma test could benefit patients who are unable to provide a tumor tissue sample on account of the risks associated with the procedure. The overall performance status of a patient and the anatomical location of the site suitable for a biopsy are the most common limiting factors in performing these biopsies.30 and 31 Recent studies estimate that as many as 40% of relapsed patients with NSCLC may be unable to provide a contemporaneous tumor tissue sample suitable for molecular analysis.30 and 32 Performing the cobas plasma test is faster and cheaper than obtaining and processing tumor tissue samples, and this benefits both the patient and the health care provider.31 This is in line with the increasing trend toward the development of liquid-based companion diagnostics.33 For example, the cobas EGFR Mutation Test v2 is paired with erlotinib and osimertinib for NSCLC.24 A limitation of ctDNA plasma testing is the higher probability of a false-negative result. Therefore, after a plasma T790M negative cobas test result, it is advisable to obtain new biopsy material if possible to do so and perform a tissue test to determine the T790M status.

Other PCR-based assays that can be used to detect EGFR sensitizing mutations in plasma samples with varying specificity and sensitivity include the amplification refractory mutation system assay (therascreen EGFR RGQ PCR Kit [Qiagen, Hilden, Germany]); droplet digital PCR (ddPCR); and the beads, emulsions, amplification, and magnetics (BEAMing) ddPCR technique (Sysmex Inostics, Inc., Mundelein, IL).34 The ddPCR and BEAMing methods are enhancements of traditional PCR with improved sensitivity and specificity.14 and 35 These tests can determine absolute quantification of mutant EGFR levels in plasma samples and could be used to serially monitor treatment response or failure and disease progression.36 Comparison of the cobas test with the BEAMing test for detection of the T790M mutation in plasma samples has revealed that the BEAMing test has a higher sensitivity whereas the cobas test has a higher specificity than tissue testing.14

Detection of the T790M mutation in patients with advanced NSCLC is important to guide clinical decisions. Collectively, the results of this study support the clinical utility of the cobas plasma test for detection of the EGFR T790M mutation in patients with advanced NSCLC.

Acknowledgments

The AURA extension (NCT01802632) and AURA2 (NCT02094261) studies were funded by AstraZeneca, Cambridge, United Kingdom, the manufacturer of osimertinib. Mutation testing was performed at the central testing laboratories of Carolinas Pathology Group, Histogenex, Quintiles, and Roche Molecular Systems Laboratories. The authors would like to thank all the patients and their families. The authors would also like to acknowledge the Roche Molecular Systems team, Xiangning Huang, and Dr Kenneth Thress for their input, as well as Natalie Griffiths, PhD, of iMed Comms, Macclesfield, UK, an Ashfield Company and part of UDG Healthcare plc for medical writing support funded by AstraZeneca.

Supplementary Data

Download file

Supplementary Tables 1–4

References

  • 1 S. Novello, F. Barlesi, R. Califano, et al. Metastatic non-small-cell lung cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2016;27:v1-v27 Crossref
  • 2 NCCN Clinical Practice Guidelines in Oncology. Non–small cell lung cancer, Version 5. 2017. https://www.nccn.org/store/login/login.aspx?ReturnURL=https://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf. Accessed April 1, 2017.
  • 3 M. Maemondo, A. Inoue, K. Kobayashi, et al. Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med. 2010;362:2380-2388 Crossref
  • 4 T. Mitsudomi, S. Morita, Y. Yatabe, et al. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. Lancet Oncol. 2010;11:121-128 Crossref
  • 5 Y.L. Wu, C. Zhou, C.K. Liam, et al. First-line erlotinib versus gemcitabine/cisplatin in patients with advanced EGFR mutation-positive non-small-cell lung cancer: analyses from the phase III, randomized, open-label, ENSURE study. Ann Oncol. 2015;26:1883-1889 Crossref
  • 6 Y.-L. Wu, C. Zhou, C.-P. Hu, et al. Afatinib versus cisplatin plus gemcitabine for first-line treatment of Asian patients with advanced non-small-cell lung cancer harbouring EGFR mutations (LUX-Lung 6): an open-label, randomised phase 3 trial. Lancet Oncol. 2014;15:213-222 Crossref
  • 7 T.S. Mok, Y.L. Wu, S. Thongprasert, et al. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med. 2009;361:947-957 Crossref
  • 8 R. Rosell, E. Carcereny, R. Gervais, et al. Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial. Lancet Oncol. 2012;13:239-246 Crossref
  • 9 L.V. Sequist, J.C. Yang, N. Yamamoto, et al. Phase III study of afatinib or cisplatin plus pemetrexed in patients with metastatic lung adenocarcinoma with EGFR mutations. J Clin Oncol. 2013;31:3327-3334 Crossref
  • 10 M. Xu, Y. Xie, S. Ni, H. Liu. The latest therapeutic strategies after resistance to first generation epidermal growth factor receptor tyrosine kinase inhibitors (EGFR TKIs) in patients with non-small cell lung cancer (NSCLC). Ann Transl Med. 2015;3:96
  • 11 G.R. Oxnard, M.E. Arcila, C.S. Sima, et al. Acquired resistance to EGFR tyrosine kinase inhibitors in EGFR-mutant lung cancer: distinct natural history of patients with tumors harboring the T790M mutation. Clin Cancer Res. 2011;17:1616-1622 Crossref
  • 12 H.A. Yu, M.E. Arcila, N. Rekhtman, et al. Analysis of tumor specimens at the time of acquired resistance to EGFR-TKI therapy in 155 patients with EGFR-mutant lung cancers. Clin Cancer Res. 2013;19:2240-2247 Crossref
  • 13 M.J. Overman, J. Modak, S. Kopetz, et al. Use of research biopsies in clinical trials: are risks and benefits adequately discussed?. J Clin Oncol. 2013;31:17-22 Crossref
  • 14 K.S. Thress, R. Brant, T.H. Carr, et al. EGFR mutation detection in ctDNA from NSCLC patient plasma: a cross-platform comparison of leading technologies to support the clinical development of AZD9291. Lung Cancer. 2015;90:509-515 Crossref
  • 15 J.Y. Douillard, G. Ostoros, M. Cobo, et al. Gefitinib treatment in EGFR mutated Caucasian NSCLC: circulating-free tumor DNA as a surrogate for determination of EGFR status. J Thorac Oncol. 2014;9:1345-1353 Crossref
  • 16 E.A. Punnoose, S. Atwal, W. Liu, et al. Evaluation of circulating tumor cells and circulating tumor DNA in non-small cell lung cancer: association with clinical endpoints in a phase II clinical trial of pertuzumab and erlotinib. Clin Cancer Res. 2012;18:2391-2401 Crossref
  • 17 M. Qiu, J. Wang, Y. Xu, et al. Circulating tumor DNA is effective for the detection of EGFR mutation in non-small cell lung cancer: a meta-analysis. Cancer Epidemiol Biomarkers Prev. 2015;24:206-212 Crossref
  • 18 H. Schwarzenbach, D.S. Hoon, K. Pantel. Cell-free nucleic acids as biomarkers in cancer patients. Nat Rev Cancer. 2011;11:426-437 Crossref
  • 19 K.L. Aung, R.E. Board, G. Ellison, et al. Current status and future potential of somatic mutation testing from circulating free DNA in patients with solid tumours. Hugo J. 2010;4:11-21 Crossref
  • 20 D.A. Cross, S.E. Ashton, S. Ghiorghiu, et al. AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated resistance to EGFR inhibitors in lung cancer. Cancer Discov. 2014;4:1046-1061 Crossref
  • 21 J.C. Yang, M.J. Ahn, D.W. Kim, et al. Osimertinib in pretreated T790M-positive advanced non-small-cell lung cancer: AURA study phase II extension component. J Clin Oncol. 2017;12:1288-1296
  • 22 G. Goss, C.-M. Tsai, F.A. Shepherd, et al. Osimertinib for pretreated EGFR Thr790Met-positive advanced non-small-cell lung cancer (AURA2): a multicentre, open-label, single-arm, phase 2 study. Lancet Oncol. 2016;17:1643-1652 Crossref
  • 23 E.L. Chin, C. da Silva, M. Hegde. Assessment of clinical analytical sensitivity and specificity of next-generation sequencing for detection of simple and complex mutations. BMC Genet. 2013;14:6 Crossref
  • 24 U.S. Food and Drug Administration. FDA summary of safety and effectiveness data (SSED). Premarket approval (PMA) P120019, cobas EGFR Mutation Test v2. http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpma/pma.cfm?id=P120019S007. Accessed July 8, 2016.
  • 25 C. Karlovich, J.W. Goldman, J.M. Sun, et al. Assessment of EGFR mutation status in matched plasma and tumor tissue of NSCLC patients from a phase I study of rociletinib (CO-1686). Clin Cancer Res. 2016;22:2386-2395 Crossref
  • 26 A.G. Sacher, C. Paweletz, S.E. Dahlberg, et al. Prospective validation of rapid plasma genotyping for the detection of EGFR and KRAS mutations in advanced lung cancer. JAMA Oncol. 2016;2:1014-1022 Crossref
  • 27 G.R. Oxnard, K.S. Thress, R.S. Alden, et al. Association between plasma genotyping and outcomes of treatment with osimertinib (AZD9291) in advanced non-small-cell lung cancer. J Clin Oncol. 2016;34:3375-3382 Crossref
  • 28 T.K. Sundaresan, L.V. Sequist, J.V. Heymach, et al. Detection of T790M, the acquired resistance EGFR mutation, by tumor biopsy versus noninvasive blood-based analyses. Clin Cancer Res. 2016;22:1103-1110 Crossref
  • 29 H.A. Wakelee, S.M. Gadgeel, J.W. Goldman, et al. Epidermal growth factor receptor (EGFR) genotyping of matched urine, plasma and tumor tissue from non-small cell lung cancer (NSCLC) patients (pts) treated with rociletinib. J Clin Oncol. 2016;34(suppl):9001 [abstract]
  • 30 C. Chouaid, C. Dujon, P. Do, et al. Feasibility and clinical impact of re-biopsy in advanced non small-cell lung cancer: a prospective multicenter study in a real-world setting (GFPC study 12-01). Lung Cancer. 2014;86:170-173 Crossref
  • 31 L.A. Diaz Jr., A. Bardelli. Liquid biopsies: genotyping circulating tumor DNA. J Clin Oncol. 2014;32:579-586 Crossref
  • 32 H.J. Yoon, H.Y. Lee, K.S. Lee, et al. Repeat biopsy for mutational analysis of non-small cell lung cancers resistant to previous chemotherapy: adequacy and complications. Radiology. 2012;265:939-948 Crossref
  • 33 A. Agarwal, D. Ressler, G. Snyder. The current and future state of companion diagnostics. Pharmacogenomics Pers Med. 2015;8:99-110 Crossref
  • 34 N. Normanno, M.G. Denis, K.S. Thress, M. Ratcliffe, M. Reck. Guide to detecting epidermal growth factor receptor (EGFR) mutations in ctDNA of patients with advanced non-small-cell lung cancer. Oncotarget. 2016;8:12501-12516
  • 35 G. Zhu, X. Ye, Z. Dong, et al. Highly sensitive droplet digital PCR method for detection of EGFR-activating mutations in plasma cell-free DNA from patients with advanced non-small cell lung cancer. J Mol Diagn. 2015;17:265-272 Crossref
  • 36 T.K. Yung, K.C. Chan, T.S. Mok, J. Tong, K.F. To, Y.M. Lo. Single-molecule detection of epidermal growth factor receptor mutations in plasma by microfluidics digital PCR in non-small cell lung cancer patients. Clin Cancer Res. 2009;15:2076-2084 Crossref

Footnotes

a AstraZeneca, Alderley Park, United Kingdom

b National Taiwan University Hospital, Taipei, Republic of China

c Emory University, Winship Cancer Institute, Atlanta, Georgia

d Roche Molecular Systems, Inc., Pleasanton, California

e AstraZeneca, Cambridge, United Kingdom

f Dana-Farber Cancer Institute, Boston, Massachusetts

g Kindai University Faculty of Medicine, Osaka-Sayama, Japan

h Ottawa Hospital Research Institute, Centre for Cancer Therapeutics, Ottawa, Canada

Corresponding author. Address for correspondence: Suzanne Jenkins, DPhil, AstraZeneca, Alderley Park, Cheshire, SK10 4TG, United Kingdom.

Readers of this article may receive CME credit. Further information can be found at https://www.iaslc.org/journal-based-cme.

Disclosure: Dr. Jenkins, Mrs. Weston, Miss Hodge, and Dr Cantarini are current or recent employees and shareholders of AstraZeneca. Dr. Goss reports honoraria from Pfizer, Bristol-Myers Squibb, AstraZeneca, Celgene, and Lilly. Dr. Jänne reports grants from Astellas Pharmaceuticals and AstraZeneca during the conduct of the study, as well as personal fees from AstraZeneca, Boehringer Ingelheim, Pfizer, Merrimack, Roche/Genentech, Chugai, AceaBiosciences, Ariad Pharmaceuticals, Ignyta, and LOXO Oncology outside the submitted work. In addition, Dr. Jänne has a patent owned by LabCorp with royalties paid. Dr. Mitsudomi reports personal fees from AstraZeneca during the conduct of the study, as well as grants and personal fees from Boehringer Ingeheim, Chugai, Pfizer, Ono Pharmaceutical, Eli Lilly, and Taiho and personal fees from Bristol-Myers Squibb and Merck Sharp and Dohme outside the submitted work. Dr. Patel reports being an employee and shareholder of AstraZeneca and former employee of Qiagen Manchester Ltd. outside the submitted work. Dr. Ramalingam reports personal fees from AstraZeneca, AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Novartis, Genentech, and Lilly outside the submitted work. Dr. Yang reports personal fees from Boehringer Ingelheim, AstraZeneca, Roche/Genentech/Chugai, Eli Lilly, Clovis Oncology, Pfizer, Novartis, Merck Sharp and Dohme, Merck Serono, Ono Pharmaceutical, Astellas, Bayer, Yuhan Pharmaceutical, and Celgene outside the submitted work. The remaining authors declare no conflict of interest.