Abstract
Purpose: Reliable detection of drug-sensitive activating EGFR mutations is critical in the care of advanced non–small cell lung cancer (NSCLC), but such testing is commonly performed using a wide variety of platforms, many of which lack rigorous analytic validation.
Experimental Design: A large pool of NSCLC cases was assayed with well-validated, hybrid capture–based comprehensive genomic profiling (CGP) at the request of the individual treating physicians in the course of clinical care for the purpose of making therapy decisions. From these, 400 cases harboring EGFR exon 19 deletions (Δex19) were identified, and available clinical history was reviewed.
Results: Pathology reports were available for 250 consecutive cases with classical EGFR Δex19 (amino acids 743–754) and were reviewed to assess previous non-hybrid capture–based EGFR testing. Twelve of 71 (17%) cases with EGFR testing results available were negative by previous testing, including 8 of 46 (17%) cases for which the same biopsy was analyzed. Independently, five of six (83%) cases harboring C-helical EGFR Δex19 were previously negative. In a subset of these patients with available clinical outcome information, robust benefit from treatment with EGFR inhibitors was observed.
Conclusions: CGP identifies drug-sensitive EGFR Δex19 in NSCLC cases that have undergone prior EGFR testing and returned negative results. Given the proven benefit in progression-free survival conferred by EGFR tyrosine kinase inhibitors in patients with these alterations, CGP should be considered in the initial presentation of advanced NSCLC and when previous testing for EGFR mutations or other driver alterations is negative. Clin Cancer Res; 22(13); 3281–5. ©2016 AACR.
Prospective testing for EGFR alteration in the course of clinical care is a critical component of NSCLC management, as EGFR inhibitors have been shown to confer great clinical benefit. In this series, hybrid capture–based comprehensive genomic profiling (CGP) utilizing next-generation sequencing detected EGFR exon 19 deletions that were not identified by standard focused molecular testing, leading to changes in therapy. The missed clinical benefit rate for patients with classical 743–754 deletions, as well as the less common C-helical deletions, is 17% and 83%, respectively, in this series. These patients are likely to benefit from EGFR inhibitors, given the extensive evidence for the clinical efficacy of EGFR inhibitors for patients with EGFR-mutant lung carcinoma and as reported here with patient examples. CGP performed in the course of clinical care should be considered for patients with NSCLC testing wild-type for EGFR and other oncogenic drivers by hotspot molecular testing.
Introduction
Activating mutations in the EGFR were the first clinically relevant genomic biomarkers identified in non–small cell lung cancer (NSCLC) and have been reported in 10%–30% of patients with NSCLC. The majority of EGFR mutations (45%–50%) in NSCLC are deletions in exon 19 (Δex19), clustering in the 746–750 amino acid range. EGFR exon 19 deletion mutations are well characterized, known to be activating, and are associated with responses to treatment with EGFR tyrosine kinase inhibitors (TKI) in approximately 75% of cases, with median progression-free survival exceeding 1 year in many series (1–3). Exon 19 deletions within the C-helix of EGFR (amino acids 753–761) have been identified at significantly lower frequencies but have also been reported to respond to both reversible and irreversible EGFR TKIs (4–6). Methodology for assessing EGFR alterations in clinical specimens is left to laboratory discretion, and assay identity and importantly, limitations and performance characteristics, is typically not readily apparent to the treating physician. Given the large clinical benefit demonstrated for EGFR TKIs in patients with NSCLC whose tumors harbored EGFR mutations, assessing the limitations of EGFR testing typically used in clinical care is essential for the thoracic oncology and pathology communities (7).
To this end, a comprehensive review of NSCLC cases harboring EGFR Δex19 assayed in the course of clinical using a hybrid capture–based comprehensive genomic profiling (CGP) assay was conducted for both history of prior testing for EGFR mutations, as well as available response data to treatment with EGFR-TKIs in a PHI compliant fashion for this subset of patients.
Materials and Methods
DNA was extracted from 40 μm of formalin-fixed paraffin-embedded sections, and CGP was performed on hybridization captured, adaptor ligation–based libraries to a mean coverage depth of 678X for at least 3,769 exons of 236 cancer-related genes plus 47 introns from 19 genes frequently rearranged in cancer. CGP is defined as a well-validated molecular assay and interpretive procedure that simultaneously sequences the entire coding sequence of genes known to be somatically altered and biologically relevant in human cancer, including all clinically relevant classes of genomic alterations (i.e., base pair substitutions, insertions/deletions, copy number alterations, and rearrangements), and then matches clinically relevant alterations to targeted treatment options (8, 9). Results from prior molecular EGFR testing using “hotspot” (assaying for common EGFR alterations without additional broader interrogation of the gene) PCR-based assays were obtained when available by review of available medical records.
Results
From a larger series of NSCLC cases assayed with CGP in the course of clinical care, 400 consecutive cases harboring EGFR Δex19 alterations were reviewed. In each of these cases, the EGFR mutation identified was the sole activating EGFR alteration, and 96.5% were in the classical 743–754 amino acid range, including 62.5% of cases with the canonical E746_A750del mutation. Fourteen cases harbored an EGFR Δex19 within the C-helix of EGFR exon 19 (amino acids 753–761), and the majority of these deletions were amino acids 752–759 (Table 1; Fig. 1).
Frequency of EGFR exon 19 deletion mutations identified by CGP
| EGFR Mutation | No. of cases (%) |
|---|---|
| EGFR classical Δex19 range (aa 743–754) | 386 (96.5) |
| E746_A750del | 250 (62.5) |
| EGFR Δex19 C-helix (aa 753–761) | 14 (3.5) |
| S752_I759del | 8 (2.0) |
| A750_E758>P | 1 (0.3) |
| T750_L759>NLD | 1 (0.3) |
| T751_L795>N | 1 (0.3) |
| T751_L759>NC | 1 (0.3) |
| T751_L760>NL | 2 (0.5) |
| EGFR Mutation | No. of cases (%) |
|---|---|
| EGFR classical Δex19 range (aa 743–754) | 386 (96.5) |
| E746_A750del | 250 (62.5) |
| EGFR Δex19 C-helix (aa 753–761) | 14 (3.5) |
| S752_I759del | 8 (2.0) |
| A750_E758>P | 1 (0.3) |
| T750_L759>NLD | 1 (0.3) |
| T751_L795>N | 1 (0.3) |
| T751_L759>NC | 1 (0.3) |
| T751_L760>NL | 2 (0.5) |
Consort diagram detailing 400 NSCLC cases with EGFR exon 19 deletions identified by CGP. #, In 46 of these 71 cases, we confirmed that the same specimen was tested by prior non-hybrid–based capture testing and by CGP. Of these, eight samples had prior negative EGFR test results, giving a true “false-negative” rate of 8/46 cases (17%).
Consort diagram detailing 400 NSCLC cases with EGFR exon 19 deletions identified by CGP. #, In 46 of these 71 cases, we confirmed that the same specimen was tested by prior non-hybrid–based capture testing and by CGP. Of these, eight samples had prior negative EGFR test results, giving a true “false-negative” rate of 8/46 cases (17%).
Of 14 NSCLC cases with an EGFR Δex19 C-helical deletion, previous non-hybrid capture–based EGFR sequencing results were available for six cases, and of these, five (83%) had negative prior testing (Table 2, patients 1–5). Clinical history was provided by the treating physician for one of these cases. Case #1 is a 44-year-old never-smoker female diagnosed with lung adenocarcinoma who presented with lymph node and intracranial metastases. Radiotherapy resulted in significant improvement in the intracranial disease, and a cervical node excisional biopsy was sent for focused molecular testing. As reported by a commercially available “hotspot” test, the patient was “pan-negative” with a report of no mutations in EGFR exons 19–21, or kinase fusions. The patient received two cycles of carboplatin and pemetrexed with no response. The previously tested specimen was submitted for CGP and EGFR T751_I759>N was observed. The patient had an early partial response to afatinib with significant shrinkage of all measurable lesions that persisted for 8 months (Table 2 patient 3; Fig. 2). On relapse, a biopsy of the progressive disease was performed, and the resulting specimen was also submitted for CGP. EGFR T751_I759>N was present, but this postprogression specimen also harbored EGFR T790M. On the basis of this genomic profile, the patient is currently enrolled in a trial for a third-generation EGFR TKI, as this class of agents have activity against EGFR T790M (10, 11).
Characteristics of cases with prior negative EGFR test results
| Patient no. | EGFR mutation detected by CGP | Previous EGFR test result | Same specimen tested by prior EGFR assay and CGP | Additional genes negative by hotspot testing | Response to EGFR targeted therapy |
|---|---|---|---|---|---|
| 1 | S752_I759del | Negative | N | ALK | NA |
| 2 | T750_L759>NLD | Negative | NA | ALK, KRAS, PIK3CA | NA |
| 3 | T751_I759>N | Negative | Y | ALK, ROS1 | PR, afatinib |
| 4 | T751_L760>NL | Negative | N | NA | |
| 5 | T751_L760>NL | Negative | Y | ALK | NA |
| 6 | E746_A750del | Negative | Y | ALK, KRAS | PR, erlotinib |
| 7 | E746_A750del | Negative | Y | ALK, ROS1 | NA |
| 8 | E746_A750del | Negative | N | NA | |
| 9 | E746_A750del | Negative | N | NA | |
| 10 | E746_A750del | Negative | Y | KRAS | NA |
| 11 | E746_A750del | Negative | N | NA | |
| 12 | E746_A750del | Negative | Y | NA | |
| 13 | E746_A750del | Negative | Y | NA | |
| 14 | E746_A750del | Negative | Y | ALK, KRAS | Afatiniba |
| 15 | L747_A750>P | Negative | Y | ALK, KRAS, BRAF | NA |
| 16 | L747_A750>P | Negative | N | NA | |
| 17 | L747_K754>G | Negative | Y | ALK, KRAS | NA |
| Patient no. | EGFR mutation detected by CGP | Previous EGFR test result | Same specimen tested by prior EGFR assay and CGP | Additional genes negative by hotspot testing | Response to EGFR targeted therapy |
|---|---|---|---|---|---|
| 1 | S752_I759del | Negative | N | ALK | NA |
| 2 | T750_L759>NLD | Negative | NA | ALK, KRAS, PIK3CA | NA |
| 3 | T751_I759>N | Negative | Y | ALK, ROS1 | PR, afatinib |
| 4 | T751_L760>NL | Negative | N | NA | |
| 5 | T751_L760>NL | Negative | Y | ALK | NA |
| 6 | E746_A750del | Negative | Y | ALK, KRAS | PR, erlotinib |
| 7 | E746_A750del | Negative | Y | ALK, ROS1 | NA |
| 8 | E746_A750del | Negative | N | NA | |
| 9 | E746_A750del | Negative | N | NA | |
| 10 | E746_A750del | Negative | Y | KRAS | NA |
| 11 | E746_A750del | Negative | N | NA | |
| 12 | E746_A750del | Negative | Y | NA | |
| 13 | E746_A750del | Negative | Y | NA | |
| 14 | E746_A750del | Negative | Y | ALK, KRAS | Afatiniba |
| 15 | L747_A750>P | Negative | Y | ALK, KRAS, BRAF | NA |
| 16 | L747_A750>P | Negative | N | NA | |
| 17 | L747_K754>G | Negative | Y | ALK, KRAS | NA |
NOTE: Cases 1–5 represent EGFR exon 19 deletions affecting the C-helix.
Abbreviation: NA, clinical history or prior biopsy specimen information was not available.
aPatient passed away before beginning treatment with afatinib.
Afatinib response in a patient with EGFR T751_I759>N mutation. Patient 3 in Table 2. Pretreatment (left) and 7-month scan following treatment with afatinib (right).
Afatinib response in a patient with EGFR T751_I759>N mutation. Patient 3 in Table 2. Pretreatment (left) and 7-month scan following treatment with afatinib (right).
Pathology reports for 250 consecutive NSCLC cases harboring classic EGFR Δex19 (amino acids 743–754) identified by CGP were systematically reviewed. Of these, previous EGFR test results were available for 71 cases, and 12 of 71 (17%) had previously tested negative for EGFR mutation (Table 2, patients 6–17). For a subset of cases, we confirmed that the same specimen was tested both by a prior non-hybrid capture–based assay and by CGP, and 8 of 46 (17%) of those cases were determined to be “false negatives” by prior testing. A detailed clinical history was available for selected cases via discussion with the treating physicians in the context of appropriate consent.
Case #2 was that of a 55-year-old male with lung adenocarcinoma with negative results for EGFR, ALK, and KRAS using a commercially available “hotspot” test. The patient underwent six lines of treatment including carboplatin/paclitaxel/bevacizumab, pemetrexed, carboplatin/gemcitabine, erlotinib, cisplatin/vinorelbine/cetuximab, and docetaxel. The patient had a response to fourth-line empiric erlotinib treatment, lasting for 9 months. After a sixth line of therapy, the same specimen from the primary tumor as was tested previously was demonstrated to harbor EGFR E746_A750del via CGP (Table 2, patient 6). This case is described in detail as part of another report in which an additional EGFR false-negative case was also identified by CGP, and multiple lines of empiric therapy were administered (12).
Case #3 was a 57-year-old never-smoker male diagnosed with advanced lung adenocarcinoma, which tested negative for mutation of EGFR and KRAS and ALK rearrangement using a commercial laboratory hotspot test. The patient received six cycles of carboplatin and pemetrexed, followed by maintenance pemetrexed, but the disease remained active. One year following initial diagnosis, the patient sought a second opinion, at which point the physician submitted the same specimen that was tested previously for CGP, revealing the tumor harbored EGFR E746_A750del. The patient passed away before initiating EGFR TKI therapy (Table 2, case 14).
Discussion
The use of rationally applied targeted therapy has revolutionized the care of NSCLC, beginning with the identification of EGFR-mutant lung cancers that respond to erlotinib, and more recently the identification of kinase fusions that responds to tyrosine kinase inhibitors, such as EML4-ALK rearranged NSCLC. For such therapies to be optimally delivered, there is an inherent mandate for specific and sensitive clinical testing that detects genomic alterations (GA) that could affect the use of targeted therapy. A recent study from Drilon and colleagues demonstrated that CGP applied prospectively in the course of clinical care can identify GA that guide targeted therapy for a group of NSCLC patients who has previously tested “negative” by standard-of-care molecular testing, that is, EGFR and ALK were assessed as wild-type (13). CGP identified GA such as RET and ROS fusions that led to patients receiving alteration-matched therapy. Interesting, although this study was small, two previously “negative” cases, one each of EGFR and ALK alterations, were identified. This study was limited, however, by a small population who was treated at an elite, tertiary medical center.
The current study demonstrates that for a population of NSCLC patients treated at multiple institutions, hybrid capture–based CGP can detect EGFR Δex19 where commercial non-hybrid capture hotspot tests were previously negative. Specifically, CGP revealed that EGFR Δex19 alterations were missed by omission in 83% (5/6) of patients with deletions affecting the C-helix who had previously tested negative for EGFR mutation. This result highlights the importance of using CGP with full exon coverage as opposed to next-generation sequencing assays that are limited to small regions of the gene, and thus may miss less well characterized but still actionable alterations. Prior negative results were also reported in 12 of 71 of patients with more common EGFR Δex19 alterations, including the classic E746_A750del mutation, representing a missed clinical benefit rate of 17%. Consistent with the EGFR Δex19 being a well-characterized oncogenic driver, these patients lacked other oncogenic drivers, such as ALK and ROS1 fusions, and KRAS mutation (Table 2). In 8 of 46 of these cases with classical EGFR Δex19 alterations, we confirmed that the same biopsy was tested by CGP and by prior non-hybrid–based capture testing, giving a true false-negative rate of 17% in this study.
The cases reported here suggest that EGFR Δex19 cases as detected only by CGP likely obtain clinical benefit from erlotinib consistent with historical means. Moreover, when such patients progress on erlotinib, the advent of therapies with activity against T790M, or other therapy paradigms for acquired resistance, suggests additional options (10, 11). The possibility of a long natural history for these EGFR-mutant patients adds weight to the imperative for reliable detection of EGFR exon 19 deletions to allow what is often an exceptional outcome in the setting of advanced disease.
Published guidelines strongly encourage of the use of EGFR mutational assays, but without explicit mention of the needed qualities of an analytic validation, and also do not require the peer-reviewed publication of detailed validation data. Several hypotheses might explain the striking rate at which PCR-based EGFR mutation tests fail to detect activating EGFR alterations. Such commercial non-hybrid capture–based hotspot assays generally require a proportion of tumor nuclei to normal nuclei of 50% or greater to reduce the risk of false negatives (7), but many clinical specimens do not harbor this great a proportion of tumor nuclei and thus could yield false negatives. However, among the 77 EGFR Δex19 cases with previous test results in this study, there was no significant difference in estimated tumor purity between specimens assessed by CGP, as the percentage of tumor nuclei in EGFR prior negative samples of 33% ±17% compared with 35% ±19% tumor nuclei in samples for which EGFR mutation was detected by a hotspot test. Given this observation, insufficient tumor purity does not account for the high rate of discordance between the non-hybrid capture hotspot tests and CGP in this study. Another hypothesis explaining the false negative rate is that of intratumoral heterogeneity. In 8 of the 12 cases described here for which prior testing did not detect a classic EGFR Δex19, we confirmed that the hotspot testing and the hybrid capture–based CGP were performed on the same specimen. For the remaining four cases, CGP was performed on a different specimen, including one case where the samples were collected at the same time, but from different sites, and another case where hotspot testing was performed on a sample with low cellularity and reported a negative result, prompting the treating physician to request that another sample be collected and submitted for CGP. In general, oncogenic drivers have been demonstrated to be highly concordant in primary and metastatic biopsies across a variety of tumor types (14–17). Explanation of the discordance between hotspot results and CGP observed here awaits further analysis.
The results of this study highlight the importance of using a hybrid capture–based CGP assay that employs full coverage sequencing to allow for sensitive detection of clinically relevant mutations. Significant and durable responses following treatment with EGFR-targeted therapies in patients with EGFR Δex19 not detected by hotspot testing but identified by CGP highlight the need for consistent accurate detection of these alterations by a well-validated sensitive assay. Importantly, understanding the limitations inherent to performance characteristics of the deployed assay is critical to the management of patients with NSCLC, particularly when hotspot molecular testing is negative. The cases presented here and elsewhere suggest that comprehensive, adequately sensitive, and well-validated testing for EGFR Δex19 is absolutely essential, as it can lead directly to well-established clinical benefit for the patient.
Disclosure of Potential Conflicts of Interest
A.B. Schrock, J.A. Elvin, R. Erlich, J.S. Ross, G.M. Frampton, J. Greenbowe, D. Lipson, R. Yelensky, Z. Chalmers, J. Chmielecki, P.J. Stephens, S.M. Ali and V.A. Miller have ownership interests (including patents) in Foundation Medicine. M. Wollner is a consultant/advisory board member for Boehringer Ingelheim. F. Braiteh reports receiving speakers bureau honoraria from Amgen, Bayer, Bristol-Myers Squibb, Celgene, Incyte, Ipsen, Insys, Pfizer and Sanofi, and is a consultant/advisory board member for Amgen, Astra Zeneca, Bristol-Myers Squibb, INYSIS and Pfizer. J.S. Ross reports receiving commercial research grants from and has ownership interest (including patents) in Foundation Medicine. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
Conception and design: A.B. Schrock, J. Chmielecki, J.A. Elvin, N. Peled, F. Braiteh, S.-H.I. Ou, J.S. Ross, S.M. Ali, V.A. Miller
Development of methodology: A.B. Schrock, G.M. Frampton, R. Yelensky, Z. Chalmers, J.S. Ross, S.M. Ali, V.A. Miller
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.B. Schrock, G.M. Frampton, J.A. Elvin, M. Wollner, A. Dvir, L. Soussan-Guttman, R. Bordonii, N. Peled, L.E. Raez, S.-H.I. Ou, M. Mohamed, J.S. Ross, S.M. Ali, V.A. Miller
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.B. Schrock, G.M. Frampton, J. Greenbowe, K. Wang, D. Lipson, Z. Chalmers, F. Braiteh, S.-H.I. Ou, J.S. Ross, P.J. Stephens, S.M. Ali
Writing, review, and/or revision of the manuscript: A.B. Schrock, G.M. Frampton, D. Herndon, K. Wang, J. Chmielecki, J.A. Elvin, R. Bordonii, N. Peled, F. Braiteh, L.E. Raez, R. Erlich, S.-H.I. Ou, M. Mohamed, J.S. Ross, P.J. Stephens, S.M. Ali, V.A. Miller
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G.M. Frampton, K. Wang, Z. Chalmers, J.S. Ross, S.M. Ali
Study supervision: F. Braiteh, R. Erlich, J.S. Ross, S.M. Ali
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
RSS Feeds