Brief Report: S768I mutation in EGFR in patients with lung cancer

Abstract

Introduction

Epidermal growth factor receptor (EGFR) mutations are relatively common oncogenic drivers in non-small cell lung cancer (NSCLC). The outcomes of patients who present with less common EGFR mutations or more than one EGFR mutation are uncertain. We reviewed our experience with the S768I mutation of exon 20 of EGFR in order to provide insight to the clinical significance of this mutation.

Methods

We searched our electronic medical record system using a natural language search program and every EGFR mutation analysis performed in our Department of Molecular Genetics of patients with NSCLC treated at Mayo Clinic to identify patients with EGFR S768I mutation. Relevant clinical and laboratory data were abstracted for selected cases, including evaluation of response after treatment with TKIs.

Results

1,527 cases of NSCLC that underwent EGFR testing were reviewed and the S768I mutation was present in 9 patients (0.59%), 4 of whom were female. Only 3 cases had an isolated S768I mutation, 4 cases had a concurrent G719 mutation and 2 cases had a concurrent L858R mutation. Among patients with stage IV disease treated with erlotinib (n=4), one had isolated S768I mutation and 3 had additional (2 patients with G719 and 1 patient with L858R) mutations. The tumor response of patients with stage IV disease was highly variable to erlotinib (progression free survival: 3-30months and overall survival 5 to more than 51 months).

Conclusion

S768I mutations in exon 20 of the EGFR are rare and are commonly seen in conjunction with sensitizing EGFR mutations. Due to its rarity and the variability of responses of treated cases, its exact prognostic and predictive role is not fully understood. In our experience, S768I mutations in isolation do not necessarily confer sensitivity to erlotinib but in conjunction with sensitizing EGFR mutations, S768I mutations do not restrict efficacy.

Introduction

Non-small-cell lung cancers (NSCLCs) are the most prevalent form of lung cancer and an understanding of their oncogenic drivers has led to the development of effective molecularly-targeted therapeutics. In particular, mutations in epidermal growth factor receptor (EGFR) and translocations involving anaplastic lymphoma kinase (ALK) or the proto-oncogene tyrosine kinase c-ROS1 (ROS1) play an important role in carcinogenesis of many lung cancers. Alterations in these actionable kinases predict the clinical effectiveness of tyrosine kinase inhibitors (TKIs) that target these genetic aberrations[1].

The most frequent, and clinically-significant mutations in EGFR include in-frame deletions involving amino acids LREA of exon 19, and the exon 21 substitution L858R. Together, these two “classic” mutations account for almost 85% of all EGFR mutations and are associated with favorable outcome when patients are treated with EGFR TKIs[2]. Nevertheless, there are a multitude of other somatic mutations in EGFR that have been described in NSCLC samples, and most cluster within the tyrosine kinase domain. Some of these mutations are associated with enhanced efficacy of EGFR TKIs [3] whereas others are not [4].

It is not uncommon for patients to have compound (i.e., double or complex) EGFR mutations. In different cohorts the frequencies of compound EGFR mutations range from less than 4% to 17% of patients.[5-7] Therefore, it seems that the frequency of EGFR compound mutations is neither fully elucidated nor insignificant. The pattern of response of the patients with compound mutations is not fully understood.

At Mayo Clinic, we have been testing for EGFR mutations in all patients with advanced NSCLC and that gives us the opportunity to try to further investigate the role of these less common mutations. Herein, we review our experience with S768I, a missense mutation in exon 20 resulting in substitution of serine for isoleucine in order to provide further insight into its prevalence and clinical significance.

Methods

We identified patients with an EGFR S768I mutation treated at Mayo Clinic through searching our electronic medical record system using a natural language search program and every EGFR mutation analysis performed in our Department of Molecular Genetics of patients with NSCLC. Relevant clinical and laboratory data were abstracted for these cases, including evaluation of response after treatment with TKIs using the Response Evaluation Criteria In Solid Tumors (RECIST) version 1.1 [8]. Pathological diagnosis and clinical staging (according to American Joint Committee on Cancer Guidelines[9]) was performed by a pathologist and oncologist, respectively, in blinded fashion. Inclusion criteria were cases with EGFR S768I mutation seen at Mayo Clinic before December 2014. EGFR testing was performed following microscopic examination by a pathologist to identify and select areas of tumor for macrodissection and confirm sufficient tumor percent. The EGFR test is a PCR based assay employing allele specific amplification and is used to test for the presence mutations within exons 18-21 of the EGFR gene, most recently using the FDA approved platform. When Sanger sequencing confirmation was done, universally tagged sequencing primers were designed using IDT software (Coralville, IA) and SNPCheck (NGRL, Manchester, UK). Genomic DNA was amplified by singleplex PCR using the FailSafe PCR System (Epicentre, Madison, WI). Unincorporated primers and nucleotides were removed using the Ampure XP magnetic bead purification kit (Beckman Coulter, Brea, CA). Amplicons were bidirectionally sequenced using Big Dye Terminator version 1.1 technology on an ABI 3730 system (Applied Biosystems, Foster City, CA). Sequence analysis was performed using Mutation Surveyor software version 4.0.9 (SoftGenetics, State College, PA), Alamut Visual, and visual inspection. When testing was performed elsewhere, bidirectional sequencing was most commonly used. This study was approved by the Institutional Review Board of Mayo Clinic.

Results

There were 1,527 cases of NSCLC that underwent EGFR testing and the S678I mutation was present in 9 of these patients (0.59%), 4 of which were female. Median age at diagnosis was 61 years (range 49-68 years), 5 patients were never smokers and no subjects were current smokers at the time of diagnosis. The stage of disease at diagnosis was stage I for 2 patients, stage III for 3 patients and stage IV for 4 patients. All specimens were adenocarcinomas with 5 of them being grade 3. Three cases had an isolated S768I mutation, 4 cases had a concurrent G719 mutation and 2 cases had a concurrent L858R mutation. Patients with metastatic disease were treated with erlotinib and their tumor responses are summarized in Table 1. Disease control was achieved in three of the four patients (see Table, Supplemental Digital Content 1, which illustrates the RECIST 1.1 measurements of the lesions of the patients), and the one patient who progressed immediately on therapy had an isolated S768I mutation. One patient, not included in Table 1, with a concurrent S768I and L858R mutation with stage IIIa disease received curative intent lobectomy after neoadjuvant treatment with erlotinib (Figure 1). His resected specimen identified the S768I mutation in addition to an exon 19 deletion, suggesting that this patient had multiple clones, although both detected clones had S768I present (See Figure, Supplemental Digital Content 2, which illustrates the Sanger sequencing confirmation of individual sample with EGFR G719C, S768I, and L858R mutations). This patient developed metastatic disease to his bones after adjuvant chemotherapy therapy. Erlotinib was generally well tolerated with the most common side effects being rash and fatigue and was only discontinued in one case due to fatigue.

Table 1.

Response of S768I mutations to erlotinib by RECIST 1.1 in patients with stage IV disease at diagnosis

Mutation(s)	Best response on erlotinib	PFS (months)	Overall survival (months)
S768I alone	Progressive disease	3	5
S768I + G719	Partial response	6	23
S768I + G719S±	Stable disease	12	33
S768I + L858R	Stable disease	30	51+

^±This patient's tissue was tested outside Mayo Clinic with bidirectional direct DNA sequencing.

Figure 1.

Radiologic response of a patient with stage IIIa adenocarcinoma to neoadjuvant treatment with erlotinib that allowed for curative intent lobectomy. A. CT of the chest at diagnosis. B. CT of the chest after 2 months of treatment with erlotinib.

Discussion

S768I mutations in exon 20 of the EGFR are rare (0.49% in our series) and are commonly seen in conjunction with common EGFR mutations. Patients with this mutation are reported to have a favorable response to erlotinib [10] or to be resistant to it [11]. In our study, we observed a highly variable tumor response to erlotinib in patients with stage IV disease (progression free survival: 3-30months and overall survival 5 to more than 51 months).

Meta-analyses have recently explored the clinical significance of less common EGFR mutations. One analysis of patients with rare EGFR mutations treated with erlotinib in 7 trials (MERIT, SATURN, TITAN, TRUST, ATLAS, BeTa, and FASTACT-2) identified 5 patients with S768I (2 with isolated mutation, 2 with a second exon 20 point mutation and 1 with compound exon 18 mutation) and suggested that S768I appeared to confer sensitivity to erlotinib. In this analysis, 3 patients had stable disease and two patients had partial response and their progression free survival (PFS) ranged between 83 and 371 days. Interestingly, the two patients with isolated S678I mutation had a PFS of 371 and 169 days. [12]. In a combined post-hoc analysis of LUX-Lung 2, LUX-Lung 3, and LUX-Lung 6, the investigators identified 8 cases of S768I mutation (1 isolated, 5 with G719 and 2 with L858R)[13]. In this analysis, objective response to afatinib was 100%, (95% CI 63.1-100.0), median PFS was 14.7 months in patients with S768I mutation and suggested that it had a favorable outcome. Similarly patients with exon 19 deletions or L858R mutation were found to have a median PFS of 11.0 or 13.8 months respectively in these studies.[14]

In our single institution study we saw variable responses to erlotinib in 4 patients with metastatic disease and S768I EGFR mutations. Our patient with an isolated S768I mutation had disease progression, whereas other patients with compound L858R or G719 substitutions retained sensitivity to erlotinib. Although our study is small, when seen in the context of other reports and trials, the clinical evidence suggests that S768I mutations in isolation do not necessarily confer sensitivity to erlotinib but in conjunction with sensitizing EGFR mutations, S768I mutations do not restrict efficacy of EGFR TKIs. A limitation of this study is the small size of the uncommon EGFR mutation cohort, which does not allow us to further clarify what is the optimal first line EGFR inhibitor for this mutation. Additionally, all of our patients were treated with erlotinib, so we are not certain as to what effect other TKIs such as afatinib would have on the S768I mutation. We strongly believe that the most effective way to help practicing oncologists to guide their patient's treatment is the implementation of a central molecular repository of these cases. CancerLinq, as proposed by the American Society of Clinical Oncology, is one such resource that may further our understanding of the clinical significance of rare mutations.