Identification of osimertinib-resistant EGFR L792 mutations by cfDNA sequencing: oncogenic activity assessment and prevalence in large cfDNA cohort
Cell-free DNA (cfDNA) next-generation sequencing has the potential to capture tumor heterogeneity and genomic evolution under treatment pressure in a non-invasive manner. Here, we report the detection of EGFR L792 mutations, a non-covalent mechanism of osimertinib resistance, using Guardant360 cfDNA testing in a patient with metastatic EGFR-mutant non-small cell lung cancer (NSCLC) whose disease progressed on osimertinib. We subsequently analyzed a large cohort of over 1800 additional patient samples harboring an EGFR T790M mutation and identified a concomitant L792 mutation in a total of 22 (1.2%) cases. In vitro functional assays demonstrated that the EGFR L858R/T790M/L792F/H mutations conferred intermediate-level resistance to osimertinib. Further understanding of potential acquired resistance mechanisms to targeted therapy may help inform treatment strategy in EGFR-mutant NSCLC.
KeywordsCell-free DNA cfDNA Genomic evolution Acquired resistance Osimertinib Guardant360 Functional studies
Epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) is a distinct molecular subtype with sensitivity to EGFR-selective tyrosine kinase inhibitors (TKIs) [1, 2, 3, 4]. However, tumors invariably develop resistance to these EGFR TKIs, mediated by on-target genetic alterations within the EGFR tyrosine kinase domain, EGFR-independent mechanisms, or small cell transformation [5, 6]. In initial reports of acquired resistance to first-generation EGFR TKIs erlotinib and gefitinib, 50–60% of cases harbored an EGFR T790M gatekeeper mutation [5, 6]. Osimertinib, an irreversible, third-generation EGFR inhibitor, was developed to target T790M mutation-positive, first-generation TKI-refractory tumors and demonstrated robust efficacy with objective response rates of 61–71% among T790M-positive NSCLC patients [7, 8, 9]. More recently, osimertinib became the new standard initial therapy in advanced EGFR-mutant NSCLC . Despite its efficacy, patients acquire resistance to osimertinib through various mechanisms including EGFR C797S mutations which eliminate the covalent bonding site for osimertinib, and amplification of MET or ERBB2 (HER2), among others [11, 12, 13]. The prevalence of C797S mutations may differ depending on the clinical setting and is more common in patients with a pre-existing T790M mutation [14, 15]. Serial assessment of the molecular characteristics of EGFR-mutant NSCLC with each line of therapy will assist in understanding the evolution of on- and off-target mechanisms of resistance and can help guide the development of new therapeutic strategies for patients with resistant disease.
Historically, tumor tissue biopsies have been standard for detection of resistance mechanisms. However, tissue biopsies are inevitably limited by their invasive procedural risk, high cost, treatment delays related to procedure and processing, and inability to capture spatial tumor heterogeneity. In contrast, plasma cell-free DNA (cfDNA) next-generation sequencing (NGS) from peripheral blood allows for safe, global, and repeated longitudinal assessment of mutation dynamics throughout the course of disease and treatment. Therefore, this approach has the potential to accelerate our understanding of TKI resistance.
Here, we report the detection of EGFR L792 resistance mutations via cfDNA sequencing in a patient progressing on osimertinib, their prevalence in a large clinically tested NSCLC cfDNA cohort, and in vitro functional characterization.
Given their genomic proximity, the T790M and C797S mutations were phased to determine allelic origin and found to be in cis, and the F795C mutation appeared on that allele. In contrast, the L792H and L792F variants were in cis to T790M but arose in trans to C797S and to each other. While multiple tissue biopsies over time were not available to determine the temporal sequence of mutational emergence, when mapped against the patient’s treatment history the clonal phylogeny of these EGFR alleles suggested that at least the L792H and L792F mutations arose during osimertinib treatment at the same branch point as the known osimertinib resistance mutation C797S (Fig. 1e). Moreover, structural modeling indicated that each mutation affects a residue that impinges on the ATP-binding pocket (Fig. 1f, g).
Prevalence in a large cfDNA cohort
Nonsynonymous EGFR L792 alterations co-occurring with EGFR T790M mutations identified in the Guardant360 database of patients with lung cancer
Number of patients
L792H and L792V
L792V and L792F
L792H and L792F
Besides the initial case described above, only one other patient was found to have a nonsynonymous EGFR F795 alteration in conjunction with an L792 mutation; this patient’s sample had 14 nonsynonymous EGFR alterations (Additional file 1: Table S2). One additional patient’s sample harbored an EGFR L792R alteration in the absence of a co-occurring EGFR T790M mutation; six other nonsynonymous EGFR alterations were detected in this sample (L717V, L718Q, G796S, C797S, G796R, and S1036R).
Phasing analysis was performed on 27 samples from the 22 unique patients containing an L792F/H/V/P/R or F795C/L mutation. As in the initial case described above, the L792 F/H/V/P/R and F795C/L alterations were invariably present subclonal to and frequently in cis with EGFR T790M, but independent of one another, C797S, and other osimertinib resistance alterations (Additional file 1: Table S3). The recurrence of these mutations across multiple patients supports the hypothesis that these variants confer a selective advantage compatible with osimertinib resistance. However, the relatively low frequency with which these variants are observed and lower VAFs at which they occur suggest that this advantage may be less potent than that conferred by C797S.
In this report, through clinical cfDNA NGS we identify EGFR L792 mutations in 22 of 1851 (1.2%) NSCLC patient cases with an EGFR T790M mutation. These L792 mutations appear to be a non-covalent mechanism of osimertinib resistance in which alterations in the EGFR ATP binding pocket diminish, but do not entirely prevent, osimertinib binding. In vitro assays suggest that increasing doses of osimertinib may overcome this resistance and inhibit EGFR activity, compatible with steric hindrance of drug binding or altered affinity to the drug or ATP rather than elimination of the binding site.
These results are consistent with recent reports of EGFR L792 mutations. Chen et al.  reported L792 mutations identified through cfDNA testing of plasma or pleural effusion in three patients with NSCLC progressing on osimertinib, with a follow-up study from the same group  identifying mutations at this residue in 11/93 (12%) of Chinese patients with osimertinib-resistant lung cancer.
There are inherent limitations to examining the prevalence of EGFR L792 mutations in the context of co-occurring T790M mutations. This approach was used due to the unavailability of treatment history details for genomic data from a commercial laboratory. With the recent approval of osimertinib for first-line use, this genomic context may not apply moving forward. Nishino et al.  found that EGFR L792 mutations in combination with L858R but in the absence of T790M conferred moderate resistance to osimertinib in vitro. Future studies examining the prevalence and functional effect of EGFR L792 mutations in the absence of T790M may clarify how broadly this data may be extrapolated in the dynamic landscape of drug approvals and treatment sequences.
Notably, the case described in detail above was found to have multiple EGFR mutations on cfDNA NGS, as did many other cases subsequently identified in the cohort prevalence analysis (Additional file 1: Table S2). The emergence of multiple alterations across the course of disease and treatment makes it increasingly difficult to delineate the isolated impact of any individual mutation in the acquired resistance process; this limitation of traditional analysis heightens the need for repeated comprehensive genomic profiling in the setting of clinical progression to capture the full context of changes under treatment pressure. The evolution of multiple on-target alterations underscores the complexity of the genomic landscape that can emerge in the setting of TKI resistance and highlights the importance of repeat genomic analysis, and in particular cfDNA NGS to non-invasively capture heterogeneous resistance, in detecting potentially targetable genomic alterations over the disease course.
SRF and LAK: data acquisition/analysis for patient case and cfDNA cohort, manuscript drafting. JJL, ATS, and TES: data acquisition/analysis for patient cases. OZ and GT: functional studies. JO, RBL, and RJN: data acquisition/analysis for the cfDNA cohort. All authors contributed to the conception and design of the work and edited the final manuscript. All authors read and approved the final manuscript.
The authors have no funding sources to disclose.
Ethics approval and consent to participate
This study has obtained appropriate institutional review board approval for analysis of deidentified and limited data sets which waived the need for individual patient informed consent.
Consent for publication
SRF, AK, RBL, and RJN are employees and shareholders of Guardant Health, Inc. OZ and GT are employees of NovellusDx. JJL received honoraria from or served as a consultant for Chugai and Boehringer-Ingelheim and received institutional research funding from Loxo Oncology. ATS has served as a compensated consultant or received honoraria from Pfizer, Novartis, Genentech/Roche, Ariad/Takeda, Ignyta, LOXO, Bayer, Chugai, Blueprint Medicines, KSQ Therapeutics, Daiichi Sankyo, EMD Serono, Taiho Pharmaceutical, TP Therapeutics, Foundation Medicine, Guardant, Natera, Servier, and Syros; has received institutional research funding from Pfizer, Novartis, Roche/Genentech, Ariad, Ignyta, and TP Therapeutics; and has received travel support from Pfizer and Roche/Genentech. The remaining authors do not report relevant competing interests.
- 2.Rosell R, Carcereny E, Gervais R, Vergnenegre A, Massuti B, Felip E, 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;13:239–46.CrossRefGoogle Scholar
- 3.Zhou C, Wu Y-L, Chen G, Feng J, Liu X-Q, Wang C, et al. Erlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): a multicentre, open-label, randomised, phase 3 study. Lancet Oncol. 2011;12:735–42.CrossRefGoogle Scholar
- 4.Mitsudomi T, Morita S, Yatabe Y, Negoro S, Okamoto I, Tsurutani J, 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;11:121–8.CrossRefGoogle Scholar
- 14.Ramalingam SS, Cheng Y, Zhou C, Ohe Y, Imamura F, Cho BC, et al. LBA50 Mechanisms of acquired resistance to first-line osimertinib: preliminary data from the phase III FLAURA study. Ann Oncol. 2018;29. https://academic.oup.com/annonc/article/doi/10.1093/annonc/mdy424.063/5142018. Accessed 18 Nov 2018.
- 15.Papadimitrakopoulou VA, Wu Y-L, Han J-Y, Ahn M-J, Ramalingam SS, John T, et al. LBA51 analysis of resistance mechanisms to osimertinib in patients with EGFR T790M advanced NSCLC from the AURA3 study. Ann Oncol. 2018;29. https://academic.oup.com/annonc/article/doi/10.1093/annonc/mdy424.064/5142034. Accessed 18 Nov 2018.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.