Comparison of Clinicopathogenomic Features and Treatment Outcomes of EGFR and HER2 Exon 20 Insertion Mutations in Non–Small Cell Lung Cancer: Single-Institution Experience
Article information
Abstract
Purpose
Exon 20 insertion mutations (E20ins) in epidermal growth factor receptor (EGFR) or human epidermal growth factor receptor 2 (HER2) in non–small cell lung cancer (NSCLC) patients has become more important with emergence of novel agents targeting E20ins.
Materials and Methods
Advanced/Metastatic NSCLC patients with E20ins were included. EGFR E20ins was identified by two methods, next-generation sequencing (NGS) or real-time polymerase chain reaction (PCR), while HER2 E20ins was done by NGS only.
Results
Between December 2013 and July 2021, E20ins were identified in 107 patients at Asan Medical Center; 67 EGFR E20ins and 40 HER2 E20ins. Out of 32 patients with EGFR E20ins who had tested both PCR and NGS, 17 were identified only through NGS and the other 15 through both tests, giving a discordance rate of 53.1%. There was no clinically significant difference in clinicopathologic features between EGFR and HER2 E20ins; both were observed more frequently in adenocarcinoma, female and never-smokers. Brain metastases were evident at diagnosis in 31.8% of EGFR E20ins and 27.5% of HER2 E20ins, respectively. Platinum-based doublets demonstrated objective response rates (ORR) of 13.3% with a median progression-free survival (PFS) of 4.2 months for EGFR E20ins and 35.3% with 4.7 months for HER2 E20ins, respectively. In contrast, novel EGFR E20ins-targeted agents exhibited an ORR of 46.2% with a median PFS of 5.4 months, while HER2-targeted agents showed an ORR of 50% with that of 7.0 months.
Conclusion
Identification of EGFR and HER2 E20ins is more important as their targeted therapies improved outcomes. Upfront NGS test as a comprehensive molecular approach is strongly warranted.
Introduction
With the discovery of driver oncogenes and the relevant targeted therapies in non–small cell lung cancer (NSCLC), the treatment has shown significant improvements in survival outcomes [1-5]. Despite these advances, the prognosis of patients with some actionable alterations including epidermal growth factor receptor (EGFR/HER1) exon 20 insertion (E20ins) mutations and human epidermal growth factor receptor 2 (HER2/ERBB2) E20ins mutations seemed unsatisfactory. EGFR E20ins mutations account for 0.3%-3.7% of NSCLC [6] and are characterized by in-frame insertion and duplication near the C-helix of the EGFR kinase domain. They showed limited response rates (8%) to conventional EGFR-targeted agents such as gefitinib, erlotinib, and afatinib [7,8]. HER2 E20ins mutations account for 2.0%-4.0% of NSCLC, occurring as in-frame insertion with the duplication of amino acids [6,9]. Fortunately, novel agents targeting EGFR E20ins such as poziotinib, pyrotinib, mobocertinib (TAK-788), and amivantamab were recently introduced and showed a promising efficacy [10-12]. In addition, trastuzumab deruxtecan, a HER2-targeted antibody-drug conjugate, has also shown a promising anti-tumor activity in patients with HER2-mutant NSCLC [13].
In line with these recent advances, we tried to elaborate clinicopathogenomic features and treatment outcomes of NSCLC patients with EGFR and HER2 E20ins mutations in order not only to build proper diagnosis and treatment plan but also to overcome resistance mechanisms and develop more efficacious drugs.
Materials and Methods
1. Patients
Patients diagnosed with histologically confirmed metastatic NSCLC between December 2013 and July 2021 at Asan Medical Center, a tertiary referral center in the Republic of Korea, were retrospectively reviewed and those who were identified to have EGFR and HER2 E20ins mutations were included in the analysis.
2. Bioinformatics analysis
The presence of EGFR and HER2 E20ins mutations was identified using real-time polymerase chain reaction (PCR)– based analysis or targeted next-generation sequencing (NGS). Regarding PCR, the PANAMutyper EGFR kit (PANAGENE Inc., Daejeon, Korea) was used to detect EGFR mutations. Targeted sequencing was performed using an in-house panel (ver. 3 and 4, OncoPanel AMC) of the Asan Medical Center, Seoul, Republic of Korea, with the MiSeq and NextSeq platforms (Illumina, San Diego, CA). The panel covered a total of 382 genes, including the entire exons of 199 genes, introns of eight genes, and 184 hotspots involved in rearrangements for ver. 3. For version 4, the panel targeted 323 genes, including the entire exons of 225 genes, introns of six genes, and 99 hotspots involved in rearrangements. The Panel of Normals approach was also used for variant calling in somatic variant analysis. This approach uses a set of matched normal samples to determine the baseline level from which to call variants. Tumor mutational burden (TMB) was calculated as the number of nonsynonymous alterations per megabase (Mb) of genome examined. Therefore, EGFR E20ins mutations were identified through PCR or NGS, while HER2 E20ins mutations were identified by NGS only.
3. Assessment and statistical analysis
Baseline characteristics were analyzed by descriptive method. Treatment outcomes were assessed for patients with measurable or evaluable diseases who received at least one systemic treatment. The assessment included all treatments administered after the documentation of E20ins mutations, representing the initial agent within their respective drug classes, including platinum-based doublets, pemetrexed monotherapy, EGFR- or HER2-targeted agents, and immunotherapy either as monotherapy or in combination with chemotherapy. Tumor response was retrospectively assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) v1.1. Objective response rates (ORR) were defined as the proportion of patients who achieved a partial or completed response (PR or CR), and disease control rates (DCR) were defined as the proportion of patients who achieved stable disease, PR or CR among patients with measurable diseases. Survival outcomes were estimated using the KaplanMeier curve and compared with log-rank tests. Progression-free survival (PFS) was defined as the time period between the start of the treatment and the date of disease progression or death whichever occurred first. Overall survival (OS) was defined as the time period between the confirmation of metastatic disease and the death of any cause. A two-sided p-value of < 0.05 was considered statistically significant. All analyses were performed using R Foundation statistical software, ver. 4.1.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
1. Clinical characteristics of the patients
A total of 67 patients with advanced/metastatic NSCLC harboring EGFR E20ins and 40 patients with HER2 E20ins were identified to be eligible and were included in the analysis. The baseline characteristics of patients are summarized in Table 1. The median ages at diagnosis of metastatic diseases were 62 years (range, 25 to 83 years) and 60 years (range, 24 to 81 years) in EGFR and HER2 E20ins mutation groups, respectively. The predominant histologic type was adenocarcinoma, with the exception of one patient with squamous cell carcinoma in the EGFR E20ins group. The majority of patients in both groups had never smoked (61.2%, 41/67 and 71.8%, 28/39 in EGFR and HER2 E20ins groups, respectively). Females comprised 56.7% (38/67) and 75.0% (30/40) of EGFR and HER2 E20ins mutation groups, respectively. About 59.7% (40/67) and 50% (20/40) of EGFR and HER2 E20ins patients had metastatic disease at diagnosis, while the others were recurrent after curative-intent treatment. The proportion of patients who had brain or central nervous system (CNS) metastases at diagnosis was 31.8% (21/67) and 27.5% (11/40) in the EGFR and HER2 E20ins groups, respectively, and that at the first recurrence was 22.2% (6/27) and 20% (4/20), respectively. However, there was no difference in terms of clinicopathologic features between the two groups. In addition, their clinicopathologic features seemed similar to other mutation types reported in the literature, such as EGFR exon 19 or 21 mutations [14,15].
Baseline characteristics
2. Diagnostic methods and genomic characteristics
Regarding identification of EGFR E20ins mutations, NGS and PCR identified E20ins mutations in 43 patients (64.1%) and 39 patients (58.2%), respectively. To be more specific, out of 32 patients who underwent both tests for EGFR E20ins, only 15 patients showed EGFR E20ins by the first PCR test, meaning that a discordance rate between PCR and NGS was 53.1%. On the other hand, there was no patient who showed negative for E20ins mutation by NGS but positive by PCR (Table 2). All 40 patients with HER2 E20ins mutations were diagnosed by NGS.
The detection of EGFR E20ins mutations using NGS or PCR-based analysis
Details of each mutation are illustrated in the lollipop plot (Fig. 1). The most common mutation was S768_D770dup (n=9, 20.9%) in the EGFR E20ins group. Mutations in the near-loop regions of EGFR E20ins were found in 34 patients (79.1%), while mutations in the far-loop regions were found in nine patients (20.9%). No patient had mutations in C-helix regions. In the HER2 E20ins group, Y772_A775dup (n=31, 77.5%) was the most prevalent mutation. The previously unreported mutations were D770_N771delinsERG (n=1) in EGFR group, and G778_S779insFPG (n=1), V773_M774insWTYV (n=1), V777dup (n=1) in HER2 group. As for other actionable alterations, one patient in the EGFR E20ins group had concurrent exon 19 deletion mutation, and one patient in the HER2 E20ins group had HER2 amplification (copy number 9). Other actionable alterations were not observed in patients with EGFR and HER2 E20ins mutations (Fig. 2).
Genomic characteristics. (A) Lollipop of EGFR E20ins mutations. (B) Lollipop of HER2 E20ins mutations. E20ins, exon 20 insertion; EGFR, epidermal growth factor receptor; HER2, human epidermal growth factor receptor gene 2.
Oncoplot for most commonly mutated genes and targetable alterations in patients with EGFR and HER2 E20ins mutations. Analysis based on NGS testing of 43 patients with EGFR E20ins and 40 patients with HER2 E20ins mutations. E20ins, exon 20 insertion; EGFR, epidermal growth factor receptor; HER2, human epidermal growth factor receptor gene 2; NA, not available; NGS, next-generation sequencing; PD-L1, programmed death-ligand 1; TMB, tumor mutational burden.
Programmed death-ligand 1 (PD-L1) expression status (SP263 immunohistochemistry) was also assessed in 40 EGFR E20ins patients and 26 HER2 E20ins patients. Among them, 10.0% and 15.4% of each group had high PD-L1 expression of ≥ 50% by tumor proportion score, respectively (Table 1). The median TMB of these groups were 10.9 (9.4-15.6) and 7.8 (6.2-15.6), respectively.
3. Treatment pattern and response rates
Table 3 displays a summary of systemic treatments administered following the identification of E20ins mutations, featuring the primary agents within their respective drug classes. Within the EGFR E20ins group, comprising 67 patients, a total of 105 treatments were administered, with a median treatment line of one (range, 1 to 5). For the HER2 E20ins group, which included 40 patients, 65 treatments were administered, with a median treatment line of two (range, 1 to 6). Platinum-based doublets and conventional EGFR tyrosine kinase inhibitors (TKIs) were predominantly administered as first-line treatments in the majority of cases, accounting for 66.7% (28/42) for platinum-based doublets and 84.2% (16/19) for conventional EGFR TKIs. On the other hand, the median treatment line for novel EGFR or HER2-targeted agents was two, with 72.7% (8/11) of patients having received prior platinum-based chemotherapy, including four out of seven patients in the second-line setting and all four patients who received these agents as third-line treatments or beyond.
Palliative systemic treatments administered
Responses to systemic treatments were evaluated among those who had measurable diseases (n=40 in the EGFR and n=23 in the HER2 E20ins groups, respectively) and summarized in Table 4. ORRs for platinum-based doublets were 13.3% (4/30) and 35.3% (6/17) for patients with EGFR and HER2 E20ins mutations, respectively, while DCRs were 70% (21/30) and 76.5% (13/17), respectively. Conventional first-and second-generation EGFR TKIs, including erlotinib, gefitinib, and afatinib, yielded an ORR of 8.3% (1/12). A third-generation EGFR TKI was given to one patient who had co-mutation of T790M and showed a PR. Interestingly, within the cohort of 17 patients enrolled in the clinical trial for novel E20ins-targeted agents, six out of 13 with EGFR E20ins patients demonstrated a PR, resulting an ORR of 46.2%, while two out of three with HER2 E20ins patients exhibited a PR, yielding an ORR of 66.7%. Furthermore, HER2-targeted agents, including anti–HER2-monoclonal antibody such as trastuzumab or HER2-targeted antibody-drug conjugates like trastuzumab deruxtecan, also demonstrated an ORR of 50%. Anti–programmed death-1/PD-L1 inhibitor therapy was administered in 22.4% (n=15/67) of EGFR E20ins and 60% (n=24/40) of HER2 E20ins patients, respectively; however, it rarely resulted in a response or a meaningful response rate.
Response rates to treatments
4. Survival outcomes
With a median follow-up duration of 44.2 months (95% confidence interval [CI], 36.8 to not estimated [NE]) for EGFR E20ins and 49.3 months (95% CI, 34.2 to NE) for HER2 E20ins patients, the median OS was reportedly 23.2 months (95% CI, 14.6 to 26.1) in EGFR E20ins and 25.2 months (95% CI, 15.8 to 37.4) in HER2 E20ins (Fig. 3A), respectively. For systemic treatments, the median PFS to platinum-based doublets for EGFR E20ins and HER2 E20ins patients was 4.2 months (95% CI, 2.7 to 7.1) and 4.7 months (95% CI, 2.5 to 9.9), respectively (Fig. 3B). No significant differences in the PFS were observed among patients receiving an immune checkpoint inhibitor, irrespective of PD-L1 expression status (Figs. 3C and 4). The median PFS to the conventional EGFR TKIs was 2.3 months (95% CI, 2.1 to 6.5) in patients with EGFR E20ins mutation. With novel agents targeting EGFR E20ins mutation, however, the median PFS was 5.4 months (95% CI, 2.8 to NE). In patients with HER2 E20ins mutation, the median PFS to HER2-targeted agents was 7.0 months (95% CI, 3.3 to NE) (Fig. 3D).
Survival outcomes. (A) Overall survival to platinum-based doublets. (B) Progression-free survival to platinum-based doublets. (C) Progression-free survival to PD-1/PD-L1 inhibitors as monotherapy. (D) Progression-free survival to conventional first- and second-generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (blue line) and novel EGFR exon 20 insertion (E20ins)–targeted agents (red line) in patients with EGFR E20ins, and progression-free survival to human epidermal growth factor receptor gene 2 (HER2)-targeted agents (green line) in patients with HER2 E20ins. CI, confidence interval; NE, not evaluable; PD-1, programmed death-1; PD-L1, programmed death-ligand 1; TKI, tyrosine kinase inhibitor.
Progression-free survival to immune checkpoint inhibitors according to the programmed death-ligand 1 (SP263) expression status. (A) EGFR E20ins. (B) HER2 E20ins. CI, confidence interval; E20ins, exon 20 insertion; EGFR, epidermal growth factor receptor; HER2, human epidermal growth factor receptor gene 2; NE, not evaluable.
5. CNS progression
During the disease course, 42 out of 54 patients (77.8%) with EGFR E20ins and 25 out of 35 patients (71.4%) with HER2 E20ins experienced disease progression while undergoing systemic treatments, primarily occurring after the first-line treatments, except for one patient in the EGFR E20ins group. Among them, 18 patients (42.8%) with EGFR E20ins and eight (32.0%) with HER2 E20ins experienced CNS disease progression. Notably, six out of 18 (33.3%) patients with EGFR E20ins and two out of eight (25%) with HER2 E20ins exhibited isolated CNS disease progression without evidence of extracranial systemic disease progression. The majority of patients underwent progression of brain parenchymal disease (EGFR E20ins, 14/18, 77.8%; and HER2 E20ins, 8/8, 100.0%), while four patients with EGFR E20ins had experienced leptomeningeal carcinomatosis. The median time from the start of treatment to the first CNS progression was 2.5 months (95% CI, 1.5 to 8.0) in EGFR E20ins group and 3.8 months (95% CI, 3.5 to NE) in HER2 E20ins group, respectively.
Discussion
In this study, we identified the limited treatment outcomes with conventional or available systemic therapies in NSCLC patients with EGFR or HER2 E20ins. However, those who have participated into clinical trials involving novel EGFR E20ins- or HER2 E20ins-targeted agents demonstrated promising clinical efficacies. Therefore, the identification of E20ins mutation becomes more important than before as novel agents have come to clinic.
With regard to detection methods, we found that real-time PCR is not enough to detect EGFR E20ins mutation, despite its advantages in terms of turnaround time, accessibility, and cost. Our findings revealed a substantial discordance rate of 53.1% between real-time PCR and NGS. Interestingly, no patients were identified as positive for PCR but negative for NGS, emphasizing the superior sensitivity of NGS in capturing EGFR E20ins mutations. The observed discrepancy in the detection rates of E20ins mutations between PCR and NGS can be attributed to several factors. The genetic region where EGFR E20ins mutations manifest is notably wide, encompassing various locations such as the C-helix, near the loop, and regions distinct from the loop, in contrast to being concentrated in major hotspots. This broad distribution makes it challenging for PCR, which typically relies on a limited set of primers targeting specific regions. In an effort to enhance the role of PCR test or to adopt an enrichment strategy for E20ins mutation identification, we also tried to find clinicopathologic features. However, both EGFR and HER2 E20ins exhibited similarities with other mutation subtypes in the literature, such as EGFR exon 19 or 21 mutations [14,15]. Moreover, there were no differences between EGFR and HER2 E20ins mutations. As genomic sequencing technologies continue to evolve, NGS has facilitated the identification of E20ins mutations [16,17]. Therefore, rather than enrichment strategies based on clinicopathologic features, upfront and active NGS testing is warranted as a comprehensive molecular approach to identify actionable mutations in NSCLC patients. Actually, the prevalence of the mutations may have been underestimated due to the limited availability of NGS during the study period for a variety of reasons, including its high cost and accessibility.
Patients harboring EGFR and HER2 E20ins mutations have not been recognized as a population with targetable alterations. We observed unfavorable outcomes in patients with EGFR E20ins treated with conventional first- and second-generation EGFR TKIs, with an ORR of 8.3% and a median PFS of 2.3 months. Although some studies have reported encouraging responses to platinum-based chemotherapy in this patient cohort [8,18,19], its efficacy was also found to be limited, showing an ORR of 13.3% and a median PFS of 4.2 months, which are comparable to the prognoses observed in other NSCLC patients without actionable alterations [19-21]. Of note, immune checkpoint inhibitors did not induce any tumor responses in both EGFR and HER2 E20ins, regardless of the PD-L1 expression status, consistent with findings from previous studies [22].
We observed median OS of 23 months and 25 months for EGFR and HER2 E20ins, respectively, which seemed inferior to those with common EGFR mutations or other actionable mutations but superior to those without any targetable alterations [8,19,23,24]. This difference may be attributed partly to the different tumor biology or the availability of more effective targeted treatments. In the current study, patients treated with novel EGFR E20ins– and HER2 E20ins–targeted agents exhibited promising outcomes, with an ORR of 46.2% and 50.0%, and a median PFS of 5.4 months and 7.0 months, respectively, highlighting the necessity for novel treatment strategies in these patients. Recently, two novel EGFR-targeted agents, amivantamab and mobocertinib, granted Food and Drug Administration approval for the treatment of patients with EGFR E20ins mutations whose diseases have progressed after platinum-based chemotherapy [11,12,25,26]. Early data of amivantamab and mobocertinib showed median PFS of 8.3 months and 7.3 months, respectively [11,12]. In a recent phase 3 study, the combination of amivantamab plus platinum-based chemotherapy demonstrated significantly greater efficacy compared to chemotherapy alone (median PFS, 11.4 months vs. 6.7 months) for first-line treatment, while mobocertinib failed to exhibit superiority over chemotherapy leading to the withdrawal of this drug [27,28]. In addition, trastuzumab deruxtecan has emerged as a promising targeted options for HER2-mutant NSCLC (DESTINY-Lung01 trial), with an ORR of 55% and a median duration of response of 9.3 months [13].
It is noteworthy that brain metastases were more prevalent at the time of diagnosis in 31.8% of EGFR E20ins patients and 27.5% of HER2 E20ins patients, and during the disease course, 42.8% and 32.0% of patients with progressive disease experienced CNS progression, respectively [29]. In the phase I/II study of novel agents, the ORR to mobocertinib was 25% in patients with baseline brain metastases compared to 56% in those without brain metastases, with a median PFS of 3.7 months versus 10.2 months [30]. Amivantamab is expected to have low CNS activity due to its large molecular size, therefore, a combination with a small molecule EGFR inhibitor is being evaluated in several clinical trials [31]. Improving intracranial efficacy would be crucial for the future development of new drugs for NSCLC patients harboring EGFR and HER2 E20ins mutations.
Recent studies suggest that the location of EGFR E20ins mutations (C-helix, near-loop, far-loop) may influence responses to EGFR-targeted agents [32]. Among patients tested by NGS, we identified a relative prevalence of EGFR E20ins mutation locations and variations in TKI efficacy according to their respective regions. In this regard, adopting deeper genomic sequencing approaches to precisely define the mutation region would be vital for advancing precision medicine. The accurate classification of patient population, guided by detailed genomic information, needs to be in line with the development of treatments in the future.
The study has limitations. First, a direct comparison between the group receiving specific E20ins-targeted drugs and those receiving other treatments was not feasible. It is worth mentioning, however, that treatment outcomes demonstrated superiority to those achieved with conventional approaches like platinum-based doublets or conventional EGFR TKIs, even though novel EGFR or HER2 E20ins-targeted agents were administered as a later line of therapy. Second, the inclusion of a small number of patients in this study is due to the rarity of these mutations. While the small sample size is a limitation, we aimed to conduct a comprehensive data collection over a period of more than 7 years. Importantly, the prevalence of these mutations may increase with the widespread use of NGS, and additional data from a larger cohort will contribute to more robust real-world evidence.
Novel agents targeting either EGFR or HER2 E20ins in NSCLC have demonstrated improved clinical outcomes not only compared to standard platinum-based doublets but also in comparison to pre-existing EGFR TKIs or immunotherapeutic agents. This underscores the increasing importance of identifying these molecular aberrations. With this regard, the use of NGS emerges as a much more appropriate and warranted approach for upfront comprehensive molecular evaluation of newly diagnosed advanced/metastatic NSCLC patients, rather than adopting enrichment approach based on clinicopathologic features. Moreover, the higher prevalence of CNS disease at the time of diagnosis or progression suggests that strategies for overcoming CNS disease should be mapped out before treating E20ins mutations.
Notes
Ethical Statement
This study was approved by the International Review Board of the Asan Medical Center, Seoul, Korea (IRB approval number: 2020-1204), and patient data was anonymized. As a retrospective study, additional informed consent was not required.
Author Contributions
Conceived and designed the analysis: Yoon S, Lee DH.
Collected the data: Lee SH, Jeong H, Kim DH, Jang SJ, Kim SW, Yoon S, Lee DH.
Contributed data or analysis tools: Lee SH, Jeong H, Kim DH, Jang SJ, Kim SW, Yoon S, Lee DH.
Performed the analysis: Lee SH, Jeong H.
Wrote the paper: Lee SH, Jeong H, Yoon S, Lee DH.
Conflicts of Interest
Dae Ho Lee declares honoraria from Abion, AstraZeneca, Boehringer-Ingelheim, Bristol-Myers Squibb, CJ Healthcare, ChongKunDang, Eli Lilly, Janssen, Merck, MSD, Mundipharma, Novartis, Ono, Pfizer, Roche, Samyang Biopharm and ST Cube. The other authors declare that they have no conflict of interest.