The 2 most common types of EGFR alterations in NSCLC are exon 19 deletions, which account for 47% of all EGFR alterations, and L858R mutations, which account for 32%.¹ The third most common type of EGFR alterations are EGFR exon 20 insertions, which account for 12% of all EGFR mutations. This corresponds to approximately 1% to 2% of nonsquamous NSCLC overall, which is comparable to other targetable oncogenic drivers we test for in advanced NSCLC as standard of care, including ROS1 fusions and, more recently, RET fusions.

The heterogeneity of EGFR exon 20 insertions in NSCLC was highlighted in this 2018 study of 263 patients.¹ All patients had comprehensive genomic profiling to identify the location and frequency of each insertion. In total, 64 unique insertion mutations were identified, illustrating the range of nucleotides affected by exon 20 insertions in patients with NSCLC. It is imperative to identify which nucleotides are affected by the insertion because this may impact the efficacy of targeted therapies. For instance, A763_Y764insFQEA insertions are sensitive to standard EGFR TKIs, but most other exon 20 insertions confer resistance.

Tumors with an EGFR exon 20 insertion mutation may co-occur with other mutations, which—although not oncogenic—may impact treatment efficacy. For example, co-occurring EGFR exon 20 insertions and TP53 mutations are a negative prognostic marker associated with shorter efficacy from EGFR TKIs.²

The clinical features of patients with NSCLC presenting with EGFR exon 20 insertions appear to differ from those with common EGFR mutations. In 2 retrospective analyses of Korean and Chinese patients with EGFR exon 20 insertions,^3,4 approximately 25% to 30% are current or former smokers, which is higher than what is observed in patients with common EGFR mutations. There also appears to be an increase in the proportion of men and a higher age at diagnosis.

These data also reflect the aggressiveness of EGFR exon 20 insertion–positive NSCLC. Bone and central nervous system (CNS) metastases, associated with negative prognosis, were common among these patients, highlighting why it is important to have EGFR exon 20 insertion–targeted agents to improve patient survival.

This retrospective analysis compared the real-world survival of patients with EGFR exon 20 insertions (n = 181) with patients with common EGFR mutations (n = 2833) using data from the Flatiron Health Database.⁵ Patients with exon 20 insertions had a 75% increased risk of death and 60% increased risk of shorter time to next therapy compared with patients with common EGFR mutations. In addition, 5‑year OS was 8% vs 19% and median real-world OS was 16.2 months vs 25.5 months (P <.0001), with the better outcomes seen among patients with common EGFR mutations. These data indicate again how different EGFR exon 20 insertion–positive NSCLC is from disease with common EGFR mutations.

The techniques used to detect genetic mutations in NSCLC are important to ensure that patients receive the appropriate targeted therapy for their particular mutation. Considering the heterogeneity of EGFR exon 20 insertions, it is important to test for these mutations using DNA NGS, which allows for coverage of multiple insertion types, vs using single-gene assays. As can be seen in these data from the American Association for Cancer Research GENIE and Foundation Medicine databases, single-gene PCR tests are expected to miss approximately 50% of the exon 20 insertion cases that NGS can identify.⁶ Furthermore, not all DNA NGS panels are created equally—the more mutation types included in a given NGS panel, the more patients with this alteration you will be able to identify.

A recent bridging study reported at WCLC 2021 sought to validate 2 companion diagnostic tests, Guardant360 CDx and Oncomine Dx Target Test (ODxT), for the identification of candidate patients with EGFR exon 20 insertions for amivantamab therapy. Amivantamab is an EGFR-MET bispecific antibody that—as we will discuss later—was recently approved by the FDA for the treatment of patients with this alteration after progression on platinum-based chemotherapy.⁷ This analysis used samples from 81 patients with EGFR exon 20 insertion–positive NSCLC in the efficacy patient population from the CHRYSALIS trial, which evaluated amivantamab.⁸ The Guardant360 CDx test analyzes circulating tumor DNA from plasma obtained from peripheral whole blood, and the ODxT test analyzes DNA obtained from formalin-fixed paraffin-embedded tumor samples.

When comparing overall response rates (ORRs) among patients with identified EGFR exon 20 insertions, the results with the 2 companion diagnostic tests were comparable both with each other and with the assays used in CHRYSALIS. The ORR with Guardant360 CDx was 39% (95% CI: 26.6%-51.9%) compared with 46% (95% CI: 30.1%-62.8%) with the ODxT test and 40% (95% CI: 28.8%-51.0%) with the clinical trial assays. These data indicate that either Guardant360 CDx or ODxT can accurately identify patients with EGFR exon 20 insertions for amivantamab treatment.

To help us measure the educational impact of this activity, please provide an answer to the following question, which was presented to you in the pre-education section of the activity. As a reminder, your responses will be aggregated for analysis, and your specific responses will not be shared.

Platinum-based chemotherapy is currently the first-line standard of care for patients with EGFR exon 20 insertion–positive NSCLC. In a 2020 retrospective analysis of 119 patients from China who received first-line chemotherapy, pemetrexed improved median PFS from 3.0 to 5.5 months (P = .0026) and trended toward an improvement in median OS (19.6 vs 25.0 months; P = .0769) compared with other chemotherapy regimens.⁹

A 2019 retrospective analysis of 59 patients from China also evaluated pemetrexed-containing regimens in patients who received first-line systemic therapy.¹⁰ Similarly, pemetrexed-containing regimens improved median PFS from 2.7 months to 6.2 months (P <.001) and median OS from 15.4 months to 28.0 months (P = .009) compared with regimens that did not contain pemetrexed.

We know that in patients with nonsquamous NSCLC and common EGFR mutations, the addition of bevacizumab, an anti‑VEGF antibody, to platinum-based chemotherapy is associated with improved outcomes vs chemotherapy alone.¹¹ But what impact does adding antiangiogenic agents to chemotherapy have in patients with EGFR exon 20 insertion mutations?

This retrospective analysis of real-world outcomes in 165 patients with EGFR exon 20 insertion–positive NSCLC from China compared the efficacy of first- or second-line chemotherapy plus bevacizumab with chemotherapy alone.⁴ There was a PFS benefit with the addition of bevacizumab to chemotherapy as a second-line treatment (P = .013) but no benefit in those receiving first-line treatment (P = .221). This result may relate to the general condition and eligibility of patients for bevacizumab, as the magnitude of PFS benefit is in line with what we have seen in the global population of patients with NSCLC. Based on these data, if patients are eligible for bevacizumab, I think it is a pretty good option for second-line treatment.

The researchers also compared the efficacy of first-line chemotherapy with first-line EGFR TKI therapy in these patients by CNS status. First‑line chemotherapy improved PFS compared with TKIs regardless of CNS metastases in patients with exon 20 insertion–positive disease.

Returning to the retrospective analysis of real-world outcomes in 181 patients with EGFR exon 20 insertion mutations from the Flatiron Health Database, we see further evidence that our standard EGFR TKIs do not do as well in our patients with exon 20 insertions.⁵ The median real-world PFS was 2.9 months in patients with exon 20 insertions vs 10.5 months with common EGFR mutations. I would posit that the efficacy we do see is being driven by patients with the FQEA insertion.

Let’s now consider some data specifically for osimertinib, a third-generation EGFR TKI that is the current first-line standard of care for the treatment of advanced NSCLC with common EGFR mutations.¹⁴ In a retrospective analysis of 21 patients with EGFR exon 20 insertion–positive NSCLC in the Netherlands and a median of 1 previous line of therapy (range 0-3), treatment with osimertinib at the standard dose of 80 mg QD¹⁵ or the higher dose of 160 mg QD showed limited benefit, with 1 patient achieving a partial response (PR), 16 patients having stable disease (SD), and 4 patients with PD as the best response.¹⁶

Further data on high‑dose osimertinib for patients with EGFR exon 20 insertions from the ECOG-ACRIN EA5162 study were presented at the 2020 American Society of Clinical Oncology (ASCO) annual meeting. In this phase II study, there was a 24% confirmed ORR with high-dose osimertinib among 15 evaluable patients, with the majority having SD. Looking at the swimmer’s plot, only 1 patient had a deep and prolonged response, which again I postulate is likely due to them having an FQEA insertion.

At the 2021 ESMO Congress, the phase II POSITION20 study corroborated the above data on high-dose osimertinib for EGFR exon 20 insertion–positive NSCLC.¹⁷ The ORR was 28% in 25 patients, with all responding patients (7/25) having a PR. In total, 52% of patients in this trial had SD, and 20% had disease progression.

In my opinion, further studies looking at standard EGFR TKIs for the treatment of EGFR exon 20 insertion–positive NSCLC would be best served by selecting for patients with FQEA insertions.

Finally, let’s look at some data for immunotherapy in this population. At this year’s WCLC meeting, Bazhenova and colleagues¹⁸ presented a retrospective analysis of the efficacy of immune checkpoint inhibitors in patients with EGFR exon 20 insertion–positive advanced NSCLC. Patients who had EGFR exon 20 insertion mutations and received checkpoint inhibitors had a 58% increased risk of shorter time to next therapy compared with patients with wild-type NSCLC (median real-world time to next therapy: 3.7 months vs 5.8 months, respectively). Most patients in this analysis (approximately 75% with an exon 20 insertion and approximately 60% with wild-type NSCLC) received immune checkpoint inhibitors as second- or subsequent‑line therapy.

This result is not surprising considering that the microenvironment of EGFR‑mutant lung carcinomas, regardless of the specific mutation, is immunosuppressive—there are many T-regulatory cells and high constitutive expression of PD‑L1 that does not reflect an antitumor immune response.

As such, I do not recommend immune checkpoint inhibitor therapy for patients EGFR exon 20 insertion–positive advanced NSCLC.

To conclude our discussion of standard therapies for the treatment of EGFR exon 20 insertion–positive NSCLC, I will again return to the retrospective analysis of real-world outcomes in this subset of patients from the Flatiron Health Database.

Most patients received platinum-based chemotherapy, which is the current standard of care. Platinum-based chemotherapy was given to 61.3% of patients in the first‑line setting, with a median real-world PFS of 6.6 months and median real-world OS of 17.4 months.⁵ In the second-line setting, 23.5% of patients received a platinum-based regimen, with a median real-world PFS of 5.0 months and median real-world OS of 14.2 months.

In the first line, 21.5% of patients received EGFR TKIs as monotherapy, with a median real-world PFS of only 2.9 months and median real-world OS of 9.5 months, far lower than for patients with common EGFR mutations receiving EGFR TKIs. As a second-line therapy, TKIs were used in 21.7% of patients for a median real-world PFS of 2.5 months and median real-world OS of 14.0 months. These real-world data indicate that far more patients with exon 20 insertions are receiving EGFR TKIs than would be indicated by the data showing that our standard EGFR TKIs are only effective in patients with FQEA insertions, which make up only 6% of patients with exon 20 insertion mutations. This is probably because prescribers are unaware of differences between EGFR mutation subtypes.

Finally, immunotherapy also is being used in these patients, with 8.8% and 28.7% receiving immunotherapy alone in the first- and second-line settings, respectively. However, as I mentioned above, I do not recommend immunotherapy for these patients.

To help us measure the educational impact of this activity, please provide an answer to the following question, which was presented to you in the pre-education section of the activity. As a reminder, your responses will be aggregated for analysis, and your specific responses will not be shared.

I will now discuss novel therapies being developed specifically for the treatment of EGFR exon 20 insertion–positive NSCLC. We have exciting data on several treatment strategies in this setting and will start our discussion with the TKI poziotinib.

The open-label phase II ZENITH20 study is evaluating poziotinib in patients with EGFR or HER2 exon 20 insertion–positive advanced NSCLC (planned N = 603).^19,20 In patients with an EGFR exon 20 insertion receiving poziotinib in the second line (cohort 1), the response rate was 14.8% (95% CI: 8.9-22.6) with a median PFS of 4.2 months (95% CI: 3.7-6.6), which is a little disappointing for targeting of an oncogene addiction. Still, the aggressiveness of exon 20 insertion–positive disease is something to consider when looking at these data.

Data on the CNS activity of poziotinib were reported at ASCO 2021 for cohorts 1-3 of the ZENITH20 trial.²¹ This analysis included 36 patients with baseline CNS metastasis who had either an EGFR (n = 12; cohort 1) or HER2 (n = 14; cohort 2) exon 20 insertion mutation and received poziotinib in the second line or had an EGFR exon 20 mutation and received poziotinib in the first line (n = 10; cohort 3). Of importance, 12 patients underwent whole-brain radiation ≤12 weeks before poziotinib treatment. The ORR in patients with CNS metastases was 22.2%, with 8 patients having a PR. By independent review committee (IRC), 3 patients had a complete intracranial response (defined by 2 consecutive MRI scans with disappearance of CNS metastases), and an additional 24 patients had SD (defined as 1 MRI without disease progression or new lesions).

In addition, poziotinib is associated with significant toxicity, with 30% of patients developing a grade 3 rash, 26% grade 3 diarrhea, and 22% grade 3 stomatitis/mucosal inflammation. These adverse events were likely so common due to the drug also targeting wild‑type EGFR.

The second targeted drug with data in patients with EGFR exon 20 insertion–positive NSCLC is mobocertinib (TAK-788), an oral TKI designed selectively to target exon 20 insertion mutations.^22-24 Mobocertinib recently received accelerated approval from the FDA for the treatment of patients with EGFR exon 20 insertions who have progressed on or after platinum-based chemotherapy.²⁵ This approval was based on data from this multicohort phase I/II study, which had a cohort dedicated to refractory EGFR exon 20 insertion mutations without active/measurable CNS metastases, along with a single-arm extension cohort of patients with previously treated advanced NSCLC with EGFR exon 20 insertion mutations (EXCLAIM cohort).^22-24,26,27

The response rate with mobocertinib in patients previously treated with platinum-based chemotherapy (n = 114) was a little higher than what we saw with poziotinib at 28% by IRC.²² The ORR in patients in the expansion cohort, which included 96 previously treated patients, was similar at 25%.

The median duration of response (DoR) was 17.5 months in patients previously treated with platinum chemotherapy and is not yet reached in the expansion cohort. The median PFS was 7.3 months in both groups of patients. Overall, the efficacy of mobocertinib appears a little higher than poziotinib, although cross‑trial comparison is very difficult.

At WCLC 2021, Jӓnne and colleagues²⁷ reported on the efficacy of mobocertinib in platinum-pretreated patients who did or did not receive prior anti–PD-1/PD-L1 therapy. Of the 114 patients in the platinum-pretreated population, 43% received prior immunotherapy, and efficacy was similar regardless of prior anti–PD-1/PD-L1 therapy. The confirmed ORR by IRC was 25% (95% Cl: 14%-40%) for patients with prior anti–PD-1/PD-L1 therapy vs 30% (95% CI: 20%-43%) for patients without prior anti–PD-1/PD-L1 therapy. Of note, the median DoR was slightly higher for those with prior anti–PD-1/PD-L1 therapy than for those who did not receive an immune checkpoint inhibitor at 17.5 months (95% CI: 3.7 to not estimable) vs 11.0 months (95% CI: 3.8-17.5), respectively. When looking at disease progression rates, the median PFS was almost identical between the 2 groups: 7.4 months (95% CI: 5.5-21.1) with prior anti–PD-1/PD-L1 therapy and 7.3 months (95% CI: 5.4-10.2) without prior anti–PD-1/PD-L1 therapy.

Mobocertinib also is associated with adverse events related to the inhibition of wild‑type EGFR such as nausea, diarrhea, stomatitis, and rash, but digestive tract toxicities were the most common adverse events leading to treatment discontinuation.²² In both cohorts, 99% of patients experienced any-grade treatment-related adverse events (TRAEs), with 41% and 46% developing a grade ≥3 TRAE in the platinum-pretreated and EXCLAIM cohorts, respectively. One platinum-treated patient had a TRAE leading to death.

The safety with mobocertinib was similar regardless of whether patients received prior anti–PD(L)-1 therapy, with all patients having any-grade TRAEs and 39% and 58% developing a grade ≥3 TRAE with and without prior anti–PD(L)-1 therapy, respectively.²⁷

At WCLC 2021, Spira and colleagues²⁴ reported efficacy results of mobocertinib for cohort 5, which comprised 20 patients with refractory EGFR exon 20 insertion–positive metastatic NSCLC with disease progression after response or SD ≥6 months on prior EGFR TKI therapy. Prior TKIs included poziotinib, osimertinib, afatinib, erlotinib, and investigational TKIs.

The confirmed ORR for this cohort of patients was 40% (95% CI: 19.1%-63.9%), with all responding patients achieving a PR. The confirmed disease control rate was 90%, and the median DoR was 13.0 months (95% CI: 5.6-13.5).

When looking at survival in this cohort, the median PFS was 7.3 months (95% CI: 3.6-13.0), and the median OS was not reached (NR) (95% CI: 14.7-NR). The OS rate was 94.7% at 6 months and 78.6% at 12 months.

Additional investigational EGFR inhibitors also are being evaluated in clinical trials. Data from the phase I trials WU-KONG1 and WU-KONG2 were presented at ASCO 2021 and WCLC 2021.^28,29 In this trial, the selective, irreversible EGFR inhibitor DZD9008 is being evaluated in patients with previously treated advanced NSCLC and EGFR or HER2 mutations, including expansion cohorts with exon 20 insertions (N = 102). Treatment-emergent adverse events of any grade occurred in all patients, and grade ≥3 adverse events occurred in 39.2%.

The response rate per investigator was 39.6%, with a disease control rate of 86.8%. There was efficacy across different exon 20 insertion mutation subtypes, and some patients responded after failing another treatment specific to exon 20 insertions (eg, amivantamab), suggesting we may be able to sequence DZD9008 with our other agents that target EGFR. PFS data are not yet mature, but DZD9008 seems promising based on the data we have so far.

CLN‑081, formerly known as TAS6417, is an oral EGFR inhibitor with a unique pyrrolopyrimidine scaffold and broad activity against exon 20 insertions and other EGFR mutations. Preliminary efficacy data from a phase I/IIa trial of 42 patients with EGFR exon 20 insertion–positive metastatic NSCLC were presented at ASCO 2021.³⁰ A best response of PR was achieved in 21 patients (50%), with 20 patients (48%) achieving SD and the remaining patient (2%) having PD. The response rate was 44% in patients who had received prior EGFR TKI therapy, including mobocertinib and poziotinib, suggesting there is a possibility of response with CLN-081 after the failure of other EGFR exon 20 insertion–specific treatments.

Amivantamab is a bispecific antibody targeting EGFR and MET.^31,32 Amivantamab has demonstrated activity as monotherapy against a diversity of activating and resistance EGFR mutations, including exon 20 insertions.^33-35 It causes deep inhibition of ligand binding to EGFR, degradation of the EGFR and MET receptors through trogocytosis, and immune cell–directed activity toward the tumor cell.

CHRYSALIS is a dose escalation and expansion phase I study of amivantamab in 460 patients with EGFR exon 20 insertion mutations who progressed on prior platinum-based chemotherapy.^34,36,37 The current analysis is from the 81 patients in the efficacy population who received amivantamab at the recommended phase II dose of 1050 mg (1400 mg for patients ≥80 kg). Evaluable patients received ≥3 disease assessments at clinical cutoff. The primary endpoints are ORR per Response Evaluation Criteria in Solid Tumors version 1.1, clinical benefit rate (CBR), DoR, and adverse events. The secondary endpoints are PFS, OS, pharmacokinetics, and immunogenicity.

As I mentioned previously, all patients in this analysis were previously treated with platinum-based chemotherapy.^34,36 In addition, 46% of patients had received previous immunotherapy, and 25% had received a first-, second-, or third-generation EGFR TKI. The study reported an ORR of 40% (95% CI: 29%-51%) and median DoR of 11.1 months (95% CI: 6.9-NR) at a median follow-up of 9.7 months.

This waterfall plot shows the efficacy of amivantamab in patients based on the location of their exon 20 insertion (helical region, near loop, far loop, or not detected by circulating tumor DNA).^34,36 In total, 54 patients had an insertion in the near loop, and in this group ORR was 41% and CBR was 70%. Patients with insertions in the far loop (n = 8) had an ORR of 25% and CBR of 75%. Only 1 patient was found to have an insertion in the helical region, and that patient responded to therapy. The remaining 18 patients who had an unknown insertion location had an ORR of 39% and CBR of 83%. Overall, these data suggest promising efficacy regardless of the location of insertion in exon 20.M

CHRYSALIS is still ongoing, but at the time these data were published, median PFS was 8.3 months (95% CI: 6.5-10.9) and median OS was 22.8 months (95% CI: 14.6-NR).^34,36 The patients who responded did so rapidly and deeply, with 63% (20/32) having responses of at least 6 months. Because the patients received an average of 2 prior lines of therapy, the efficacy of amivantamab is impressive.

Amivantamab was associated with rash, stomatitis, and cutaneous reactions, mostly grades 1 and 2, but 4% of the patients developed grade ≥3 rash.³⁶ This safety profile is different than that seen with mobocertinib or poziotinib and may even be considered better.

Infusion‑related reactions with amivantamab occurred in 66% of patients, primarily during the first infusion (94% of cases). This did not impact patient ability to receive subsequent treatments. So, although patients must be monitored for infusion-related reactions with delivery of amivantamab, it was not a reason for treatment discontinuation.

Based on these data, amivantamab recently received accelerated approval from the FDA for the treatment of patients with EGFR exon 20 insertions who have progressed on or after platinum-based chemotherapy.⁷

To help us measure the educational impact of this activity, please provide an answer to the following question, which was presented to you in the pre-education section of the activity. As a reminder, your responses will be aggregated for analysis, and your specific responses will not be shared.

In another cohort from the CHRYSALIS trial, amivantamab was tested in combination with carboplatin/pemetrexed in eligible patients (n = 20).³⁸ There were no mutation requirements for this cohort, but 95% had an EGFR mutation and 35% had an EGFR exon 20 insertion. In this cohort, most patients (45%) received >2 prior lines of therapy, 30% received 1 prior line of therapy, and 25% were treatment naive.

The combination of amivantamab and carboplatin/pemetrexed was found to be tolerable. In this cohort, 65% of patients experienced any-grade infusion-related reaction or nausea. The most frequent grade ≥3 treatment-emergent adverse events were neutropenia (35%) and thrombocytopenia (15%)—adverse events anticipated with chemotherapy—followed by diarrhea (5%) and nausea (5%). TRAEs led to discontinuation for 2 patients and dose reductions for 4 patients.

Amivantamab and chemotherapy had antitumor activity in patients with different EGFR mutations and in patients who were previously treated or treatment naive. More specifically, 80% of treatment-naive patients (n = 5) with an EGFR exon 20 insertion mutation had the best PR response and remained on treatment at the time of this analysis. Overall, the median duration of treatment in this patient population was 5.6 months (range: 0.1-9.2).

The combination of amivantamab plus chemotherapy for the treatment of EGFR exon 20 insertion–positive advanced NSCLC is being further evaluated in the ongoing randomized phase III PAPILLON study. PAPILLION is comparing amivantamab in combination with carboplatin/pemetrexed vs carboplatin/pemetrexed alone in 300 patients with untreated locally advanced/metastatic NSCLC and EGFR exon 20 insertion mutations.³⁹ Patients are receiving 4 cycles of therapy in 21-day cycles before ≥5 cycles of maintenance therapy with amivantamab plus pemetrexed or pemetrexed alone. The primary endpoint is PFS by blinded independent central review, and key secondary endpoints include ORR, OS, and safety.

There are now several options available for the treatment of patients with pretreated EGFR ex20ins+ NSCLC, including mobocertinib and amivantamab. The efficacy and safety profile of mobocertinib seems less favorable than that of amivantamab based on available data but it may be considered easier for patients given the oral delivery. However, mobocertinib will still have the challenge of patient adherence. The early identification of EGFR ex20ins mutations will allow us to pinpoint patients who are candidates for these new second-line therapies. For patients with EGFR ex20ins+ mutations in the first line setting, I recommend enrollment in a clinical trial.