In this module, I discuss key clinical trials evaluating new therapeutic strategies for patients with EGFR-mutated advanced NSCLC in the newly diagnosed setting, second-line following osimertinib, and third-line following osimertinib and platinum-based therapy.

To start with a brief overview on EGFR mutations, approximately 10% of US patients with NSCLC have EGFR mutations.¹EGFR mutations are more common in those who have never smoked, in East Asians, and in women. Most EGFR mutations are deletions in exon 19 or an L858R mutation in exon 21. These classical mutations comprise 85‑90% of all EGFR mutations, with the remainder being atypical EGFR mutations and exon 20 insertions.

Since the early 2000s, management of metastatic EGFR‑mutated NSCLC has relied on EGFR TKIs, starting with such first-generation TKIs as gefitinib and erlotinib.^1,2 Whereas the first-generation TKIs were reversible inhibitors that block ATP binding to the TK domain of EGFR, the second-generation TKIs (eg, afatinib, dacomitinib) are irreversible inhibitors that covalently bind to EGFR. The third-generation TKI osimertinib is an irreversible inhibitor that is more selective for EGFR mutations. Osimertinib is now the standard of care for first‑line treatment of metastatic EGFR-mutated NSCLC.

Multiple clinical trials have demonstrated that first-line osimertinib has a very high response rate (80%) with a median progression-free survival (PFS) of 17.7 months and median overall survival (OS) of 38.6 months. Importantly, rash and diarrhea are less common with osimertinib vs first-generation and second-generation EGFR TKIs. Although rash and diarrhea are generally mild to moderate, I should point out that even grade 2 chronic rash still negatively affects quality of life for patients, and just dismissing this toxicity because it is “only” grade 2 does not do justice for a patient’s treatment experience.

How can we improve upon the favorable outcomes observed with osimertinib in the first-line setting? One possibility is with dual EGFR‑VEGF pathway inhibition, for several notable reasons. First, both pathways share common downstream signaling molecules including PI3K, AKT, and mTOR as well as RAS, RAF, MEK, and ERK.³ Second, EGFR signaling is upregulated in EGFR-mutated NSCLC, which may lead to increased VEGF expression. Third, acquired resistance to EGFR TKIs is associated with increased levels of VEGF. Taken together, this evidence suggests that dual EGFR‑VEGF inhibition may delay acquired resistance to EGFR TKIs.

Dual inhibition may be achieved with therapeutic agents targeting EGFR (eg, EGFR TKIs or antibodies such as amivantamab or necitumumab) combined with those targeting VEGF, such as bevacizumab or ramucirumab.

Multiple clinical trials have examined combination therapy with an EGFR TKI plus an antibody to VEGF in EGFR-mutated advanced NSCLC.³ J025567 was one of the first randomized phase II trials to compare erlotinib plus bevacizumab vs erlotinib alone.^3-5 J025567 reported no difference in median OS, but median PFS was significantly increased from 9.7 months to 16.0 months with the addition of bevacizumab. These positive PFS results supported development of the phase III NEJ026 trial, which also showed no change in median OS and a significantly improved median PFS with the addition of bevacizumab to erlotinib.^3,6,7

Turning to studies of combination therapy with third-generation EGFR TKIs, a phase I/II trial conducted by Memorial Sloan Kettering Cancer Center reported that osimertinib plus bevacizumab was associated with an overall response rate (ORR) of 89% and a median PFS of 19 months.^3,8 The randomized phase II RAMOSE trial is investigating the combination of osimertinib with ramucirumab.

The Hoosier Cancer Research Network is conducting the open-label, randomized phase II RAMOSE trial, which is evaluating osimertinib with ramucirumab vs without ramucirumab in EGFR-mutated advanced NSCLC (NCT03909334).⁹ Patients are eligible if they have EGFR exon 19 deletion or L858R mutation; asymptomatic brain metastases are permitted, but patients cannot have had prior EGFR or VEGF therapy.

Patients are randomized 2:1 to osimertinib plus ramucirumab or osimertinib alone until progression or unacceptable toxicity. The primary endpoint is PFS. The secondary endpoints include OS, response rate, disease control rate, and safety.

RAMOSE is currently open to accrual and is available at 12 sites in the United States.⁹ As of December 10, 2020, 47 patients had been accrued, and 43 were randomized. There are currently 110 patients enrolled with an interim analysis projected for the second quarter of 2023 (personal communication, Dr. Xiuning Le).

In the safety analysis performed on a subset of 39 patients, the safety profile for osimertinib plus ramucirumab did not seem to differ greatly from osimertinib alone, although there were the typical VEGF-related adverse events (AEs) such as hypertension. We also should note that patients randomized to the combination arm must come to clinic every 3 weeks for their ramucirumab infusion, which is quite different from taking osimertinib as a pill daily. It is important to bear in mind that even if a trial may demonstrate a small benefit with a novel regimen, we must determine whether that benefit outweighs the inconvenience for patients to come in for infusions every few weeks.

Turning now to the second‑line setting, it can be important to biopsy the tumor at progression on osimertinib to search for mechanisms of acquired resistance. Third-generation EGFR TKIs are associated with more off‑target resistance mechanisms, such as transformation to small-cell carcinoma or squamous cell carcinoma.¹⁰ Other resistance mechanisms include acquired gene alterations (eg, RET or BRAF fusions) as well as on‑target resistance such as the EGFR C797X mutation. MET amplification represents another important and therapeutically targetable resistance mechanism, with an incidence of 7% in the setting of first-line osimertinib and 6% in later lines. As shown here, the incidence and types of acquired resistance mechanisms vary between first‑line and later‑line osimertinib.

For the next few studies, I want to focus on MET amplification—one of the most important mechanisms of acquired resistance to osimertinib.

The initial clinical data supporting the strategy of targeting MET amplification come from the international, open-label, multiple-arm phase Ib TATTON trial, which evaluated safety and preliminary efficacy of osimertinib in combination with varying doses of investigational agents in patients with advanced EGFR-mutated, MET-amplified NSCLC who progressed after EGFR TKI treatment.¹¹ Patients were eligible for different cohorts based on whether they had been treated with a third-generation EGFR TKI and tumor T790M mutation status.

In the analysis shown here, the combination of osimertinib plus savolitinib, a MET inhibitor, was associated with an ORR of 30% in 69 patients previously treated with a third-generation EGFR TKI. This was essentially the first signal for targeting MET amplification after osimertinib.

The findings from TATTON led to the ORCHARD trial, a biomarker‑directed phase II study in patients with advanced EGFR-mutated NSCLC progressing on first‑line osimertinib.^12,13 ORCHARD is still open, but the various arms do open and close as they fill.

In ORCHARD, patients are biopsied after progressing on first‑line osimertinib, and the biopsy results direct their assignment into various baskets. There are baskets for MET alterations, EGFR C797X, other EGFR alternations, and ALK or RET rearrangements. I consider the most interesting arm to be evaluating the combination of osimertinib plus savolitinib in patients with MET alterations. This combination was associated with an ORR of 41% in a small analysis of 20 patients.¹⁴

Continuing with therapeutic targeting of MET amplification, the phase II SAVANNAH trial enrolled patients with advanced EGFR-mutated NSCLC and MET overexpression or MET amplification following progression on osimertinib.¹⁵ Using central assessment of tumor tissue collected after progression on osimertinib, MET amplification was identified by fluorescence in situ hybridization (FISH) with a MET copy number that was ≥5 and/or the MET:CEP signal ratio ≥2 (FISH5+). MET overexpression was determined by IHC with 3+ MET overexpression in ≥50% of tumor cells (IHC50+).

SAVANNAH initially was designed as a single-arm study where patients received savolitinib at varying doses in combination with osimertinib. After a protocol amendment, the study design was changed to a blinded, randomized trial in which patients were randomized to savolitinib plus osimertinib vs savolitinib plus placebo.

The primary endpoint was ORR, and the secondary endpoints were ORR stratified by MET IHC/FISH, PFS, OS, duration of response, and safety.

An analysis of SAVANNAH was presented at the 2022 World Conference on Lung Cancer.¹⁵ Among 193 patients with IHC90+ (defined as 3+ MET overexpression in ≥90% of tumor cells) and/or FISH10+ status (MET copy number ≥10), osimertinib plus savolitinib was associated with an ORR of 49%, which is really impressive. Among 108 patients with IHC50+ and/or FISH5+, the ORR was slightly lower at 32%; among those without IHC90+ and/or FISH 10+, the ORR was much lower, at 9%.

The safety profile of this combination was consistent with the established profiles for osimertinib and savolitinib as single agents.

I find these results quite exciting. These positive results have led to a phase III trial called SAFFRON.

SAFFRON is a randomized, open-label phase III trial in which patients with advanced EGFR-mutated NSCLC and MET amplification (FISH10+) or MET overexpression (IHC90+) are randomized to savolitinib plus osimertinib vs chemotherapy after progression on first-line or second-line osimertinib (NCT05261399).¹⁶ The primary endpoint is PFS, with secondary endpoints including OS, response, and pharmacokinetics.

This was a nice story showing how an initial signal from a phase I trial can lead to a phase II trial, which can then help us refine the biomarker, leading us to a phase III trial that uses that biomarker for selection. Hopefully, the results of SAFFRON will help us decide standard of care after progression on osimertinib in patients with MET amplification or MET overexpression.

Resistance to EGFR TKIs also can be mediated by expression of HER3, a member of the ERRB/HER protein kinase family that nonetheless lacks tyrosine kinase activity.^17-19 HER3 is not itself an oncoprotein but can heterodimerize with other receptor tyrosine kinases to activate oncogenic signaling, particularly through the PI3K/AKT/mTOR pathway. This signaling inhibits apoptosis and leads to cell proliferation. Although HER3 mutations and amplifications are rare in NSCLC, tissue microarray studies have identified HER3 expression in 83% of NSCLC specimens.²⁰

A promising agent under development is patritumab deruxtecan, a novel HER3‑targeted ADC.²¹ This agent consists of a human anti-HER3 IgG1 antibody conjugated via a tumor-selectable cleavable linker to a topoisomerase I inhibitor payload that is membrane permeable and has a short systemic half-life.

A multicenter, multicohort phase I trial of patritumab deruxtecan was published in Cancer Discovery in 2022.^22,23 In this analysis of the phase I trial, patients with advanced EGFR-mutated NSCLC were eligible if they had progressed on previous EGFR TKI; stable brain metastases were permitted. A tumor biopsy was required, but selection was not based on HER3 expression.

The dose-escalation phase evaluated patritumab deruxtecan at doses ranging from 3.2 to 6.4 mg/kg Q3W, with the dose-expansion phase using patritumab deruxtecan at 5.6 mg/kg Q3W.

Most patients in this analysis were female and approximately one half had a history of central nervous system (CNS) metastases.^22,23 The number of prior systemic lines of treatment ranged from 1‑9 with a median of 4, indicating that this is a heavily pretreated population. All patients had received an EGFR TKI, predominantly osimertinib (86%). Almost all (91%) also had received platinum‑based chemotherapy with or without osimertinib, and slightly more than one third also had received prior immunotherapy.

In the efficacy analysis of 57 patients treated with patritumab deruxtecan at 5.6 mg/kg Q3W, the confirmed ORR by blinded independent central review (BICR) was 39% (95% CI: 26% to 52%).^22,23 The disease control rate was quite high at 72%. Antitumor activity was comparable between the overall efficacy population and the subset treated with prior osimertinib plus platinum‑based chemotherapy.

I consider these to be impressive results, given how heavily pretreated this population was.

This analysis also examined the activity of patritumab deruxtecan across diverse mechanisms of EGFR TKI resistance.^22,23 As shown in this waterfall plot, responses were observed in patients with known and unknown resistance mechanisms. HER3 was expressed in all evaluable tumors, but responses did not clearly correlate with the degree of HER3 expression.

Regarding safety, 74% of patients who received patritumab deruxtecan at 5.6 mg/kg had a grade ≥3 AE, and 44% experienced a serious treatment‑emergent AE.^22,23 The most frequent grade ≥3 AEs were thrombocytopenia, neutropenia, and fatigue. Thrombocytopenia and neutropenia generally occurred early and were transient, and no patients had to discontinue treatment due to thrombocytopenia.

Overall, I would consider patritumab deruxtecan to be fairly well tolerated. There were treatment‑emergent AEs associated with death in 6% of patients, but none were judged by the investigators to be related to study treatment. Interstitial lung disease (ILD) was observed in 18 patients, but only 4 patients had ILD that was deemed treatment-related and none of these treatment-related ILD cases were grade 4/5.

This table highlights a selection of the ongoing studies of patritumab deruxtecan in EGFR‑mutated advanced NSCLC. The top row lists the phase I trial that we just discussed, which is still recruiting patients (NCT03260491). The second row highlights a phase I trial evaluating patritumab deruxtecan in combination with osimertinib in the first‑line setting as well as after progression on osimertinib (NCT04676477). The phase II HERTHENA‑Lung01 trial is evaluating patritumab deruxtecan in patients who have progressed after ≥1 EGFR TKIs and platinum‑based chemotherapy (NCT04619004). Finally, the phase III HERTHENA‑Lung02 trial is evaluating patritumab deruxtecan plus platinum‑based chemotherapy in patients who progressed after 1-2 EGFR TKIs, including a third-generation EGFR TKI (NCT05338970).

Next, we discuss third‑line clinical trials, such as the phase I CHRYSALIS-2 study involving the EGFR-MET bispecific antibody amivantamab plus lazertinib, a third-generation EGFR TKI.²⁴ I should note that although CHRYSALIS-2 is considered a third‑line clinical trial, this does not mean that we lack efficacy data for amivantamab plus lazertinib in earlier settings. In fact, the phase I CHRYSALIS trial had evaluated this combination in 45 patients with EGFR-mutated NSCLC who had progressed on osimertinib but were chemotherapy naive.²⁵ The response rate was 36% with a median duration of response of 9.6 months without any biomarker selection. Our discussion will focus on the CHRYSALIS-2 cohort that had been previously treated with platinum chemotherapy and osimertinib, but first I want to review amivantamab and the rationale for combining this agent with lazertinib.

Amivantamab is an EGFR-MET bispecific antibody.²⁶ It has multiple mechanisms of action, including receptor degradation, immune cell–directing activity, and inhibition of ligand binding. Amivantamab has demonstrated activity as monotherapy against a variety of activating and resistant EGFR mutations, including exon 20 insertions.²⁷ Based on positive results from another CHRYSALIS cohort, in 2021 the FDA granted accelerated approval to amivantamab for treatment of adults with locally advanced or metastatic NSCLC harboring EGFR exon 20 insertion mutations who had progressed on or after platinum-based chemotherapy.²⁸

Why combine amivantamab with lazertinib? Lazertinib is a highly selective, CNS-penetrant, third-generation EGFR TKI with efficacy against activating EGFR mutations, EGFR T790M, and CNS disease.^29,30 Furthermore, preclinical studies have reported improved antitumor activity when both the extracellular and the catalytic EGFR domains are simultaneously targeted by amivantamab and lazertinib.²⁶

As I already mentioned, amivantamab plus lazertinib has activity in patients who only received osimertinib.²⁵ These data, in addition to the desire to have good CNS penetration, led to further investigation of amivantamab in combination with lazertinib.

CHRYSALIS‑2 is a multicohort, dose-escalation/-expansion phase I trial evaluating amivantamab plus lazertinib in various populations assigned to dose-expansion cohorts.²⁴ There is a cohort for patients harboring EGFR exon 20 insertions, for uncommon EGFR mutations, and for biomarker validation in those with classical EGFR mutations. We will be focusing on cohort A, which has patients with classical activating EGFR mutations after osimertinib and platinum‑based chemotherapy. This cohort is fully accrued.

In all cohorts, patients were treated with the recommended phase II dose of amivantamab at 1050 mg IV (or 1400 mg for those ≥80 kg) plus lazertinib at 240 mg PO daily. The primary endpoint was ORR with key secondary endpoints including duration of response, clinical benefit rate, OS, PFS and safety.

Looking at the baseline characteristics of the 162 patients enrolled on CHRYSALIS-2 cohort A, the median age was 61.5 years, and 65% of the patients were female.²⁴ The study started off in Asia, and 61% of participants were Asian, with 69% being nonsmokers. Approximately 40% of patients had brain metastases with a mix of treatment status, as the trial initially allowed for patients with untreated and stable/asymptomatic brain metastases (19% at baseline), but a later amendment changed this to permit only those with treated brain metastases (22%).

I again want to highlight the median number of prior lines of treatment, which was 3 with a range from 2 to 14, and 28% of the patients had ≥4 lines of therapy. Almost one quarter of patients had received either first-line osimertinib followed by chemotherapy and 42% had received a first-generation/second-generation EGFR TKI followed by osimertinib and then chemotherapy. Notably, 35% were heavily pretreated or out of sequence. Taken together, these baseline data show that this is a very heavily pretreated, real‑world population.

The ORR per BICR was 33% with a median duration of response of 9.6 months.²⁴ There were 2 patients with complete response (1%), 52 with partial responses (32%), and 69 with stable disease (43%), with a clinical benefit rate of 57%. With a median follow-up time of 10 months, median PFS was 5.1 months and median OS was 14.8 months.

The forest plot here indicates a consistent ORR across all subpopulations. Notably, there was no marked difference in efficacy between those harboring an EGFR exon 19 deletion vs L858R mutation.

This waterfall plot depicts best antitumor response in target lesions and the table reports ORR by prior therapy group. The ORRs per BICR were fairly consistent, ranging from 21% to 39%, across the prior therapy groups.²⁴

No new safety signals were identified in this analysis.²⁴ AEs were predominantly low grade with most being EGFR related and MET related. EGFR-related AEs included rash (all grade in 44%), diarrhea (22%), and stomatitis (39%), and MET-related AEs included edema (27%) and hypoalbuminemia (43%). Of note, 80% of patients experienced cumulative grouped rash‑related AEs, of which 10% were grade ≥3.

Another notable AE is infusion‑related reaction, which occurred in 67% of patients—consistent with what we generally see with amivantamab. I have found that if you counsel the patient, infusion staff, and the nurses and also premedicate the patient carefully, then these infusion‑related reactions are quite manageable and should not prevent successful drug administration.

Because amivantamab is such a promising new agent with antitumor activity in this setting, there are multiple ongoing trials evaluating amivantamab in various combinations in EGFR-mutated advanced NSCLC. For example, a nonrandomized phase II trial is evaluating the CNS activity of amivantamab plus lazertinib in cohorts with CNS metastases or leptomeningeal disease (NCT04965090). The fully accrued phase III MARIPOSA trial is comparing first-line amivantamab plus lazertinib vs osimertinib (NCT04487080). The phase III MARIPOSA-2 trial is a 3-arm study comparing amivantamab plus lazertinib plus platinum-based chemotherapy vs amivantamab in patients who progressed on or after an EGFR TKI and platinum-based chemotherapy (NCT04988295).

A phase III trial is also comparing 2 different subcutaneous formulations of amivantamab vs the standard IV formulation (NCT05388669). Receiving IV amivantamab is fairly time intensive for patients because they must come in for infusion weekly to start and then every other week thereafter; hopefully, a subcutaneous formulation would ease that patient burden.

I also want to note that the ongoing phase II METalmark trial is evaluating amivantamab plus capmatinib, a MET inhibitor, in patients with metastatic NSCLC harboring MET amplification or exon 14 skipping mutations (NCT05488314). This combination could potentially represent another approach against MET-mediated resistance to osimertinib, which we discussed earlier. I am very interested to see these results as well.