Savolitinib/Osimertinib Yields Responses in EGFR-Mutant, MET-Amplified, Osimertinib-Resistant NSCLC

James Chih-Hsin Yang, MD, PhD

The addition of the selective MET TKI savolitinib (Orpathys) to osimertinib (Tagrisso) demonstrated consistent safety and higher clinical activity vs savolitinib plus placebo in patients with EGFR-mutated, MET-amplified advanced non–small cell lung cancer (NSCLC) whose disease had previously progressed on osimertinib, according to findings from a proteogenomic analysis presented at the 2024 AACR Annual Meeting.¹

Findings showed that savolitinib plus osimertinib produced an objective response rate (ORR) of 57% (95% CI, 29%-82%) compared with 13% (95% CI, 2%-38%) with savolitinib plus placebo in the overall patient population (n = 30). This was entirely composed of partial responses. The percentage of patients who achieved stable disease (SD) for 5 weeks or longer was 36% with the combination and 31% with placebo plus savolitinib. Disease progression occurred in 7% and 50% of patients in these respective groups.

The median progression-free survival (PFS) with savolitinib plus osimertinib in the overall population was 7.4 months (95% CI, 5.6-not calculable (NC)] compared with 1.6 months (95% CI, 1.3-4.1) with savolitinib plus placebo (HR, 0.07; 95% CI, 0.01-0.25). Importantly, 2 patients who crossed over to the combination arm due to early study termination did not show progression at the data cutoff of December 1, 2022, and were accordingly censored. The median duration of response (DOR) was 30.6 weeks (95% CI, 18.9-NC) in the experimental arm vs NC (95% CI, 12.4-NC) in the control arm.

Further efficacy analysis of patients with higher MET amplification and/or overexpression cutoffs (n = 15) experienced greater clinical activity with the combination vs placebo regimen, with respective ORRs of 63% (95% CI, 24%-91%) and 29% (95% CI, 4%-71%). SD for 5 weeks or more was achieved by 38% and 29% of patients in the experimental and control arms, respectively. No patients in this group experienced disease progression with savolitinib plus osimertinib, as opposed to 43% of patients in the control arm. The median PFS was 8.2 months (95% CI, 4.1-NC) with savolitinib plus osimertinib vs 4.0 months (95% CI, 1.3-NC) with savolitinib plus placebo; the median DOR in these respective arms was 30.6 weeks (95% CI, 23.0-NC) and NC (95% CI, 12.4-NC).

“Savolitinib plus osimertinib demonstrated clinical activity, with savolitinib plus placebo showing lower clinical activity,” lead study author James Chih-Hsin Yang, MD, PhD, of the National Taiwan University Hospital and National Taiwan University Cancer Centre in Taipei, Taiwan, and colleagues, wrote in a poster presentation of the data. “[However], due to small sample sizes, these data need to be interpreted with caution.”

Notably, this study was ended early based on early antitumor activity seen in the ongoing phase 2 SAVANNAH study (NCT03778229) of savolitinib in this same population. The SAVANNAH study used a twice-daily, 300-mg dose of savolitinib and higher MET overexpression cutoffs.

Although the third-generation, central nervous system–active TKI osimertinib is considered the first-line standard of care in advanced, EGFR-mutant NSCLC, most patients will eventually develop osimertinib resistance, decreasing the agent’s efficacy. As one of the most common acquired mechanisms of resistance to osimertinib is MET amplification, the addition of the oral, potent, and highly selective MET TKI savolitinib to osimertinib may help overcome this acquired resistance.

Accordingly, this phase 2 double-blind study was conducted to assess savolitinib’s individual efficacy and safety contribution in combination with osimertinib in EGFR-mutant and MET-amplified NSCLC. The trial enrolled patients 18 years of age or older, or those at least 20 years of age in Japan, with locally advanced or metastatic NSCLC who displayed a known EGFR-sensitizing mutation alongside MET amplification. MET amplification and/or overexpression was defined as a MET gene copy number of at least 5 per fluorescence-in-situ (FISH) and/or a MET/CEP7 ratio of 2 or more with central confirmation by FISH. Prior progression on osimertinib was required, although patients did not need to receive osimertinib as their most recent therapy. Other inclusion criteria were treatment with 3 or fewer prior therapies for advanced NSCLC and an ECOG performance status (PS) of 0 or 1.

Once enrolled, patients were stratified according to whether they had received osimertinib monotherapy in the first-line setting or beyond. Patients were randomly assigned 1:1 to receive 300 mg of savolitinib alongside 80 mg of osimertinib every day vs 300 mg of daily savolitinib plus placebo. Plasma samples were taken on day 1 prior to initial dosing, week 3, and week 6.

The primary end point was investigator-assessed ORR per RECIST v1.1 criteria. Key secondary end points included best DOR, PFS, overall survival, and safety. Efficacy in patients with higher MET amplification/overexpression cutoffs of FISH 10+ and/or at least 90% of tumor cells with 3+ staining per immunohistochemistry was an exploratory end point.

Assessments continued every 6 weeks for the first 24 weeks, at which point they continued until a patient experienced objective disease progression, unacceptable toxicity, withdrew consent, or met another discontinuation criterion. Notably, patients in the control arm were permitted to cross over to receive the experimental regimen if they experienced disease progression on the study; this was permitted for all patients after early unblinding.

Of the 30 patients recruited onto the study, 14 received the savolitinib combination and 16 received the placebo regimen. The median age of patients was 60.5 years (range, 37-77) in the combination arm and 61.5 years (range, 45-78) in the placebo arm. Over half of patients were female (50% with combination; 56% with placebo), and most of the patients were Asian (86%; 94%), had an ECOG PS of 1 (86%; 81%), and had received 2 or more prior lines of therapy (86%; 81%). Almost all patients had received osimertinib as their immediate prior therapy (100%; 94%). Brain metastases were reported in 36% and 19% of patients in the combination and placebo arms, respectively.

At the data cutoff, 29% of patients were still on study treatment. All patients in the placebo arm had discontinued their initial treatment, 10 of whom crossed over to the experimental arm.

Overall, 29 patients had evaluable baseline plasma EGFR mutation results. Among these patients, 12 in the experimental arm (n = 14) and 13 in the control arm (n = 15) had detectable plasma EGFR-mutated circulating tumor DNA (ctDNA) at baseline. At week 3, ctDNA clearance indicated benefit with the savolitinib combination. Six patients treated with savolitinib plus osimertinib and 2 with savolitinib plus placebo achieved a clearance response, for an ORR of 63% (95% CI, 24%-91%). Moreover, molecular response rates at this time point were 50% (95% CI, 21%-79%) with the savolitinib combination and 15% (95% CI, 2%-45%) with the placebo regimen. However, interpretation of ctDNA clearance was limited due to a lack of samples at week 6. Two patients who crossed over to the experimental arm and had valid post-crossover plasma results demonstrated improved ctDNA responses following treatment initiation.

Regarding safety, the toxicity profiles for savolitinib and osimertinib in this study were consistent with known data, and no new safety signals were reported. All patients experienced 1 or more treatment-emergent adverse effects (AEs). In the experimental arm, Grade 3 or higher AEs were reported in 21% of patients, and serious AEs occurred in 19% of patients. These percentages were 31% and 19%, respectively, in the control arm.

AEs leading to death or treatment discontinuation in the savolitinib plus osimertinib arm were seen in 7% of patients each; these rates were 19% and 13%, respectively, with savolitinib plus placebo. No AEs leading to death were considered treatment related; however, both a grade 2 hepatic enzyme increase and a grade 3 amylase increase leading to discontinuation were considered treatment related.

The most common any-grade AEs in the combination and placebo arms were nausea (50%; 19%), peripheral edema (36%; 25%), vomiting (21%; 31%), amylase increase (14%; 25%), decreased appetite (7%; 25%), hypoalbuminemia (21%; 13%), anemia (14%; 13%), dyspnea (7%; 19%), headache (29%; 0%), hyponatremia (14%; 13%), insomnia (7%; 19%), pneumonia (0%; 25%), arthralgia (14%; 6%), constipation (14%; 6%), cough (0%; 19%), dermatitis acneiform (21%; 0%), fatigue (21%; 0%), musculoskeletal chest pain (0%; 19%), and stomatitis (14%; 6%).

Reference

Chih-Hsin Yang J, Chen YM, Batra U, et al. Savolitinib (savo) + osimertinib (osi) vs savo + placebo (PBO) in patients (pts) with EGFR-mutated (EGFRm), MET-amplified advanced NSCLC with progression on osi. Presented at: 2024 AACR Annual Meeting; April 5-10, 2024; San Diego, CA. Abstract CT251.