Phase I Study of AMG 337, a Highly Selective Small-molecule MET Inhibitor, in Patients with Advanced Solid Tumors

MET is a receptor tyrosine kinase that regulates cell survival, proliferation, migration, matrix invasion, and branching morphogenesis (1–5). Binding of its ligand hepatocyte growth factor (HGF) induces receptor dimerization and autophosphorylation and activates intracellular signaling cascades, including PI3K–AKT, CDC42, RAP1, and RAS–MAPK pathways (6–8). MET gene amplification and/or activating mutations are observed in many solid tumors, including glioblastoma and esophageal, colorectal, and gastric cancers, and often correlate with poor prognosis (9–14). Altered MET expression also correlates with drug resistance in some tumor types (15, 16).

Evidence supporting a role for MET/HGF signaling in tumor biology led to the investigation of MET as a cancer therapy target. Two main approaches have been tested: MET inhibition with monoclonal antibodies (mAbs) (e.g., onartuzumab; ref. 17) or small-molecule inhibitors [e.g., capmatinib (18), volitinib (19), crizotinib (20), AMG 208 (21)], and HGF blockade with mAbs [e.g., rilotumumab (22), ficlatuzumab (23)]. Although early-phase studies have demonstrated clinical activity with both approaches (17, 19, 22, 24–26), subsequent trials showed rilotumumab and onartuzumab did not improve outcomes in patients with MET overexpression (27, 28). Identifying patients with alterations in the MET pathway (including MET amplification) has been proposed as a method to identify cancer subtypes that may be dependent on MET signaling and sensitive to MET inhibition (29–31).

AMG 337 is a highly selective small-molecule inhibitor of the MET receptor signaling pathway (32). The objective of this first-in-human phase I study of AMG 337 was to evaluate the maximum tolerated dose (MTD), safety, tolerability, and pharmacokinetics of AMG 337 in patients with advanced solid tumors (ClinicalTrials.gov Identifier: NCT01253707).

Included patients were ≥18 years of age and had documented advanced solid tumors refractory to standard treatment, or for which no standard therapy was available; measurable disease per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (33); Eastern Cooperative Oncology Group (ECOG) performance status ≤2 (34); adequate organ function; and no previous MET-targeted therapy except MET pathway–directed antibody therapy. Full inclusion/exclusion criteria are provided in the Supplementary Material. Patients with nonmeasurable but evaluable tumors were allowed in the dose-escalation phase only. Tumors with MET amplification were required for inclusion in the dose-expansion phase. This study was conducted in accordance with the principles of the applicable country, FDA, and International Conference on Harmonization (ICH) Good Clinical Practice (GCP) regulations/guidelines. Compliance with ICH GCP guidelines provides public assurance that the rights, safety, and well-being of trial subjects are protected, consistent with the principles that have their origin in the Declaration of Helsinki. Study procedures were approved by institutional review boards at each participating center. All patients provided written informed consent.

This was a first-in-human, open-label, phase I study of oral AMG 337 comprising dose-escalation and dose-expansion phases conducted between December 8, 2010 (first patient enrolled) and January 26, 2016 (data cutoff) at 6 centers in the United States. Primary endpoints were assessment of treatment-emergent adverse events (AEs), establishment of the MTD, and determination of AMG 337 pharmacokinetics. Secondary endpoints were clinical response, objective response rate per RECIST version 1.1 (33), and duration of response. Exploratory endpoints included assessment of MET mutation/amplification status, MET expression, and biomarkers of MET signaling.

The dose-escalation phase had a modified 3+3+3 design (Fig. 1A). Sequential enrollment of 8 cohorts for once-daily oral treatment (25–500 mg) was planned. AMG 337 was self-administered orally in a fasted state. For the once-daily dose-escalation cohorts, patients self-administered AMG 337 on day 1, waited 48 hours, restarted dosing on day 3, and then continued daily self-administration through day 28. A weeklong treatment-free period began on day 29. For the twice-daily dose-escalation cohorts, the first dose of AMG 337 was self-administered on day 1 and continued every 12 ± 2 hours through day 28 without a drug-free break. Sequential dose escalation continued until the MTD was determined. A dose-limiting toxicity (DLT) was defined as any grade ≥3 nonhematologic or grade 4 hematologic event that occurred during the first 28 days of treatment (the DLT evaluation window) with additional qualifications (Supplementary Material).

Dose-limiting toxicities of headache and hypertension occurred in 3 of 8 patients initially enrolled in the 200-mg cohort. The protocol was amended as described in the Supplementary Material to allow medical management of hypertension and headache, and the dose level was immediately de-escalated (per protocol) to 150 mg once daily. Reescalation to 200 mg (7 additional patients enrolled) and then 300 mg was performed thereafter. Twice-daily dosing was explored to potentially decrease maximum plasma concentration (C_max)–related AEs (e.g., headache) observed in the once-daily cohorts; 4 twice-daily treatment cohorts (ranging from 100 to 250 mg) were assessed.

In the dose-expansion phase, patients with MET amplification (Supplementary Material) were enrolled at the MTD (300 mg once daily; Fig. 1B). Dose expansion consisted of 3 cohorts: gastric, gastroesophageal, or esophageal cancer (E1), non–small cell lung cancer (NSCLC; E2), and other tumor types (E3). AMG 337 doses were adjusted or delayed per prespecified criteria if patients experienced toxicity (Supplementary Material).

Patients began a week-long treatment-free period on day 29 for assessment of tumor response. Patients with complete response (CR), partial response (PR), or stable disease (SD), and without DLTs, resumed treatment on day 36 and continued until disease progression, withdrawal of consent, or intolerable toxicity. Patients who achieved a PR, CR, or who had clinical benefit, but experienced a grade ≥3 AE on or after day 29 suspended treatment until the AE resolved and then resumed AMG 337 with appropriate dose reduction.

For patients who received at least 1 dose of AMG 337, AEs were evaluated from day 1 of dosing through the last day of dosing, and at 30 to 37 days after the last dose during the safety follow-up visit. AEs were graded per Common Terminology Criteria for Adverse Events version 4.0.

Baseline imaging (CT or MRI) was performed within 28 days before day 1. Follow-up scans were performed at week 5 (before day 36 visit), week 9 (before day 64 visit), and every 8 weeks thereafter. Radiologic tumor response was assessed per RECIST version 1.1.

Pharmacokinetic and biomarker analyses are described in the Supplementary Material.

Analyses of safety, tolerability, efficacy, pharmacokinetics, and biomarker data were descriptive and hypothesis generating. The safety/response analysis set included all patients who received at least 1 dose of AMG 337. Pharmacokinetic analyses included all enrolled patients for whom at least 1 pharmacokinetic parameter or endpoint could be adequately estimated. Analysis of efficacy data was descriptive.

A total of 221 patients were screened and 111 were enrolled in the study (Supplementary Fig. S1); the first patient was enrolled on December 8, 2010, and the last patient completed the investigational product on December 3, 2015. The development of AMG 337 was prematurely terminated because of lower than expected efficacy in a phase II trial of the drug. Further enrollment in this phase I study was terminated on November 24, 2015; however, treatment of the 111 patients on study continued. All enrolled patients received at least 1 dose of AMG 337 and were included in the safety analyses. The most common solid tumor type was gastrointestinal tumor [gastroesophageal junction (GEJ)/gastric/esophageal (21%); colorectal (19%)]. Patient characteristics are listed in Table 1. Twenty-seven patients (24%) had MET amplification by local testing (defined as MET:CEN7 ratio ≥2.0 as determined by fluorescence in situ hybridization [FISH] or next-generation sequencing). Patients received a median (range) of 53 (1–1,270) AMG 337 doses. One hundred and eight patients (97%) discontinued treatment [disease progression, n = 75 (68%); AEs, n = 17 (15%); consent withdrawal, n = 9 (8%); noncompliance, n = 3 (3%); administrative decision, n = 2 (2%); and death, n = 2 (2%)]; 3 patients (3%) remained on treatment as of the cutoff date of January 26, 2016. Unless otherwise noted, safety, efficacy, and pharmacokinetic analyses are reported as of the data cutoff date, at which point 45 patients (41%) had completed the study.

No DLTs were reported in the 3, 4, and 14 patients in the 25-, 50-, and 100-mg once-daily dose cohorts. DLTs were reported in 3 of the initial 8 patients enrolled at the 200-mg once-daily dose (grade 3 headache, n = 2; grade 3 hypertension, n = 1), resulting in immediate dose de-escalation (per protocol) and a protocol amendment to modify the DLT definition. One patient in the 150-mg once-daily group (n = 9) developed a DLT (grade 3 headache); no other DLTs were observed and the dose was re-escalated. No DLTs were reported in the 7 additional patients enrolled at 200 mg once daily. Dose escalation to 300 mg once daily (n = 8) yielded 1 DLT (grade 3 headache), and escalation to 400 mg once daily (n = 6) yielded 3 DLTs (grade 3 headache, n = 2; increased amylase, n = 1). Therefore, the MTD of AMG 337 was determined to be 300 mg once daily and this dose was evaluated in the dose-expansion phase.

No DLTs were reported in the twice-daily dose cohorts except for the 200-mg twice-daily cohort, in which 2 of 6 patients experienced DLTs (generalized edema, increased blood creatine phosphokinase, and back pain in 1 patient; pain in extremity in 1 patient). Enrollment continued in this cohort and none of the additional 4 patients experienced DLTs. After the 28-day DLT window for the 250-mg twice-daily cohort (n = 4), overall safety and pharmacokinetic data were reviewed and it was decided not to escalate further. At doses above 250 mg twice daily, the C_max was predicted to be well above that with 300 mg once daily, the MTD for daily dosing, and twice-daily dose escalation beyond this point would therefore not have helped overcome C_max-related AEs. In addition, at doses beyond 200 mg twice daily, AEs were observed that were thought to possibly limit long-term tolerability (e.g., edema); thus, dose escalation was stopped at 250 mg twice daily. The MTD for twice-daily dosing was not reached and the maximum administered dose was 250 mg twice daily.

Overall, 13 patients had a total of 23 DLTs. Headache was the most frequent (6 patients, 8 events). Other DLTs (1 event each) were abdominal pain, amylase increased, arthralgia, ascites, back pain, blood creatine phosphokinase increased, fatigue, generalized edema, hypertension, liver function test abnormal, nausea, peripheral edema, pain in extremity, vomiting, and weight increased.

Treatment-emergent AEs were reported for all patients (Supplementary Table S1). The most common treatment-related AEs were headache (63%) and nausea (31%; Table 2). Grade ≥3 treatment-related AEs occurred in 23 patients (21%); the most common were headache (n = 6) and fatigue (n = 5). AE incidences were similar between the once-daily and twice-daily dosing groups (Table 2); however, patients in the twice-daily cohorts experienced less severe headache (Supplementary Table S1). Forty-four patients (40%) had serious AEs; 10 patients (9%) had serious treatment-related AEs. Nine patients (8%) had fatal AEs; most (n = 7) were related to underlying disease. Other fatal AEs were cardiac arrest (n = 1) and dyspnea and confusional state (n = 1); none were deemed related to treatment by investigators. One patient died during the 30-day follow-up period but did not have a fatal AE reported. A total of 47 patients (42%) had doses altered or withheld because of AEs.

Adverse events of interest included headache and edema. Seven of 8 patients (88%) in the 300-mg once-daily group (escalation) reported headache; 1 patient experienced a grade ≥3 event. Peripheral edema occurred in 24% of patients, including 1 patient at the 300-mg once-daily (escalation) dose. No grade 3/4 events of peripheral edema occurred; 1 patient had grade 3 generalized edema (200 mg once daily).

Nine patients (8.1%) exhibited absolute heart rate–corrected QT interval (Fridericia's formula; QTcF) >480–500 msec, and 5 (4.5%) exhibited QTcF >500 msec. Thirty-nine patients (35.1%) reported maximum QTcF increase from baseline of >30–60 msec, and 10 patients (9.0%) reported maximum increase from baseline >60 msec. This prolongation weakly correlated with the higher serum drug concentration. A linear mixed-effects model relating AMG 337 concentration and change in QTcF (ΔQTcF) indicated that 300 mg once daily for 28 days (mean C_max, 4,140 ng/mL) could result in a ΔQTcF of approximately 9.4 msec.

Area under the AMG 337 plasma concentration–time curve from 0 to 24 hours (AUC_0–24) on day 1 increased dose proportionally in the once-daily dose-escalation groups up to 200 mg and peaked at 300 mg (Table 3). Estimates of AUC_0–24 on day 1 were similar to estimates on day 28, indicating AMG 337 did not accumulate over time. Elimination half-life (t_½) estimates in the once-daily escalation groups on day 1 ranged from 5.9 to 7.0 hours. The 300-mg once-daily dose achieved higher AMG 337 exposure levels than the 200-mg once-daily dose (Supplementary Fig. S2A and S2B).

Among patients who received 300 mg once daily, the mean estimate of C_max on day 1 was 4.01 μg/mL; mean time to C_max was 2.9 hours. Following twice-daily treatment, C_max estimates were lower than those observed at comparable daily doses in the once-daily dose-escalation phase, and C_max on day 1 peaked at a dose of 150 mg twice daily.

Among patients who received AMG 337 200 mg once daily, mean trough concentration (C_trough) was 416 ng/mL on day 1; this corresponds to a mean unbound C_trough of 312 ng/mL, a concentration that is approximately 10-fold higher than the unbound concentration (27 nmol/L) associated with a 90% inhibition of MET signaling in the TPR-MET mouse tumor model (32).

Eighty-two patients had data evaluable for tumor response per RECIST version 1.1 by central review; the remaining 29 patients either had nonmeasurable disease or discontinued before the postbaseline scan. Best responses of CR (n = 1), PR (n = 10), and SD (n = 53) were observed. The objective response rate was 9.9% (11/111; 95% CI, 5.1%–17.0%) in all patients (Fig. 2A) and 29.6% (8/27; 95% CI, 13.8%–50.2%) in patients with known MET-amplified tumors (Fig. 2B). The median (range) duration of response was 202 (51–1,430+) days in all patients (Fig. 2C) and 197 (64–1,430+) days in patients with MET-amplified tumors. The CR occurred in a 63-year-old man with stage IV MET-amplified (FISH ratio, 25.0) distal esophageal adenocarcinoma who was treated on the 200-mg once-daily dosing schedule (dose reduced to 150 mg once daily after 35 weeks). He achieved a CR at week 33 (imaging shown in Fig. 2D). Another patient achieved 100% tumor reduction by local testing but was classified as PR because no target lesion was identified in central testing.

Patients with PR received 200 mg once daily (n = 2), 300 mg once daily (n = 6), 400 mg once daily (n = 1), and 200 mg twice daily (n = 1). Of these 10 PRs, 7 occurred in MET-amplified patients: 5 with GEJ/gastric/esophageal tumors, 1 with NSCLC, and 1 with carcinoma of unknown origin. The remaining PRs occurred in 1 patient with mesothelioma with MET overexpression by immunohistochemistry (IHC) local testing (50% MET staining, intensity 2) and unknown MET amplification status (failed test), and in 2 patients with gastric and breast cancers, respectively, with unknown MET amplification status. The MET/CEP7 ratio for known MET-amplified responders ranged from 4.22 to 25. The median (range) duration of treatment until PR was 189 (28–1,070) days; the median (range) time to PR was 119 (29–730) days; and the median (range) duration of PR was 147 (51–835) days for all 10 patients who experienced PR and 91 (64–700) days for the 7 patients with MET-amplified tumors. Fifty-three patients had a best response of SD, including patients with colon/colorectal cancer (n = 13), GEJ/gastric/esophageal tumors (n = 8), and soft-tissue sarcoma (n = 5); duration of SD ranged from 2 to 89 weeks. Among the 19 patients with MET-amplified tumors who did not achieve a CR/PR, 9 achieved SD, 3 had progressive disease, and 7 had unknown best tumor response. The patients with unknown response may have had a response that was not measurable, or they may not have had a postbaseline scan.

Best tumor response by central testing ,MET amplification or overexpression by local testing, and the relationship between MET ratio and MET gene copy number and response are shown in Table 4 and Supplementary Fig. S3. Median [interquartile range (IQR)] FISH ratio was 1.14 (1.07–1.35); median (IQR) MET gene copy number by FISH was 3.50 (2.75–4.90). Among all patients, median FISH ratios for responders versus nonresponders were 5.30 versus 1.13.

MET is central in the regulation of tumor cell growth, survival, and metastasis (35, 36) and is an oncogenic driver in a variety of aggressive solid tumors (37–41). AMG 337 blocks MET signaling and inhibits growth of MET-dependent cancer cell lines, including gastric cancers (32).

In this phase I study, the MTD of AMG 337 was determined to be 300 mg once daily and the MTD for twice-daily dosing was not reached. AMG 337 was associated with a dose-dependent increase in headache incidence, possibly due to its potent inhibition of adenosine transporters (refs. 42, 43; Amgen Inc., data on file). Twice-daily treatment was added to reduce C_max and headache occurrence, as was caffeine (as an A2A receptor antagonist; ref. 44). Consistent with other MET inhibitors (17, 26, 45–48), low-grade peripheral edema occurred in 24% of patients who received AMG 337; however, no grade ≥3 peripheral edema was observed. One case of grade 3 generalized edema was reported.

Pharmacokinetic analyses showed that AMG 337 exposures increased up to 300 mg once daily. Consistent with a relatively short plasma t_1/2, AMG 337 did not accumulate with multiple-dose administration. At 200 mg once daily, mean AMG 337 plasma concentrations exceeded the average concentration associated with tumor regression in the U87 MG xenograft model (0.530 μg/mL; Amgen, data on file).

Preclinical studies showed that AMG 337 inhibits the phosphorylation of MET and downstream effectors in various MET-amplified cancer cell lines, blocking MET-dependent cell proliferation and inducing apoptosis (32). In this study, however, a direct relationship between AMG 337 exposure and clinical response was not observed. Unfortunately, posttreatment tumor samples that would have allowed for the assessment of target engagement in tumor tissue were not collected during this study. Within the esophagogastric cancers, the subgroup of MET-amplified patients has a reportedly worse prognosis; thus, the response and disease control rates in this subgroup are particularly intriguing (14, 49). Three patients were still receiving AMG 337 as of the data cutoff date and had ongoing responses (1 CR and 2 PR). Total study duration for these patients at the time they ended the study and were transferred to an expanded access protocol was 1,892, 953, and 1,421 days, respectively.

Patient selection is an essential component for the development of potent and selective targeted therapies, as was observed in studies with rilotumumab (RILOMET-1) and onartuzumab (METGastric) that showed that MET overexpression was not a predictive biomarker of response (27, 50). On the basis of these results, we hypothesized that MET amplification would be a better marker for tumors that are dependent on the MET pathway, and hence, we enrolled only patients with MET-amplified tumors into the dose-expansion cohorts to receive AMG 337. In this study, although evidence of durable clinical responses with 200- or 300-mg once-daily doses of AMG 337 was demonstrated in patients with refractory MET-amplified GEJ/gastric/esophageal tumors and other tumor types, we concluded that the presence of MET amplification alone was not sufficient to predict response across the wide range of doses tested. In cases where the MET pathway is indeed acting as a driver of tumor growth, inhibiting the pathway can lead to dramatic responses; however, identifying a predictive biomarker for all patients, or the appropriate level of MET amplification, has been challenging.

Despite failure of several MET pathway inhibitors (27, 51–54) to produce a significant response in patients with MET amplification, some small-molecule MET inhibitors have shown activity in patients with highly MET-amplified tumors. In a study evaluating crizotinib in patients with NSCLC, PRs were observed in 4 of 12 evaluable patients, with a 35-week median duration of response (55). In a trial of ABT 700, an anti–c-MET antibody, PRs were observed in 3 of 4 patients with MET-amplified gastric or esophageal cancer (56). The antibody–drug conjugate ABBV-399 showed promising preclinical antitumor activity and is currently recruiting for phase I/Ib and II trials (57–60). The criteria by which tumors are defined as MET amplified vary across studies; thus, potential efficacy in MET-amplified patient subgroups may be obscured by variable or less stringent MET amplification criteria. A recent review (61) suggests that appropriate determination of the threshold for MET amplification is key in its use as a predictive biomarker in MET inhibitor studies. High-level MET amplification, defined as a MET/CEP7 ratio of at least 5, correlates to higher MET protein levels, poorer patient prognosis, and absence of other oncogenic driver mutations (61, 62). More stringent criteria for MET amplification, regardless of the methodology used for scoring, may better define the patient population that would respond to anti-MET therapies (61).

Molecular heterogeneity complicates the detection and significance of gene amplification as an oncogenic driver (63); therefore, monotherapy with a MET pathway inhibitor may only be effective in some clones. The de novo or acquired heterogeneity of MET amplification is a limitation to the overall effectiveness of treatment that requires further study. Furthermore, the samples analyzed for biomarkers in this study were from archival formalin-fixed, paraffin-embedded specimens, which may not be representative of patients' MET status at enrollment or during the course of treatment. Ultimately, a larger and more comprehensive dataset is required to identify additional biomarker(s) beyond MET amplification that may indicate which patients are most likely to benefit from therapy. Because MET amplification may occur concomitantly with other genetic aberrations (e.g., HER2 amplification) in tumor cells, combination therapy with cytotoxic agents or other targeted therapies (e.g., antiangiogenic agents, checkpoint inhibitors) could overcome this limitation. Finally, MET amplification defined by FISH and MET overexpression defined by IHC do not always correlate (31, 41) and may represent differing biologic functions with incompletely defined implications for MET-directed cancer therapy.

In conclusion, this phase I study suggested that orally administered AMG 337 has manageable toxicity, with an MTD and recommended phase II dose of 300 mg once daily. AMG 337 produced meaningful responses in patients with MET-amplified tumors, most notably in a subset of patients with GEJ/gastric/esophageal tumors. Most responders were either highly MET amplified or had MET protein overexpression. Although durable responses were observed in some MET-amplified tumors, many did not respond, indicating the complexity of the MET signaling pathway and likely requirement of combination strategies. Further biomarker refinement is necessary to improve patient selection in future clinical trials. A phase II study of AMG 337 in patients with MET-amplified GEJ/gastric/esophageal adenocarcinoma or other MET-amplified solid tumors has been conducted and is reported separately (64).

D.S. Hong reports receiving commercial research grants from AbbVie, Adaptimmune, Amgen, AstraZeneca, Bayer, Bristol-Myers Squibb, Daiichi-Sankyo, Eisai, Fate Therapeutics, Genentech, Genmab, Ignyta, Infintiy, Kite, Kyowa, Lilly, LOXO, Merck, MedImmune, Mirati, MiRNA, Molecular Templates, Mologen, NCI-CTEP, Novartis, Pfizer, Seattle Genetics, and Takeda; receives speakers bureau honoraria from Bayer; holds ownership interest (including patents) in Molecular Match, OncoResponse, and Presagia Inc.; is a consultant/advisory board member for Bayer, Alpha Insights, Axiom, Adaptimmune, Baxter, Genentech, GLG, Group H, Guidepoint Global, Infinity, Janssen, Merrimack, Medscape, Numab, Pfizer, Seattle Genetics, Takeda, Trieza Therapeutics; and reports other remuneration in the form of travel accommodations and expenses from LOXO and MiRNA. P.M. LoRusso is a consultant/advisory board member for AbbVie, Roche/Genentech, Agenus, Ipsen, Pfizer, Genmab, Bayer, Agios, Menarini, SOTIO, IMAB, CytomX, Five Prime, and Takeda. O. Hamid is a consultant/advisory board member for and reports receiving speakers bureau honoraria from Amgen. F. Janku reports receiving commercial research grants from Agios, BioMed Valley Discoveries, Novartis, Symphogen, Astellas, FujiFilm Pharma, and Piqur, and is a consultant/advisory board member for Deciphera. D.V.T. Catenacci reports receiving speakers bureau honoraria from Amgen, Lilly, Merck, Bristol-Myers Squibb, Taiho, Astellas, Roche, Gritstone, Guardant Health, and Five Prime, and is a consultant/advisory board member for Astellas, Merck, and Amgen. E. Chan is an employee of and holds ownership interest (including patents) in Amgen. T. Bekaii-Saab is a consultant/advisory board member for Amgen, Bayer, Genentech, Roche, Seattle Genetics, Imugene, Immuneering, and Merck, and reports receiving other remuneration from Armo DSMB and Silajen DSMB. S. Gadgeel reports receiving other commercial research support from Merck and is a consultant/advisory board member for Bristol-Myers Squibb, Genentech/Roche, Takeda/Ariad, AbbVie, and AstraZeneca. R.D. Loberg is an employee of and holds ownership interest (including patents) in Amgen. B.M. Amore holds ownership interest (including patents) in Amgen. Y.C. Hwang is an employee of and holds ownership interest (including patents) in Amgen. G. Ngarmchamnanrith is an employee of and holds ownership interest (including patents) in Amgen. E.L. Kwak is an employee of Novartis Institutes for Biomedical research and reports receiving other commercial research support from Amgen. No potential conflicts of interest were disclosed by the other authors.

There is a plan to share data. This may include deidentified individual patient data for variables necessary to address the specific research question in an approved data-sharing request; also related data dictionaries, study protocol, statistical analysis plan, informed consent form, and/or clinical study report. Data sharing requests relating to data in this manuscript will be considered after the publication date and (i) this product and indication (or other new use) have been granted marketing authorization in both the US and Europe, or (ii) clinical development discontinues and the data will not be submitted to regulatory authorities. There is no end date for eligibility to submit a data sharing request for these data. Qualified researchers may submit a request containing the research objectives, the Amgen product(s) and Amgen study/studies in scope, endpoints/outcomes of interest, statistical analysis plan, data requirements, publication plan, and qualifications of the researcher(s). In general, Amgen does not grant external requests for individual patient data for the purpose of reevaluating safety and efficacy issues already addressed in the product labeling. A committee of internal advisors reviews requests. If not approved, requests may be further arbitrated by a Data Sharing Independent Review Panel. Requests that pose a potential conflict of interest or an actual or potential competitive risk may be declined at Amgen's sole discretion and without further arbitration. Upon approval, information necessary to address the research question will be provided under the terms of a data sharing agreement. This may include anonymized individual patient data and/or available supporting documents, containing fragments of analysis code where provided in analysis specifications. Further details are available at https://www.amgen.com/science/clinical-trials/clinical-data-transparency-practices/.

Conception and design: D.S. Hong, R.D. Loberg, B.M. Amore, R. Tang, G. Ngarmchamnanrith, E.L. Kwak

Development of methodology: D.S. Hong, R.D. Loberg, E.L. Kwak

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.S. Hong, P.M. LoRusso, O. Hamid, F. Janku, M. Kittaneh, D.V.T. Catenacci, T. Bekaii-Saab, S. Gadgeel, R.D. Loberg, R. Tang, G. Ngarmchamnanrith, E.L. Kwak

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.S. Hong, P.M. LoRusso, O. Hamid, D.V.T. Catenacci, E. Chan, T. Bekaii-Saab, S. Gadgeel, R.D. Loberg, B.M. Amore, Y.C. Hwang, R. Tang, G. Ngarmchamnanrith, E.L. Kwak

Writing, review, and/or revision of the manuscript: D.S. Hong, P.M. LoRusso, O. Hamid, F. Janku, M. Kittaneh, D.V.T. Catenacci, E. Chan, T. Bekaii-Saab, S. Gadgeel, R.D. Loberg, B.M. Amore, Y.C. Hwang, R. Tang, G. Ngarmchamnanrith, E.L. Kwak

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Kittaneh

Study supervision: D.S. Hong, P.M. LoRusso, O. Hamid, M. Kittaneh, E. Chan

Other (supervised at her site): P.M. LoRusso

Other (consulting): O. Hamid

The authors would like to acknowledge Min Yi and Hui Yang (Biostatistical Sciences, Amgen Inc.) for providing biostatistical support. The authors would also like to acknowledge Meghan E. Johnson, PhD, and James Balwit, MS, CMPP (both Complete Healthcare Communications, LLC, an ICON plc company, North Wales, PA), whose work was funded by Amgen Inc., and Jenilyn J. Virrey, PhD, and Micah Robinson, PhD (both Amgen Inc.), for assistance in writing this manuscript, as well as Isabelle Dussault, PhD, and Darrin Beaupre, MD, PhD (both Amgen Inc.), for their contributions to the preclinical and early development of AMG 337. This study was sponsored by Amgen Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.