Differential Outcomes in Codon 12/13 and Codon 61 NRAS-Mutated Cancers in the Phase II NCI-MATCH Trial of Binimetinib in Patients with NRAS-Mutated Tumors

This article is featured in Highlights of This Issue, p. 2959

RAS mutations result in upregulation of the MAPK pathway and are thought to be key driver mutations in many malignancies (1). The importance of RAS mutations is underscored by their high prevalence in human malignancies. RAS, which has three highly homologous isoforms (KRAS, NRAS, and HRAS), is mutated in approximately 19% of human malignancies (2). NRAS-mutated tumors are less common than KRAS-mutated malignancies and are found in approximately 8% of human cancers (3, 4). Although the NRAS mutation is most frequently found in melanoma, these mutations are seen in many other solid malignancies including colorectal, thyroid, biliary tract, endometrial, and ovarian cancers (5–11).

Although there has recently been some progress in developing G12C KRAS inhibitors, by and large, efforts to directly target RAS have been unsuccessful (4, 12, 13). Clinical trials testing numerous strategies have not demonstrated biological activity (4). Interestingly, preclinical studies have shown NRAS-mutated cell lines to be more sensitive to MEK inhibition than KRAS-mutated cell lines (14–16). For example, one study found that five of six tested NRAS-mutated lung cancer cell lines were sensitive to MEK inhibitors (14).

In agreement with in vitro observations, the most successful effort in targeting RAS mutations has been in NRAS-mutated melanoma. The randomized phase III NEMO trial (17) compared the efficacy of binimetinib (MEK162, ARRY-162), a potent oral inhibitor of MEK1 and MEK2 (18, 19), with that of dacarbazine in chemotherapy-naïve patients with advanced melanoma with a codon 61 NRAS-mutated tumor. The NEMO trial demonstrated that binimetinib-treated patients, compared with those treated with dacarbazine, had an improved median progression-free survival (PFS; 2.8 vs. 1.5 months) and objective response rate (ORR; 15% vs. 9%; ref. 17). However, there was no difference in overall survival (OS) between the two arms (17).

On the basis of these preclinical and clinical data, we hypothesized that the binimetinib MEK inhibitor might be efficacious in other NRAS-mutated malignancies. Here, we report the results of the subprotocol of the NCI-MATCH basket trial that evaluated the antitumor efficacy of single-agent binimetinib in patients with a refractory NRAS-mutated malignancy.

The NCI-MATCH trial (NCT02465060) is a multicenter open-label phase II trial evaluating targeted therapy directed by molecular profiles. Eligible patients have histologically documented solid tumor, lymphoma, or multiple myeloma who require therapy after progression on at least one line of standard systemic therapy. Patients were required to have an Eastern Cooperative Oncology Group (ECOG) performance status ≤1 and measurable disease. Adequate renal, hematologic, and liver function were required.

Patients enrolled in the NCI-MATCH trial master protocol underwent a tumor biopsy and molecular profiling with the following investigational assays: an adapted Oncomine AmpliSeq panel (Thermo Fisher Scientific) and IHC assays for PTEN, MLH2, MSH2, and Rb expression in protocol-designated Clinical Laboratory Improvement Amendments–accredited laboratories (20, 21). Patients whose tumor had a molecular alteration targeted by one of the treatments included in the trial were offered enrollment onto a subprotocol according to the NCI-MATCH treatment assignment algorithm (MATCHbox). The treatment-assignment algorithm was designed to enroll patients in the treatment subprotocol that had the highest level of evidence for their therapeutic agent and alteration.

This study was conducted according to the principles of the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines. The study was approved by the NCI Central Institutional Review Board (CIRB). All patients were provided with and signed CIRB-approved consent forms before enrollment.

Patients eligible for the NRAS subprotocol had a tumor harboring a codon 12, 13, or 61 NRAS mutation. Patients with melanoma were excluded from this subprotocol because binimetinib has already been extensively investigated in this population. Exclusion criteria included prior treatment with an MEK inhibitor, a history of retinal pathology and left ventricular ejection fraction < 50%. Patients needed to have completed chemotherapy, radiation, or surgery ≥ 4 weeks before starting this subprotocol.

Patients assigned to the NRAS subprotocol were treated with open label, orally administered binimetinib 45 mg twice daily continuously until disease progression or the development of unacceptable toxicity. A cycle was defined as 28 days. Decrease in the binimetinib dose below 30 mg twice daily was not allowed. Safety assessments performed included a multigated acquisition/echocardiography at the end of the second cycle and every four cycles thereafter. Retinal examinations were performed after the first cycle and then every two cycles.

Radiological tumor assessments, evaluated according to RECIST, version 1.1, were performed every 8 weeks (two cycles) for the first four cycles and every three cycles thereafter (22). Adverse events assessment was performed according to the Common Toxicity Criteria for Adverse events, version 4.0.

This subprotocol was designed to accrue 35 patients; to ensure the enrollment of 31 eligible patients. Unexpectedly, 18 of the first 25 subjects enrolled in this subprotocol were patients with colorectal cancer. Because of the predominance of patients with colorectal cancer on this subprotocol, enrollment of patients with this diagnosis was halted after 24 had been registered to increase the number of patients with noncolorectal cancer. The subprotocol was amended so that after the first 35 patients were enrolled, noncolorectal cancer patient accrual was allowed to continue for up to 6 more months to enroll a maximum of 35 additional patients or until response data were available on at least 31 patients, whichever came first. A maximum of 10 patients per tumor type was enforced during accrual beyond the first 35 patients.

This subprotocol's primary objective was assessment of the ORR [complete response + partial response (PR)] according to RECIST, version 1.1. If an objective response was detected in ≥ 5/31 patients (16%), the agent was considered promising and worthy of further testing. The subprotocol had 91.8% power to conclude that the agent is promising if the true ORR was 25%. The type I error rate (one sided) was 1.8% under a null response rate of 5%.

Secondary objectives included assessment of PFS, 6-month PFS, and OS. PFS was defined as the time between the start of binimetinib treatment and disease progression or death from any cause, censored at the date of last disease assessment for patients who had not progressed. OS was defined as the time between initiation of binimetinib and death or the patient was censored at the date of last contact. Kaplan–Meier methodology was used to estimate survival distributions (23).

Exploratory analyses to assess differences in ORR and PFS by mutation codon were conducted using Fisher exact test (ORR) or log-rank (PFS) test. Testing was two sided, level 0.05 without adjustment for multiple comparisons due to the exploratory nature of these analyses.

Survival data on patients with NRAS-mutated colorectal cancer was obtained from The Cancer Genome Atlas (TCGA) database on cBioPortal (www.cbioportal.org; refs. 24, 25). MEK inhibitor sensitivity data from the Genomics of Drug Sensitivity in Cancer Database and the Cancer Cell line Encyclopedia database was obtained from the Cancer Dependency Map portal (https://depmap.org/portal/; refs. 26–28). Data from codon 61 and codon 12/13 NRAS-mutated cell lines were analyzed for sensitivity to MEK inhibitors by analyzing the area under the fitted dose–response curve (AUC). Statistical significance was determined using two-tailed Welch t test.

Between June 6, 2016 and July 24, 2017, 4,889 patients were screened for the MATCH trial (Fig. 1). Of those patients, 114 had tumors harboring an NRAS mutation, and 53 of the 114 were subsequently enrolled in the NRAS subprotocol. Of the 53 patients enrolled, 50 patients started binimetinib treatment. Ultimately, three patients were excluded because they were found ineligible for the study, leaving 47 eligible patients who were treated on this subprotocol (Fig. 1). The median age of the 47 eligible patients was 60 years (Table 1). In total, 31 patients (66%) had an ECOG performance status of 1 (Table 1). Most patients had been heavily pretreated, and 53% had received four or more prior lines of treatment (Table 1).

More than half the patients had metastatic colorectal adenocarcinoma (24/47, 51.1%). In agreement with other studies, our data demonstrated that NRAS-mutated colorectal cancers are most frequently mutated in codon 12 (13/24, 54.1%) (Table 2), and they occur mainly in left-sided colon and rectal tumors (21/24, 87.5%; Table 1; ref. 29). There were no significant differences in age, gender, race, performance status, or prior lines of therapy between patients with colorectal cancer harboring codon 12/13 and codon 61 NRAS mutations (Supplementary Table S1).

The most common noncolorectal cancers were cholangiocarcinoma (15%), low-grade papillary serous carcinoma of the ovary (6.4%), and endometrioid endometrial adenocarcinoma (6.4%; Table 1). In contrast to colorectal adenocarcinoma, other cancers harbored predominantly codon 61 mutations (14/23, 60.9%) compared with codon 13 (5/23, 21.7%) and codon 12 (4/23, 17.4%) mutations (P = 0.03; Table 2). Analysis of genomic alterations cooccurring with an NRAS mutation revealed that the TP53 mutation was most common (23/47, 48.9%; Supplementary Fig. S1). APC genomic alterations were also frequent (19/47, 40.4%), but these were observed only in colorectal cancers (19/24, 79.2%).

As of the data cutoff of May 3, 2019, all 47 patients had discontinued study treatment. The most common reason for withdrawal was progressive disease (62%). The median follow-up for patients was 24 months (range, 4–30 months).

The study failed to demonstrate a promising level of activity and did not meet the primary endpoint. The observed ORR was 2.1% (90% CI, 0.1–9.7), and the null hypothesis of 5% ORR (deemed a nonpromising level of activity) could not be rejected (Fig. 2A). For the efficacy population (n = 47), the 6-month PFS was 29.2% (90% CI, 19.4–44.0), the median PFS was 3.5 months (95% CI, 1.8–5.8 months), and the median OS was 10.5 months (95% CI, 5.3–13.2 months; Supplementary Fig. S2A and S2B).

The median PFS for patients with NRAS-mutated colorectal cancer was 1.8 months (90% CI, 1.7–3.7 months) compared with 4.4 months (90% CI, 3.6–8.5 months) for patients who had other cancer types (P = 0.07; Supplementary Fig. S3). The median PFS for patients with cholangiocarcinoma (n = 7) was 3.6 months (90% CI, 1.8–not reached months; Supplementary Fig. S4).

The sole confirmed PR was observed in a patient with a Q61R mutation NRAS-mutated malignant ameloblastoma. This patient had developed the ameloblastoma 19 years prior to enrollment and had metastatic lung lesions measuring up to 6 cm. The patient received binimetinib for 26 months before discontinuing the subprotocol because of grade 2 myalgias. There was also one unconfirmed PR in a patient with colorectal cancer whose tumor harbored an NRAS codon Q61R mutation along with a TP53 and APC mutation. This patient had a 48.2% tumor reduction according to RECIST after cycle 4 (Supplementary Fig. S5). Restaging scans at cycle 7 showed new metastatic lesions, and the patient was removed from this subprotocol. Two additional patients with colorectal cancer harboring NRAS Q61K mutations remained on this subprotocol for 12 and 17 months with stable disease before ultimately developing progressive disease.

In a post hoc analysis, patients with colorectal cancer harboring NRAS codon 61 mutations who were treated with binimetinib (n = 8) had a significantly longer OS (HR, 0.34; 95% CI, 0.12–0.95; P = 0.03) and PFS (HR, 0.23; 95% CI, 0.07–0.74; P = 0.007) than those with NRAS codon 12/13 mutations (n = 16; Fig. 3A and B). Similarly, when all tumor types were examined, binimetinib-treated patients with codon 61–mutated tumors also had significantly longer OS and PFS than those with codon 12/13–mutated tumors. Binimetinib-treated patients with codon 61 NRAS-mutated tumors had a median OS of 13.1 (90% CI, 9.1–not reached) months compared with a median OS of 5.5 (90% CI, 4.7–11.6) months for tumors harboring a codon 12 or 13 NRAS mutation (P = 0.04; Supplementary Fig. S6A). The median PFS for binimetinib-treated patients with codon 12/13 NRAS-mutated tumors was 1.8 months (90% CI, 1.8–3.7 months) compared with 5.8 months (90% CI, 2.5–9.1 months) for binimetinib-treated patients whose tumor harbored an NRAS codon 61 mutation (P = 0.006; Supplementary Fig. S6B). In contrast to patients with colorectal cancer, binimetinib-treated patients with other tumor types with codon 61 NRAS mutations (n = 14) did not have a significantly longer PFS (HR, 0.67; 95% CI, 0.25–1.76; P = 0.4) and OS (HR, 0.84; 95% CI, 0.31–2.27; P = 0.70) than those with codon 12/13 NRAS mutations (n = 9; Supplementary Fig. S7A and S7B).

One explanation for the improved clinical outcomes of binimetinib-treated patients with codon 61 NRAS-mutated colorectal cancer is that codon 61 NRAS-mutated colorectal cancers have a more indolent natural history than codon 12/13 NRAS-mutated colorectal cancers. We examined survival data of patients with NRAS-mutated colorectal cancer (Supplementary Table S2) in TCGA database (15, 24) and found that patients with codon 61 NRAS-mutated colorectal cancer (n = 14) did not have an improvement in OS compared with patients with codon 12/13 NRAS-mutated colorectal cancer (n = 16; HR, 1.06; 95% CI, 0.23–4.94; P = 0.94; Supplementary Fig. S8).

We next examined the hypothesis that codon 61 NRAS-mutated tumors are more sensitive to MEK inhibition by examining MEK inhibitor drug sensitivity testing performed in large cancer cell line collections. Using NRAS-mutated cell lines from the Genomics of Drug Sensitivity in Cancer database (27), we found that codon 61 NRAS-mutated cancer cell lines were significantly more sensitive than codon 12/13 NRAS-mutated cancer cell lines to five different MEK inhibitors (refametinib, selumetinib, trametinib, PD-0325901, and PD-184352; Fig. 4), although more limited data from the Cancer Cell Line Encyclopedia did not demonstrate a statistically significant difference (Supplementary Fig. S9; ref. 26).

Toxicities that are known to be caused by MEK inhibitors, including rash, diarrhea, retinal abnormalities, decreased ejection fraction, elevated creatinine phosphokinase levels, and hypertension, occurred in 43 (86%) of the 50 patients who started treatment. Grade 3 or 4 toxicities that were considered to be possibly, probably, or definitely binimetinib related are shown in Table 3. One death, which was from multiorgan failure, was assessed as possibly related to binimetinib.

Among the 50 patients treated on this subprotocol, 30% discontinued treatment because of adverse events. Binimetinib dose reduction was required in 44% of patients. The first dose reductions occurred at a median of 2.6 weeks into the trial. There was no difference in the rate of early discontinuation of binimetinib, for reasons other than progressive disease or death, between patients with colorectal cancer harboring codon 12/13 and codon 61 NRAS mutations. Among the 24 binimetinib-treated patients with colorectal, the vast majority of patients (21) discontinued therapy due to disease progression or death. Two patients with colorectal cancer discontinued therapy because of adverse events (both tumors harbored a codon 12/13 NRAS mutation) and one patient with colorectal cancer withdrew consent (this patient's tumor harbored a codon 61 NRAS mutation).

Targeted therapy for RAS-mutated malignancies represents an enormous yet critical challenge in clinical oncology. This subprotocol evaluated whether a single-agent MEK inhibitor, binimetinib, might be an effective therapy in patients with nonmelanoma NRAS-mutated malignancies. Similarly to the lack of efficacy of MEK inhibitor monotherapy in KRAS-mutated cancers (19, 30, 31), single-agent binimetinib did not show a promising ORR in patients with NRAS-mutated solid tumors. The modest efficacy of binimetinib monotherapy was disappointing. In the NEMO trial, while binimetinib did not improve OS compared with dacarbazine in codon 61 NRAS-mutated melanoma, it did result in a modest increase in PFS over dacarbazine, and it had an ORR of 15% (17). Possible reasons for the lower ORR in this subprotocol compared with the NEMO trial include lineage or allelic differences between the two study populations. Patients with melanoma who participated in the NEMO trial were treatment naïve, with the exception of immunotherapy, while more than half the patients registered on this subprotocol had received four or more prior lines of therapy. However, taken together, data from the NEMO trial and this subprotocol clearly indicate that MEK inhibitor monotherapy is inadequate to provide clinical benefit to most patients with NRAS-mutated tumors.

Preclinical data suggest that NRAS mutation allelic differences can generate functionally significant phenotypic distinctions (32–34). Mechanistically, codon 61 NRAS mutations activate NRAS differently than codon 12 and 13 NRAS mutations. Codon 61 NRAS mutations block the hydrolysis of GTP to GDP (35), while codon 12 and 13 mutations interfere with the binding of GTPase-activating proteins, which ordinarily would accelerate the hydrolysis of GTP to GDP (36). We explored whether the outcomes of binimetinib-treated patients with codon 61 NRAS-mutated tumors were different than binimetinib-treated patients with codon 12/13 NRAS-mutated tumors. Binimetinib-treated patients with colorectal cancer whose tumor harbored a codon 61 NRAS-mutated tumor had a significantly longer OS and PFS than binimetinib-treated patients with colorectal cancer with a codon 12/13 NRAS-mutated tumor. Similarly, an analysis of all binimetinib-treated patients, regardless of tumor type, also demonstrated that patients with codon 61 NRAS-mutated tumors had a superior OS and PFS compared with patients with a codon 12/13 NRAS-mutated tumor. However, when patients with colorectal cancer were excluded from the analysis, the improvement in OS and PFS was no longer statistically significant. Given the colorectal cancer patient predominance in this trial, it is unclear whether this improvement in OS and PFS was solely attributable to colorectal cancer or that it is more generalizable to other tumor types.

One possible explanation for the superior OS and PFS of binimetinib-treated patients with colorectal cancer with codon 61 NRAS-mutated tumors is that these tumors have a more indolent natural history than codon 12/13 NRAS-mutated cancers. Using TCGA database, where 80% of patients had pathologically staged I to III colorectal cancers, we found no difference in the OS of patients with colorectal cancer with codon 12/13 NRAS mutations and codon 61 NRAS mutations. Cercek and colleagues performed a similar analysis and found that patients with NRAS-mutated metastatic colorectal cancer with exon 3 (codon 60 and 61) NRAS mutations had a shorter OS than patients with metastatic colorectal cancer with exon 2 (codon 12 and 13) NRAS mutations (37). Potential reasons for the differences observed between these two analyses include the different stage distribution of the two datasets and the relatively small sample size.

Because codon 61 NRAS-mutated colorectal cancer does not appear to have a more indolent natural history than codon 12/13 NRAS-mutated colorectal cancer, an alternative explanation is that codon 61 NRAS-mutated colorectal cancers are more susceptible to binimetinib. All the patients with colorectal cancer that either achieved an unconfirmed PR or stable disease for longer than 12 months harbored a codon 61 NRAS mutation. Consistent with the findings observed in our trial, we found that codon 61 NRAS-mutated cell lines were more sensitive to MEK inhibitors compared with codon 12/13 NRAS-mutated cell lines in the Genomics of Drug Sensitivity in Cancer Database. While analysis of MEK inhibitor sensitivity in more limited data from the Cancer Cell Line Encyclopedia cell line collection did not replicate these findings, previous work has noted inconsistent results between these two large pharmacogenomic databases (38). To more rigorously assess the hypothesis that cancers with NRAS codon 61 mutations are more sensitive to MEK inhibition than cancers with NRAS codon 12/13 mutations, future preclinical studies could explore whether the differential sensitivity of MEK inhibitors is observed in experimental systems, such as isogenic cell lines or transgenic mice, that only differ in the NRAS codon that is mutated. Interestingly, a phosphoproteomic study demonstrated that while NRAS codon 61–mutated tumors exhibit hyperactivation of the MAPK pathway, tumors harboring NRAS codon 12 mutations were more reliant on the PI3K/AKT pathway (39). If this finding is confirmed, this differential downstream signaling could explain how NRAS codon 61–mutated cancer are more sensitive to MEK inhibition than NRAS codon 12/13 mutated cancers.

Although MEK inhibitor monotherapy is insufficient to treat NRAS-mutated cancers, the modest activity observed in this subprotocol and in NRAS codon 61–mutated melanoma suggests that MEK inhibitor combinations may potentially be an effective treatment strategy. In an NRAS-mutated murine model of melanoma, single-agent MEK inhibition was cytostatic, whereas combined MEK and CDK4/6 inhibition resulted in significant tumor regression (40). In this model, MEK inhibitor monotherapy induced apoptosis but did not block cell-cycle arrest, while combined MEK and CDK4/6 inhibition was more potent because it stimulated both apoptosis and cell-cycle arrest. Consequently, the strategy of combined MEK and CDK4/6 inhibition is undergoing clinical testing. In addition to MEK inhibitor–based strategies, encouraging antitumor activity has recently been reported in trials using MRTX849 and AMG510, which are novel covalent inhibitors of KRAS G12C (13, 41). These data raise the possibility that similar compounds might have activity in NRAS G12C- or G13C-mutated tumors. Data from this subprotocol indicate that NRAS G12C mutations were present in one of 17 (5.8%) codon 12–mutated tumors and two of eight (25%) codon 13–mutated tumors.

A patient with a codon 61 NRAS-mutated malignant ameloblastoma had a durable 26-month PR to binimetinib. Malignant ameloblastoma is a rare odontogenic tumor that appears to be driven by the MAPK pathway. More than 80% of malignant ameloblastomas harbor either an RAS or a BRAF mutation; most (46%–62%) are V600E BRAF mutations, and the minority (6%) are NRAS mutations (42, 43). V600E BRAF-mutated malignant ameloblastomas have been reported to have durable responses to BRAF inhibition (44, 45). Although this patient had indolent tumor for 19 years prior to accrual to the trial, the patient's 26-month response to binimetinib further supports the sensitivity of this tumor type to MAPK inhibition (44, 45).

High percentages of patients in this subprotocol discontinued treatment (30%) and required dose reductions (44%) because of binimetinib-associated toxicity. This finding is similar to the experience reported for patients with melanoma treated with binimetinib in the NEMO trial, in which 25% of patients discontinued binimetinib because of toxicity (17).

In conclusion, MEK inhibitor monotherapy is an insufficient treatment for NRAS-mutated solid tumors. A post hoc analyses suggested that binimetinib-treated patients with colorectal cancer with an NRAS codon 61–mutated tumor had a significantly longer OS and PFS than binimetinib-treated patients with colorectal cancer with an NRAS codon 12– or 13–mutated tumor. The design of future clinical trials targeting NRAS should carefully consider differentiating between codon 12/13 and codon 61 NRAS-mutated tumors as these two populations appear to have important biological differences and therapeutic sensitivities.

J.M. Cleary reports grants from Merck and Tesaro; nonfinancial support from AstraZeneca and Esperas Pharma; and personal fees from BMS outside the submitted work. V. Wang reports grants from the NIH during the conduct of the study. R.S. Heist reports other from Novartis, Daichii Sankyo, EMD Serono, Tarveda, and Apollomics; grants from Novartis, Corvus, Incyte, Genentech Roche, Mirati, Turning Point, BMS, Daichii Sankyo, AbbVie, Agios, Exelixis, and Lilly outside the submitted work. E.S. Kopetz reports personal fees from Roche, Genentech, Merck, Karyopharm Therapeutics, Amal Therapeutics, Navire Pharma, Symphogen, Holy Stone, Biocartis, Amgen, Novartis, Lilly, Boehringer Ingelheim, Boston Biomedical, AstraZeneca/MedImmune, Bayer Health, Pierre Fabre, EMD Serono, Redx Pharma, Jacobio, Natera, Repare Therapeutics, Daiichi Sankyo, Lutris, Pfizer, Ipsen, and HalioDx outside the submitted work. R.J. Gray reports grants from the NCI during the conduct of the study. F. Meric-Bernstam reports personal fees from Aduro BioTech Inc., Alkermes, AstraZeneca, DebioPharm, eFFECTOR Therapeutics, F. Hoffmann-La Roche, Genentech, IBM Watson, Jackson Laboratory, Kolon Life Sciences, OrgiMed, PACT Pharma, Parexel International, Pfizer Inc., Samsung Bioepis, Seattle Genetics, Tyra Biosciences, Xencor, Zymeworks, Immunomedics, Inflection Biosciences, Mersana Therapeutics, Puma Biotechnology, Seattle Genetics, Silverback Therapeutics, Spectrum Pharamaceuticals, Zentalis, Chugai Biopharmaceuitcals, Mayo Clinic, and Rutgers Cancer Institute of New Jersey; grants from Aileron Therapeutics, AstraZeneca, Bayer Healthcare, Calithera, Curis, CytomX, Daiichi Sankyo, DebioPharm, eFFECTOR Therapeutics, Genentech, Guardant Health, Klus Pharma, Millennium Pharmaceuticals Inc, Novartis, Puma Biotechnolgy, and Taiho Pharmaceutical; and nonfinancial support from Beth Israel Deaconess Medical Center outside the submitted work. A.J. Aguirre reports personal fees from Merck, Arrakis Therapeutics, and Oncorus; grants from Deerfield Management, Novo Ventures, Syros, and Mirati outside the submitted work. P.J. O'Dwyer reports personal fees from Array outside the submitted work. C.L. Arteaga reports grants from Lilly, Pfizer, and Takeda and personal fees from Novartis, Lilly, Immunomedics, Merck, Daiichi Sakyo, TAIHO Oncology, AstraZeneca, OrigiMed, Arvinas, and Clovis outside the submitted work; holds minor stock options in Provista and used to hold minor stock options in Y-TRAP; serves in the scientific advisory board of the Susan G. Komen Foundation; and reports that none of these relationships has any relationship with or relevance to his role as coauthor of this article. K.T. Flaherty reports personal fees from Loxo Oncology, Clovis Oncology, Strata Oncology, Checkmate Pharmaceuticals, Kinnate Pharmaceuticals, X4 Pharmaceuticals, PIC Therapeutics, Amgen, Asana Biosciences, Adaptimmune, Aeglea, Shattuck Labs, Tolero Pharmaceuticals, Apricity, Oncoceutics, Fog Pharma, Neon Therapeutics, Tvardi, xCures, Monopteros, Vibliome, Lilly, Genentech, Bristol Myers Squibb, Merck, Takeda, Verastem, Boston Biomedical, Pierre Fabre, and Debiopharm; and personal fees and nonfinancial support from Sanofi and Novartis during the conduct of the study. No disclosures were reported by the other authors.

J.M. Cleary: Conceptualization, formal analysis, investigation, writing–original draft, writing–review and editing. V. Wang: Formal analysis, investigation, writing–review and editing. R.S. Heist: Conceptualization, formal analysis, writing–original draft, writing–review and editing. E.S. Kopetz: Conceptualization, formal analysis, writing–original draft, writing–review and editing. E.P. Mitchell: Writing–review and editing. J.A. Zweibel: Formal analysis, writing–review and editing. K.S. Kapner: Formal analysis, investigation, writing–review and editing. H.X. Chen: Formal analysis, investigation, writing–review and editing. S. Li: Formal analysis, investigation, writing–review and editing. R.J. Gray: Formal analysis, investigation, writing–review and editing. L.M. McShane: Formal analysis, investigation, writing–original draft, writing–review and editing. L.V. Rubinstein: Formal analysis, investigation, writing–original draft, writing–review and editing. D.R. Patton: Investigation, writing–review and editing. F. Meric-Bernstam: Investigation, writing–original draft, writing–review and editing. M.S. Dillmon: Writing–review and editing. P.M. Williams: Investigation, writing–original draft, writing– review and editing. S.R. Hamilton: Formal analysis, investigation, writing–original draft, writing–review and editing. B.A. Conley: Conceptualization, formal analysis, writing–original draft, writing–review and editing. A.J. Aguirre: Formal analysis, investigation, writing–original draft, writing–review and editing. P.J. O'Dwyer: Conceptualization, resources, formal analysis, investigation, writing–original draft, writing–review and editing. L.N. Harris: Conceptualization, formal analysis, investigation, writing–original draft, project administration, writing–review and editing. C.L. Arteaga: Conceptualization, formal analysis, validation, writing–original draft, project administration, writing–review and editing. A.P. Chen: Conceptualization, supervision, writing–original draft, writing–review and editing. K.T. Flaherty: Conceptualization, resources, investigation, writing–original draft, writing–review and editing.

This study was coordinated by the ECOG-ACRIN Cancer Research Group (P.J. O'Dwyer and Mitchell D. Schnall) and was supported by the NCI of the NIH under the following award numbers: CA180820, CA180794, CA180858, CA180867, CA180870, and CA189997. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. government. Array BioPharma provided the study drug for this study. The work of J.M. Cleary is supported by DFCI Hale Family Center for Pancreatic Cancer, Team Evan Schumacher, Haya Linde Memorial Fund, and the Pancreatic Cancer Collective New Therapies Challenge Grant, an initiative of the Lustgarten Foundation and Stand Up To Cancer, grant number SU2C-AACR-PCC-02-18. Stand Up To Cancer is a division of the Entertainment Industry Foundation. Research grants are administered by the American Association for Cancer Research, the Scientific Partner of SU2C. S.R. Hamilton is the recipient of the Frederick F. Becker Distinguished University Chair in Cancer Research at the University of Texas.

The Editor-in-Chief of Clinical Cancer Research is an author on this article. In keeping with AACR editorial policy, a senior member of the Clinical Cancer Research editorial team managed the consideration process for this submission and independently rendered the final decision concerning acceptability.

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.