A Phase II Study of the c-Met Inhibitor Tivantinib in Combination with FOLFOX for the Treatment of Patients with Previously Untreated Metastatic Adenocarcinoma of the Distal Esophagus, Gastroesophageal Junction, or Stomach

The 5-year survival rates for metastatic gastroe- sophageal (GE), GE junction, and gastric cancers are less than5% (1). Approximately 50% of patients with gastric cancer present with advanced disease (2). With no clear consensus on the best regimen, multiple reg- imens are used in the treatment of metastatic GE can- cers. These include cisplatin and 5-FU (CF); epirubicin, cisplatin, and 5-FU (ECF); epirubicin, oxaliplatin, and capecitabine (EOX); cisplatin and irinotecan; and doc- etaxel, cisplatin, and 5-FU (DCF). Trials of infusional 5-FU and oxaliplatin for the treatment of patients with advanced GE cancer have shown response rates (RRs) of approximately 45% (3, 4). A recent randomized phase II study showed that FOLFOX-based chemother- apy (5-FU/leucovorin/oxaliplatin) demonstrated equivalent outcomes to the more traditional ECF regimen, with RRs of 51% versus 58%, respectively (5). The c-Met receptor is a tyrosine kinase that binds to hepatocyte growth factor. Activation of Met results in downstream activation of multiple signaling pathways, including PI3 kinase/AKT, Ras/Raf/MAPK, and PLC (6). C-Met is most amplified in not only gastric can- cers but also esophageal cancers (7, 8). C-Met is also involved in activation of other tyrosine kinase recep- tors, including epidermal growth factor receptor and others (9), making c-Met a rational target in the treatment of GE cancers.Tivantinib is an orally bioavailable small molecule inhibitor of c-Met with potential antineoplastic activ- ity. Tivantinib binds to the c-Met protein and disrupts c-Met signal transduction pathways, which may induce cell death in tumor cells overexpressing c-Met protein or expressing constitutively activated c-Met protein.Clinical studies of tivantinib have shown some effi- cacy in patients with high c-Met expression in colorec- tal cancer, liver cancer, and nonsmall cell lung cancer (10–13) either as a monotherapy or in combination treatment. Here we present the results of a phase I trial of tivantinib plus FOLFOX for the treatment of patients with advanced solid tumors followed by a phase II por- tion for patients with first-line metastatic GE cancer.

This study was an open-label, multicenter phase I/II study of the combination of tivantinib and FOL- FOX in patients with advanced solid tumors (phase I) or first-line metastatic adenocarcinoma of the esophagus, GE junction or stomach (NCT01611857). The primary objective of the phase I portion of the study was to determine the maximum tolerated dose (MTD)/recommended phase II dose of tivantinib plus FOLFOX in patients with advanced solid tumors. The primary objective of the phase II portion of the study was to determine the RR of the combination of regimen. Secondary objectives were to describe the dose-limiting toxicities and adverse event profile of this drug combination in the phase I portion; to evaluate other efficacy measurements, including time to progression (TTP), progression-free survival (PFS), and overall survival (OS); and to further describe toxicities associated with this regimen in patients with metastatic GE cancers.
The study was conducted according to the ethi- cal principles of the Declaration of Helsinki and in accordance with the International Conference on Harmonisation Guideline for Good Clinical Practice. The protocol was approved by the Institutional Review Boards of participating sites and patients were enrolled following written informed consent.

In the phase I portion, adult patients were required to have histologically confirmed metastatic or unre- sectable solid tumor malignancy for which standard therapy would include FOLFOX or for which standard curative or palliative measures do not exist or are no longer effective. Phase II patients had histolog- ically confirmed metastatic adenocarcinoma of the esophagus, GE junction, or stomach for which they have received no prior chemotherapy for metastatic disease. Prior treatment for localized disease was allowed. The following were additional inclusion criteria: Karnofsky performance status ?70%; mea- surable disease per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (6); adequate bone marrow function (absolute neutrophil count?1500/µL; hemoglobin (Hgb) ?9 g/dL; platelets?100,000/µL), hepatic function (aspartate transami- nase, alanine transaminase, and alkaline phosphatase≤2.5 × the upper limit of normal (ULN) or ≤3 × ULN in patients with liver metastases; total biliru- bin ≤1.5 × the institutional ULN or ≤3 × ULN in patients with documented Gilbert Syndrome, and renal function (serum creatinine ≤1.5 × ULN or 24-h creatinine clearance ?40 mL/min); written informed consent.Patients were considered ineligible for any of the following: untreated central nervous system metas- tases; history of current cardiac abnormalities includ- ing, but not limited to, a mean QTc ? 450 msec on baseline ECG, history of clinically manifested ischemic heart disease (i.e., myocardial infarction and/or unsta- ble angina) or clinically significant atherosclerosis vas- cular disease ≤6 months prior to study start.

Before initiation of treatment, patients underwent a complete medical history, physical examination, assessment of performance status, complete blood count, including differential and platelets, carcinoem- bryonic antigen (CEA), comprehensive metabolic pro- file, prothrombin time/ partial thromboplastin time/ international normalization ratio, and a serum or urine pregnancy test for women of child-bearing potential. Archived tumor tissue was collected from patients at baseline and analyzed for c-Met expression by immunohistochemistry (IHC). Tissue was considered positive for c-Met expression if >50% of cells showed c-Met expression by IHC (14). The levels of c-Met were stratified by intensity of staining (weak: 1+, moderate: 2+, strong: 3+). If c-Met-positive tissue had ? moder- ate (2+) IHC staining, it was considered to have high c-Met expression. All patients had computed tomogra- phy scans of the chest, abdomen, and pelvis. In the phase I portion of the study, a standard 3 + 3 design was used to determine the MTD of the combination of tivantinib (PO BID) plus FOLFOX (5-FU 400 mg/m2, 5-FU continuous IV 2400 mg/m2 over 46 h, leucovorin 400 mg/m2 IV, and oxaliplatin 85 mg/m2), all Day 1. Three different dose levels of tivantinib (120, 240, and 360 mg BID) were evaluated. All patients (phase I and phase II) continued treat- ment until disease progression or intolerable toxicity. If patients in the treatment escalation portion of the study had to discontinue one or more of the drugs in the treatment regimen (tivantinib, 5-FU, leucovorin, and oxaliplatin) due to toxicity, they could continue on the remaining drugs.

Patients who were evaluable for dose escalation received at least 85% of required doses of study drug in cycle 1. Dose limiting toxicities (DLTs) for the phase I portion of the study were defined as grade 4 neu- tropenia for ?7 days or febrile neutropenia, grade 4 thrombocytopenia or grade 3 thrombocytopenia with bleeding, grade 3 or 4 nausea, vomiting, or diarrhea lasting ?4 days despite adequate supportive measures, grade 3 or 4 nonhematologic toxicity (excluding alope- cia or allergic reaction) thought to be possibly related to treatment, any treatment-related death or hospitaliza- tion, any toxicity that prevented a patient from receiv- ing at least 85% of the required doses of study drug due to toxicity, and the inability to start cycle 2 dosing within 3 days of the planned dose due to toxicity. The MTD was defined as the highest dose at which ≤1 of6 patients experienced DLT during the first cycle of treat- ment. The maximally administered dose was defined as the highest dose planned in the phase I dose escalation portion of the study if the MTD was not reached.
All patients who received at least one dose of protocol treatment were followed for safety. All adverse events and serious adverse events (SAEs), regardless of rela- tionship to study medication, were collected from the day of the first dose to 30 days after the last dose of study medication. Toxicities were graded according to the National Cancer Institute’s (NCI) Common.

Response evaluations, based on RECIST 1.1 (16), were performed every treatment 8 weeks until progres- sion, the start of an alternative treatment regimen, or death. Final response presented for these patients is the best response experienced during treatment.This study was designed to establish the MTD of tivantinib in combination with FOLFOX in patients with advanced solid tumors and to evaluate the RR, defined as the proportion of patients achieving objec- tive response, of the drug combination in patients with previously untreated metastatic adenocarcinoma of the distal esophagus, GE junction, or stomach. For the pur- pose of efficacy analysis, the intent-to treat popula- tion was defined as all patients who met the eligibility requirements and gave written informed consent. PFS was defined as the interval of time (in months) between the date of the first administration of study treatment and the earlier of the date of disease progression or the date of death due to any cause. TTP was defined as the interval of time (in months) between the date of first administration of study treatment until the earlier of the date of tumor progression or the date of last ade- quate tumor assessment. OS was defined as the inter- val from first study treatment until the earlier of the date of death or date last known alive. Analysis for PFS, TTP, and OS were summarized using the Kaplan-Meier methods (17). The safety analysis population consisted of all patients who received at least one dose of study treatment.Sample size was based on the historic overall response rate (ORR) of 45% for first-line FOLFOX. A sample size of 29 (plus 10% to account for potential nonevaluable patients) achieved 80% power to detect an increase in the ORR to 65% based on a one-sided test of proportion at an alpha level of 0.10.

Between August 2008 and August 2014, 49 patients (15 patients on phase I and 34 patients on phase II) enrolled on the study. All 15 phase I patients and 32 of the 34 phase II patients received at least one dose of study drug. Summaries of baseline patient character- istics are found in Tables 1 (phase I) and 2 (phase II). I patients had colorectal cancer, four had esophageal cancer, and the remaining four patients had pancreatic, GE junction, cholangiocarcinoma, or unknown pri- mary tumors. Eleven phase I patients had at least two regimens of prior therapy and five had prior radiation. Ten phase I patients had prior treatment with FOLFOX: four patients in the adjuvant setting, five patients in the metastatic setting, and one patient in both the adjuvant and metastatic settings.Phase II patients had a median age 65 years (range, 34–88) and 76% were male. Forty-seven percent of these patients underwent prior surgery, 32% received prior radiation, and 15% received prior systemic treat- ment in the adjuvant setting. Forty-one percent of phase II patients had esophageal cancer, while 32% had gastric cancer, and 27% had tumors of the GE junction. Forty-four percent of phase II patients had high c-Met expressing tumors.The dose escalation is summarized in Table 3. Grade 3 neutropenia occurred in one patient at dose level 3 (tivantinib 360 mg BID plus FOLFOX) and was the only DLT observed. With only one DLT, the MTD was not exceeded and tivantinib 360 mg BID plus FOLFOX (5-FU 400 mg/m2, 5-FU continuous IV 2400 mg/m2 over 46 h, leucovorin 400 mg/m2 IV, and oxaliplatin 85 mg/m2) was the maximally administered dose. This dose was used for the phase II dose expansion.Table 4 summarizes treatment-related adverse events by dose level that occurred in the phase I patients of patients. The most common treatment- related adverse events were nausea (40%) and throm- bocytopenia (33%).Thirty-four patients were accrued to the phase II dose expansion. Patients received a median of eight cycles of treatment (range 1–45). Two patients did not receive treatment due to patient request and anemia prior to cycle 1 day 1. Thirty of the 32 treated patients have discontinued treatment. Seventeen patients (50%) discontinued due to disease progression.

Four (12%) discontinued due to toxicity (one patient each with neuropathy, mucositis, neutropenia, and gastric ero- sion with associated anemia). Four patients died on treatment of unrelated pneumonia, aspirational pneu- monia, myocardial infarction, and cerebrovascular accident. An additional four patients discontinued due to patient request or investigator discretion and one patient discontinued treatment due to unrelated abdominal pain and an optic nerve disorder. At the time of data cutoff, two patients remained on treat- ment. One patient with low c-Met expression was on cycle 36, receiving tivantinib plus 5-FU and leucovorin while another patient with high c-Met expression was on cycle 45, receiving tivantinib and 5-FU.Phase II safetyAll phase II patients that received at least dose of the treatment regimen were included in the safety popu- lation. A summary of all treatment-related toxicities is found in Table 5. The most common treatment- related toxicities included (all grades) neutropenia (63%), fatigue (56%), diarrhea (50%), nausea (50%), and peripheral neuropathy (50%). There were seven treatment-related SAEs in seven patients (grade 4 diarrhea; grade 3 sepsis, anemia, neutropenia, febrile neutropenia, and vomiting; grade 2 noncardiac chest pain). Four patients (12%) discontinued due to treatment-related adverse events (gastric mucosal erosions and associated anemia [one patient], neu- ropathy (one patient), mucositis (one patient), and neutropenia (one patient)). Four patients died on study, though none of these deaths were considered related to treatment.Response to treatment is found in Tables 6 and 7. One patient (3%) achieved a complete response and partial response was seen in 13 patients (38%), for an ORR of 41% among all patients (54% among evaluable patients). Stable disease was reached in nine patients (26%). Three patients (9%) had progressive disease as a best response. Six patients were not evaluable for response due to clinical progression (two patients), patient request (one patient), treatment-related gastric mucosal erosion (one patient), intercurrent illness of abdominal pain/optic nerve disorder (one patient), unrelated death due to aspiration pneumonia (one patient).

The ORR of patients with high c-Met tumor expression, low c-Met expression, and unknown expression was 33% (5 out of 15 patients), 45% (5out of 11 patients), and 50% (4 out of 11 patients), respectively.Twenty-one of the 34 intent-to-treat patients had CEA levels measured at baseline and on treatment. Eight of the 10 patients with elevated CEA at baseline (defined as ?5 ng/mL) showed a decrease of CEA levels during treatment.With a median follow up of 15 months (range 3–21 months), the median PFS was 6.1 months (95% CI 3.6–7.8) months, the median TTP was 7.0 months (95% CI 5.7–11.5 months), and the median OS wasTable . Overall response summary.aReason for unevaluable: patient request (); gastric mucosal erosion (); clinical progression (); intercurrent illness of abdominal pain/optic nerve disorder (); death on study due to aspiration pneumonia (). 9.6 months (95% CI 7.2 months – not reached). The median PFS of patients with high c-Met expressing tumors (Figure 1A) was 5.8 months (95% CI 1.9–7.0 months), compared to a median PFS of 7.7 months (95% CI 1.7–12.3 months) for patients with low c-Met tumor expression. A similar trend was seen in median TTPs (c-Met high: 6 months, c-Met low: 7.7 months) (Figure 1B) and OS (c-Met high: 8.8 months, c-Met low: 12.3 months) (Figure 1C) when stratified by c-Met expression levels.DiscussionIn this study, we evaluated safety and efficacy of the c- Met inhibitor tivantinib in combination with FOLFOX as first-line treatment for patients with metastatic ade- nocarcinoma of the distal esophagus, GE junction, or stomach.The combination of tivantinib plus FOLFOX was safe in this treatment population. With only one DLT seen during dose escalation, the maximally adminis- trated dose of tivantinib 360 mg BID was safe when combined with FOLFOX, and this dose level was Texpression.

The lack of correlation between c-Met expression could be due to residual, low-level c-Met activity in high c-Met expressing tumors treated with tivantinib.C-Met is an attractive tumor target for patients with gastric cancer. Recent TCGA analysis of gastric cancer subtypes showed that 50% of gastric cancers have chromosomal instability and that these tumors were more likely to have Met amplification (25). How- ever, clinical trials using monoclonal antibodies and RTK inhibitors against c-Met have been disappointing despite the expectation of this being a rational target.The HGF-targeting monoclonal antibody (mAb), rilotumumab, showed promise when combined with epirubicin, cisplatin, and capecitabine for the treat- ment of gastric/GEJ tumors. Median PFS was longer with c-Met-positive tumors in this phase II study (5.7 months for rilotumumab-treated patients com- pared to 4.2 months for placebo-treated patients) (26). However, the randomized phase III study of this drug combination was discontinued prematurely after show- ing no response benefit and because of deaths on the rilotumumab arm (27).The addition of the c-Met mAb, onartuzumab, did not improve PFS in an unselected patients (median PFS of 6.77 months for patients on onatuzumab and6.97 months for patients on placebo), or in a c-Met- positive population (median PFS of 5.95 months for patients on onatuzumab and 6.8 months for patients on placebo) when added to a mFOLFOX regimen (28).

The c-Met kinase inhibitor, AMG337, appeared promising in the initial results of a study of the drug in patients with gastric, esophageal, and GE tumors with c-Met amplification (24). The study was later ter- minated though after an interim review of safety and activity (29).
One possible reason for the disappointing results seen with drugs targeting Met might be due to high variability of c-Met expression. TCGA analysis of Met expression found amplification rates close to 8% (25). Immunohistochemistry analysis shows anywhere from a 22% to 90% increase in expression (27, 28). Some c- Met-positive patients may only have a low level of c- Met amplification and thus may not fully benefit from inhibition of this target.Development of next generation c-Met inhibitors may include novel agents like ABBV-399, an antagonist anti-c-Met antibody (ABT-700) linked to monomethyl auristatin E. It utilizes a different approach to targeting c-Met in that instead of pathway inhibition, the drug delivers cytotoxin directly to tumor via c-Met. Strick- ler et al. (30) presented a phase I study of ABBV 399 that demonstrated evidence of early activity in patients with nonsmall cell lung carcinoma.Future use of c-Met-targeted inhibitors and mAbs may be more effective if conducted only in patients with a higher threshold of c-Met expression than previously used or novel agents targeting c-Met linked to cytotoxic Tivantinib agents.