A randomized, phase II, multicenter, double-blind, placebo-controlled study evaluating onartuzumab (MetMAb) in combination with mFOLFOX6 plus bevacizumab in patients with metastatic colorectal cancer

Background: Dysregulation of the HGF/Met (Met) pathway has been linked with poor prognosis in colorectal cancer. Crosstalk between the Met and vascular endothelial growth factor (VEGF) pathways may be important during tumorigenesis. Aberrant activation of the HGF/Met pathway may promote angiogenesis via tumor cell secretion of angiogenic factors or directly activating endothelial cells. Onartuzumab (MetMAb) is a monovalent, monoclonal antibody that specifically binds to the Met receptor. The combination of onartuzumab and VEGF inhibition in preclinical models resulted in enhanced antitumor activity over either treatment alone. Preclinical efficacy data support the combination of onartuzumab with platinum agents. In phase I studies, onartuzumab has been generally well tolerated alone and in combination with bevacizumab. Adverse events most commonly associated with onartuzumab are peripheral edema and fatigue. Methods: This is a randomized, two-arm, phase II study in patients with previously untreated metastatic colorectal cancer. Patients (n=188) will be randomized (1:1) to either mFOLFOX6/bevacizumab/placebo or mFOLFOX6/bevacizumab/onartuzumab. Oxaliplatin will be discontinued after 8 cycles with remaining drugs continued until progression. The primary endpoint of this study is PFS in all patients. PFS by Met IHC diagnostic status (Met positive vs Met negative) will also be analyzed. Secondary endpoints include OS, ORR, safety, and biomarker analyses. Primary and secondary analyses will include all randomized patients and will be conducted according to assigned treatment arm. Kaplan-Meier methodology will be used to estimate median PFS for each treatment arm. An estimate of HR with 95% CI will be determined using a Cox regression model. Safety will be assessed in all patients receiving at least one dose of any treatment.