YOSEMITE: A 3 arm double-blind randomized phase 2 study of gemcitabine, paclitaxel protein-bound particles for injectable suspension, and placebo (GAP) versus gemcitabine, paclitaxel protein-bound particles for injectable suspension and either 1 or 2 truncated courses of demcizumab (GAD).

Abstract

Background: Delta-like ligand 4 (DLL4) is a ligand that activates the Notch pathway which is important for cancer stem cell (CSC) survival. Demcizumab is a humanized, anti-DLL4 antibody that has been shown using an in vivo tumorigenicity limiting dilution assay to inhibit tumor growth and decrease CSC frequency in minimally passaged human xenograft models. In addition, inhibition of DLL4 has also been shown in preclinical studies to cause dysfunctional sprouting of new vessels resulting in an antiangiogenic effect. Encouraging data from a Phase 1b study of paclitaxel protein-bound particles for injectable suspension (Abraxane), gemcitabine and demcizumab in patients with 1st line metastatic pancreatic cancer led to this double blind randomized 3 arm placebo-controlled Phase 2 study. Method(s): Patients with metastatic pancreatic cancer were randomized (1:1:1) to 1stline therapy with either Arm 1 - GAP, Arm 2 - GAD with a single 70 day truncated course of demcizumab or Arm 3 - GAD with two 70 day truncated courses of demcizumab (second course starting on Day 168). GA were given at usual dose & schedule, P/D was given IV on days 1 and 15 in cycles 1-3 & 7-9. The primary endpoint was progression-free survival and secondary endpoints included response, survival, safety, immunogenicity, pharmacokinetics, and biomarkers of Notch signaling and CSCs in blood, hair follicles and tumor cells. The primary study analyses compared GAP to the two pooled GAD arms. Result(s): 207 patients were randomized and 204 were treated. The median age was 63, the male/female ratio was 116/88, the ECOG 0 vs 1 distribution was 98/106, the median # metastatic sites was 2 and 74% had hepatic metastases. The response/clinical benefit rates were 41.2%/70.6% and 33.1%/74.3% in the GAP and pooled GAD arms, respectively. The median progression-free survival (PFS) (mPFS) was 5.5 months in the GAP and pooled GAD arms. The interimmedian overall survival (OS) for the GAP and pooled GAD arms were not reached and 13.2 months (HR=1.02), respectively. Geographic differences in OS were observed. GAD was generally well tolerated with nausea, diarrhea, anemia, peripheral edema and fatigue being the most common reported toxicities. The incidence of the Grade 3 or greater toxicities of special interest with demcizumab therapy were hypertension (7.4% vs 16.2%), pulmonary hypertension (0% vs 0.7%), heart failure (0% vs. 3.7%), and bleeding (1.5% vs. 8.1%) in the GAP and pooled GAD arms, respectively. No cases of Grade 3 heart failure or pulmonary hypertension occurred during the 2nd 70 day course of demcizumab. Conclusion(s): The addition of either 1 or 2 truncated courses of demcizumab to 1st-line gemcitabine and Abraxane did not improve the efficacy compared to GAP in patients with 1st line metastatic pancreatic cancer. GAD therapy was generally well tolerated.