Abstract Purpose: Tivozanib (T), is a potent oral selective inhibitor of VEGF receptors 1, 2, and 3. Bevacizumab (B) is an antibody targeting VEGF A mediated signaling. The randomized, open label phase 2 BATON-CRC (Biomarker Assessment of Tivozanib in ONcology), trial compared the progression free survival (PFS) and safety of T + mFOLFOX6 vs. B + mFOLFOX6 in mCRC patients. Data from the pre-specified interim analysis demonstrated futility for superiority; however the addition of T to mFOLFOX6 resulted in comparable PFS (9.4 vs 10.7 months, p=0.706), overall response rate (45.2% vs 43.2%, p=0.718), and safety to B + mFOLFOX6 and resulted in discontinuation of the study. At the time of the interim analysis (September, 2013) 62 patients out of 163 progressed to generate a PFS event and none of the pre-specified biomarkers demonstrated statistically significant association with PFS (ESMO 2014). At the close of the study (February, 2014), there were 91 PFS events out of 164 patients with baseline serum samples. The final analysis included the updated events which led to the observation that serum NRP1 may be a predictive biomarker for T. Serum NRP1, a pre-specified biomarker in BATON CRC, is associated with improved PFS and overall survival (OS) in multiple T studies in renal cell carcinoma (RCC). This strong association and clinical benefit in RCC prompted a post-hoc analysis of the final BATON-CRC PFS data to examine the differential association between base line serum NRP1 levels and outcomes in the two arms. Methods: Similar to the RCC trials, baseline serum from patients in BATON-CRC had NRP1 quantified using the Myriad RBM ELISA assay. The median NRP1 level from this dataset was used to define the cut off for NRP1 high and NRP1 low. A Cox proportional hazard model was used to assess the association between serum NRP1 levels and PFS. Results: In BATON-CRC, 265 patients were enrolled and randomized 2:1 to T + mFOLFOX6 (n = 177) or B + mFOLFOX6 (n = 88). A total of 164 baseline serum samples (T, n = 109 and B, n = 55) were available for NRP1 analysis. In both the T and B arms, patients with NRP1 low showed an improved PFS vs patients with NRP1 high; for T HR (95% CI): 0.224 [0.120, 0.42], p=3.56e-7 and for B HR [95% CI]: 0.487 [0.237, 1.00], p=0.046) supporting the value of NRP1 as a potential prognostic marker for angiogenesis inhibitors. T treated NRP1 low patients demonstrated a statistically significantly improvement in PFS vs B, PFS =17.9 vs 11.2 mos (multivariate analysis stratified for cancer origin, metastatic sites, LDH levels HR [95% CI]: 0.325 [0.147, 0.717], p=0.005). In NRP1 high patients, there was no difference in PFS between the two arms (HR (95% CI): 0.889 [0.639, 1.980], p=0.683). The interaction p-value is 0.027. Conclusions: The data suggest that NRP1 might be a predictive biomarker of T PFS benefit over B in combination with mFOLFOX 6 in advanced CRC. This finding needs to be further validated in another prospective trial. The differential activity observed with T vs B, in NRP1 low patients is potentially due to the broader VEGF pathway inhibitory activity of T. Citation Format: Al Benson, III, Andrew K. Krivoshik, Chip Van Sant, Jeno Gyuris, Bin Feng. Neuropilin 1 (NRP1) as a potential biomarker for tivozanib + mFOLFOX6 versus bevacizumab + mFOLFOX6 in metastatic colorectal cancer (mCRC): Post-hoc biomarker analysis of BATON-CRC phase 2 trial. [abstract]. In: Proceedings of the AACR Special Conference: Tumor Angiogenesis and Vascular Normalization: Bench to Bedside to Biomarkers; Mar 5-8, 2015; Orlando, FL. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl):Abstract nr A24.
Read full abstract