Lenvatinib treatment of advanced RAI-refractory differentiated thyroid cancer (DTC): Cytokine and angiogenic factor (CAF) profiling in combination with tumor genetic analysis to identify markers associated with response.

Abstract

5518 Background: Lenvatinib is an oral tyrosine kinase inhibitor targeting VEGFR1-3, FGFR1-4, RET, KIT and PDGFRβ. In a phase II study of lenvatinib, 58 patients (pts) with DTC were enrolled and a response rate of 50% was observed (Sherman, ASCO 2011). Methods: Pts received lenvatinib at a starting dose of 24 mg oral once daily in 28-day cycles. Serum was collected at baseline (BL), on day 8 and on day 36 and concentrations of 47 CAFs were measured using multiplex bead arrays and enzyme-linked immunosorbent assay (ELISA). 33 genes (443 mutations) were examined in archival tumor tissues (n=25). Association of baseline CAF, changes in CAF levels upon treatment and gene mutation status with treatment outcomes was investigated. Results: Combination of low baseline VEGF and ANG-2 (p=0.02 and HR=0.386) correlated with longer PFS. Both baseline and changes in CAF levels demonstrated an association with gene mutation status. High baseline levels of VEGF were observed in pts with wild type RAS and BRAF (p=0.035) whereas high baseline sTIE-2 levels associated with RAS mutation (p=0.023). Supporting pre-clinical mouse xenograft tumor model developed using ANG2-overexpressing human thyroid cancer cell line FTC-286 demonstrated reduced sensitivity to lenvatinib treatment. Increased levels of IL-10 and FGF-2 8-day post-treatment (8d post-tx) significantly associated with RAS (N- or K-) and BRAF mutation. Combination of gene mutation status with baseline CAF levels provided better prediction of longer PFS compared to gene mutation alone. Hierarchical clustering modeling using changes in CAF levels 8d post-tx with tumor shrinkage identified a set of CAFs that could predict two categories of lenvatinib response: longer PFS and greater tumor shrinkage or longer PFS in the absence of significant tumor shrinkage. Conclusions: CAF profiling identified potential predictive biomarkers of lenvatinib treatment outcomes in DTC. Combination of RAS and BRAF mutation with baseline VEGF and ANG-2 or treatment-associated changes in FGF-2 and IL-10 level correlated with lenvatinib treatment response.