Identification of proangiogenic plasmatic biomarkers in patients with advanced grade 1/2 pancreatic (pan) and gastrointestinal (gi) neuroendocrine tumors (NETs) treated with Lenvatinib: A subanalysis from the TALENT phase II clinical trial

Abstract

Introduction:TALENT phase II trial provided evidence of the efficacy of Lenvatinib in patients with panNETs and giNETs. However, no predictive biomarker has been associated to antiangiogenic treatment benefit to date. Aim(s):Identify predictive biomarkers by evaluating a subset of proangiogenic molecules found in circulating plasma of patients included in the TALENT trial. Materials and methods:Plasma samples were obtained from 84 patients prior to the start of Lenvatinib. Quantitative determinations of VEGF-A, FGF2, FGF4, Ang2, IL8, PlGF, VEGF-C, VEGF-D and VEGFR2 were analysed by multiplex ELISA with a custom made Quantibody Array. Protein levels of each biomarker were correlated with PFS, OS and ORRin the different subgroups included in the trial. Results:In the whole population, a significant association of high-Ang2 and low-FGF2 plasma levels with higher ORR was observed. This association was independent of primary tumor origin, giNETs (p=0.003) and panNETs (p=0.024). In the panNET cohort (prior mTOR or multikinase inhibitors mandatory) 8 patients received Sunitinib previously. In this subgroup, Ang2 and VEGFR2 were significantly decreased compared to the rest of panNET patients (-73% p=0.022 and -62% p=0.042). Low levels of Ang-2 and VEGFR2 were associated to better ORR (p=0.029 and p=0.029) and PFS (log rank p=0.027 and p=0.007) in Sunitinib pre-treated patients. Prediction value capacity determined by ROC curve resulted in cutoffs for Ang2 and VEGFR2 being 415pg/ml and 1770pg/ml (p=0.021). Conclusion:High-Ang2 and low-FGF2 levels are associated to better response to lenvatinib in panNETs and giNETs and emerge as possible predictive factors of response to Lenvatinib. In Sunitinib-pretreated patients, Ang2 and VEGFR2 levels significantly predicted PFS and ORR in panNETs.