A phase II trial of axitinib plus pembrolizumab for patients with advanced alveolar soft part sarcoma (ASPS) and other soft tissue sarcomas (STS).

Abstract

11547 Background: Inhibition of programmed-death 1 (PD1) by pembrolizumab (P) produced overall response rates (ORR) of 19% in advanced STS [SARC028]. Vascular endothelial growth factor (VEGF) suppresses the immune microenvironment and may contribute to checkpoint inhibitor resistance. Axitinib (Ax) is a pan-VEGFR inhibitor that prolongs progression-free survival (PFS) in advanced STS [Axi-STS], with striking antitumor activity in combination with P in renal cell carcinoma. We report toxicity, efficacy, and correlative endpoints of combination Ax/P for patients (pts) with advanced STS. Methods: NCT02301039 is an open-label Phase II trial of Ax/P in 30 pts with advanced or metastatic STS, requiring RECIST 1.1 progressing disease, adequate end-organ function and performance status. Pts received Ax 5 mg PO twice daily with concurrent P 200mg IV q21 days. Primary endpoint: 3-month PFS rate (PFS3mo); secondary endpoints: toxicity, ORR and clinical benefit rate (CBR) by RECIST 1.1, and overall survival. Tumor biopsies and peripheral blood samples were obtained for correlative studies of TIL and PBMC at baseline, 3 months on treatment, and at progression. Results: 33 pts received at least one dose of study drugs, with 30 pts evaluable for primary endpoint. Enrolled subtypes: ASPS (36%), UPS (15%), LMS (18%), other (30%). PFS3mo was 70.3% [95% CI 50.7-83.3], with median PFS of 5.4 months [95% CI 3.02-11.6]. Best ORR was 21.9% [95% CI 5.9-33.5]. PFS3mo in ASPS pts was 90.9% [95% CI 50.8-98.7], with best ORR of 45.5% [95% CI 18.1-75.4] and CBR of 72.7% [95% CI 39.3-92.7]. Ax/P was well-tolerated, with similar immune-related toxicity rates to prior studies. Correlative markers were analyzed by a penalized logistic regression model with bootstrapping to predict PFS3mo. Baseline high plasma angiogenic activity, circulating neutrophil:lymphocyte ratio < 4.1, low naïve fraction CD4+ TIL, and low PD1+CD8+ PBMC were associated with lack of progression (AUC 0.878 [95% CI 0.743-1, p = 0.0002]). Conclusions: Combination Ax/P is well-tolerated with promising activity in ASPS pts. Identified immune correlates associated with PFS should be validated in prospective immunotherapy trials. Clinical trial information: NCT02301039.