Durvalumab plus pazopanib combination in patients with advanced soft tissue sarcomas: a phase II trial

Abstract

We aimed to determine the activity of the anti-VEGF receptor tyrosine-kinase inhibitor, pazopanib, combined with the anti-PD-L1 inhibitor, durvalumab, in metastatic and/or recurrent soft tissue sarcoma (STS). In this single-arm phase 2 trial (NCT03798106), treatment consisted of pazopanib 800 mg orally once a day and durvalumab 1500 mg once every 3 weeks. Primary outcome was overall response rate (ORR) and secondary outcomes included progression-free survival (PFS), overall survival, disease control rate, immune-related response criteria, and safety. The ORR was 30.4% and the trial met the pre-specified endpoint. The median PFS was 7.7 months (95% confidence interval: 5.7-10.4). The common treatment-related adverse events of grades 3-4 included neutropenia (9 [19.1%]), elevated aspartate aminotransferase (7 [14.9%]), alanine aminotransferase (5 [10.6%]), and thrombocytopenia (4 [8.5%]). In a prespecified transcriptomic analysis, the B lineage signature was a significant key determinant of overall response (P = 0.014). In situ analysis also showed that tumours with high CD20⁺ B cell infiltration and vessel density had a longer PFS (P = 6.5 × 10^-4) than those with low B cell infiltration and vessel density, as well as better response (50% vs 12%, P = 0.019). CD20⁺ B cell infiltration was identified as the only independent predictor of PFS via multivariate analysis. Durvalumab combined with pazopanib demonstrated promising efficacy in an unselected STS cohort, with a manageable toxicity profile.

Fig. 1. Efficacy of durvalumab and pazopanib combination in the treatment of STS (n = 46).

A Waterfall plot of the maximum change in tumour size. From top to bottom, panels indicate: Waterfall plot representing the percentage of maximum tumour reduction as assessed according to RECIST 1.1 criteria, including histological subtype and PD-L1 expression. Others indicate clear cell sarcoma (n = 2), endometrial stromal sarcoma (n = 2), malignant glomus tumor (n = 1), and hemangioendothelioma (n = 1). B Representative spider plot illustrating changes in tumour burden from the baseline. ASPS Alveolar soft part sarcoma, ANG angiosarcoma, MPNST malignant peripheral nerve sheath tumour, UPS undifferentiated pleomorphic sarcoma, DDLPS dedifferentiated liposarcoma, DSRCT desmoplastic small round cell tumour, LMS leiomyosarcoma, SS synovial sarcoma, CR complete response, PR partial response, SD stable disease, PD progressive disease. Source data are provided as a Source Data file.

Fig. 2. Survival analyses of durvalumab and pazopanib combination (n = 46).

Kaplan–Meier curves associated with PFS (A) and OS (B) for all patients. C Swimmer plot. Each lane represents a single patient’s data. X-axis represents the duration of treatment for each patient. PFS progression-free survival, OS overall survival. Source data are provided as a Source Data file.

Fig. 3. Molecular landscape and response to durvalumab and pazopanib combination treatment (n = 28).

A Integrated plot of clinical and molecular features with whole-exome sequencing results From top to bottom, panels indicate: Waterfall plot representing the percentage of maximum tumour reduction as assessed according to RECIST 1.1 criteria; the number of somatic mutations; clinical characteristics including best response, histological subtype, and landscape of alterations (LOH of HLA genes, gene fusions, mutations, and somatic copy number alterations). B Transcriptomic correlates of clinical response to durvalumab and pazopanib combination treatment. Heat maps describing tumour microenvironment cell infiltration estimated by MCP-counter. From top to bottom, heat maps indicate: MCP-counter scores of immune and stromal cells; single sample gene set enrichment analysis (ssGSEA) scores for immune-associated gene signatures; and gene expression levels for immune-checkpoint genes. The colour scale indicates Z-normalised values of each gene signature for gene expression across samples. The colour bars above the heatmap indicate the tumor histology (top), best response (middle; CR, purple; PR, yellow; SD, green; PD, blue), and responders (pink) and non-responders (blue) to the pazopanib-durvalumab combination (bottom). C Comparison of TME and immune-checkpoint genes between responders (n = 9) and non-responders (n = 19). Composition of the TME estimated using MCP-counter scores were compared between responders and non-responders (top). Expression levels of immune-checkpoint genes were compared between responders and non-responders (bottom). Center lines, upper and lower bounds of boxplots indicate the median, 25th, and 75th quantile, respectively. The whiskers of boxplots indicate 1.5 times of the interquartile range. P-value was derived from the two-sided Wilcoxon rank-sum test and not adjusted for multiple tests. TME tumour microenvironment, TMB tumour mutation burden, HLA human leukocyte antigen, MPNST malignant peripheral nerve sheath tumour, UPS undifferentiated pleomorphic sarcoma, DSRCT desmoplastic small round cell tumour, LMS leiomyosarcoma, SS synovial sarcoma, DDLPS dedifferentiated liposarcoma, BR best response, CR complete response, PR partial response, SD stable disease, PD progressive disease, LOH loss of heterozygosity, Treg regulatory T cells, Class_I_MHC major histocompatibility complex class I, TLS tertiary lymphoid structures, HAVCR2 Hepatitis A Virus Cellular Receptor 2, LAG3 Lymphocyte-Activation Gene 3. Source data are provided as a Source Data file.

Fig. 4. Distinct quantification of the immune microenvironment by multiple IHC analysis (n = 39).

PFS of patients according to CD20⁺ B cell infiltration (A) and vessel density (B). C PFS and response of patients based on CD20⁺ cell infiltration and vessel density. The right-bottom panel indicates the number of samples grouped by response to the treatment and CD20⁺ B cell infiltration and vessel density, and the association of response and CD20⁺/vessel density group was evaluated via two-sided fisher’s exact test. P < 0.05 was considered statistically significant. Hi indicates patients with a density higher than or equal to the median, and Lo indicates patients with a density lower than the median. IHC immunohistochemistry, PFS progression free survival, R responder, NR non-responder. Source data are provided as a Source Data file.