Response to anti-PD1 therapy with nivolumab in metastatic sarcomas

Patients with metastatic soft tissue sarcomas generally have a poor prognosis, with low response rates after first line chemotherapy [3]. Of note, the tyrosine kinase inhibitor pazopanib was approved by the FDA in 2012 on the basis of a phase III randomized trial showing improved PFS in the second line setting; the overall response rate was only 6% [8].

Multiple recent genomic studies have provided better insight into sarcoma biology through a more accurate classification by molecular subtype, identification of recurrent mutations in oncogenic pathways and evidence of epigenetic dysregulation [9]. Barretina et al. [10] for example, recently provided a comprehensive database of sarcoma genome alterations in 207 samples of STS; despite their elucidation of genes and signaling pathways not previously associated with STS, we still lack appropriate pharmacologic tools for targeting specific genomic alterations.

Several subtypes of STS are characterized by specific chromosomal translocations which result in unique fusion proteins; while many of them function as transcription factors, making their therapeutic targeting quite challenging, these proteins may represent attractive targets from an immunotherapy standpoint [11]. Immunogenicity of sarcoma is supported by several preclinical studies and some clinical data with human sarcoma specimens.

Immunotherapeutic strategies in sarcoma have included cytokine-based immunotherapies, treatment with muramyl tripeptide phosphatidyl ethanolamine in osteosarcoma, vaccines and adoptive immunotherapy to cite a few examples, although none have appeared promising to date [12]. Our retrospective analysis shows the potential clinical benefit from treatment of soft tissue and bone sarcomas with the anti-PD1 antibody nivolumab.

This is not a prospective study, and given the retrospective nature of this series, it has several limitations; data on patients who received either nivolumab alone (N = 10) or nivolumab + pazopanib (N = 18), were pulled together in order to capture a possible signal of activity from immunotherapy (alone or in combination) that may be helpful for a following prospective study. Additionally, this is small study with multiple hystologies included: the largest group of patient had a diagnosis of leiomyosarcoma (N = 7), but most subtypes are represented by only 1 or 2 patients.

In our series we showed disease improvement or stabilization in 12/24 patients evaluable for response. Eighteen out of twenty-eight patients concomitantly received pazopanib, however 1 partial response was observed in a dedifferentiated chondrosarcoma on nivolumab alone. Another response was seen in a patient with an unresectable maxillary OS who received four cycles of nivolumab and only one month of pazopanib that was started after the 4 cycles of nivolumab. A third patient with an epithelioid sarcoma, progressing on pazopanib, had a partial response after only four cycles of nivolumab; unfortunately he progressed after four additional cycles. Of note, overall responses were observed in some subtypes that are generally resistant to traditional chemotherapy such as dedifferentiated chondrosarcoma and epithelioid sarcoma. Interestingly, all the three aforementioned patients received adjuvant radiation therapy up to 20 years before, bringing up the possibility of a distant abscopal effect as hypothesized for other diseases such as melanoma [13].

At least three prospective phase II studies are exploring the role of the checkpoint inhibitors pembrolizumab and nivolumab in metastatic STS/bone sarcomas and/or uterine LMS; preliminary data were recently presented at the ASCO 2016 conference for two studies. Pembrolizumab showed some interesting responses in undifferentiated pleomorphic sarcoma (4/9), liposarcoma (2/9), synovial (1/9), chondrosarcoma (1/6) and osteosarcoma (1/19); no responses were seen in LMS (0/10) and Ewing sarcoma (1/13) [14]. Interestingly in our series, we also observed a partial response in one patient with a dedifferentiated chondrosarcoma and a PD-L1 expression that was higher compared to all other tested patients (20% versus less than 5%); additionally, Kostine et al. [15] recently showed that this specific subtype of bone sarcoma expresses PD-L1 in association with immune-infiltrating cells and HLA class I in nearly 50% of cells. Immunotherapy with check-point inhibitors seems a particularly promising approach for the treatment of this rare and challenging histology but more data is needed.

A second prospective study is exploring nivolumab in 12 patients with uterine LMS and showed no responses [16]. In our series, among seven patients with LMS we observed 4 PD and 3 SD. LMS is characterized by a significant degree of morphologic and molecular heterogeneity and different molecular subtypes may respond differently to immunotherapy [17].

The combination of nivolumab and pazopanib is interesting but needs dose optimization to prevent, in particular, excessive liver toxicity. Nivolumab at 2 mg/kg every 3 weeks has been combined with pazopanib at 800 mg po daily in patients with renal cell carcinoma; about 70% of patients experienced grade 3–4 side effects, mainly LFT abnormalities, fatigue and diarrhea [18].