Pembrolizumab is a potentially viable treatment for a subset of sarcomas, with promising activity in undifferentiated pleomorphic sarcoma (UPS) and liposarcoma in a phase II clinical trial. Melissa A. Burgess, MD, of the University of Pittsburg, presented final results and biomarker analyses from the SARC028 multicenter trial (Abstract 11008) at the Sarcoma Oral Abstract Session on June 2.
Response in UPS correlated with PD-L1 status, and baseline infiltrating cytotoxic T cells also correlated with response, Dr. Burgess said.
SARC028 is a multicenter trial evaluating single-agent pembrolizumab in patients with advanced soft tissue sarcomas (STS) and bone sarcomas. The primary endpoint was objective response rate (ORR) by RECIST 1.1, and secondary objectives included safety and tolerability, progression-free survival (PFS), overall survival (OS), immune-related response rate, and analysis of pretreatment PD-L1 tumor expression and response.
The STS arm included 10 patients in four cohorts: UPS, dedifferentiated liposarcoma, synovial sarcoma, and leiomyosarcoma. The bone sarcoma arm had 40 patients, including 22 with osteosarcoma, 13 with Ewing sarcoma, and five with chondrosarcoma. Most patients in each arm received two or more previous treatments.
Patients received 200 mg of pembrolizumab intravenously every 3 weeks. Biopsies were taken before treatment and while patients were receiving treatment. Tumors were assessed for PD-L1 expression and immune infiltrates using multicolor immunohistochemistry (IHC).
In the overall study, with a median follow-up of 19 months, the projected overall response was 18%, Dr. Burgess said. In the STS arm, one patient had a complete response (CR; 2.5%) and six patients (15%) had a partial response (PR); one patient (10%) with UPS had CR and three had PR (30%). Two patients (20%) with liposarcoma and one (10%) with synovial sarcoma had PR. The median response duration in the STS arm was 33 weeks. Median PFS was 18 weeks, and median OS was 49 weeks.
In the bone sarcoma arm, two patients (5%) experienced PR: one (20%) with chondrosarcoma and one (5%) with osteosarcoma. Nine patients overall had stable disease. The median duration of response was 43 weeks in the bone sarcoma arm. Median PFS was 8 weeks, and median OS was 52 weeks.
Serious adverse events were “infrequent and not unexpected,” Dr. Burgess said. Overall, grade 3 and 4 adverse events were seen in 7% of the STS arm and 12% of the bone sarcoma arm.
In 70 biopsies analyzed for pretreatment PD-L1 expression, three were positive. The PD-L1–positive cases were restricted to UPS; of the two that were evaluable for response, one had CR and one had PR. No post-treatment biopsies were PD-L1–positive. In 28 paired tumor samples, multicolor IHC showed that responders had an increased number of infiltrating cytotoxic T cells at baseline. Further biomarker analysis with a multiplex panel is planned, as are confirmatory trials in UPS and leiomyosarcoma.
Discussant Jean-Yves Blay, MD, PhD, of Centre Léon Bérard, in France, noted that there are multiple levels of complexity in attempting to understand immunotoxicity in sarcoma, and he congratulated the study authors for “looking at biomarkers with a multidimensional approach.” Even though most patients in SARC028 did not experience tumor reduction, he noted, some had significant responses. He said biomarkers in sarcoma must be further investigated “because it could be a link to better guide therapy in the future.”
Commenting on Dr. Burgess’s presentation along with two other abstracts presented in the session, Dr. Blay said, “We can say that immunotherapy works in sarcoma. But it works in immune subsets of molecular subsets of histological subgroups. We must do better. Biomarkers are the key question.” He called for a worldwide collaborative effort by the sarcoma community to undertake basket and umbrella trials exploring combination treatments in subtypes of sarcomas.