<b>Objectives:</b> Although ovarian cancer is a rational target for immune therapy, immune checkpoint blockade has not demonstrated the efficacy seen in other solid tumors. In preclinical studies, PARP inhibition (PARPi) synergized with CTLA4 immune checkpoint blockade, resulting in durable anti-tumor immunity and long-term survival in ovarian cancer models. This combination was tested in women with germline BRCA1 or BRCA2 mutations (gBRCAm) and recurrent ovarian cancer in a phase I/II clinical trial (NCT0257125). Correlative studies were performed to identify predictive biomarkers of response. <b>Methods:</b> Women with recurrent ovarian cancer, confirmed gBRCAm, and measurable disease were eligible for participation. There were no restrictions on the number or type of prior treatment regimens. Prior PARPi treatment was allowed. Subjects received olaparib at 300mg BID and the CTLA4 antibody tremelimumab at 10mg/kg IV every 4 weeks for six doses followed by maintenance dosing every 12 weeks. Treatment continued for up to two years. The Phase I lead-in allowed dose modifications for toxicity. The phase II component used Simon's 2-stage design based on the primary endpoint of overall response rate. RNAseq was performed on archived formalin-fixed, paraffin embedded pre-treatment tumor samples from 30 subjects, and candidate biomarkers were selected based on differential expression in subjects with and without clinical benefit. <b>Results:</b> Forty-nine patients received study treatment (mean age 56 yrs, range 39-79) at five sites. Twenty-eight subjects had received ≥3 prior chemotherapy regimens, and 29 were previously treated with a PARPi. After 14 patients received the study regimen, the recommended phase 2 schedule for tremelimumab was changed to 10mg/kg infusions every 4 weeks for four doses followed by maintenance dosing every 12 weks due to immune-related adverse events, primarily colitis. Fifty-nine percent of subjects experienced at least one non-hematologic ≥ grade 3 toxicity. Among 44 evaluable patients, the overall response rate was 39%, with a clinical benefit rate of 48%. The median PFS was 3.4 months (95% CI 2.7-6.1). In a curated analysis of interferon response genes, lack of IFIT1B expression was associated with significantly enhanced PFS (median 11.4 vs 2.7 months; log-rank p=0.009; Fig 1). In an unbiased analysis of all genes, VSTM5 was the most differentially expressed marker (AUC=0.852), and lack of VSTM5 expression was significantly associated with PFS (8.12 vs 2.56 months; log-rank p=0.001). <b>Conclusions:</b> The combination of olaparib and tremelimumab achieved durable treatment benefit in a subset of gBRCAm recurrent ovarian cancer patients. Immune toxicity was moderate. Biomarkers based on gene expression in archived tumor samples were associated with clinical benefit. Future studies are needed to assess the utility of these biomarkers to identify candidates for immune therapy.
Read full abstract