746MO Randomized phase II trial of durvalumab in combination with olaparib and cediranib (DOC) compared to olaparib and cediranib (OC) or durvalumab and cediranib (DC) or standard of care chemotherapy (SOC) in platinum-resistant ovarian cancer with prior bevacizumab (NRG-GY023)

In platinum-resistant ovarian cancer, non-platinum chemotherapy is the standard of care after progression on combination bevacizumab and chemotherapy. We assessed the activity of DOC, DC or OC compared to SOC in bevacizumab pre-exposed population. NRG-GY023 is an open-label, randomized, phase II trial conducted in the United States with a targeted sample size of 164 patients (pts). Eligibility included high-grade serous/endometrioid, or clear cell platinum-resistant ovarian cancer with 2-5 prior systemic therapies. Prior use of bevacizumab is required. Treatments were randomly assigned 1:2:2:2 to SOC (weekly paclitaxel, topotecan or liposomal doxorubicin), DOC, DC or OC. The primary end point was progression-free survival (PFS). Secondary end points included overall survival and overall response rate. Interim futility analysis was planned when 28 PFS events (60% information) were available in any pair of one experimental arm with the SOC; arms with a hazard ratio (HR) estimate (ref: SOC) > 0.87 would be discontinued. By the data cutoff (10/31/2022), 120 pts were enrolled with a median age 65 yo [IQR: 58-69] and 10% were with germline or somatic BRCA mutation. All received prior bevacizumab in either the upfront or recurrent setting. Other prior therapies included PARP inhibitor (49%) and immunotherapy (11%). 103 pts were treated (15 SOC, 30 DOC, 28 DC, and 30 OC, respectively). At the data cutoff, median follow-up time was 4.3 months (IQR: 2.0-7.9). Median PFS was 4.2 (95% CI, 1.9-5.0), 2.6 (95% CI, 2-4.2), 2.3 (95% CI, 2-4.6), and 2.3 (95% CI, 2.1-3.9) months with SOC, DOC, DC and OC, respectively. The HR estimates (ref: SOC) were 1.01 (95% CI, 0.44-2.31), 1.21 (95% CI, 0.52-2.84), and 1.34 (95%CI, 0.58-3.08) for DOC, DC and OC, respectively. Hematologic adverse events were more common with SOC. Safety signals were consistent with past experience. None of the experimental arms improved PFS compared with SOC. SOC was active in this heavily pretreated population.