PURPOSE: We assessed the efficacy of cediranib, olaparib, and cediranib/olaparib compared with standard-of-care chemotherapy (SOC) in platinum-resistant or platinum-refractory epithelial ovarian cancer (PROC). PATIENTS AND METHODS: NRG-GY005 is an open-label, four-arm, phase II/III superiority trial enrolling patients with high-grade serous/endometrioid PROC and one to three previous therapies. Key exclusion criteria included previous receipt of poly(ADP-ribose) polymerase inhibitor or receipt of antiangiogenic therapy in the recurrent setting. Treatment arms (SOC [once weekly paclitaxel, topotecan, or pegylated liposomal doxorubicin], cediranib, olaparib, or cediranib/olaparib) were equally randomized. A preplanned interim futility analysis on the basis of progression-free survival (PFS) selected treatment arms to advance to phase III. PFS and overall survival (OS) were phase III coprimary end points, with hierarchical testing of PFS followed by OS to preserve type 1 error control, designed to have 90% power for a 0.625 PFS hazard ratio (HR). OS was tested after PFS in the multiple hierarchical testing procedure. Secondary end points included objective response rate (ORR) and patient-reported outcomes. RESULTS: Five hundred sixty-two eligible patients were enrolled for phase II/III. Three arms met PFS criteria to carry forward to phase III (SOC, cediranib/olaparib, and cediranib). Median PFS was 3.4, 5.2, and 4 months with SOC, cediranib/olaparib, and cediranib, respectively, with a median follow-up duration of 42.2 months. PFS HR estimates for cediranib/olaparib and cediranib ( SOC) were 0.796 (98.3% CI, 0.597 to 1.060) and 0.972 (98.3% CI, 0.726 to 1.300), respectively. Median OS was 13.6, 12.8, and 10.5 months, and of 443 patients with measurable disease, ORR was 8.6%, 24.7%, and 13.1% for SOC, cediranib/olaparib, and cediranib, respectively. No new safety signals were identified. In patients receiving cediranib/olaparib, no statistically significant difference was observed on the NFOSI-DRS-P subscale compared with SOC (98.3% CI, -1.3 to 1.5, = .8725). CONCLUSION: The cediranib-containing arms demonstrated clinical activity on the basis of PFS but were not superior compared with SOC.