A randomized phase 3 trial of paclitaxel (P) plus carboplatin (C) versus paclitaxel plus ifosfamide (I) in chemotherapy-naive patients with stage I-IV, persistent or recurrent carcinosarcoma of the uterus or ovary: An NRG Oncology trial.

5500 Background: Gynecologic carcinosarcomas (CS) are rare yet aggressive epithelial malignancies for which optimal therapy is debated. PI was shown to be superior to I. PC demonstrated compelling phase 2 activity with improved safety and convenience. Methods: Main inclusion: ≥18 y; chemotherapy naïve stage I-IVB or recurrent uterine (U) or ovarian (O) CS. Treatment randomised 1:1 to PC (P 175mg/m2 with C: AUC 6 or 5 if prior RT on D1) or PI (P: 135 mg/m2; I 1.6 g/m2 D1-3; G-CSF support with dose escalation & de-escalation based on nadir counts) q21days for 6-10 cycles. Quality of life (QOL) (FACT-En TOI) and neurotoxicity (FACT/GOG-Ntx subscale) administered at 4 timepoints. A stratified log-rank test compared primary endpoint of overall survival (OS) from entry between treatment groups for non-inferiority (NI) of PC to PI. With 264 events, power was 80% for a null hazard ratio of 1.2 against a 13% greater death rate on PI when type I error is limited to 5% for a one-tail test. NCT00954174. Results: 637 pts accrued with a median follow-up of 61 months. The primary (U, n = 536) and secondary (O, n = 101) cohorts are analyzed separately and included 449 and 90 pts eligible pts, respectively. For the U cohort:PC and PI were randomly assigned to 228 and 221 eligible pts. Stage distribution: I (40%), II (6%), III (31%), IV (15%) and recurrent (8%). The study met its primary objective withPC not inferior to PI (intention-to-treat analysis;Median OS 37 vs. 29 mo, HR = 0.87; 90% CI = 0.70 to 1.075; p < 0.01 for NI, p > 0.1 for superiority (S)).PFS (median on PC 16mo vs PI 12mo; HR = 0.73; p = < 0.01 for NI, p < 0.01 for S). Toxicity (grade 1/2/3/4/5: PC 1/8/40/48/2%; PI 1/32/39/25/1%). Most of increase toxicity for PC was hematologic with G-CSF rarely used (N = 6). Confusion and genitourinary hemorrhage were significantly worse with PI. Both groups had decline in QOL and neurotoxicity scores. Similar trends were noted for the O cohort (OS: PC 30mo vs PI 25mo; and PFS: 15 mo vs 10 respectively). Conclusions: PC was not inferior to PI for OS with longer PFS and similar QOL and neurotoxicity. These results establish a new standard regimen for women with CS. Clinical trial information: NCT00954174.

Duke Scholars

Author Angeles Alvarez Secord Obstetrics and Gynecology, Gynecologic Oncology