Survival (S) of ovarian cancer (OC) patients (pts) treated on SWOG9701/GOG178: 12 versus (v) 3 cycles (C) of monthly single-agent paclitaxel (PAC) following attainment of a clinically-defined complete response (CR) to platinum (PLAT)/PAC

Abstract

5005 Background: This phase 3 trial was closed by the SWOG Data Safety and Monitoring Committee (DSMC) when a prospectively-defined interim analysis revealed a highly statistically significant difference in progression-free S (PFS) (28 v 21 months [M]; p = 0.0023; hazard ratio [HR] 2.31) (JCO 2003; 21:2460). At study closure, there were insufficient deaths to describe the impact of 12-C of PAC on overall S (OS). Methods: Eligibility included OC or primary peritoneal cancer, attainment of a CR to PLAT/PAC, and ≤ grade 1 neuropathy. Pts entered into the trial (n = 296) received either 12-C or 3-C of PAC (175 mg/m2 q-28 days), and were then observed until progression. Results: Of the 146 pts on the 3-C arm, 9 (6%) actually received prolonged PAC (> 3-C) following closure of the study by the DSMC. Median (12-C v 3-C; intention-to-treat analysis): updated PFS (all pts) 22 v 14 M, p = 0.01; OS (all pts) 53 v 46 M, p = 0.27. Exploratory analysis (Cox models adjusting for stratification factors): 12-C v 3-C HR 0.59 (p = 0.03) if baseline CA-125 ≤ 10 (n = 175), and 1.25 (p = 0.34) for baseline CA-125 >10 (n = 121). The treatment v baseline CA-125 interaction was statistically significant (Cox model p = 0.03). Conclusion: OS was not significantly different between 12-C v. 3-C of PAC, either because (a) treatment at relapse equalized OS; (b) the sample size was insufficient to reveal a difference; (c) the crossover of pts from 3-C to longer treatment masked a potential difference; or (d) treatment was possibly prematurely discontinued at 12-C in pts who had not progressed (suggested by a previously reported exploratory analysis -JCO 2003; 21:2460). Of interest, an exploratory analysis strongly suggested an improvement in S for pts receiving 12-C of PAC if the baseline CA-125 level was ≤10, those individuals likely to have the smallest volume of clinically-undetectable residual OC when single-agent “maintenance” PAC was initiated. No significant financial relationships to disclose.