ECOG 4599 phase III trial of carboplatin and paclitaxel ±bevacizumab: Subset analysis of survival by gender

Abstract

7036 Background: E4599 compared carboplatin and paclitaxel (PC) versus carboplatin, paclitaxel, and bevacizumab (PCB) in patients with advanced stage non-small cell lung cancer. Survival was superior with PCB. However, an unplanned subset analysis did not show a survival benefit for women in the PCB arm. Methods: Patients (pts) in the E4599 database were divided into male (M) and female (F) cohorts by assigned treatment. Survival was calculated separately for each cohort. Known prognostic factors and toxicities were compared by gender. Proportional hazards models (PHM) of survival with multiple factor combinations were fit to adjust for treatment effect for sex, performance status, stage, liver, bone and adrenal involvement, measurable disease, prior radiation therapy, weight loss ≤ 5% and to examine the difference in treatment effect by sex. Results: The analysis includes 850 patients. The median survival for M is 8.7 months (mo) (PC) versus (vs) 11.7 mo (PCB) (p=0.001). The median survival for F is 13.1 mo (PC) vs 13.3 mo (PCB) (p=0.87). On PCB, progression free survival (PFS) for M and F is 6.3 mo and 6.2 mo respectively. Response rate (RR) is 23.6% for M and 38.5% for F. Duration of response is 6.8 mo for both. No demographic differences exist between the two arms for M. A higher proportion of F on the PCB arm have liver metastasis (PCB 23.2% vs PC 11.7%, p=0.003). On the PC arm, the febrile neutropenia rate is higher for M (M 3.1% vs F 0%, p=0.02). On the PCB arm, the adverse events with a gender difference are severe hypertension (M 4.2%, F 9.9%, p=0.02), constipation (M 1.4%, F 4.7%, p=0.05), and abdominal pain (M 0.9%, F 5.2%, p=0.01). The test for a treatment by sex interaction in a PHM for survival has p = 0.04. In the PHM adjusting for the other factors, the estimated treatment hazard ratios are 0.73 for M and 0.97 for F and the test for a sex by treatment interaction has p=0.09. Conclusions: Women on the PCB arm vs the PC arm appear not to have a survival advantage. Reasons for this remain unclear. However, the addition of bevacizumab does result in significant improvements in RR and PFS, arguing in favor of an overall treatment benefit. Further analysis of other demographic information (tumor type and second-line therapy) is ongoing. [Table: see text]