Standard chemotherapy (CMF or AC) versus capecitabine in early-stage breast cancer (BC) patients aged 65 and older: Results of CALGB/CTSU 49907

Abstract

507 Background: Older women with BC are underrepresented in clinical trials. Scant data exist on the effects of adjuvant chemotherapy (CRx) in these patients (pts). CALGB/CTSU 49907 was designed to compare the efficacy of capecitabine (X) with standard treatment (S) in BC pts ≥ 65 yrs old. Methods: Pts were randomized to X or S (CMF or AC per physician choice). Eligibility: Stage T1–4, N0–3, M0; PS 0–2, ≤ 84 days from surgery; adequate organ function. Regimens (all doses per M2): CMF (C, 100 mg d1–14 oral; M 40 mg d1,8; FU 600 mg d1,8) q4wk×6 cycles; AC (A 60 mg; C 600 mg) q3wk×4; X (2,000 d1–14) q3wk×6. Endocrine therapy was recommended after completion of CRx for hormone receptor positive (HR+) pts. A Bayesian design addressed equivalence of X to S; total sample size planned 600–1,800 pts with serial monitoring. The primary endpoint was relapse-free survival (RFS) defined as locoregional or distant relapse or death. We calculate Bayesian measures, including the probability of inferiority of X (that the hazard ratio of X to S is greater than 1) and relate these to conventional p values. In unplanned subset analyses we repeat these calculations by HR status. Results: Pts were randomized between 9/2001 and 12/2006. Per protocol, accrual stopped with 633 pts (326 on S, 307 on X) based on predictive probability that with longer follow-up X was unlikely to be equivalent to S. Median follow-up is 2 yrs. Pt and tumor characteristics were similar in the 2 arms. Overall: 61% age 70+, 54% T > 2cm, 69% N+, 66% HR+. Adherence to X was excellent. Toxicity was moderate, with more myelosuppression on S, more hand-foot syndrome on X, and 2 drug-related deaths on X. Pts randomized to X were 2.4 (95% CI: 1.5–3.8) times more likely to experience an RFS event (adjusted p=0.0003) and 2.1 (95% CI: 1.2–3.7) times more likely to die (p=0.02). Probability of inferiority of X vs S was > 99% for both RFS and OS. Unplanned subset analysis showed a highly significant interaction between HR and CRx arm (p=0.0032 for RFS; p=0.0067 for OS). All advantage was for S in HR negative tumors; hazard ratios 5.1 (95% CI: 2.4–10.6) for RFS, 6.1 (95% CI: 2.1–18.0) for OS. Conclusions: X is inferior to S in older pts, particularly in those with HR-negative tumors. Analyses to be updated at ASCO 2008, including by HER-2 status. Author Disclosure Employment or Leadership Consultant or Advisory Role Stock Ownership Honoraria Research Expert Testimony Other Remuneration Hoffman-La Roche Hoffman-La Roche