The primary outcome was progression-free survival, which was defined as the time from random assignment to disease progression or death from any cause. Data from patients who were alive and progression-free at the time of cutoff of the data (September 29, 2011) were censored at the last follow-up visit at which progression had not yet been observed. Overall survival, which was a secondary outcome, was defined as the time from random assignment to death from any cause. We calculated the rates of clinical benefit using the number of patients with a complete or partial response or stable disease as the numerator and the number of all patients (even those in whom a response could not be assessed or for whom response data were missing) as the denominator. The rate of objective response was calculated only for patients with measurable disease, whereas the rate of clinical benefit applied to all patients. Both the primary analysis of progressionfree survival and the analysis of overall survival were specified as log-rank tests stratified according to prior receipt or no prior receipt of adjuvant tamoxifen therapy. Kaplan–Meier methods were used to construct survival plots and to estimate the survival percentages and the median times to progression-free and overall survival. Cox regression was used to estimate hazard ratios and 95% confidence intervals.