Adjuvant ovarian suppression in premenopausal breast cancer

Abstract

Background: Suppression of ovarian estrogen production reduces the recurrence of hormone-receptor-positive early breast cancer in premenopausal women, but its value when added to tamoxifen is uncertain.

Methods: We randomly assigned 3066 premenopausal women, stratified according to prior receipt or nonreceipt of chemotherapy, to receive 5 years of tamoxifen, tamoxifen plus ovarian suppression, or exemestane plus ovarian suppression. The primary analysis tested the hypothesis that tamoxifen plus ovarian suppression would improve disease-free survival, as compared with tamoxifen alone. In the primary analysis, 46.7% of the patients had not received chemotherapy previously, and 53.3% had received chemotherapy and remained premenopausal.

Results: After a median follow-up of 67 months, the estimated disease-free survival rate at 5 years was 86.6% in the tamoxifen-ovarian suppression group and 84.7% in the tamoxifen group (hazard ratio for disease recurrence, second invasive cancer, or death, 0.83; 95% confidence interval [CI], 0.66 to 1.04; P=0.10). Multivariable allowance for prognostic factors suggested a greater treatment effect with tamoxifen plus ovarian suppression than with tamoxifen alone (hazard ratio, 0.78; 95% CI, 0.62 to 0.98). Most recurrences occurred in patients who had received prior chemotherapy, among whom the rate of freedom from breast cancer at 5 years was 82.5% in the tamoxifen-ovarian suppression group and 78.0% in the tamoxifen group (hazard ratio for recurrence, 0.78; 95% CI, 0.60 to 1.02). At 5 years, the rate of freedom from breast cancer was 85.7% in the exemestane-ovarian suppression group (hazard ratio for recurrence vs. tamoxifen, 0.65; 95% CI, 0.49 to 0.87).

Conclusions: Adding ovarian suppression to tamoxifen did not provide a significant benefit in the overall study population. However, for women who were at sufficient risk for recurrence to warrant adjuvant chemotherapy and who remained premenopausal, the addition of ovarian suppression improved disease outcomes. Further improvement was seen with the use of exemestane plus ovarian suppression. (Funded by Pfizer and others; SOFT ClinicalTrials.gov number, NCT00066690.).

Figure 1. Randomization and Primary Analysis Populations

The flow diagram shows the intention-to-treat population of 2033 patients included in the primary analysis (shaded) of tamoxifen plus ovarian suppression, as compared with tamoxifen alone, and the analogous population of patients assigned to receive exemestane plus ovarian suppression. Additional details are provided in Figure S1 in the Supplementary Appendix.

Figure 2. Primary Analysis Comparisons of Tamoxifen plus Ovarian Suppression (OS) with Tamoxifen Alone

Panel A shows Kaplan–Meier estimates of disease-free survival, and Panel B shows the results of the Cox proportional- hazards models for the comparisons of disease- free survival, freedom from recurrence of breast cancer, freedom from recurrence of breast cancer at a distant site, and overall survival, according to treatment group, among all the patients and according to chemotherapy cohort. The solid vertical lines at 0.83, 0.81, 0.88, and 0.74 indicate the overall hazard-ratio estimates for the four end points, respectively. In the analysis of disease-free survival, the hazard ratio is for breast-cancer recurrence, a second invasive cancer, or death. In the analyses of freedom from breast cancer and freedom from recurrence of breast cancer at a distant site, the hazard ratios are for recurrence. In the overall survival analysis, the hazard ratio is for death. The P value for the comparison among all patients was obtained by means of a stratified log-rank test; the P value for the assessment of treatment-effect heterogeneity according to chemotherapy cohort was calculated by means of a test of treatment by chemotherapy cohort interaction from a stratified Cox proportional-hazards model. The x axis is scaled according to the natural logarithm of the hazard ratio. The size of the squares is inversely proportional to the standard error of the hazard ratio. The median follow-up was 67 months.

Figure 3. Kaplan–Meier Estimates of Freedom from Recurrence of Breast Cancer and Freedom from the Recurrence of Breast Cancer at a Distant Site after a Median Follow-up of 67 Months, According to Treatment Assignment

The estimates for the primary analysis population and the exemestane–ovarian suppression group are summarized for all patients (Panels A and B) and according to chemotherapy cohort (Panels C through F). The 5-year values are based on Kaplan–Meier estimates of the time to an event. In the analyses of freedom from recurrence of breast cancer and freedom from recurrence of breast cancer at a distant site, the hazard ratios are for recurrence of breast cancer and recurrence of breast cancer at a distant site, respectively.