Cyclophosphamide-Free Adjuvant Chemotherapy for Ovarian Protection in Young Women With Breast Cancer

A Randomized Phase 3 Trial

Ke-Da Yu, MD, PhD; Jing-Yu Ge, MD; Xi-Yu Liu, MD; Miao Mo, MD; Min He, MD; Zhi-Ming Shao, MD


J Natl Cancer Inst. 2021;113(10):1352-1359. 

In This Article


Study Design and Participants

The Substitution of Paclitaxel for Cyclophosphamide on Survival Outcomes and Resumption of Menses in Young Women with ER-Positive Breast Cancer trial is a randomized, open-label, multicenter, phase III trial performed at 8 hospitals in China. It was designed to compare the difference in menses resumption rates, as well as in DFS, between the EC-wP and EP-wP groups in young women.

Women aged 18 to 40 years with unilateral operable primary invasive ER-positive HER2-negative breast cancer were eligible for enrollment following definitive surgery. Because platinum-based adjuvant chemotherapy might be an effective treatment for triple-negative breast cancer,[8] and anti-HER2 treatments are pivotal for HER2-positive disease, we excluded young patients with triple-negative or HER2-positive cancers and focused on women with ER-positive and HER2-negative disease.

Patients were required to have pathologically confirmed regional node-positive disease or node-negative disease with high-risk factors (primary tumor diameter >10 mm when histological grade III or tumor diameter >20 mm when histological grade II). ER, progesterone receptor, and HER2 statuses were identified locally at each participating center based on immunohistochemistry of tumor sections. The immunohistochemical cutoff for ER-positive or progesterone receptor–positive statuswas 1% or more staining in nuclei.[9] HER2-negative status was defined as immunohistochemistry score 0 or 1 or the absence of HER2 amplification by fluorescence in situ hybridization analysis.[10] The Eastern Cooperative Oncology Group performance status was required to be 0 or 1.

To assess menstrual status accurately, eligible participants should take no estrogens, antiestrogens, selective estrogen-receptor modulators, aromatase inhibitors, LHRHa, or hormonal contraceptives within the month before enrollment. Patients had regular menstrual cycles and normal menses before surgery. The regular menstrual cycles should occur every 21 to 35 days, with menstruation lasting 2 to 7 days.[11–13] Patients who underwent hysterectomy or bilateral salpingectomy, oophorectomy, or salpingo-oophorectomybefore enrollment were ineligible because they were not evaluable for menses. A complete description of the inclusion and exclusion criteria is provided in the protocol (available in Supplementary Methods, available online).

The independent institutional review boards of the participating centers approved the study protocol. All patients provided written informed consent. We performed the study according to the International Conference on Harmonisation Good Clinical Practice guidelines and the ethical principles of the Declaration of Helsinki. This study is registered at (NCT01026116).

Randomization and Masking

Patients were randomly assigned (1:1) using a web-response system. A permuted-block randomization scheme was used with stratification according to pathological node status (negative vs positive), tumor size (pT1 vs pT2-3), and age (≤35 years vs 35–40 years). The stratification factors were used for stratified analyses unless indicated otherwise.


The participants' baseline characteristics were recorded at random assignment. The concurrent use of taxane and anthracycline might increase the toxicity, such as febrile neutropenia.[14] Therefore, we modified the dose of epirubicin to 75 mg/m2 rather than the standard dosage of 90–100 mg/m2. Epirubicin 75 mg/m2 was considered an acceptable dosage based on the EORTC 10994/BIG 1–00 trial, where the survival efficacy of epirubicin 75 mg/m2 was not compromised and the toxicity was reduced compared with epirubicin 100 mg/m2.[15] We also balanced the total dose of epirubicin between the 2 regimens to increase the comparability.

Patients were randomly assigned to receive epirubicin (75 mg/m2) and paclitaxel (175 mg/m2) every 3 weeks for 4 cycles followed by weekly paclitaxel (80 mg/m2) for 12 weeks (EP-wP) or epirubicin (75 mg/m2) and cyclophosphamide (600 mg/m2) every 3 weeks for 4 cycles followed by weekly paclitaxel (80 mg/m2) for 12 weeks.

After chemotherapy, all patients were recommended to receive 20 mg/d of tamoxifen for at least 5 years. If an ovarian suppression treatment was administered after adjuvant chemotherapy, the patients should have experienced at least 1 menstruation and be diagnosed as premenopausal according to the follicle-stimulating hormone and estradiol levels.[16] In 2015, the SOFT and TEXTtrials showed that, for women who were at sufficient recurrence risk and remained premenopausal, the use of ovarian suppression (LHRHa) with tamoxifen or an aromatase inhibitor would improve survival.[17] For ethical reasons, after 2015, the high-risk young patients were permitted to be treated with ovarian suppression before the apparent menses resumption after chemotherapy. Patients were told of their potential choice of endocrine therapy regimens before random assignment, and the type of endocrine therapy was mainly determined by physicians according to the patients' risk. However, we were near the end of our patient recruitment in 2015, and only approximately 5% of patients used upfront ovarian suppression with an aromatase inhibitor.


The 2 coprimary endpoints were the rate of menstrual resumption at 12 months after chemotherapy and the 5-year DFS in the intention-to-treat population.

Resumption of menses was defined as at least 2 consecutive menstruations or at least 1 menstruation with a confirmed premenopausal level of follicle-stimulating hormone and estradiol after chemotherapy.[18] Patients with no results for menstrual resumption (because of a loss to follow-up, early intervention of ovarian suppression, or any recurrence event if it occurred first) were treated as nonresumed in the menses analysis. DFS was defined as the time from random assignment to breast cancer recurrence, second primary breast and other cancers, and death from any cause.

Secondary endpoints included distant DFS, overall survival, and toxicity. Toxicity was graded according to National Cancer Institute Common Toxicity Criteria version 4.0. All reported serious adverse events were judged by an independent data safety and monitoring board.

The posthoc exploratory analysis of pregnancy within 48 months was performed in May 2020. Pregnancy outcomes were assessed by patient-reported outcome questionnaires via telephone survey because of the coronavirus disease 2019 pandemic.

Statistical Analysis

Two coprimary endpoints were investigated, and the study was considered positive if either the DFS and/or the menstrual resumption results were statistically significant.

To address multiple endpoint-related multiplicity problems, we used the unequally weighted Bonferroni method by dividing the overall α into unequal portions.[19] The type I error (α = 0.05) was controlled and split between the analyses of DFS (α = 0.04) and menstrual resumption rate (α = 0.01). Interim analysis was not planned, because the EP-wP and EC-wP had been previously investigated in the Loesch study,[7] and it showed that the toxicity of the 2 regimens is manageable and tolerable. Long-term observation time is needed to observe menses resumption and pregnancy.

To detect an absolute 15% improvement in the menstrual resumption rate between EC-wP (assumed 65%) and EP-wP (assumed 80%), a 2-sided test of resumption rates with 80% power required 440 patients to achieve a .01 level of statistical significance. To detect an absolute 8% improvement in the 5-year DFS rate between the EC-wP group (assumed 80%) and EP-wP group (assumed 88%), a 2-sided log-rank test with 80% power required 480 patients (240 for each arm) to show a .04 statistical significance level, with approximately 90 DFS events expected after a median 5-year follow-up. Considering a 5% loss to follow-up, we calculated approximately 500 patients required.

The cumulative incidence estimates of menstrual resumption were calculated, and the between-group differences were compared by the stratified Miettinen and Nurminen method.[20] Stratified logistic regression analysis was used to estimate the odds ratio (OR) of menses resumption. Fine-Gray competing risk regression model (Stata command: stcrreg) was used to calculate the odds ratio of menses recovery, accounting for DFS events as competing risk events.[21] In the sensitivity analysis, the proportion of menses resumption between groups was compared using the stratified Mantel-Haenszel test. The Kaplan-Meier method was used to estimate DFS, and survival rates were compared using the stratified log-rank test. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using a stratified Cox proportional hazards model.

The findings of secondary endpoints should be interpreted as exploratory because of the potential for type I error from multiple comparisons. For other continuous and categorical variables, the Wilcoxon rank-sum test and χ 2 test were used to evaluate differences between the 2 groups, respectively. All statistical tests were 2-sided, a P value less than .05 was considered statistically significant, and analyses were performed using STATA 16.0 software (StataCorp LLC).