Timing of CMF chemotherapy in combination with tamoxifen in postmenopausal women with breast cancer: role of endocrine responsiveness of the tumor

M. Colleoni1,*, S. Li2, R. D. Gelber2, A. S. Coates3, M. Castiglione-Gertsch4, K. N. Price2, J. Lindtner5, C.-M. Rudenstam6, D. Crivellari7, J. Collins8, O. Pagani9, E. Simoncini10, B. Thürlimann11, E. Murray12, J. Forbes13, D. Erzen5, S. Holmberg14, A. Veronesi7, A. Goldhirsch1,9 On behalf of the International Breast Cancer Study Group (IBCSG){dagger}

1 European Institute of Oncology, Milan, Italy; 2 IBCSG Statistical Center, Dana-Farber Cancer Institute and Frontier Science and Technology Research Foundation, Boston, MA, USA; 3 The Cancer Council Australia and University of Sydney, Sydney, Australia; 4 IBCSG Coordinating Center, Bern, Switzerland; 5 Institute of Oncology, Ljubljana, Slovenia; 6 West Swedish Breast Cancer Study Group, Sahlgrenska University Hospital, Göteborg, Sweden; 7 Centro di Riferimento Oncologico, Aviano, Italy; 8 Department of Surgery, The Royal Melbourne Hospital, Melbourne, Australia; 9 Oncology Institute of Southern Switzerland Lugano, Switzerland; 10 Oncologia Medica-Spedali Civili, Brescia, Italy; 11 Kantonsspital, St Gallen, Switzerland; 12 Groote Schuur Hospital and University of Cape Town, Cape Town, South Africa; 13 Australian New Zealand Breast Cancer Trials Group, Newcastle, Australia; 14 Department of Surgery, SU/Moelndal's Hospital, Moelndal, Sweden

*Correspondence to: Dr M. Colleoni, Division of Medical Oncology, European Institute of Oncology, Via Ripamonti 435, 20141, Milan, Italy. Tel: +39-02-57489439; Fax: +39-02-7489212; Email: marco.colleoni{at}ieo.it


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
Background:: Controversy persists about whether chemotherapy benefits all breast cancer patients.

Patients and methods:: In the International Breast Cancer Study Group (IBCSG) trial VII, 1212 postmenopausal patients with node-positive disease were randomized to receive tamoxifen for 5 years or tamoxifen plus three concurrent courses of cyclophosphamide, methotrexate and 5-fluorouracil (‘classical’ CMF) chemotherapy, either early, delayed or both. In IBCSG trial IX, 1669 postmenopausal patients with node-negative disease were randomized to receive either tamoxifen alone or three courses of adjuvant classical CMF prior to tamoxifen. Results were assessed according to estrogen receptor (ER) content of the primary tumor.

Results:: For patients with node-positive, ER-positive disease, adding CMF either early, delayed or both reduced the risk of relapse by 21% (P=0.06), 26% (P=0.02) and 25% (P=0.02), respectively, compared with tamoxifen alone. There was no difference in disease-free survival when CMF was given prior to tamoxifen in patients with node-negative, ER-positive tumors.

Conclusions:: CMF given concurrently (early, delayed or both) with tamoxifen was more effective than tamoxifen alone for patients with node-positive, endocrine-responsive breast cancer, supporting late administration of chemotherapy even after commencement of tamoxifen. In contrast, sequential CMF and tamoxifen for patients with node-negative, endocrine-responsive disease was ineffective.

Key words: breast cancer, chemoendocrine therapy, estrogen receptors, postmenopausal, tamoxifen, timing of chemotherapy


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
Although the Early Breast Cancer Trialists' Collaborative Group overview of randomized trials supports the value of combination chemotherapy in postmenopausal patients with operable breast cancer [1Go], controversy persists about whether the benefit applies equally for all patients [2Go]. Furthermore, the debate continues on the optimal timing for such treatments in postmenopausal women with endocrine-responsive disease who receive adjuvant tamoxifen, especially those with node-negative disease. The use of chemoendocrine therapy, or endocrine therapy alone are considered standard options in both node-negative and node-positive disease, as stated at the recent St Gallen Meetings [3Go, 4Go], which also emphasized the endocrine responsiveness of the tumor as a critical factor for treatment selection.

Historically, cytotoxics and tamoxifen were administered concurrently because their effects were assumed to be independent. More recently, preclinical [5Go] and clinical trials data indicated a possible detrimental effect in the concomitant administration of chemotherapy and tamoxifen, suggesting that sequential administration should be preferred [6Go, 7Go]. However, there are only limited data on the optimal timing of chemoendocrine therapy, and specifically whether giving delayed chemotherapy to patients who already received tamoxifen may still be effective.

International Breast Cancer Study Group (IBCSG) trials VII [8Go] and IX [9Go] focused on chemoendocrine therapy in postmenopausal patients with early breast cancer. We analyzed these trials according to the endocrine responsiveness of the tumor and considered the question of whether adding chemotherapy concomitantly (early and/or delayed) or sequentially could influence disease-free survival (DFS) in subsets of postmenopausal breast cancer patients within defined estrogen receptor (ER) categories (absent any expression, low, moderate or high levels of expression).


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
Data were analyzed from 2881 eligible patients with breast cancer who entered the IBCSG (formerly Ludwig Group) trials VII and IX [8Go, 9Go] from 1986 to 1999.

From July 1986 to April 1993, 1212 eligible postmenopausal patients with node-positive disease were stratified by ER status, and randomized to receive: (A) tamoxifen alone 20 mg daily for 5 years; (B) tamoxifen plus three courses of concurrent early cyclophosphamide, methotrexate and 5-fluorouracil (‘classical’ CMF) on months 1, 2 and 3; (C) tamoxifen plus delayed single courses of CMF given on months 9, 12 and 15; or (D) tamoxifen plus early and delayed CMF given on months 1, 2, 3, 9, 12 and 15 (trial VII). From May 1989 to August 1999, 1669 eligible postmenopausal breast cancer patients with node-negative disease were randomized to receive either three courses of classical CMF prior to 57 months of tamoxifen compared with tamoxifen alone for 5 years (trial IX).

Clinical, hematological and biochemical assessments of each patient were required every 3 months for 2 years, every 6 months until the end of the fifth year and yearly thereafter until death. Chest X-rays and bone scans (trial VII) were required at baseline and further tests were performed only following clinical indications. All patient data, including all disease and survival related events, were reviewed and classified by the medical study coordinator (M.C.-G.).

DFS and overall survival (OS) percentages were estimated using the Kaplan–Meier method [10Go]. Greenwood's formula for the calculation of standard errors [11Go] and log-rank tests for the comparisons of treatment effects [12Go] were also used. OS results are not included because the DFS end point adequately illustrates the main message of prediction of responsiveness according to population and timing of chemoendocrine therapy.

For the simpler two-arm trial IX, early CMF followed by tamoxifen was compared with tamoxifen alone. For trial VII, three pair-wise analyses were considered: tamoxifen plus early CMF versus tamoxifen alone (early); tamoxifen plus delayed CMF versus tamoxifen alone (delay); tamoxifen with early and delayed CMF versus tamoxifen alone (both). Cox proportional hazards regression models [13Go] were used to estimate risk ratios (RRs) and confidence intervals (CIs) for the pair-wise treatment comparisons to display the magnitude of the treatment effects. Analyses were conducted both for all patients and according to subgroups defined by ER and progesterone receptor (PgR) status of the primary tumor. Results for PgR subroups are not included because they are similar to those observed for ER, but with somewhat less discrimination than ER regarding responsiveness to CMF among postmenopausal women. Cox models were also used to test for interactions between treatment effects and ER status. All probability values were two-sided and were not adjusted for multiple comparisons. The median follow-up was 10 years for trial VII and 6 years for trial IX.

To further explore the trends in treatment effect differences according to receptor levels, we used the non-parametric subpopulation treatment effect pattern plot (STEPP) methodology [14Go, 15Go]. STEPP involves defining several overlapping subgroups of patients on the basis of a covariate of interest and studying the resulting pattern of the treatment effects estimated within each subgroup. In this report, ER value (based on ligand-binding assay) was the covariate of interest, and the treatment effects estimated within each ER subgroup were measured in terms of 5-year DFS percentages.

Patient characteristics are shown in Tables 1 (trial VII) and 2 (trial IX). The median age of the patients was 60 years (range 35–84) in trial VII and 60 years (range 34–81) in trial IX. All eligible patients had ER status available according to stratified group (trial VII: positive or negative; trial IX: positive, negative or unknown) at the time of randomization, but not all were determined by ligand-binding. The STEPP analysis according to ligand-binding assay of ER included 97% of the patients in trial VII and 71% of the patients in trial IX, owing to increasing use of immunohistochemistry during the final years of accrual to trial IX.


View this table:
[in this window]
[in a new window]
 
Table 1. Patient characteristics according to treatment in International Breast Cancer Study Group trial VII

 

View this table:
[in this window]
[in a new window]
 
Table 2. Patient characteristics according to treatment in International Breast Cancer Study Group trial IX

 

    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
Treatment comparisons for all patients
The 5-year DFS percentage±standard error (SE) for the 306 patients assigned to tamoxifen alone in trial VII was 56 ± 3% compared with 62 ± 3% for the 302 patients assigned to tamoxifen plus early cycles of CMF, 60 ± 3% for the 308 patients assigned to tamoxifen plus delayed cycles of CMF and 64 ± 3% for the 296 patients assigned to tamoxifen plus both early and delayed cycles of CMF. The 5-year DFS percentage ± SE for 846 patients assigned to tamoxifen alone in trial IX was 81 ± 2% compared with 85 ± 1% for the 823 patients assigned to three cycles of initial CMF prior to tamoxifen. The risk ratios, 95% CIs and P values for the DFS comparisons are shown in Figure 1.



View larger version (13K):
[in this window]
[in a new window]
 
Figure 1. Disease-free survival risk ratios for all patients. CI, confidence interval; RR, risk ratio; Tam, tamoxifen; CMF, cyclophosphamide, methotrexate, 5-fluorouracil; N+, node-positive; N–, node-negative; Del., delayed.

 
STEPP analyses of 5-year DFS according to ER value
STEPP analyses were used to explore the pattern of treatment effect differences in terms of 5-year DFS according to ER content of the primary tumor (Figure 2AD). For this sliding-window STEPP analysis, each subpopulation contained ~120 patients for the trial VII analyses and ~200 patients for the trial IX analysis, and each subsequent subpopulation was formed moving from left to right by dropping ~10 patients with the lowest ER value and adding ~10 patients with the next higher ER value. The x coordinate indicates the median ER value (log scale) for the patients in each subpopulation. The y coordinate indicates the 5-year DFS percent estimated using the Kaplan–Meier method on data from patients in each subpopulation. These plots demonstrate a pattern of differential effectiveness of the addition of chemotherapy for varying levels of ER. In trial VII, which included patients with node-positive disease and in which the CMF was administered concurrently with tamoxifen, the benefit of chemotherapy was seen for intermediate values of ER. By contrast, for trial IX, which included patients with node-negative disease and in which the CMF was administered prior to starting tamoxifen, the benefit appeared only with low values of ER.



View larger version (22K):
[in this window]
[in a new window]
 
Figure 2. Subpopulation treatment effect pattern plots (STEPP) showing 5-year disease-free survival (DFS) estimates on the y-axis for subpopulations of patients with median values of estrogen receptor (ER) (measured in fmol/mg cytosol protein by ligand-binding assay) shown on the x-axis. For this sliding-window STEPP analysis, each subpopulation contained ~120 patients for trial VII and ~200 patients for trial IX, and each subsequent subpopulation was formed moving from left to right by dropping ~10 patients with the lowest ER value and adding ~10 patients with the next higher ER. (A) Comparison of tamoxifen + cyclophosphamide, methotrexate and 5-fluorouracil (CMF) (early) (n=298) versus tamoxifen alone (n=294) in trial VII. (B) Comparison of tamoxifen + CMF (delayed) (n=301) versus tamoxifen alone (n=294) in trial VII. (C) Comparison of tamoxifen + CMF (both early and delayed) (n=286) versus tamoxifen alone (n=294) in trial VII. (D) Comparison of CMF followed by tamoxifen (n=580) versus tamoxifen alone (n=598) in trial IX. TAM, tamoxifen.

 
Treatment comparisons according to ER cohort
The relative risks of relapse with regard to the treatment groups within ER-positive and ER-negative subgroup are shown on Figure 3.



View larger version (23K):
[in this window]
[in a new window]
 
Figure 3. Disease-free survival risk ratios (RR) for patients with estrogen receptor (ER)-negative and ER-positive disease. CI, confidence interval; N+, node-positive; N–, node-negative; Del., delayed; Tam, tamoxifen; CMF, cyclophosphamide, methotrexate, 5-fluorouracil.

 
For patients with ER-positive tumors, administration of chemotherapy together with the tamoxifen provided a DFS advantage regardless of the timing and duration of chemotherapy in patients with node-positive disease (trial VII). Compared with tamoxifen alone, the relative risk of relapse was reduced by 21% by adding early CMF (P=0.06), by 26% by adding delayed CMF (P=0.02) and by 25% by adding both early and delayed CMF (P=0.02). By contrast, DFS was not significantly improved by the addition of concurrent CMF (early, delayed or both) for patients with node-positive, ER-negative disease (trial VII). The tests for interaction between the ER-positive and ER-negative cohorts and the effect of early CMF (P=0.88), delayed CMF (P=0.072) or both early and delayed CMF (P=0.25) did not reach statistical significance.

There was no indication of difference when giving CMF prior to tamoxifen for patients with ER-positive tumors in trial IX (node-negative cohort). For patients with ER-negative tumors, however, administration of CMF prior to tamoxifen had a statistically significant advantage (trial IX). Compared with tamoxifen alone, the relative risk of relapse was reduced by 48% (P<0.01). In trial IX, the test for interaction comparing the effect of CMF followed by tamoxifen versus tamoxifen alone between the ER-negative and ER-positive cohorts was statistically significant (P=0.007).

In exploratory, hypothesis-generating analyses, trends in treatment effects were examined across four ER subgroups (absent, low, moderate and high), as shown in Figure 4. Within the ER-absent subgroup in trial VII, delayed CMF appeared detrimental compared with tamoxifen alone (RR 1.74; P=0.17), although the difference was not statistically significant. The test for interaction comparing the effect of delayed CMF + tamoxifen versus tamoxifen alone between ER-absent and ER-present cohorts (all remaining cases) yielded P=0.017, but cautious interpretation is required due to multiple comparisons. The effect of early concurrent CMF (node-positive disease) appeared to be greatest for tumors expressing intermediate levels of ER (ER-low or ER-moderate), but interaction tests were not significant. In trial IX (node-negative disease), both the ER-absent and ER-low cohorts (based on ligand-binding assay performed at each participating center) had substantial benefit from CMF administered prior to tamoxifen compared with tamoxifen alone, while no benefit was observed for cohorts with ER-moderate or ER-high assay values.



View larger version (28K):
[in this window]
[in a new window]
 
Figure 4. Disease-free survival risk ratios (RR) according to four estrogen receptor (ER) groups defined by quantitative ligand-binding assay. ER-absent: 0 fmol/mg cytosol protein; ER-low: 1–9 fmol/mg cytosol protein; ER-moderate: 10–49 fmol/mg cytosol protein; ER-high: ≥50 fmol/mg cytosol protein. CI, confidence interval; N+, node-positive; N–, node-negative; Del., delayed; Tam, tamoxifen; CMF, cyclophosphamide, methotrexate, 5-fluorouracil.

 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
The current analysis indicates that the effect of timing of chemotherapy varies according to the endocrine responsiveness of the disease for postmenopausal women with node-negative and node-positive breast cancer. In fact, we observed differences in the magnitude of chemotherapy effect when given with tamoxifen (either concurrent or sequential) compared with tamoxifen alone. Postmenopausal patients with endocrine responsive tumors (ER-moderate or ER-high) and node-positive disease obtained substantial benefit from the combination of chemotherapy and tamoxifen regardless of when the concurrent chemotherapy was administered. STEPP plots (Figure 2) clearly show the pattern of differential effectiveness of the addition of chemotherapy for varying levels of ER, as they reveal the benefit of chemotherapy for intermediate values of ER.

The advantage of early chemotherapy administered together with tamoxifen has been observed by others, with anthracycline-containing chemotherapy. The National Surgical Adjuvant Breast and Bowel Project (NSABP) trial B-16 showed that a short duration of chemoendocrine therapy using doxorubicin and cyclophosphamide (four courses) combined with tamoxifen was superior to tamoxifen alone for patients aged 50 years or older with steroid hormone receptor-positive primaries [16Go]. Similarly, in IBCSG trial VII, we showed that adding three early courses of concurrent CMF chemotherapy to tamoxifen was beneficial compared with tamoxifen alone. Surprisingly, delayed chemotherapy provided a significant advantage in DFS if combined with tamoxifen among patients with ER-positive tumors. Although it is better to start tamoxifen after completing chemotherapy [7Go], the present study is important for patient care because it shows that patients who are started on tamoxifen immediately after surgery can benefit from CMF administered during tamoxifen several months after surgery.

Other trials with longer duration chemotherapy regimens showed a significant advantage for chemotherapy given with tamoxifen as compared with tamoxifen alone [17Go–19Go]. The beneficial effect observed might represent a simple additive effect of two treatment modalities or the results of an optimal interference with micrometastases according to a sophisticated pattern of tumor growth and possible tumor cell dormancy [20Go]. In 1984, Speer et al. [21Go, 22Go] described a stochastic numerical model of breast cancer growth simulating clinical data. The assumption was that breast tumor growth does not follow a Gompertzian curve, but rather a pattern of growth characterized by spurts of fast proliferation followed by periods of slower or no progression. Since cancer cells would be more likely to respond to chemotherapy during phases of fast growth, and to be resistant during periods of quiescence, the delayed timing of cytotoxics administration might prove beneficial.

A different pattern of effect for chemoendocrine therapy was observed among patients with node-negative disease in trial IX, for whom the treatment modalities were given sequentially. As previously published [9Go], postmenopausal patients with endocrine responsive (ER-moderate or ER-high), node-negative disease did not benefit from CMF prior to tamoxifen. A recent re-analysis of NSABP B-20 focused on the ER-positive, node-negative, postmenopausal (instead of age 50 years or older) cohort and found a similar lack of chemotherapy effect when added to tamoxifen [23Go]. Thus, for postmenopausal patients with endocrine-responsive, node-negative breast cancer, a prolonged endocrine therapy with tamoxifen might be considered to be the standard treatment [4Go]. The different results observed in the node-negative and node-positive cohorts might be related to the higher risk of relapse for the latter, which increases the chance of improving outcome with an effective treatment. An alternative hypothesis is that node-positive disease is more heterogeneous so that intermediate levels of steroid hormone receptor expression do not rule out the presence of chemosensitive clones. It is also possible that a detrimental effect of the delay in commencement of tamoxifen in this endocrine-responsive group was just counterbalanced by a beneficial effect of the CMF.

Postmenopausal patients with node-negative disease and endocrine non-responsive disease (using, for simplicity, ER expression only: ER-low and ER-absent cohorts) were likely to obtain substantial benefit by adding CMF prior to tamoxifen (Figure 2). The observed benefit with three cycles of CMF in this patient population may depend on the specific timing of chemotherapy prior to endocrine therapy, which we suggest may have avoided detrimental interference of concurrent tamoxifen for the ER-negative cohort.

For node-positive, endocrine non-responsive tumors (ER-absent or ER-low disease), chemotherapy given together with tamoxifen made no substantial difference and may even have been associated with a detrimental effect compared with tamoxifen alone. In particular, delayed chemotherapy administered with tamoxifen showed a statistically non-significant detrimental effect in the population of patients with ER-absent tumors. One could consider two different hypothetical explanations. In tumors with no expression of steroid hormone receptors, tamoxifen is known to exert no beneficial effect on tumor growth and might itself be detrimental [24Go]. In tumors with some expression of steroid hormone receptors, tamoxifen might have some beneficial effect [25Go] and any interaction with chemotherapy might be more complex.

A detrimental effect of concurrent cytotoxics and tamoxifen was also observed by the NSABP investigators in trial B-09. The combination of melphalan, fluorouracil (MF) and tamoxifen compared with MF alone was associated with a shorter DFS and OS in the subpopulation of patients < 49 years old with ER-negative, PgR-negative primary tumors [6Go]. It has also been shown in in vitro experiments that the cytocidal effects of several alkylating agents, as well as of fluorouracil, are inhibited in the presence of tamoxifen [5Go].

The current evaluation also distinguishes the immediate, concomitant association of tamoxifen and CMF from the delayed administration of chemotherapy to a patient who already started tamoxifen several months earlier. Our study indicates that delayed chemotherapy should not be added either in patients with ER-absent tumors, or in patients with low expression of ER already receiving tamoxifen [25Go]. Conversely, early concomitant chemotherapy might provide a benefit in the subset of patients with ER-low tumors. These results are in line with previous data, which indicate possible different chemo-responsiveness [26Go] and endocrine responsiveness [27Go] in patients with some, albeit low ER and/or PgR expression compared with patients with no expression of steroid hormone receptors. The reason the effects of combined tamoxifen and concurrent CMF differ according to the timing of added chemotherapy might be due to the different effect of tamoxifen on target cells after short- versus long-lasting exposure [28Go, 29Go].

Despite the observed interaction of chemoendocrine therapy effects according to the degree of ER expression, and the biological explanations for this effect, there is a potential for bias due to the retrospective nature of these evaluations. We view these results as hypothesis-generating and hope they will stimulate future exploration of treatment effects for the cohort of patients with tumors having no expression of ER (ER-absent).

The current study is unique for two reasons. First, the question of chemoendocrine therapy compared with endocrine therapy alone was studied for postmenopausal patients with low or no hormone receptors expression in the tumors. Because this subgroup analysis according to ER category was defined prospectively and used as a stratification factor for the randomization (i.e. ER not expressed and ER 1–9 fmol/mg cytosol protein), it provides relevant and reliable information on the chemoendocrine combination therapy for the subgroup of patients with ER-low tumors. These tumors are assumed to retain some responsiveness to endocrine agents [25Go] and are thus candidates for sequential treatment while concomitant use should be avoided. Secondly, the benefit observed may be ascribed to the rigorous use of the classical CMF regimen, which, even if used for only three courses, provided a significant improvement in outcome [30Go]. Others, who did not observe similar results with CMF-like regimens of longer duration, might have missed the positive effect on outcome associated with classical CMF [31Go].

The question of timing of chemoendocrine therapy was addressed in three studies recently presented on the administration of anthracycline-containing chemotherapy either concurrent or sequential with tamoxifen [7Go, 32Go, 33Go]. The results of the two larger studies favored sequential administration [7Go, 32Go]. These trials were not designed to evaluate the effect of chemoendocrine therapy separately according to the degree of presumed endocrine-responsiveness of the tumor. Furthermore, no other trial except IBCSG trial VII has addressed the question of delayed timing of chemotherapy after starting tamoxifen.

In conclusion, our results indicate that the effect of chemotherapy administered with tamoxifen is largely dependent on the endocrine-responsiveness of the tumor. The positive effect of three different timings of chemotherapy (early, delayed and both) on outcome for patients with ER-positive, node-positive disease, sustain a role for chemotherapy even several months after diagnosis in this patient population. It is therefore reasonable to propose starting chemotherapy if the patient presents having already started tamoxifen. If the patient presents without already having starting tamoxifen, there is evidence that the chemotherapy should be completed before commencing tamoxifen. Studies of effects of delayed chemotherapy for patients with endocrine-responsive disease and high risk of relapse should be integrated into tailored treatment trials using novel endocrine agents.


View this table:
[in this window]
[in a new window]
 
 

    Appendix
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
International Breast Cancer Study Group trials VII and IX participants and authorsGo


    Acknowledgements
 
We thank the patients, physicians, nurses and data managers who participate in the International Breast Cancer Study Group trials. We gratefully acknowledge the initial support provided by the Ludwig Institute for Cancer Research and the Cancer League of Ticino, and the continuing support for central coordination, data management and statistics provided by the Swedish Cancer Society, The Cancer Council Australia, Australian New Zealand Breast Cancer Trials Group, the Frontier Science and Technology Research Foundation, the Swiss Group for Clinical Cancer Research, the Swiss Cancer League and the United States National Cancer Institute (CA-75362). We also acknowledge support for the Cape Town participants from the Cancer Association of South Africa and for the St Gallen participants from the Foundation for Clinical Research of Eastern Switzerland. In addition, we acknowledge IBCSG participating institutions and investigators for trials VII and IX.


    Notes
 
{dagger}IBCSG trial VII and IX participants and authors are listed in the Appendix. Back

Received for publication August 25, 2004. Revision received January 3, 2005. Accepted for publication January 7, 2005.


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Appendix
 References
 
1. Early Breast Cancer Trialists' Collaborative Group. Polychemotherapy for early breast cancer: an overview of the randomised trials. Lancet 1998; 352: 930–942.[CrossRef][ISI][Medline]

2. Coates AS, Goldhirsch A, Gelber RD, for the International Breast Cancer Study Group. Overhauling the breast cancer overview: are subsets subversive? Lancet Oncol 2002; 3: 525–526.[CrossRef][ISI][Medline]

3. Goldhirsch A, Glick JH, Gelber RD et al. Meeting highlights: International Consensus Panel on the Treatment of Primary Breast cancer. Seventh International Conference on Adjuvant Therapy of Primary Breast Cancer. J Clin Oncol 2001; 19: 3817–3827.[Free Full Text]

4. Goldhirsch A, Wood WC, Gelber RD et al. Meeting highlights: updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol 2003; 21: 3357–3365.[Abstract/Free Full Text]

5. Osborne CK. Interactions of tamoxifen with cytotoxic chemotherapy for breast cancer. In Jordan VC (ed.): Long Term Tamoxifen For Breast Cancer. Madison, WI: University of Wisconsin Press 1994; 181–198.

6. Fisher B, Redmond C, Brown A et al. Adjuvant chemotherapy with and without tamoxifen: five-year results from the National Surgical Adjuvant Breast and Bowel Project Trial. J Clin Oncol 1986; 4: 459–471.[Abstract]

7. Albain KS. Adjuvant chemo-endocrine therapy for breast cancer: combined or sequential? Breast 2003; 12: 12.

8. International Breast Cancer Study Group. Effectiveness of adjuvant chemotherapy in combination with tamoxifen for node-positive postmenopausal breast cancer patients. J Clin Oncol 1997; 15: 1385–1394.[Abstract]

9. International Breast Cancer Study Group. Endocrine responsiveness and tailoring adjuvant therapy for postmenopausal lymph node-negative breast cancer: a randomized trial. J Natl Cancer Inst 2002; 94: 1054–1065.[Abstract/Free Full Text]

10. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481.[ISI]

11. Greenwood M. The Natural Duration of Cancer. London: HM Stationery Office 1926.

12. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–170.[Medline]

13. Cox DR. Regression models and life-tables (with discussion). J R Stat Soc 1972; 34: 187–220.[ISI]

14. Bonetti M, Gelber RD. A graphical method to assess treatment-covariate interactions using the Cox model on subsets of the data. Stat Med 2000; 19: 2595–2609.[CrossRef][ISI][Medline]

15. Bonetti M, Gelber RD. Patterns of treatment effects in subsets of patients in clinical trials. Biostatistics 2004; 5: 465–481.[Abstract/Free Full Text]

16. Fisher B, Redmond C, Legault-Poisson S et al. Postoperative chemotherapy and tamoxifen compared with tamoxifen alone in the treatment of positive-node breast cancer patients aged 50 years and older with tumors responsive to tamoxifen: Results from the National Surgical Adjuvant Breast and Bowel Project B-16. J Clin Oncol 1990; 8: 1005–1018.[Abstract]

17. Wils JA, Bliss JM, Marty M et al. Epirubicin plus tamoxifen versus tamoxifen alone in node-positive postmenopausal patients with breast cancer: a randomized trial of the International Collaborative Cancer Group. J Clin Oncol 1999; 17: 1988–1998.[Abstract/Free Full Text]

18. Goldhirsch A, Gelber R. Adjuvant treatment for early breast cancer: the Ludwig breast cancer studies. J Natl Cancer Inst Monogr 1986; 1: 55–70.

19. Fisher B, Dignam J, Wolmark N et al. Tamoxifen and chemotherapy for lymph node-negative, estrogen receptor-positive breast cancer. J Natl Cancer Inst 1997; 89: 1673–1682.[Abstract/Free Full Text]

20. Demicheli R, Retsky MW, Swartzendruber DE, Bonadonna G. Proposal for a new model of breast cancer metastatic development. Ann Oncol 1997; l8: 1075–1080.[CrossRef]

21. Speer JF, Petrosky VE, Retsky MW, Wardwell RH. A stochastic numerical model of breast cancer growth that simulates clinical data. Cancer Res 1984; 44: 4124–4130.[Abstract]

22. Gompertz B. On the nature of the function expressive of the law of human mortality and on a new mode of determining the value of life contingencies. Phil Trans R Soc Lond 1825; 115: 513–585.

23. Fisher B, Jeong J-H, Bryant J et al. Findings from two decades of National Surgical Adjuvant Breast and Bowel Project clinical trials involving breast cancer patients with negative axillary nodes. Breast Cancer Res Treat 2002; 76(Suppl. 1): 16.

24. IBIS Investigators. First results from the International Breast Cancer Intervention Study (IBIS-I): a randomised prevention trial. Lancet 2002; 360: 817–824.[CrossRef][ISI][Medline]

25. Harvey JM, Clark GM, Osborne CK, Allred DC. Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 1999; 17: 1474–1481.[Abstract/Free Full Text]

26. Colleoni M, Minchella I, Mazzarol G et al. Response to primary chemotherapy in breast cancer patients with tumors not expressing estrogen and progesterone receptors. Ann Oncol 2000; 11: 1057–1059.[Abstract]

27. Colleoni M, Gelber S, Coates AS et al. Influence of endocrine-related factors on response to perioperative chemotherapy for patients with node-negative breast cancer. J Clin Oncol 2001; 19: 4141–4149.[Abstract/Free Full Text]

28. Martin EA, Carthew P, White IN et al. Investigation of the formation and accumulation of liver DNA adducts in mice chronically exposed to tamoxifen. Carcinogenesis 1997; 18: 2209–2215.[Abstract]

29. Johnston SR, Haynes BP, Sacks NP et al. Effect of oestrogen receptor status and time on the intra-tumoural accumulation of tamoxifen and N-desmethyltamoxifen following short-term therapy in human primary breast cancer. Breast Cancer Res Treat 1993; 28: 241–250.[CrossRef][ISI][Medline]

30. Colleoni M, Litman HJ, Castiglione-Gertsch M et al. Duration of adjuvant chemotherapy for breast cancer: a joint analysis of two randomised trials investigating three versus six courses of CMF. Br J Cancer 2002; 86: 1705–1714.[CrossRef][ISI][Medline]

31. Goldhirsch A, Coates AS, Colleoni M et al. Adjuvant chemoendocrine therapy in postmenopausal breast cancer: cyclophosphamide, methotrexate, and fluorouracil dose and schedule may make a difference. International Breast Cancer Study Group. J Clin Oncol 1998; 16: 1358–1362.[Abstract]

32. Pico C, Martin M, Jara C. Epirubicin-cyclophosphamide adjuvant chemotherapy plus tamoxifen administered concurrently versus sequentially: randomized phase III trial in postmenopausal node-positive breast cancer patients. A GEICAM 9401 study. Ann Oncol 2004; 15: 79–87.[Abstract/Free Full Text]

33. Sertoli MR, Pronzato P, Venturini M et al. A randomized study of concurrent versus sequential adjuvant chemotherapy and tamoxifen in stage II breast cancer. Proc Am Soc Clin Oncol 2002; 21: 182.