Tamoxifen Versus Control After Adjuvant, Risk-Adapted Chemotherapy in Postmenopausal, Receptor-Negative Patients With Breast Cancer: A Rando
http://www.100md.com
《临床肿瘤学》
the Universit?ts-Frauenklinik, Frankfurt
Medizinische Biometrie und Statistik, Universit?tsklinikum Freiburg
Universit?ts-Frauenklinik, Kiel
Rot-Kreuz-Krankenhaus, München
Universit?ts-Frauenklinik, Heidelberg
Universit?ts-Frauenklinik, Marburg
Universit?tsklinikum, Magdeburg
Senologisches Zentrum, Kassel
Prosper Hospital, Recklinghausen, Germany
ABSTRACT
PURPOSE: To investigate the effect of adjuvant sequential tamoxifen after chemotherapy in postmenopausal patients with hormone receptor–negative breast cancer.
METHODS: Patients were randomly assigned to oral tamoxifen (30 mg daily for 5 years; n = 421) or no additional treatment (n = 408) after risk-adapted polychemotherapy consisting of three 28-day cycles of CMF (cyclophosphamide, 500 mg/m2, methotrexate, 40 mg/m2, and fluorouracil, 600 mg/m2) in patients with negative or one to three positive lymph nodes and four 21-day cycles of epirubicin 90 mg/m2, cyclophosphamide 600 mg/m2 followed by three cycles of CMF in patients with four to nine positive lymph nodes.
RESULTS: Thirty-six percent of the patients included were older than 60 years, 63% were node negative, 13% had four to nine positive nodes, 55% had tumor grade 3, and 41% received breast-preserving surgery. At 5.3 years’ median follow-up, the first event of failure (recurrence, secondary tumor, or death) had occurred in 123 patients in the tamoxifen group and 107 patients of the control group. Event-free survival rates after 5 years were 70.3% (95% CI, 65.5% to 75.0%) and 72.8% (95% CI, 68.2% to 77.5%) for the tamoxifen and control groups, respectively. The estimated hazard ratio of tamoxifen versus control was 1.13 (95% CI, 0.87 to 1.48; P = .34), which gives no indication of an additional benefit of tamoxifen in these patients.
CONCLUSION: This study contributes substantially to finalization of the presently emerging evidence that tamoxifen does not benefit women with receptor-negative breast cancer after chemotherapy.
INTRODUCTION
In the early 1990s, overviews by the Early Breast Cancer Trialists’ Collaborative Group had demonstrated the benefit of adjuvant systemic therapy in women with node-negative and node-positive breast cancer.1 The role of chemotherapy versus endocrine therapy or chemoendocrine therapy remained to be determined in women with an intermediate or high risk of tumor recurrence on the basis of risk factors such as menopausal status, node status, and sensitivity to steroid receptors.2 Although the benefit of using tamoxifen for receptor-positive breast cancer had been firmly established, uncertainty remained regarding receptor-negative patients.1
In 1993 the German Adjuvant Breast Cancer Group (GABG) started a series of randomized trials for various risk groups. Trial IV D-93 included postmenopausal patients with primary breast cancer and negative estrogen and progesterone receptors. Patients were either node negative or node positive and had one to nine positive nodes. The study investigates the effect of tamoxifen given sequentially, after risk-adapted chemotherapy, at a daily dose of 30 mg for 5 years according to consensus practice in Germany at that time versus no additional treatment. The baseline chemotherapy was cyclophosphamide, methotrexate, and fluorouracil (CMF), in agreement with International Consensus Panel recommendations.2 However, because the preliminary and final results of several trials3-5 indicated no benefit of the usual six over three cycles, node-negative patients and patients with one to three positive nodes received a nonstandard regimen of three cycles of CMF every 4 weeks. To account for their higher risk of recurrence, patients with four to nine positive nodes were given four cycles of an anthracycline-based regimen2 (epirubicin and cyclophosphamide every 3 weeks [EC]) followed by three cycles of CMF. With a median follow-up of approximately 5 years, this report presents the initial findings from the GABG-IV D-93 trial.
METHODS
Patients
Postmenopausal women up to 70 years of age with a histologic diagnosis of invasive breast cancer and zero to nine positive axillary lymph nodes were eligible if they complied with the following criteria: negative hormone receptor status; stage pT1-3, N0 to N3, M0; no prior surgical, systemic, or radiation therapy against breast cancer; Karnofsky index score 60 or World Health Organization performance scale score of 0 to 2; hormone-replacement therapy (HRT) to be stopped if applicable; suitability for follow-up; and written informed consent. Major exclusion criteria were distant metastases; incomplete resection of the tumor; resection of less than 10 axillary lymph nodes; simultaneous contralateral breast cancer; previous second malignancy except basal cell carcinoma of the skin or carcinoma-in-situ of the cervix uteri; insufficient organ function or significant comorbidities; and any T4 tumor. For eligibility, chemotherapy had to start within 28 days from definitive primary surgery, and randomization was possible in the interval between surgery and up to 28 days after start of the third cycle of CMF chemotherapy.
Patients were recruited between February 1993 and December 2000 from 63 centers all over Germany. Centers included university hospitals as well as regional and city hospitals (see Appendix). The trial protocol was approved by all ethical committees.
Determination of Postmenopausal Status and Negative Hormone Receptor Status
For patients without HRT before enrollment, premenopausal status was excluded either on the basis of reported menstrual periods within the last 6 months or according to follicle-stimulating hormone (< 20 IU/L) and luteinizing hormone (> 50 pg/mL) serum levels. Serum levels were assessed 1 week after discontinuation of HRT if applicable. Estrogen and progesterone hormone receptor status were determined at local laboratories either biochemically by a dextran-coated charcoal assay (positive: 20 fmol/mg according to consensus practice in Germany) and/or by scores based on an immunohistochemistry assay (positive: score 2) and/or by scores from immunologic evaluation6 (positive: score 2). Biochemistry was preferred if it was available. If the institution specified a different cut point for immunohistochemical or immunologic assays, this cut point was used. Negative hormone receptor status was defined as negative estrogen and progesterone receptors.
Study Treatments
Patients were randomly assigned to either oral tamoxifen (Nolvadex 30 mg once a day for 5 years) or no additional treatment after a risk-adapted adjuvant chemotherapy consisting of three cycles of CMF (cyclophosphamide 500 mg/m2, methotrexate 40 mg/m2, and 5-fluorouracil 600 mg/m2 given intravenously on days 1 and 8, every 4 weeks) in patients with zero to three positive lymph nodes and four cycles of EC (epirubicin 90 mg/m2 and cyclophosphamide 600 mg/m2 given intravenously every 3 weeks) followed by three cycles of CMF in patients with four to nine positive lymph nodes. The interval between the last course of EC and the first course of CMF was 3 weeks. Chemotherapy had to start within 28 days from definitive primary surgery. Optional radiotherapy was applied according to local policies.
Evaluation Criteria
Follow-up examinations were scheduled every 3 months for the first 2 years, every 6 months for the following 3 years, and annually thereafter. The primary end point was event-free survival (EFS), defined as the time from definitive primary surgery to the first event of failure (locoregional recurrence, metastases, second primaries including contralateral breast cancer, or death). The first event of failure was classified as isolated locoregional recurrence if locoregional recurrence occurred at least 4 weeks before an event at a distant site. The secondary end point was overall survival (OS), which was defined as the time from definitive primary surgery to death resulting from any cause. Results on quality of life will be reported elsewhere.
Randomization and Quality Control
Patients with confirmed eligibility were randomly assigned by phone or fax at the Department of Medical Biometry and Statistics (University Hospital Freiburg). They were stratified by clinical center and nodal status (zero to three or four to nine positive lymph nodes). Within each stratum, block randomization with randomly varying block size and a 1:1 treatment ratio was performed based on computer-generated lists. For ease of organization, no blinding was done. Starting in December 1997, source-data verification was performed by an external clinical research organization (SKM, Wiesbaden, Germany). The data were subjected to computerized plausibility checks.
Statistical Methods
On the basis of preliminary data of patients selected from previous studies,7 it was anticipated that EFS would be 60% at 5 years in the control group treated with risk-adaptive chemotherapy only. Thus, 190 events would be required to detect a hazard ratio of 0.67 for tamoxifen versus control (ie, an improvement to approximately 71% in the tamoxifen group), with a power of 80% using a log-rank test at the level = 5%.8 With 4 years of planned recruitment and 2 years additional follow-up, at least 700 patients would have to be included. Inclusion of patients was continued beyond this minimum target number to ensure that GABG-IV trials A-93, B-93, and E-939,10 would all be closed at approximately the same time.
EFS and OS rates were estimated by Kaplan-Meier curves.12 Event-free patients were censored at the last reported visit. Median follow-up was based on the estimated censoring distribution,13 and the percentage of complete follow-up was calculated.14 The treatment effect on EFS was estimated as the hazard ratio in a Cox model15 with a 95% CI. P values were based on Wald tests.16 An adjusted analysis of treatment and prognostic factors was performed in a multiple Cox model including treatment, tumor size, tumor grade, number of positive lymph nodes, and type of surgery. This analysis was carried out in complete cases (ie, in patients with complete data on size, grade, nodes, and surgery).
To investigate possible differential treatment effects on EFS (interactions), effects were estimated for each of two subgroups according to tumor size, tumor grade, and nodal status, respectively, in three Cox models with simultaneous adjustment for the prognostic factors mentioned above. The effect of nodal status was studied in cases with zero or one to three positive nodes only to avoid confounding with risk-adapted chemotherapy (see Study Treatments). Interactions and treatment effects within subgroups were tested by Wald tests. Because of multiple testing, a significance level of 1% was used for these tests.
During the study, two formal interim analyses of the treatment effect on EFS were conducted with a significance level of 0.5% each when 29 and 118 events had been observed, respectively. To maintain the overall significance level of = 5%, the present, final analysis with 230 events was carried out at the nominal significance level of ' = 4.40%, which was calculated according to the -spending approach.17 Accordingly, 95.60% CIs are reported for the hazard ratios between treatment groups in the unadjusted and adjusted analyses of EFS.
Data were processed and evaluated with SAS (SAS Institute, Cary, NC).18 Following the intention-to-treat principle, ineligible patients were not excluded, and treatment was analyzed as randomized. Analyses were carried out in accordance with a prespecified statistical analysis plan. All tests, P values, and CIs were two sided.
RESULTS
Recruitment and Patient Characteristics
Between February 1993 and December 2000, a total of 829 patients were included at 63 centers in Germany: 408 were randomly assigned to the control arm and 421 to tamoxifen. Thirty-six percent of the patients were older than 60 years, 63% were node negative, 13% had four to nine positive nodes, 55% had tumor grade 3, and 41% received breast-preserving surgery. Baseline characteristics were well balanced between treatments (Table 1).
The eligibility criteria turned out to be violated in 97 patients, mostly because of positive hormone receptor and delayed start of chemotherapy. Fifty-one patients were from the control group (receptor positive, n = 26; enrollment despite start of therapy > 28 days from surgery, n = 9; randomization > 28 days after start of third cycle of CMF chemotherapy, n = 5; previous second malignancy, n = 4; incomplete resection of primary breast tumor, n = 1; resection of < 10 axillary lymph nodes, n = 2; premenopausal, n = 1; > 70 years of age, n = 2; perioperative tamoxifen treatment, n = 1; randomly assigned twice by error, n = 2; randomly assigned erroneously without consent, n = 1; three patients met > one of these criteria). Forty-six patients were in the arm allocated to tamoxifen (receptor positive, n = 19; enrollment despite start of therapy > 28 days from surgery, n = 11; assignment > 28 days after start of the third cycle of CMF chemotherapy, n = 2; simultaneous contralateral breast cancer, n = 1; pT4 breast cancer, n = 4; metastases at assignment, n = 1; incomplete resection of primary breast tumor, n = 1; resection of < 10 axillary lymph nodes, n = 3; premenopausal, n = 1; > 70 years of age, n = 1; randomly assigned twice by error, n = 2).
In one of the four patients randomly assigned twice by error, the result of the second assignment differed from the first; the result of the second assignment was ignored in the analyses. Ineligible patients were included in the analyses; however, all data on the patient randomly assigned erroneously without consent are missing.
Compliance
The two arms were comparable with respect to baseline chemotherapy. All planned cycles were completed in 89.7% and 92.6% of the patients in the control and tamoxifen arms, respectively (Fig 1). Overall, chemotherapy was incomplete in 5.8% and not started in 2.1% (application unknown in 0.8%).
In the tamoxifen group, 359 patients started endocrine therapy as assigned. Among these, the rate of patients receiving 5 years of treatment was 55.9% (proportion derived treating patients with incomplete data on duration of tamoxifen as censored,12 n = 95). Forty patients in the tamoxifen group did not start the treatment; 16.2% of those who did were delayed (> 34 days after start of the last cycle of chemotherapy).
Follow-Up and Observed Events
The data cutoff for this analysis was the end of June 2003, leading to a median follow-up of 5.3 years. Completeness of follow-up, measured as the ratio of observed and potential follow-up information available at the data cutoff,14 was 88%. Follow-up was similar in both treatment groups (Fig 1). With regard to EFS, 230 events have been observed thus far. Table 2 lists the distribution of the different events of failure. The first event of failure (tamoxifen v control) was an isolated local recurrence (26 v 33), a distant failure (94 v 69), and death without recurrence (3 v 5). Among the distant failures, 17 were new nonbreast primaries (four endometrial cancers) in the tamoxifen group and five in the control group (no endometrial cancer). With regard to OS, 152 deaths (80 v 72) of any cause have been observed up to the time of data cut-off.
EFS
EFS rates by treatment are displayed in Figure 2. Five-year EFS rates are estimated as 70.3% (95% CI, 65.5% to 75.0%) and 72.8% (95% CI, 68.2% to 77.5%) for the tamoxifen and control groups, respectively. The unadjusted hazard ratio of tamoxifen versus control was 1.13 (95% CI, 0.87 to 1.48; calculated as 95.60% CI to account for two interim analyses), which gives no indication of a benefit of additional tamoxifen in these patients (P = .34). This result was confirmed in the analysis adjusted for tumor size, tumor grade, type of surgery, and number of positive nodes (Table 3). The adjusted hazard ratio of tamoxifen versus control is 1.16 (95% CI, 0.88 to 1.51; calculated as 95.60% CI to account for interim analyses; P = .28; complete cases, n = 811; 228 events).
Possible differences in treatment effects were investigated in subgroups defined by tumor grade, tumor size, and nodal status with a significance level of 1% for the tests on interaction between treatment and the corresponding prognostic factor. To account for the fact that the risk-adapted chemotherapy was linked by study design to the number of positive lymph nodes (zero to three or four to nine), the predictive effect of lymph nodes was studied only in patients with zero versus one to three positive nodes. However, no significant interaction was found (data not shown).
OS
OS rates are displayed in Figure 3. Five-year OS rates are estimated as 82.2% and 81.6% in the tamoxifen and control groups, respectively. With 152 (18%) deaths observed in total up to the time of data cut-off, it is too early for a definitive analysis of OS.
Tolerability and Adverse Events
Tamoxifen was discontinued for medical reasons other than recurrence or death in 41 patients. Toxicity data are based on reports of serious adverse events before recurrence. Life-threatening adverse events were reported in eight patients of the tamoxifen group. Two events occurred during EC chemotherapy (hematologic toxicity, cardiomyopathy) and six occurred during tamoxifen (five thromboembolic events, one mastitis). In contrast, no life-threatening adverse events were reported in the control group. In total, two patients died while on adjuvant treatment, both from pulmonary embolism. The first death, judged as treatment related, occurred in the control group during EC-CMF, the other, judged as unrelated to treatment, occurred during adjuvant tamoxifen.
DISCUSSION
This report provides information about the efficacy of 5 years of adjuvant tamoxifen versus control after risk-adapted chemotherapy in a population of postmenopausal, estrogen and progesterone receptor–negative breast cancer patients. Patient characteristics are well balanced between treatment groups (Table 1) and so are the length of median follow-up and completeness of follow-up (Fig 1). After approximately 5 years of median follow-up and 230 observed events, the study is sufficiently mature to give reliable results. We observed a moderate, statistically nonsignificant detrimental effect of tamoxifen on EFS.
Our findings agree with the general evidence on the efficacy of tamoxifen in receptor-negative breast cancer, which has been accumulating since the present study was initiated. The recently published update of the overview by the Early Breast Cancer Trialists’ Collaborative Group19 shows a small but nonsignificant disadvantage for approximately 5 years of tamoxifen versus none with respect to recurrence and mortality rates, both in the subgroup of estrogen receptor-poor women (see Fig 7 in the Early Breast Cancer Trialists’ Collaborative Group’s study19) and the subgroup of estrogen receptor-poor, progesterone receptor-poor women (web appendix 1, annex Fig 7 in the Early Breast Cancer Trialists’ Collaborative Group’s study19). It included three studies20-22 that, comparable to our design, reported results on the efficacy of additional adjuvant tamoxifen combined with polychemotherapy in receptor-negative, older patients (postmenopausal or age 50 years), and neither showed a benefit for tamoxifen. There remains some inconsistency in the overview with regard to the duration of tamoxifen in that some benefit for administration for one to two years was observed in estrogen receptor-poor patients. A possible explanation could be that false-negative estrogen receptor measurements were more pronounced in earlier trials that investigated the shorter duration of tamoxifen.
Regarding the internal validity of GABG-IV D-93 study results, we have to concede that there were many ineligible patients. The most noticeable is the high number of receptor-positive patients in each group. A cutoff point of 20 fmol/mg cytosol protein for the biochemical assay was generally recommended for application in Germany in the early phase of the trial, whereas immunohistochemical techniques became standard toward the end of recruitment. Therefore, even some of the eligible patients included at the beginning would be classified as false negative by current standards. Furthermore, in this unblinded trial, 40 patients (9.5%) did not start the drug, and 69 (19.2%) of those who did discontinued it prematurely for reasons other than failure or death (Fig 1). Reduced compliance is not uncommon in tamoxifen trials. For example, either failure to start or premature discontinuation were observed in approximately 20% of women allocated to approximately 5 years of tamoxifen in the three largest trials included in the 1998 overview.23,24-26 However, the benefit of shorter durations of tamoxifen in estrogen receptor-positive patients is well established,19,23 and a minor benefit was noted even in the estrogen receptor-poor patients.19 Thus, by its failure to show a benefit for tamoxifen despite impaired compliance and inclusion of some receptor-positive patients our study provides additional important data on the role of tamoxifen in receptor-negative breast cancer.
In conclusion, as Swain27 stated with reference to a preliminary version of the updated overview, "more data are needed to definitely state that tamoxifen does not benefit women with estrogen receptor–negative disease." With its large cohort of postmenopausal, receptor-negative patients and mature follow-up, this study makes a substantial contribution toward finalizing conclusive evidence on this matter.
Appendix
The acknowledgment and appendix are included in the full-text version of this article, available online at www.JCO.org. They are not included in the PDF (via Adober Acrobat Readerr) version.
The participating patients and investigators were recruited for the trial at the following sites: Kreiskrankenhaus Albstadt (G. Geier); Universit?tsklinikum Charité, Chirugie, Berlin (K.-J. Winzer); Universit?tsklinikum Charité, Gyn?kologie, Berlin (W. Lichtenegger); Vivantes Humboldt Klinikum, Berlin (L. Schneppel); Waldfriede Krankenhaus, Berlin (C. Kempter); St Josefshospital, Cloppenburg (I. Schulz-Im Busch); Klinikum Deggendorf (D. Augustin); Kreiskrankenhaus Eggenfelden (W. Siebert); J. W. Goethe Universit?t Frankfurt (M. Kaufmann); Universit?tsklinikum Freiburg (W. Kleine); St?dtisches Krankenhaus Friedrichshafen (G. de Gregorio); Klinik am Eichert, G?ppingen (A. Hettenbach); Universit?tsklinikum, Chirurgische Onkologie, G?ttingen (T. Liersch); Universit?tsklinikum Halle (H. K?lbl); Albertinen-Krankenhaus Hamburg (M. H. Carstensen); Gyn?kologische Gemeinschaftspraxis, Hamburg (E. Goepel); Universit?tsklinikum Hamburg (F. J?nicke); Klinikum Stadt Hanau (H. H. Zippel); Henriettenstiftung, Hannover (J. Hilfrich); Krankenhaus Nordstadt, Hannover (P. Hohlweg-Majert); St Josefs-Krankenhaus, Heidelberg (E. Krystek); Universit?tsklinikum Heidelberg (G. Bastert); Klinikum Hoyerswerda (K. Lürmann); Stadtkrankenhaus Idar-Oberstein (R. Gros); Universit?tsklinikum, Klinik für Innere Medizin II, Jena (K. H?ffken); Universit?tsklinikum, Frauenklinik, Jena (A. Schneider); Diakonissenkrankenhaus Karlsruhe (M. Zedelius); St?dtisches Klinikum Karlsruhe (G. Kaltenecker); St Vincentius Krankenhaus, Karlsruhe (H. G. Meerpohl); St?dtisches Klinikum Kassel (T. Dimpfl); Universit?tsklinikum Kiel (N. Maass); Frankenwaldklinik Kronach (P. Heinkele); Kreiskrankenhaus Leonberg (M. Kuglin); St Bonifatius Hospital Lingen (M. Johnscher); Universit?tsklinikum Lübeck (C. Strunck); AG Mammakarzinom des Tumorzentrums Magdeburg/Sachsen-Anhalt (G. Gademann); Universit?tsklinikum Mainz (P. Knapstein); St Vinzenz u. St Elisabeth-Hospital Mainz (C. Lei?ner); Klinikum der Stadt Mannheim (C. F?rster); Universit?tsklinikum Marburg (U. S. Albert); Klinikum Rechts der Isar, München (M. Kiechle); Frauenklinik vom Roten Kreuz München (G. Raab); Clemenshospital Münster (A. Neff); Universit?tsklinikum Münster (C. Jackisch); Staufenklinik, Mutlangen (A. Rodewald); Frauenklinik der St?dtischen Kliniken Offenbach a. M. (S. Jung); Marienhospital Osnabrück (K. Brunnert, M. Butterwegge); St Vincenz-Krankenhaus Paderborn (W. Meinerz); Krankenhaus St Trudbert, Pforzheim (R. Stiglmayer); Prosper Hospital Recklinghausen (K. Stahl); Kreiskrankenhaus Reutlingen (C. Schick); Mathias Spital, Rheine (G. Heywinkel); Frauenklinik Rheinfelden (H. Dieterich); Universit?t Rostock (T. Reimer); Stadtkrankenhaus Rüsselsheim (L. Heilmann); Martin-Luther-Krankenhaus Schleswig (H. Anger); Evangelisches Jung-Stilling Krankenhaus Siegen (M. Türker); Kreiskrankenhaus Stadthagen (J. Feltz-Sü?enbach); Universit?tsklinikum Tübingen (A. Bergmann); Universit?tsklinikum Ulm (R. Kreienberg); Ammerland Klinik GmbH Westerstede (M. Hippach); Stadtkrankenhaus Worms (T. Hitschold); and Universit?tsklinikum Würzburg (J. Dietl).
Authors’ Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed discription of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Acknowledgment
This study was funded by the Deutsche Krebshilfe and in part by Pharmacia, Germany, and AstraZenenca, Germany. We thank them for their support for the conduct of this trial. We also thank the participating patients and investigators.
NOTES
Supported by Grants from the Deutsche Krebshilfe, AstraZeneca, Germany, and Pharmacia, Germany.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Early Breast Cancer Trialists’ Collaborative Group: Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 339:1-15, 71-85, 1992
Glick JH, Gelber RD, Goldhirsch A, et al: Adjuvant therapy of primary breast cancer. 4th International Conference on Adjuvant Therapy of Primary Breast Cancer St. Gallen, Switzerland. Ann Oncol 3:801-807, 1992
Schumacher M, Bastert G, Bojar H, et al: Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients. German Breast Cancer Study Group. J Clin Oncol 12:2086-2093, 1994
International Breast Cancer Study Group: Duration and reintroduction of adjuvant chemotherapy for node-positive premenopausal breast cancer patients. J Clin Oncol 14:1885-1894, 1996
Maass H: German Adjuvant Breast Cancer Group: 3 versus 6 cycles of CMF in breast cancer patients with 1 to 9 positive nodes: Results of the German Adjuvant Breast Cancer Group (GABG) III trial. Proc Am Soc Clin Oncol 19:A-73, 2000 (abstr 283)
Leclercq G, Bojar H, Goussard J, et al: Abbott monoclonal enzyme immunoassay measurement of estrogen receptors in human breast cancer: A European multicenter study. Cancer Res 46:4233s-4236s, 1986
Schmoor C, Olschewski M, Sauerbrei W, et al: Long-term follow-up of patients in four prospective studies of the German Breast Cancer Study Group (GBSG): A summary of key results. Onkologie 25:143-150, 2002
Schoenfeld DA: Sample-size formula for the proportional-hazards regression model. Biometrics 39:499-503, 1983
von Minckwitz G, Graf E, Geberth M, et al: Goserelin versus CMF as adjuvant therapy for node-negative, hormone receptor-positive breast cancer in premenopausal patients. The GABG IV-A-93 trial. J Clin Oncol 22(14S):534, 2004 (abstr)
Kaufmann M, Graf E, Jonat W, et al: Goserelin versus control after adjuvant, risk-adapted chemotherapy in premenopausal patients with breast cancer: GABG-IV B-93. J Clin Oncol 22(14S):588, 2004 (abstr)
Eiermann W, Graf E, Raab G, et al: High-dose epirubicin vs. standard-dose epirubicin-cyclophosphamide followed by CMF in breast cancer patients with 10 or more positive lymph nodes: First results of a randomised trial of the German Adjuvant Breast Cancer Group (GABG). Proc Am Soc Clin Oncol 20:A-127, 2001 (abstr 27)
Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Schemper M, Smith TL: A note on quantifying follow-up in studies of failure time. Control Clin Trials 17:343-346, 1996
Clark TG, Altman DG, De Stavola BL: Quantification of the completeness of follow-up. Lancet 359:1309-1310, 2002
Cox DR: Regression models and life tables. J R Stat Soc B 34:187-220, 1972
Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. New York, NY, Wiley, 1980
Lan KKG, DeMets DL: Discrete sequential boundaries for clinical trials. Biometrika 70:659-663, 1983
SAS Institute Inc: SAS Language, Version 6. Cary, NC, 1990
Early Breast Cancer Trialists’ Collaborative Group (EBCTCG): Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet 365:1687-16717, 2005
Fisher B, Anderson S, Tan-Chiu E, et al: Tamoxifen and chemotherapy for axillary node-negative, estrogen receptor-negative breast cancer: Findings from National Surgical Adjuvant Breast and Bowel Project B-23. J Clin Oncol 19:931-942, 2001
Hutchins L, Green S, Ravdin P, et al: CMF versus CAF +/– tamoxifen in high-risk node-negative breast cancer patients and a natural history follow-up study in low-risk node-negative patients: Update of tamoxifen results. Breast Cancer Res Treat 57:25, 1999 (abstr 7)
Nomura Y, Shirouzu M, Takayama T: Direct comparisons of adjuvant endocrine therapy, chemotherapy, and chemoendocrine therapy for operable breast cancer patients stratified by estrogen receptor and menopausal status. Breast Cancer Res Treat 49:51-60, 1998
Early Breast Cancer Trialists’ Collaborative Group 1998: Tamoxifen for early breast cancer: An overview of the randomised trials. Lancet 351:1451-1467, 1998
Rutqvist LE, Cedermark B, Glas U, et al: Radiotherapy, chemotherapy, and tamoxifen as adjuncts to surgery in early breast cancer: A summary of the randomized trials. Int J Radiat Oncol Biol Phys 16:629-639, 1989
McDonald CC, Alexander FE, Whyte BW, et al: Cardiac and vascular morbidity in women receiving adjuvant tamoxifen for breast cancer in a randomised trial. The Scottish Cancer Trials Breast Group. BMJ 311:977-980, 1995
Fisher B, Constantino JP, Redmont CK, et al: Endometrial cancer in tamoxifen-treated breast cancer patients: Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14. J Natl Cancer Inst 86:527-537, 1994
Swain SM: Tamoxifen for patients with estrogen receptor-negative breast cancer. J Clin Oncol 19:93s-97s, 2001(Manfred Kaufmann, Erika G)
Medizinische Biometrie und Statistik, Universit?tsklinikum Freiburg
Universit?ts-Frauenklinik, Kiel
Rot-Kreuz-Krankenhaus, München
Universit?ts-Frauenklinik, Heidelberg
Universit?ts-Frauenklinik, Marburg
Universit?tsklinikum, Magdeburg
Senologisches Zentrum, Kassel
Prosper Hospital, Recklinghausen, Germany
ABSTRACT
PURPOSE: To investigate the effect of adjuvant sequential tamoxifen after chemotherapy in postmenopausal patients with hormone receptor–negative breast cancer.
METHODS: Patients were randomly assigned to oral tamoxifen (30 mg daily for 5 years; n = 421) or no additional treatment (n = 408) after risk-adapted polychemotherapy consisting of three 28-day cycles of CMF (cyclophosphamide, 500 mg/m2, methotrexate, 40 mg/m2, and fluorouracil, 600 mg/m2) in patients with negative or one to three positive lymph nodes and four 21-day cycles of epirubicin 90 mg/m2, cyclophosphamide 600 mg/m2 followed by three cycles of CMF in patients with four to nine positive lymph nodes.
RESULTS: Thirty-six percent of the patients included were older than 60 years, 63% were node negative, 13% had four to nine positive nodes, 55% had tumor grade 3, and 41% received breast-preserving surgery. At 5.3 years’ median follow-up, the first event of failure (recurrence, secondary tumor, or death) had occurred in 123 patients in the tamoxifen group and 107 patients of the control group. Event-free survival rates after 5 years were 70.3% (95% CI, 65.5% to 75.0%) and 72.8% (95% CI, 68.2% to 77.5%) for the tamoxifen and control groups, respectively. The estimated hazard ratio of tamoxifen versus control was 1.13 (95% CI, 0.87 to 1.48; P = .34), which gives no indication of an additional benefit of tamoxifen in these patients.
CONCLUSION: This study contributes substantially to finalization of the presently emerging evidence that tamoxifen does not benefit women with receptor-negative breast cancer after chemotherapy.
INTRODUCTION
In the early 1990s, overviews by the Early Breast Cancer Trialists’ Collaborative Group had demonstrated the benefit of adjuvant systemic therapy in women with node-negative and node-positive breast cancer.1 The role of chemotherapy versus endocrine therapy or chemoendocrine therapy remained to be determined in women with an intermediate or high risk of tumor recurrence on the basis of risk factors such as menopausal status, node status, and sensitivity to steroid receptors.2 Although the benefit of using tamoxifen for receptor-positive breast cancer had been firmly established, uncertainty remained regarding receptor-negative patients.1
In 1993 the German Adjuvant Breast Cancer Group (GABG) started a series of randomized trials for various risk groups. Trial IV D-93 included postmenopausal patients with primary breast cancer and negative estrogen and progesterone receptors. Patients were either node negative or node positive and had one to nine positive nodes. The study investigates the effect of tamoxifen given sequentially, after risk-adapted chemotherapy, at a daily dose of 30 mg for 5 years according to consensus practice in Germany at that time versus no additional treatment. The baseline chemotherapy was cyclophosphamide, methotrexate, and fluorouracil (CMF), in agreement with International Consensus Panel recommendations.2 However, because the preliminary and final results of several trials3-5 indicated no benefit of the usual six over three cycles, node-negative patients and patients with one to three positive nodes received a nonstandard regimen of three cycles of CMF every 4 weeks. To account for their higher risk of recurrence, patients with four to nine positive nodes were given four cycles of an anthracycline-based regimen2 (epirubicin and cyclophosphamide every 3 weeks [EC]) followed by three cycles of CMF. With a median follow-up of approximately 5 years, this report presents the initial findings from the GABG-IV D-93 trial.
METHODS
Patients
Postmenopausal women up to 70 years of age with a histologic diagnosis of invasive breast cancer and zero to nine positive axillary lymph nodes were eligible if they complied with the following criteria: negative hormone receptor status; stage pT1-3, N0 to N3, M0; no prior surgical, systemic, or radiation therapy against breast cancer; Karnofsky index score 60 or World Health Organization performance scale score of 0 to 2; hormone-replacement therapy (HRT) to be stopped if applicable; suitability for follow-up; and written informed consent. Major exclusion criteria were distant metastases; incomplete resection of the tumor; resection of less than 10 axillary lymph nodes; simultaneous contralateral breast cancer; previous second malignancy except basal cell carcinoma of the skin or carcinoma-in-situ of the cervix uteri; insufficient organ function or significant comorbidities; and any T4 tumor. For eligibility, chemotherapy had to start within 28 days from definitive primary surgery, and randomization was possible in the interval between surgery and up to 28 days after start of the third cycle of CMF chemotherapy.
Patients were recruited between February 1993 and December 2000 from 63 centers all over Germany. Centers included university hospitals as well as regional and city hospitals (see Appendix). The trial protocol was approved by all ethical committees.
Determination of Postmenopausal Status and Negative Hormone Receptor Status
For patients without HRT before enrollment, premenopausal status was excluded either on the basis of reported menstrual periods within the last 6 months or according to follicle-stimulating hormone (< 20 IU/L) and luteinizing hormone (> 50 pg/mL) serum levels. Serum levels were assessed 1 week after discontinuation of HRT if applicable. Estrogen and progesterone hormone receptor status were determined at local laboratories either biochemically by a dextran-coated charcoal assay (positive: 20 fmol/mg according to consensus practice in Germany) and/or by scores based on an immunohistochemistry assay (positive: score 2) and/or by scores from immunologic evaluation6 (positive: score 2). Biochemistry was preferred if it was available. If the institution specified a different cut point for immunohistochemical or immunologic assays, this cut point was used. Negative hormone receptor status was defined as negative estrogen and progesterone receptors.
Study Treatments
Patients were randomly assigned to either oral tamoxifen (Nolvadex 30 mg once a day for 5 years) or no additional treatment after a risk-adapted adjuvant chemotherapy consisting of three cycles of CMF (cyclophosphamide 500 mg/m2, methotrexate 40 mg/m2, and 5-fluorouracil 600 mg/m2 given intravenously on days 1 and 8, every 4 weeks) in patients with zero to three positive lymph nodes and four cycles of EC (epirubicin 90 mg/m2 and cyclophosphamide 600 mg/m2 given intravenously every 3 weeks) followed by three cycles of CMF in patients with four to nine positive lymph nodes. The interval between the last course of EC and the first course of CMF was 3 weeks. Chemotherapy had to start within 28 days from definitive primary surgery. Optional radiotherapy was applied according to local policies.
Evaluation Criteria
Follow-up examinations were scheduled every 3 months for the first 2 years, every 6 months for the following 3 years, and annually thereafter. The primary end point was event-free survival (EFS), defined as the time from definitive primary surgery to the first event of failure (locoregional recurrence, metastases, second primaries including contralateral breast cancer, or death). The first event of failure was classified as isolated locoregional recurrence if locoregional recurrence occurred at least 4 weeks before an event at a distant site. The secondary end point was overall survival (OS), which was defined as the time from definitive primary surgery to death resulting from any cause. Results on quality of life will be reported elsewhere.
Randomization and Quality Control
Patients with confirmed eligibility were randomly assigned by phone or fax at the Department of Medical Biometry and Statistics (University Hospital Freiburg). They were stratified by clinical center and nodal status (zero to three or four to nine positive lymph nodes). Within each stratum, block randomization with randomly varying block size and a 1:1 treatment ratio was performed based on computer-generated lists. For ease of organization, no blinding was done. Starting in December 1997, source-data verification was performed by an external clinical research organization (SKM, Wiesbaden, Germany). The data were subjected to computerized plausibility checks.
Statistical Methods
On the basis of preliminary data of patients selected from previous studies,7 it was anticipated that EFS would be 60% at 5 years in the control group treated with risk-adaptive chemotherapy only. Thus, 190 events would be required to detect a hazard ratio of 0.67 for tamoxifen versus control (ie, an improvement to approximately 71% in the tamoxifen group), with a power of 80% using a log-rank test at the level = 5%.8 With 4 years of planned recruitment and 2 years additional follow-up, at least 700 patients would have to be included. Inclusion of patients was continued beyond this minimum target number to ensure that GABG-IV trials A-93, B-93, and E-939,10 would all be closed at approximately the same time.
EFS and OS rates were estimated by Kaplan-Meier curves.12 Event-free patients were censored at the last reported visit. Median follow-up was based on the estimated censoring distribution,13 and the percentage of complete follow-up was calculated.14 The treatment effect on EFS was estimated as the hazard ratio in a Cox model15 with a 95% CI. P values were based on Wald tests.16 An adjusted analysis of treatment and prognostic factors was performed in a multiple Cox model including treatment, tumor size, tumor grade, number of positive lymph nodes, and type of surgery. This analysis was carried out in complete cases (ie, in patients with complete data on size, grade, nodes, and surgery).
To investigate possible differential treatment effects on EFS (interactions), effects were estimated for each of two subgroups according to tumor size, tumor grade, and nodal status, respectively, in three Cox models with simultaneous adjustment for the prognostic factors mentioned above. The effect of nodal status was studied in cases with zero or one to three positive nodes only to avoid confounding with risk-adapted chemotherapy (see Study Treatments). Interactions and treatment effects within subgroups were tested by Wald tests. Because of multiple testing, a significance level of 1% was used for these tests.
During the study, two formal interim analyses of the treatment effect on EFS were conducted with a significance level of 0.5% each when 29 and 118 events had been observed, respectively. To maintain the overall significance level of = 5%, the present, final analysis with 230 events was carried out at the nominal significance level of ' = 4.40%, which was calculated according to the -spending approach.17 Accordingly, 95.60% CIs are reported for the hazard ratios between treatment groups in the unadjusted and adjusted analyses of EFS.
Data were processed and evaluated with SAS (SAS Institute, Cary, NC).18 Following the intention-to-treat principle, ineligible patients were not excluded, and treatment was analyzed as randomized. Analyses were carried out in accordance with a prespecified statistical analysis plan. All tests, P values, and CIs were two sided.
RESULTS
Recruitment and Patient Characteristics
Between February 1993 and December 2000, a total of 829 patients were included at 63 centers in Germany: 408 were randomly assigned to the control arm and 421 to tamoxifen. Thirty-six percent of the patients were older than 60 years, 63% were node negative, 13% had four to nine positive nodes, 55% had tumor grade 3, and 41% received breast-preserving surgery. Baseline characteristics were well balanced between treatments (Table 1).
The eligibility criteria turned out to be violated in 97 patients, mostly because of positive hormone receptor and delayed start of chemotherapy. Fifty-one patients were from the control group (receptor positive, n = 26; enrollment despite start of therapy > 28 days from surgery, n = 9; randomization > 28 days after start of third cycle of CMF chemotherapy, n = 5; previous second malignancy, n = 4; incomplete resection of primary breast tumor, n = 1; resection of < 10 axillary lymph nodes, n = 2; premenopausal, n = 1; > 70 years of age, n = 2; perioperative tamoxifen treatment, n = 1; randomly assigned twice by error, n = 2; randomly assigned erroneously without consent, n = 1; three patients met > one of these criteria). Forty-six patients were in the arm allocated to tamoxifen (receptor positive, n = 19; enrollment despite start of therapy > 28 days from surgery, n = 11; assignment > 28 days after start of the third cycle of CMF chemotherapy, n = 2; simultaneous contralateral breast cancer, n = 1; pT4 breast cancer, n = 4; metastases at assignment, n = 1; incomplete resection of primary breast tumor, n = 1; resection of < 10 axillary lymph nodes, n = 3; premenopausal, n = 1; > 70 years of age, n = 1; randomly assigned twice by error, n = 2).
In one of the four patients randomly assigned twice by error, the result of the second assignment differed from the first; the result of the second assignment was ignored in the analyses. Ineligible patients were included in the analyses; however, all data on the patient randomly assigned erroneously without consent are missing.
Compliance
The two arms were comparable with respect to baseline chemotherapy. All planned cycles were completed in 89.7% and 92.6% of the patients in the control and tamoxifen arms, respectively (Fig 1). Overall, chemotherapy was incomplete in 5.8% and not started in 2.1% (application unknown in 0.8%).
In the tamoxifen group, 359 patients started endocrine therapy as assigned. Among these, the rate of patients receiving 5 years of treatment was 55.9% (proportion derived treating patients with incomplete data on duration of tamoxifen as censored,12 n = 95). Forty patients in the tamoxifen group did not start the treatment; 16.2% of those who did were delayed (> 34 days after start of the last cycle of chemotherapy).
Follow-Up and Observed Events
The data cutoff for this analysis was the end of June 2003, leading to a median follow-up of 5.3 years. Completeness of follow-up, measured as the ratio of observed and potential follow-up information available at the data cutoff,14 was 88%. Follow-up was similar in both treatment groups (Fig 1). With regard to EFS, 230 events have been observed thus far. Table 2 lists the distribution of the different events of failure. The first event of failure (tamoxifen v control) was an isolated local recurrence (26 v 33), a distant failure (94 v 69), and death without recurrence (3 v 5). Among the distant failures, 17 were new nonbreast primaries (four endometrial cancers) in the tamoxifen group and five in the control group (no endometrial cancer). With regard to OS, 152 deaths (80 v 72) of any cause have been observed up to the time of data cut-off.
EFS
EFS rates by treatment are displayed in Figure 2. Five-year EFS rates are estimated as 70.3% (95% CI, 65.5% to 75.0%) and 72.8% (95% CI, 68.2% to 77.5%) for the tamoxifen and control groups, respectively. The unadjusted hazard ratio of tamoxifen versus control was 1.13 (95% CI, 0.87 to 1.48; calculated as 95.60% CI to account for two interim analyses), which gives no indication of a benefit of additional tamoxifen in these patients (P = .34). This result was confirmed in the analysis adjusted for tumor size, tumor grade, type of surgery, and number of positive nodes (Table 3). The adjusted hazard ratio of tamoxifen versus control is 1.16 (95% CI, 0.88 to 1.51; calculated as 95.60% CI to account for interim analyses; P = .28; complete cases, n = 811; 228 events).
Possible differences in treatment effects were investigated in subgroups defined by tumor grade, tumor size, and nodal status with a significance level of 1% for the tests on interaction between treatment and the corresponding prognostic factor. To account for the fact that the risk-adapted chemotherapy was linked by study design to the number of positive lymph nodes (zero to three or four to nine), the predictive effect of lymph nodes was studied only in patients with zero versus one to three positive nodes. However, no significant interaction was found (data not shown).
OS
OS rates are displayed in Figure 3. Five-year OS rates are estimated as 82.2% and 81.6% in the tamoxifen and control groups, respectively. With 152 (18%) deaths observed in total up to the time of data cut-off, it is too early for a definitive analysis of OS.
Tolerability and Adverse Events
Tamoxifen was discontinued for medical reasons other than recurrence or death in 41 patients. Toxicity data are based on reports of serious adverse events before recurrence. Life-threatening adverse events were reported in eight patients of the tamoxifen group. Two events occurred during EC chemotherapy (hematologic toxicity, cardiomyopathy) and six occurred during tamoxifen (five thromboembolic events, one mastitis). In contrast, no life-threatening adverse events were reported in the control group. In total, two patients died while on adjuvant treatment, both from pulmonary embolism. The first death, judged as treatment related, occurred in the control group during EC-CMF, the other, judged as unrelated to treatment, occurred during adjuvant tamoxifen.
DISCUSSION
This report provides information about the efficacy of 5 years of adjuvant tamoxifen versus control after risk-adapted chemotherapy in a population of postmenopausal, estrogen and progesterone receptor–negative breast cancer patients. Patient characteristics are well balanced between treatment groups (Table 1) and so are the length of median follow-up and completeness of follow-up (Fig 1). After approximately 5 years of median follow-up and 230 observed events, the study is sufficiently mature to give reliable results. We observed a moderate, statistically nonsignificant detrimental effect of tamoxifen on EFS.
Our findings agree with the general evidence on the efficacy of tamoxifen in receptor-negative breast cancer, which has been accumulating since the present study was initiated. The recently published update of the overview by the Early Breast Cancer Trialists’ Collaborative Group19 shows a small but nonsignificant disadvantage for approximately 5 years of tamoxifen versus none with respect to recurrence and mortality rates, both in the subgroup of estrogen receptor-poor women (see Fig 7 in the Early Breast Cancer Trialists’ Collaborative Group’s study19) and the subgroup of estrogen receptor-poor, progesterone receptor-poor women (web appendix 1, annex Fig 7 in the Early Breast Cancer Trialists’ Collaborative Group’s study19). It included three studies20-22 that, comparable to our design, reported results on the efficacy of additional adjuvant tamoxifen combined with polychemotherapy in receptor-negative, older patients (postmenopausal or age 50 years), and neither showed a benefit for tamoxifen. There remains some inconsistency in the overview with regard to the duration of tamoxifen in that some benefit for administration for one to two years was observed in estrogen receptor-poor patients. A possible explanation could be that false-negative estrogen receptor measurements were more pronounced in earlier trials that investigated the shorter duration of tamoxifen.
Regarding the internal validity of GABG-IV D-93 study results, we have to concede that there were many ineligible patients. The most noticeable is the high number of receptor-positive patients in each group. A cutoff point of 20 fmol/mg cytosol protein for the biochemical assay was generally recommended for application in Germany in the early phase of the trial, whereas immunohistochemical techniques became standard toward the end of recruitment. Therefore, even some of the eligible patients included at the beginning would be classified as false negative by current standards. Furthermore, in this unblinded trial, 40 patients (9.5%) did not start the drug, and 69 (19.2%) of those who did discontinued it prematurely for reasons other than failure or death (Fig 1). Reduced compliance is not uncommon in tamoxifen trials. For example, either failure to start or premature discontinuation were observed in approximately 20% of women allocated to approximately 5 years of tamoxifen in the three largest trials included in the 1998 overview.23,24-26 However, the benefit of shorter durations of tamoxifen in estrogen receptor-positive patients is well established,19,23 and a minor benefit was noted even in the estrogen receptor-poor patients.19 Thus, by its failure to show a benefit for tamoxifen despite impaired compliance and inclusion of some receptor-positive patients our study provides additional important data on the role of tamoxifen in receptor-negative breast cancer.
In conclusion, as Swain27 stated with reference to a preliminary version of the updated overview, "more data are needed to definitely state that tamoxifen does not benefit women with estrogen receptor–negative disease." With its large cohort of postmenopausal, receptor-negative patients and mature follow-up, this study makes a substantial contribution toward finalizing conclusive evidence on this matter.
Appendix
The acknowledgment and appendix are included in the full-text version of this article, available online at www.JCO.org. They are not included in the PDF (via Adober Acrobat Readerr) version.
The participating patients and investigators were recruited for the trial at the following sites: Kreiskrankenhaus Albstadt (G. Geier); Universit?tsklinikum Charité, Chirugie, Berlin (K.-J. Winzer); Universit?tsklinikum Charité, Gyn?kologie, Berlin (W. Lichtenegger); Vivantes Humboldt Klinikum, Berlin (L. Schneppel); Waldfriede Krankenhaus, Berlin (C. Kempter); St Josefshospital, Cloppenburg (I. Schulz-Im Busch); Klinikum Deggendorf (D. Augustin); Kreiskrankenhaus Eggenfelden (W. Siebert); J. W. Goethe Universit?t Frankfurt (M. Kaufmann); Universit?tsklinikum Freiburg (W. Kleine); St?dtisches Krankenhaus Friedrichshafen (G. de Gregorio); Klinik am Eichert, G?ppingen (A. Hettenbach); Universit?tsklinikum, Chirurgische Onkologie, G?ttingen (T. Liersch); Universit?tsklinikum Halle (H. K?lbl); Albertinen-Krankenhaus Hamburg (M. H. Carstensen); Gyn?kologische Gemeinschaftspraxis, Hamburg (E. Goepel); Universit?tsklinikum Hamburg (F. J?nicke); Klinikum Stadt Hanau (H. H. Zippel); Henriettenstiftung, Hannover (J. Hilfrich); Krankenhaus Nordstadt, Hannover (P. Hohlweg-Majert); St Josefs-Krankenhaus, Heidelberg (E. Krystek); Universit?tsklinikum Heidelberg (G. Bastert); Klinikum Hoyerswerda (K. Lürmann); Stadtkrankenhaus Idar-Oberstein (R. Gros); Universit?tsklinikum, Klinik für Innere Medizin II, Jena (K. H?ffken); Universit?tsklinikum, Frauenklinik, Jena (A. Schneider); Diakonissenkrankenhaus Karlsruhe (M. Zedelius); St?dtisches Klinikum Karlsruhe (G. Kaltenecker); St Vincentius Krankenhaus, Karlsruhe (H. G. Meerpohl); St?dtisches Klinikum Kassel (T. Dimpfl); Universit?tsklinikum Kiel (N. Maass); Frankenwaldklinik Kronach (P. Heinkele); Kreiskrankenhaus Leonberg (M. Kuglin); St Bonifatius Hospital Lingen (M. Johnscher); Universit?tsklinikum Lübeck (C. Strunck); AG Mammakarzinom des Tumorzentrums Magdeburg/Sachsen-Anhalt (G. Gademann); Universit?tsklinikum Mainz (P. Knapstein); St Vinzenz u. St Elisabeth-Hospital Mainz (C. Lei?ner); Klinikum der Stadt Mannheim (C. F?rster); Universit?tsklinikum Marburg (U. S. Albert); Klinikum Rechts der Isar, München (M. Kiechle); Frauenklinik vom Roten Kreuz München (G. Raab); Clemenshospital Münster (A. Neff); Universit?tsklinikum Münster (C. Jackisch); Staufenklinik, Mutlangen (A. Rodewald); Frauenklinik der St?dtischen Kliniken Offenbach a. M. (S. Jung); Marienhospital Osnabrück (K. Brunnert, M. Butterwegge); St Vincenz-Krankenhaus Paderborn (W. Meinerz); Krankenhaus St Trudbert, Pforzheim (R. Stiglmayer); Prosper Hospital Recklinghausen (K. Stahl); Kreiskrankenhaus Reutlingen (C. Schick); Mathias Spital, Rheine (G. Heywinkel); Frauenklinik Rheinfelden (H. Dieterich); Universit?t Rostock (T. Reimer); Stadtkrankenhaus Rüsselsheim (L. Heilmann); Martin-Luther-Krankenhaus Schleswig (H. Anger); Evangelisches Jung-Stilling Krankenhaus Siegen (M. Türker); Kreiskrankenhaus Stadthagen (J. Feltz-Sü?enbach); Universit?tsklinikum Tübingen (A. Bergmann); Universit?tsklinikum Ulm (R. Kreienberg); Ammerland Klinik GmbH Westerstede (M. Hippach); Stadtkrankenhaus Worms (T. Hitschold); and Universit?tsklinikum Würzburg (J. Dietl).
Authors’ Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed discription of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Acknowledgment
This study was funded by the Deutsche Krebshilfe and in part by Pharmacia, Germany, and AstraZenenca, Germany. We thank them for their support for the conduct of this trial. We also thank the participating patients and investigators.
NOTES
Supported by Grants from the Deutsche Krebshilfe, AstraZeneca, Germany, and Pharmacia, Germany.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Early Breast Cancer Trialists’ Collaborative Group: Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. 133 randomised trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 339:1-15, 71-85, 1992
Glick JH, Gelber RD, Goldhirsch A, et al: Adjuvant therapy of primary breast cancer. 4th International Conference on Adjuvant Therapy of Primary Breast Cancer St. Gallen, Switzerland. Ann Oncol 3:801-807, 1992
Schumacher M, Bastert G, Bojar H, et al: Randomized 2 x 2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients. German Breast Cancer Study Group. J Clin Oncol 12:2086-2093, 1994
International Breast Cancer Study Group: Duration and reintroduction of adjuvant chemotherapy for node-positive premenopausal breast cancer patients. J Clin Oncol 14:1885-1894, 1996
Maass H: German Adjuvant Breast Cancer Group: 3 versus 6 cycles of CMF in breast cancer patients with 1 to 9 positive nodes: Results of the German Adjuvant Breast Cancer Group (GABG) III trial. Proc Am Soc Clin Oncol 19:A-73, 2000 (abstr 283)
Leclercq G, Bojar H, Goussard J, et al: Abbott monoclonal enzyme immunoassay measurement of estrogen receptors in human breast cancer: A European multicenter study. Cancer Res 46:4233s-4236s, 1986
Schmoor C, Olschewski M, Sauerbrei W, et al: Long-term follow-up of patients in four prospective studies of the German Breast Cancer Study Group (GBSG): A summary of key results. Onkologie 25:143-150, 2002
Schoenfeld DA: Sample-size formula for the proportional-hazards regression model. Biometrics 39:499-503, 1983
von Minckwitz G, Graf E, Geberth M, et al: Goserelin versus CMF as adjuvant therapy for node-negative, hormone receptor-positive breast cancer in premenopausal patients. The GABG IV-A-93 trial. J Clin Oncol 22(14S):534, 2004 (abstr)
Kaufmann M, Graf E, Jonat W, et al: Goserelin versus control after adjuvant, risk-adapted chemotherapy in premenopausal patients with breast cancer: GABG-IV B-93. J Clin Oncol 22(14S):588, 2004 (abstr)
Eiermann W, Graf E, Raab G, et al: High-dose epirubicin vs. standard-dose epirubicin-cyclophosphamide followed by CMF in breast cancer patients with 10 or more positive lymph nodes: First results of a randomised trial of the German Adjuvant Breast Cancer Group (GABG). Proc Am Soc Clin Oncol 20:A-127, 2001 (abstr 27)
Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Schemper M, Smith TL: A note on quantifying follow-up in studies of failure time. Control Clin Trials 17:343-346, 1996
Clark TG, Altman DG, De Stavola BL: Quantification of the completeness of follow-up. Lancet 359:1309-1310, 2002
Cox DR: Regression models and life tables. J R Stat Soc B 34:187-220, 1972
Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. New York, NY, Wiley, 1980
Lan KKG, DeMets DL: Discrete sequential boundaries for clinical trials. Biometrika 70:659-663, 1983
SAS Institute Inc: SAS Language, Version 6. Cary, NC, 1990
Early Breast Cancer Trialists’ Collaborative Group (EBCTCG): Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet 365:1687-16717, 2005
Fisher B, Anderson S, Tan-Chiu E, et al: Tamoxifen and chemotherapy for axillary node-negative, estrogen receptor-negative breast cancer: Findings from National Surgical Adjuvant Breast and Bowel Project B-23. J Clin Oncol 19:931-942, 2001
Hutchins L, Green S, Ravdin P, et al: CMF versus CAF +/– tamoxifen in high-risk node-negative breast cancer patients and a natural history follow-up study in low-risk node-negative patients: Update of tamoxifen results. Breast Cancer Res Treat 57:25, 1999 (abstr 7)
Nomura Y, Shirouzu M, Takayama T: Direct comparisons of adjuvant endocrine therapy, chemotherapy, and chemoendocrine therapy for operable breast cancer patients stratified by estrogen receptor and menopausal status. Breast Cancer Res Treat 49:51-60, 1998
Early Breast Cancer Trialists’ Collaborative Group 1998: Tamoxifen for early breast cancer: An overview of the randomised trials. Lancet 351:1451-1467, 1998
Rutqvist LE, Cedermark B, Glas U, et al: Radiotherapy, chemotherapy, and tamoxifen as adjuncts to surgery in early breast cancer: A summary of the randomized trials. Int J Radiat Oncol Biol Phys 16:629-639, 1989
McDonald CC, Alexander FE, Whyte BW, et al: Cardiac and vascular morbidity in women receiving adjuvant tamoxifen for breast cancer in a randomised trial. The Scottish Cancer Trials Breast Group. BMJ 311:977-980, 1995
Fisher B, Constantino JP, Redmont CK, et al: Endometrial cancer in tamoxifen-treated breast cancer patients: Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14. J Natl Cancer Inst 86:527-537, 1994
Swain SM: Tamoxifen for patients with estrogen receptor-negative breast cancer. J Clin Oncol 19:93s-97s, 2001(Manfred Kaufmann, Erika G)