Abstract

Background: Results from three National Surgical Adjuvant Breast and Bowel Project sequentially conducted randomized trials of postoperative chemotherapy in women with estrogen receptor–negative tumors and negative axillary lymph nodes have demonstrated that a combination of methotrexate and 5-fluorouracil (MF) is more effective than surgery alone, that cyclophosphamide with MF (CMF) is more effective than MF, and that CMF and doxorubicin (Adriamycin) with cyclophosphamide (AC) are equally beneficial. This report presents updated findings from those trials, relates the results to age and menopausal status, and estimates the extent of progress made in treating such patients. Methods: Patients were randomly assigned as follows: in B-13, 760 patients were assigned to surgery only or to MF; in B-19, 1095 patients were assigned to MF or CMF; in B-23, 2008 patients were assigned to CMF or AC. Recurrence-free survival (RFS) and overall survival (OS) were estimated according to age and menopausal status. Smoothed recurrence rates were used to evaluate patterns of recurrence as a continuous function of age. The Cox proportional hazards model was used to test for interactions between treatment and covariates and to estimate hazard ratios (HRs) for pairwise group comparisons. Results: In B-13, through 16 years of follow-up, an overall benefit was seen with MF relative to surgery alone (RFS: HR = 0.59, 95% confidence interval [CI] = 0.44 to 0.78, P <0.001; OS: HR = 0.75, 95% CI = 0.58 to 0.98, P = 0.03). In B-19, through 13 years of follow-up, an overall benefit was seen for CMF relative to MF (RFS: HR = 0.59, 95% CI = 0.45 to 0.77, P <0.001; OS: HR = 0.71; 95% CI = 0.55 to 0.92; P = 0.01). In both trials, all age and menopausal groups demonstrated an RFS benefit, and most demonstrated an OS benefit. In B-23, through 8 years of follow-up, there were no statistically significant differences between the CMF and AC groups (RFS: HR = 1.00, 95% CI = 0.79 to 1.27, P = 0.97; OS, HR = 0.92, 95% CI = 0.73 to 1.17; P = 0.51). When women in the CMF or AC groups (B-19, B-23) were compared with those who were in the surgery-alone group (B-13), through 8 years of follow-up there was a 58% reduction in recurrence and a 40% reduction in mortality as a result of the chemotherapy. Conclusions: Outcomes in CMF- or AC-treated women with estrogen receptor–negative tumors and negative axillary lymph nodes were similar in all age groups. The decreased benefit from chemotherapy observed with increasing age was a result of a better outcome associated with advancing age in women who underwent surgery alone rather than a poorer outcome resulting from the use of chemotherapy.

National Surgical Adjuvant Breast and Bowel Project (NSABP) randomized clinical trials B-13, B-19, and B-23 were conducted sequentially in women with estrogen receptor (ER)–negative tumors and negative axillary lymph nodes. The treatment given to the experimental group of each trial served as the control for the subsequent trial. Through 4 years of follow-up, the initial findings from B-13, which was implemented to determine the value of postoperative chemotherapy, demonstrated a statistically significant improvement in the disease–free survival of women who received a combination of methotrexate and 5-fluorouracil (MF) as compared with women who received no systemic therapy after surgery ( 1 ) . The 5-year findings from B-19 demonstrated that cyclophosphamide and MF (CMF) was more effective than MF with regard to disease–free survival, distant disease–free survival, and overall survival ( 2 ) . B-23 was initiated to address two questions: whether a doxorubicin–cyclophosphamide (AC) regimen is more effective than CMF with regard to recurrence-free survival, disease-free survival, and overall survival and whether women with ER-negative tumors would benefit from the addition of tamoxifen to either CMF or AC. Through 5 years, the B-23 findings demonstrated no statistically significant differences in the outcomes of AC- and CMF-treated patients and no advantage from tamoxifen with either chemotherapy regimen over the benefit achieved with chemotherapy alone ( 3 ) .

In this report we present updated data on recurrence-free and overall survival through 16 years of follow-up in the B-13 trial, 13 years in the B-19 trial, and 8 years in the B-23 trial. We also present information about the outcome of women in the three studies according to age and menopausal status. Finally, by interrelating the findings from the three studies, we estimate the extent of progress made in the treatment of women with ER-negative tumors and negative axillary lymph nodes.

P atients and M ethods

Subjects

Women at NSABP institutions in the United States and Canada who had primary breast cancer and axillary lymph nodes that were negative for cancer on histologic examination were eligible to participate in one of the three studies if their tumors were ER negative (i.e., had <10 fmol/mg cytosol protein). ER levels in all tumor specimens from patients in B-13 and B-19 were assayed by means of the sucrose–density gradient dextran-coated-charcoal (DCC) titrations with Scatchard analysis or by DCC with a single saturating dose. ER levels in tumors from patients in B-23 were assayed by DCC (45%) or were subjected to either enzyme immunoassay (19%) or immunocytochemical assay (37%). The DCC analyses were carried out in private or commercial laboratories that had complied with NSABP prerequisites for quality control. After they had undergone surgery (total mastectomy and axillary lymph node dissection or lumpectomy and axillary lymph node dissection followed by breast irradiation) and had given written informed consent, patients were randomly assigned to one of the treatment groups. The distributions of age, tumor size at clinical examination, and type of operation were balanced across treatment groups by the use of a biased-coin minimization algorithm ( 4 ) .

Between August 1, 1981, and January 25, 1988, 760 women in the B-13 trial were randomly assigned to surgery only or to postoperative treatment with MF (Table 1 ). From October 17, 1988, to July 31, 1990, 1095 women in the B-19 study were randomly assigned to postoperative treatment with either MF or CMF. Between May 12, 1991, and December 31, 1998, 2008 women in B-23 were randomly assigned to one of four treatment groups after surgery: CMF plus placebo, CMF plus tamoxifen, AC plus placebo, or AC plus tamoxifen. Detailed accounts of the dose and administration of MF in B-13; of MF and CMF in B-19; of CMF, AC, placebo, and tamoxifen in B-23; and of radiation therapy that was administered to patients who underwent lumpectomy have been presented in previous reports of each of the studies ( 13 ) . Administration of MF was similar in B-13 and B-19, with the only difference being that in B-13 MF was administered for 1 year (i.e., every 4 weeks for 12 cycles), and in B-19 MF was administered for 6 months (i.e., for six cycles). In both B-19 and B-23, CMF was administered according to the standard Milan regimen (i.e., every 28 days for six cycles). In B-23, AC was administered every 21 days for four cycles; patients treated with tamoxifen were to receive it daily for 5 years.

Characteristics controlled for in the randomization were well balanced between the treatment groups in each of the three studies; this information has been presented in the initial reports of each of the trials ( 13 ) . There was good agreement across studies relative to age at study entry: About 55% to 60% of the women were 49 or fewer years of age, about 25% to 30% were 50–59 years of age, and about 13% to 18% were 60 or more years of age. Menopausal status was self-reported by 98.8% of the women in B-13, by 99.4% of those in B-19, and by 98.9% of those in B-23 before they underwent treatment. No predefined criteria were used for determining menopausal status. When the outcomes of premenopausal and postmenopausal women were compared, the few women who reported that they were perimenopausal (5.0% in B-13, 7.6% in B-19, and 7.9% in B-23) were included with women who reported that they were premenopausal (46.2%, 49.3%, and 44.3% in the three studies, respectively). In this article, that combined group is referred to as the premenopausal group. A total of 48.8%, 43.1%, and 47.7% of the women in the three trials, respectively, were designated as being postmenopausal. In all of the studies, tumor sizes were obtained from pathology reports that had been submitted to the NSABP Biostatistical Center at patient entry. Tumor size among studies was consistent, that is, about 2.4 ± 1.3 cm (mean ± SD) in the greatest dimension. Over time, breast-conserving surgery was accepted more frequently by patients and physicians; as a consequence, the incidence of lumpectomy increased from about 35% in the initial study (B-13) to about 55% in the most recent trial (B-23).

Statistical Analyses

The end points used in the analyses were recurrence-free survival and overall survival; times to those end points were calculated from date of surgery. Events considered in the estimation of recurrence-free survival were first local, regional, and distant recurrences. All deaths and second primary cancers were censored. The analyses of overall survival included all deaths. Time-to-event distributions were estimated by use of the Kaplan–Meier method and were compared by use of two-sided log-rank tests over all available observation times ( 5 , 6 ) . Cox’s proportional hazards model was used to examine prognostic covariates, to test for interactions between treatment and the covariates, and to estimate hazard ratios for pairwise group comparisons ( 7 ) . The proportional hazards assumption was confirmed by use of a method based on weighted residuals ( 8 ) . In the B-13 study, all analyses were carried out by the use of pairwise group comparisons between the patients randomly assigned to surgery only and those assigned to postoperative MF. In the B-19 trial, analyses were limited to pairwise comparisons between the patients randomly assigned to MF and those assigned to CMF. In B-23, in which a factorial design was employed, comparisons were made between the patients randomly assigned to CMF versus those assigned to AC and also between those randomly assigned to placebo versus those assigned to tamoxifen. Both end points were evaluated according to patient age and according to menopausal status. Results are summarized as hazard ratios (HRs) with associated 95% confidence intervals (CIs) and are expressed as percentage reductions in risk of recurrence, second primary cancer, or death relative to a comparison group. All P values were derived using two-sided tests for significance, and P values less than 0.05 were considered statistically significant. Comparison of treatment effectiveness according to age groups was assessed for predefined age cohorts (i.e., ≤49, 50–59, and ≥60 years).

In exploratory analyses, patterns of change of the annual recurrence rates were assessed in each treatment arm over age as a continuous variable by the use of a simple smoothing method ( 9 ) . Within treatment groups, patients were sorted from youngest to oldest, and the rate of recurrence was calculated for patients numbered 1 through 100. This procedure was repeated for patients numbered 2 through 101, 3 through 102, and so forth and concluded with the cohort that consisted of the 100 oldest patients. The rate associated with each cohort was then plotted against its mean age. These plots were then resmoothed by the use of a standard smoothing algorithm ( 10 ) and were compared between treatment groups. Cumulative incidence curves were employed to summarize the 10-year risk for the sites of first treatment failure ( 11 ) . In addition, to characterize the relative risk of the sites of first treatment failure according to treatment groups, we calculated cause-specific hazard ratios adjusted for clinical tumor size, age at study entry, and type of surgery performed ( 12 , 13 ) . A 95% confidence interval was provided for each cause-specific hazard ratio. The findings presented in this report pertain to eligible patients in each of the three studies who had follow-up information that was received by March 31, 2003. The median follow-up time among survivors was 15.1 years (range = 0.3 to 21.1 years) in the B-13 trial, 12.8 years (range = 0.3 to 14.4 years) in B-19, and 8.2 years (range = 0.1 to 11.9 years) in B-23.

R esults

Outcome of Each Trial Overall and According to Age or Menopausal Status

NSABP B-13.

Through 16 years of follow-up, women with lymph node–negative, ER-negative primary breast cancer in the postoperative MF therapy group had statistically significantly better outcomes, on average, than women in the surgery-only group (RFS: 77% versus 63%, P <0.001; OS: 74% versus 65%, P = 0.03) (Fig. 1 ). When outcomes were examined according to age (Table 2 ; Fig. 2 ), the hazard ratios for recurrence-free survival were 0.60, 0.48, and 0.70 in women aged 49 or fewer, 50–59, and 60 or more years, respectively, and the hazard ratios for overall survival were 0.79, 0.66, and 0.77, respectively. No age-by-treatment interactions were observed for either outcome measure (RFS: P = 0.97; OS: P = 0.83). When outcomes were examined according to menopausal status (Table 2 ; Fig. 2 ), the hazard ratios for recurrence-free survival were 0.65 and 0.48 in premenopausal and postmenopausal women, respectively, and the hazard ratios for overall survival were 0.86 and 0.65, respectively. Tests for interaction of treatment by menopausal status indicated that no differential benefits were observed in outcomes for the MF group according to menopausal status (RFS: P = 0.34; OS: P = 0.31).

NSABP B-19.

Through 13 years of follow-up, women in the CMF group had statistically significantly better outcomes, on average, than women in the MF group (RFS: 83% versus 73%, P <0.001; OS: 82% versus 74%, P = 0.01 (Fig. 1 ). When the outcomes were examined according to age (Table 2 ; Fig. 2 ), hazard ratios for recurrence-free survival were 0.50, 0.88, and 0.60, respectively, in women aged 49 or fewer, 50–59, and 60 or more years, and hazard ratios for overall survival were 0.64, 1.06, and 0.63, respectively. In both premenopausal and postmenopausal women, CMF had an advantage over MF for both recurrence-free survival (HR = 0.53 and 0.68, respectively) and overall survival (HR = 0.65 and 0.79, respectively). However, there were no age-by-treatment interactions (RFS: P = 0.32; OS: P = 0.74) or menopausal status-by-treatment interactions (RFS: P = 0.37; OS: P = 0.46).

NSABP B-23.

Through 8 years of follow-up, there were no statistically significant differences in recurrence-free or overall survival between the CMF + placebo and CMF + tamoxifen groups or between the AC + placebo and AC + tamoxifen groups (data not shown). As a consequence, the two CMF and the two AC groups were combined and are referred to as the CMF and AC groups, respectively. When the CMF and AC groups were compared through 8 years of follow-up, there were no differences in recurrence-free survival (85% versus 85%, P = 0.97) or in overall survival (85% versus 86%, P = 0.51), respectively (Fig. 1 ). No recurrence-free survival advantages from AC over CMF were observed in women aged 49 or fewer years (HR = 1.13) or 50–59 years (HR = 0.94); some benefit was noted in women aged 60 or more years although the improvement was not statistically significant (HR = 0.72) (Table 2 ; Fig. 2 ). AC therapy resulted in a statistically significant overall survival benefit in women aged 60 or more years (HR = 0.56) but not in those aged 49 or fewer years (HR = 1.01) or 50–59 years (HR = 1.07). No recurrence-free or overall survival advantage from AC was observed in either premenopausal or postmenopausal women. A test for interaction between menopausal status and treatment group was statistically significant for recurrence-free survival ( P = 0.022) and approached significance for overall survival ( P = 0.089). The tests for interaction between age and treatment were not statistically significant (RFS: P = 0.20; OS: P = 0.12).

Relationship Between the B-13, B-19, and B-23 Findings

Overall.

The recurrence-free survival and overall survival of women in each of the treatment groups comprising the three trials were plotted contemporaneously through 8 years of follow-up (Fig. 3 ). The outcomes of women in the chemotherapy groups of all three trials were compared with those of the women in B-13 who were in the surgery only group. The findings for the MF groups in B-13 and B-19 were concordant and demonstrated absolute benefits in recurrence-free survival of about 9% and in overall survival of about 4%. These corresponded to reductions in recurrence and mortality of approximately 32% and 17%, respectively. The findings in women in the B-19 and B-23 CMF groups and in the B-23 AC group were also concordant and demonstrated an absolute recurrence-free survival benefit compared with surgery alone of about 16%, corresponding to a reduction in recurrence of approximately 58%. The absolute overall survival benefit in the CMF groups and in the AC group was about 8%, corresponding to a 40% reduction in mortality.

According to age.

When 8-year outcomes were compared according to age, women aged 49 or fewer years in both the CMF groups and the AC group demonstrated similar absolute benefits in recurrence-free survival: approximately 21% compared with women treated with surgery alone; that is, about a 65% reduction in treatment failure (Fig. 4 ). These women showed an absolute improvement in overall survival of approximately 11%; that is, an overall reduction in mortality of about 48% (Fig. 5 ). Among women aged 50–59 years, those who were in the CMF or AC groups demonstrated an absolute benefit in recurrence-free survival of about 14%; that is, a reduction in treatment failure of approximately 54% and an absolute overall survival benefit of approximately 5%, or about a 30% reduction in mortality. For women aged 60 or more years, both recurrence-free and overall survival were better in the CMF groups than in the surgery-alone group, but the benefits of CMF relative to surgery were less than those seen in younger women. However, women aged 60 or more years in the AC group showed absolute benefits in recurrence-free survival of approximately 5% and in overall survival of approximately 7% (i.e., reductions in recurrence and mortality of 34% and 42%, respectively). Smoothed average annual recurrence rates through 10 years of follow-up were plotted concurrently according to age at study entry for patients in four of the treatment groups (Fig. 6 ). For patients in the surgery-only group, the average recurrence rate decreased with increasing age at study entry. At about age 50 years, the decrease began to accelerate so that, by about age 60 years, the recurrence rate began to approximate that in CMF-treated patients, although it was still greater than the rate in women who had received AC. The plots indicate that the younger the woman at study entry, the greater the difference in recurrence rate between those in the CMF or AC groups and those in the surgery-only group. In patients assigned to AC, the rate of recurrence continued to decrease beyond age 50; in the CMF groups, the converse was observed.

According to menopausal status.

The distribution of menopausal status was similar in each of the three studies and among the three age groups in each of the studies. Of the 2122 women aged 49 or fewer years with known menopausal status, 82.8% were premenopausal and 17.1% were postmenopausal; of the 1015 women aged 50–59 years with known menopausal status, 21.7% were premenopausal and 78.3% were postmenopausal; of the 586 women aged 60 or more years with known menopausal status, 98.5% were postmenopausal and 1.5% were premenopausal. When distribution of age was ascertained within each of the menopausal groups, of the 1988 premenopausal women, 88.4% were 49 or fewer years of age, 11.1% were aged 50–59 years, and 0.4% were 60 or more years of age. Of the 1735 postmenopausal women, 20.9% were 49 or fewer years of age, 45.8% were aged 50–59 years, and 33.2% were 60 or more years of age. Compared with premenopausal women in the surgery-alone group, premenopausal women in the CMF or AC groups showed an absolute benefit through 8 years in recurrence-free survival of approximately 21%, or, equivalently, a reduction in recurrence of approximately 65% (Fig. 4 ), and an overall survival benefit of approximately 11%, or, equivalently, a reduction in mortality of approximately 49% (Fig. 5 ). The recurrence-free survival benefit among postmenopausal women who received CMF or AC was somewhat smaller than that observed in premenopausal women (a 12% absolute benefit or, equivalently, a 48% relative reduction). The absolute benefit in overall survival in postmenopausal women was approximately 5% (i.e., about a 28% reduction in relative mortality).

According to sites of treatment failure.

The frequency of first sites of treatment failure was estimated within each of the three studies through 10 years (Table 3 ). The cumulative incidence of local-regional treatment failure was about 5% among women who received any form of postoperative chemotherapy in addition to surgery, as compared with about 14% in women who had undergone surgery only. The cumulative incidence of distant recurrence was also lower in women in the CMF- and AC-treated groups compared with women in the surgery-only group. In the B-13 study, MF appeared to reduce local-regional recurrence (HR = 0.61, 95% CI = 0.48 to 0.78) to a greater extent than distant recurrence (HR = 0.89, 95% CI = 0.74 to 1.07). CMF in B-19, compared with MF, also had a somewhat greater effect on local-regional recurrence than on distant recurrence (HR = 0.67, 95% CI = 0.54 to 0.83 and HR = 0.83, 95% CI =.070 to 0.99, respectively). In B-23, CMF and AC had similar effects on local-regional and distant recurrence (HR = 1.00, 95% CI = 0.77 to 1.37 and HR = 0.96, 95% CI = 0.78 to 1.19, respectively). Contralateral breast cancers occurred in only approximately 5% of the women through 10 years of follow-up, regardless of their treatment, and the cumulative incidence of ipsilateral breast tumor recurrence noted following lumpectomy and breast irradiation was decreased when patients received MF, CMF, or AC therapy.

D iscussion

In this article, we have presented follow-up data from three sequentially conducted NSABP studies of surgery plus chemotherapy for women with ER-negative, lymph node–negative breast cancer. Our observations are of interest from several perspectives. First, the findings show that women with negative axillary lymph nodes are at sufficiently poor risk to justify treating them with systemic therapy after surgery. The results from B-13 demonstrate that, after 16 years of follow-up, about one-third of the women treated with surgery alone had a treatment failure or died. In fact, the prognosis of some of these women was worse than that of some women with positive lymph nodes ( 14 ) .

Second, the updated results from each of the three trials have continued to support the initial findings reported after 4 or 5 years of follow-up ( 13 ) . The B-13 findings continue to indicate that a substantial benefit can be achieved with a chemotherapy regimen that does not include an alkylating agent and that the magnitude of the benefit from MF therapy has not diminished over time. The B-13 trial is also of historic interest. Before it was instituted, the idea of giving chemotherapy, particularly an alkylating agent, for the treatment of patients with negative lymph nodes was difficult to justify. Only after an advantage from MF was noted in animal models ( 15 ) and in patients with advanced breast cancer ( 16 ) and with head and neck cancer ( 17 ) were we granted permission to implement the B-13 study so as to evaluate the worth of that regimen. Subsequently, when CMF was found to benefit patients with negative lymph nodes, we deemed it necessary to conduct the B-19 study to ascertain the need for adding the alkylating agent cyclophosphamide (C) to MF. The 13-year findings from the B-19 study, which showed about a 40% greater reduction in treatment failure and about a 30% greater reduction in mortality overall, continue to demonstrate the advantage of combining C with MF. Those results differ from findings reported by others, who found no advantage from adding C to MF in lymph node–positive patients ( 18 ) . Perhaps the greater residual tumor burden in lymph node–positive patients than in lymph node–negative patients is responsible for the difference in findings. The current results from the B-23 trial, which continue to demonstrate no advantage from giving tamoxifen with either AC or CMF over that achieved with chemotherapy alone, might be viewed as not surprising because such results should have been anticipated in light of the fact that the patients had ER-negative tumors. However, our observation is important because it provides evidence against the thesis, which has resulted from laboratory and clinical evidence, that tamoxifen might exert an antitumor effect through a variety of biologic mechanisms that are unrelated to the presence of tumor estrogen receptors ( 1924 ) . Moreover, the finding that there was no difference in recurrence-free or overall survival between patients who received either CMF or AC, with or without tamoxifen, is also important because it conflicts with results obtained from a Southwest Oncology Group (SWOG) trial that was conducted in postmenopausal women with ER-positive tumors and positive axillary lymph nodes ( 25 ) . The findings from that study led its investigators to conclude that the simultaneous administration of tamoxifen with chemotherapy is inappropriate because tamoxifen antagonizes chemotherapeutic agents used in various regimens. That conclusion, which has been responsible for a change in the pattern of tamoxifen administration, seems tenuous in light of the findings from B-23.

A third point of interest relates to our justification for interrelating the findings from the three sequential trials. Because each of the studies had different aims, the results from one trial might be considered to have little relevance to the findings from the others. Indeed, when viewed individually, the findings from each study might seem to have contributed little to the current treatment of women with ER-negative tumors and negative axillary lymph nodes. Moreover, in most sequential clinical trials, including the three trials addressed in this article, only in the initial study is there likely to be an untreated control group; in subsequent trials, the treatment determined to be best in a previous trial usually serves as the control. As a consequence, after several successive long-term studies—each of which might have demonstrated only a modest benefit—the extent of the overall advantage that had been achieved might be obscure. However, the situation is different if the benefit observed in the latest study can be related to that noted in the initial control group. The B-13, B-19, and B-23 studies exemplify such a circumstance. Indeed, only by interrelating the findings from those trials might it be possible to determine the progress that was made overall.

It should be noted that some of the analyses of the three trials used subgroups and were, thus, not elicited by direct randomized comparisons. Although for more than 40 years we have acclaimed the latter methodology as being one of the greatest advances of the twentieth century, an approach that is preferable to all others for obtaining reliable information about treatment benefit, we believe that the findings obtained by interrelating the results from the three studies are likely to be credible because of the way that the studies were carried out ( 26 ) . Our studies were conducted successively by almost all of the same investigators. In addition, the protocols for each of the trials defined eligibility criteria that were essentially the same, as were almost all other aspects of the studies. For example, in all three trials, treatment assignments were balanced according to age, clinical tumor size, and type of operation. The successive trials also shared common treatment groups (i.e., MF in B-13 and B-19 and CMF in B-19 and B-23). Because the outcomes in those common treatment groups were similar, it is likely that we obtained a valid estimate of the benefit from chemotherapy by determining the magnitude of difference in outcome between the placebo group in B-13 and the chemotherapy groups in the B-19 and B-23 studies.

As a consequence, the findings in this report, which we obtained by contemporaneously examining three studies that had common features, should not be equated with the kind of results obtained by interrelating disassociated studies, as is commonly done. Regardless of whether our studies were able to provide the kind of information that can be obtained only by direct randomized comparisons, or whether they furnished data that are suitable mainly for hypothesis generation, they did provide a perspective of the magnitude of the overall benefit achieved from the use of chemotherapy. More important are the findings that provide an explanation for the age-related differences in outcome that we observed. The benefit from chemotherapy compared with surgery alone was greatest in women aged 49 or fewer years and decreased with advancing age. The decrease in benefit from chemotherapy associated with aging was mainly caused by the better recurrence-free and overall survival rates that were observed with increasing age in those women who were treated with surgery alone rather than by those outcomes in women treated with chemotherapy becoming poorer with age. In addition, outcomes after treatment with either CMF or AC were fairly similar at comparable follow-up times. However, in women aged 60 or more years, AC seemed to be somewhat more efficacious than CMF.

A fourth point for consideration relates to the appropriateness of the age cutoff points we used for grouping patients (i.e., ≤49, 50–59, and ≥60 years). To determine their appropriateness, we estimated tumor recurrence rates as a continuous variable according to patient age at study entry. Recurrence rates in women treated with surgery alone diminished continuously with advancing age, especially among women aged about 50–60 years, in whom the decline was more precipitous than in women aged 49 or fewer years. As a consequence, within each age group there are apt to be women having tumors of varying recurrence rates. Also of interest was our observation that, among women diagnosed before about 55 years of age, recurrence rates in those treated with CMF or AC were similar and remained constant, in contrast with recurrence rates in women treated with surgery alone. Among CMF-treated women whose tumors occurred after that age, recurrence rates increased somewhat with age at diagnosis, whereas in AC-treated patients no such change in recurrence rate with age was evident. As was noted with the estimates of recurrence-free and overall survival, these findings indicate that the decreasing rate of recurrence in the surgery-only group accounted for the decrease in benefit from chemotherapy that was observed with age. Overall, these observations indicate that relating recurrence rate to age as a continuous variable more accurately depicts the spectrum of change in outcome with age than does comparing information for two or three selected age groups. In that regard, the former approach more clearly supports the theses that age is, indeed, a prognostic marker and that the degree of benefit from chemotherapy is age-related.

Because it is common to relate patient outcome to menopausal status, a fifth point for consideration relates to how our findings obtained using age compare with those obtained when patients were grouped by menopausal status. Recently, by contemporaneously viewing outcome findings in a fashion similar to that done in this study, we estimated the progress achieved in two trials conducted in patients with ER-positive tumors and negative lymph nodes ( 27 ) . In our report of the findings from that study, we pointed out that age and menopausal status have different clinical and biologic connotations and that a woman’s menopausal status before therapy might be less relevant to outcome than her status after treatment. Moreover, our findings indicated that age was preferable to menopausal status for estimating benefit from therapy. That thesis is supported by our finding that outcomes in women aged 49 or fewer and 50–59 years were similar, even though 78% of women in the latter group but only 17% in the former group were postmenopausal. Indeed, even if menopausal status had been determined more precisely, it is unlikely that it would have been found to be a better discriminant than age for ascertaining benefit from either CMF or tamoxifen. This position is consistent with our statement in 1983 that, “The imprecision of menstrual histories elicited and the fact that ovarian function may persist for a variable time following cessation of menstruation makes division by age a more practical discriminant” ( 28 ) . Although the use of three age groups is likely to be more appropriate than is the use of two menopausal groups for estimating prognosis and benefit of therapy, referring to recurrence rates assessed as a continuous function of age might be most helpful.

Our finding that AC was more effective than CMF in women aged 60 or more years raises a sixth point for consideration; that is, whether AC should be considered for the management of women in that age group. Little or no information on this question is available from other studies conducted specifically to evaluate the worth of doxorubicin-containing regimens in patients with ER-negative tumors and negative axillary lymph nodes. However, evidence from 11 heterogeneous studies in an Early Breast Cancer Treatment Collaborative Group’s meta-analysis showed that anthracycline (doxorubicin)-containing regimens produced a somewhat greater reduction in recurrence and mortality than did CMF ( 29 ) . In addition, findings from NSABP studies ( 30 ) and those by several other investigators ( 31 , 32 ) have demonstrated improvement in women who received such regimens in conjunction with tamoxifen. Although CMF appears to be of little value, compared with surgery alone, for the treatment of women aged 60 or more years who have ER-negative tumors and negative axillary lymph nodes, the use of AC in such patients cannot be totally dismissed. In older women who have pathologic and other biomarkers that are indicative of rapidly proliferating tumors, for example, unfavorable histologic or nuclear grade and high thymidine labeling index, and who are more likely to benefit from chemotherapy, the use of an anthracycline-containing regimen, and perhaps CMF, might be considered.

Finally, the information presented in this article regarding the effect of chemotherapy on the incidence of treatment failure at various sites is of interest. The reason for the greater reduction in tumor recurrence at local-regional than at distant sites is unknown. The question of whether those findings were a result of differences in the biologic nature of such cells at different sites, differences in tumor burden at such sites, or some other reason remains speculative. Whatever the explanation, the reduction in the incidence of ipsilateral breast tumor recurrence in women treated with postoperative chemotherapy beyond that achieved in women treated with lumpectomy and breast irradiation alone enhances the efficacy of lumpectomy for breast conservation.

Fig. 1.

Recurrence-free survival and overall survival with hazard ratios (HRs) and 95% confidence intervals in B-13 patients treated with surgery only (total mastectomy or lumpectomy plus breast irradiation) or with postoperative methotrexate (M) and 5-fluorouracil (F) (MF); in B-19 patients treated postoperatively with MF or cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); and in B-23 patients treated postoperatively with CMF or doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 1.

Recurrence-free survival and overall survival with hazard ratios (HRs) and 95% confidence intervals in B-13 patients treated with surgery only (total mastectomy or lumpectomy plus breast irradiation) or with postoperative methotrexate (M) and 5-fluorouracil (F) (MF); in B-19 patients treated postoperatively with MF or cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); and in B-23 patients treated postoperatively with CMF or doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 2.

Hazard ratios (HRs) and 95% confidence intervals (CIs) for recurrence-free survival and overall survival according to age and menopausal status. (Top panel) B-13 patients treated with surgery only or with postoperative methotrexate (M) and 5-fluorouracil (F) (MF); (middle panel) B-19 patients treated postoperatively with MF or cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); (bottom panel) B-23 patients treated postoperatively with CMF or doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 2.

Hazard ratios (HRs) and 95% confidence intervals (CIs) for recurrence-free survival and overall survival according to age and menopausal status. (Top panel) B-13 patients treated with surgery only or with postoperative methotrexate (M) and 5-fluorouracil (F) (MF); (middle panel) B-19 patients treated postoperatively with MF or cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); (bottom panel) B-23 patients treated postoperatively with CMF or doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 3.

Relationship among 8-year survival outcomes in B-13, B-19, and B-23 according to treatment group. (Left) Recurrence-free survival. (Right) Overall survival. B-13 patients received surgery only or 12 months of postoperative methotrexate (M) and 5-fluorouracil (F) (MF); B-19 patients received 6 months of postoperative MF or postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 3.

Relationship among 8-year survival outcomes in B-13, B-19, and B-23 according to treatment group. (Left) Recurrence-free survival. (Right) Overall survival. B-13 patients received surgery only or 12 months of postoperative methotrexate (M) and 5-fluorouracil (F) (MF); B-19 patients received 6 months of postoperative MF or postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 4.

Relationship between recurrence-free survival of women in B-13 with that of women in B-19 and B-23, according to age (≤49, 50–59, or ≥60 years), menopausal status (premenopausal or postmenopausal), and treatment group. B-13 patients received surgery only; B-19 patients received postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC). The MF groups in B-13 and B-19 are not shown because of the difference in the duration of MF administration and because MF is no longer used as standard therapy.

Fig. 4.

Relationship between recurrence-free survival of women in B-13 with that of women in B-19 and B-23, according to age (≤49, 50–59, or ≥60 years), menopausal status (premenopausal or postmenopausal), and treatment group. B-13 patients received surgery only; B-19 patients received postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC). The MF groups in B-13 and B-19 are not shown because of the difference in the duration of MF administration and because MF is no longer used as standard therapy.

Fig. 5.

Relationship between overall survival of women in B-13 with that of women in B-19 and B-23, according to age (≤49, 50–59, or ≥60 years), menopausal status (premenopausal or postmenopausal), and treatment group. B-13 patients received surgery only; B-19 patients received postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC). The MF groups in B-13 and B-19 are not shown because of the difference in the duration of MF administration and because MF is no longer used as standard therapy.

Fig. 5.

Relationship between overall survival of women in B-13 with that of women in B-19 and B-23, according to age (≤49, 50–59, or ≥60 years), menopausal status (premenopausal or postmenopausal), and treatment group. B-13 patients received surgery only; B-19 patients received postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23 patients received postoperative CMF or postoperative doxorubicin (A) and cyclophosphamide (C) (AC). The MF groups in B-13 and B-19 are not shown because of the difference in the duration of MF administration and because MF is no longer used as standard therapy.

Fig. 6.

Recurrence rates through 10 years of follow-up according to patient age at time of surgery in each of four treatment groups: B-13, surgery only (total mastectomy or lumpectomy plus breast irradiation); B-19, postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23, postoperative CMF; and B-23, postoperative doxorubicin (A) and cyclophosphamide (C) (AC).

Fig. 6.

Recurrence rates through 10 years of follow-up according to patient age at time of surgery in each of four treatment groups: B-13, surgery only (total mastectomy or lumpectomy plus breast irradiation); B-19, postoperative cyclophosphamide (C), methotrexate (M), and 5-fluorouracil (F) (CMF); B-23, postoperative CMF; and B-23, postoperative doxorubicin (A) and cyclophosphamide (C) (AC).

Table 1.

Number and percent of eligible patients in NSABP B-13, B-19, B-23 with follow-up *

Patient category  B-13 (N = 760)
 
  B-19 (N = 1095)
 
  B-23 (N = 2008)
 
   
      CMF
 
  AC
 
 
 Surg. MF MF CMF PLAC TAM PLAC TAM 
Randomly assigned 384 376 547 548 503 502 501 502 
Eligible with follow-up  369 (96.1%) 362 (96.3%) 535 (97.8%) 539 (98.4%) 496 (98.6%) 487 (97.0%) 487 (97.2%) 485 (96.6%) 
Total eligible with follow-up 731 (96.2%)  1074 (98.1%)  983 (97.8%)  972 (96.9%)  
Patient category  B-13 (N = 760)
 
  B-19 (N = 1095)
 
  B-23 (N = 2008)
 
   
      CMF
 
  AC
 
 
 Surg. MF MF CMF PLAC TAM PLAC TAM 
Randomly assigned 384 376 547 548 503 502 501 502 
Eligible with follow-up  369 (96.1%) 362 (96.3%) 535 (97.8%) 539 (98.4%) 496 (98.6%) 487 (97.0%) 487 (97.2%) 485 (96.6%) 
Total eligible with follow-up 731 (96.2%)  1074 (98.1%)  983 (97.8%)  972 (96.9%)  
*

Surg. = surgery (total mastectomy or lumpectomy plus breast irradiation) alone; MF = methotrexate, 5-fluorouracil; CMF = cyclophosphamide, methotrexate, 5-fluorouracil; AC = doxorubicin, cyclophosphamide; PLAC = placebo; TAM = tamoxifen.

Used for analyses.

Table 2.

Survival outcomes in NSABP B-13, B-19, B-23 according to treatment group and age or menopausal status *

Age (y) or menopausal status  B-13
 
      B-19
 
      B-23
 
     
  Surgery
 
   MF
 
   MF
 
   CMF
 
   CMF
 
   AC
 
  
 Pts. Events  %  Pts. Events  %  Pts. Events  %  Pts. Events  %  Pts. Events  % § Pts. Events  % § 
 Recurrence-free survival                  
≤49 214 81 58 213 54 73 323 92 70 327 52 83 535 75 86 531 84 83 
50–59 91 28 66 96 17 82 132 25 79 140 24 82 300 40 86 266 33 87 
≥60 64 15 72 53 82 80 19 74 72 11 84 148 22 83 175 21 87 
Pre 180 68 69 190 52 73 303 83 71 305 48 83 513 65 87 497 81 83 
Post 181 54 67 171 28 83 230 52 76 230 38 82 462 72 84 461 56 87 
 Overall survival                  
≤49 214 70 67 213 56 74 323 78 74 327 54 84 535 71 87 531 72 85 
50–59 91 31 65 96 26 80 132 25 80 140 28 81 300 38 87 266 35 86 
≥60 64 27 58 53 18 64 80 30 61 72 18 74 148 31 77 175 23 86 
Pre 180 57 68 190 53 74 303 73 74 305 50 84 513 60 88 497 68 85 
Post 181 69 61 171 47 74 230 59 73 230 49 79 462 79 82 461 62 86 
Age (y) or menopausal status  B-13
 
      B-19
 
      B-23
 
     
  Surgery
 
   MF
 
   MF
 
   CMF
 
   CMF
 
   AC
 
  
 Pts. Events  %  Pts. Events  %  Pts. Events  %  Pts. Events  %  Pts. Events  % § Pts. Events  % § 
 Recurrence-free survival                  
≤49 214 81 58 213 54 73 323 92 70 327 52 83 535 75 86 531 84 83 
50–59 91 28 66 96 17 82 132 25 79 140 24 82 300 40 86 266 33 87 
≥60 64 15 72 53 82 80 19 74 72 11 84 148 22 83 175 21 87 
Pre 180 68 69 190 52 73 303 83 71 305 48 83 513 65 87 497 81 83 
Post 181 54 67 171 28 83 230 52 76 230 38 82 462 72 84 461 56 87 
 Overall survival                  
≤49 214 70 67 213 56 74 323 78 74 327 54 84 535 71 87 531 72 85 
50–59 91 31 65 96 26 80 132 25 80 140 28 81 300 38 87 266 35 86 
≥60 64 27 58 53 18 64 80 30 61 72 18 74 148 31 77 175 23 86 
Pre 180 57 68 190 53 74 303 73 74 305 50 84 513 60 88 497 68 85 
Post 181 69 61 171 47 74 230 59 73 230 49 79 462 79 82 461 62 86 
*

Pts. = patients; MF = methotrexate, 5-fluorouracil; CMF = cyclophosphamide, methotrexate, 5-fluorouracil; AC = doxorubicin, cyclophosphamide; Pre = premenopausal; Post = postmenopausal. Surgery consisted of total mastectomy or lumpectomy plus breast irradiation.

Percent of patients event-free at 16 years after surgery.

Percent of patients event-free at 13 years after surgery.

§

Percent of patients event-free at 8 years after surgery.

Table 3.

First sites of treatment failure in NSABP B-13, B-19, B-23: cumulative incidence at 10 years (in percent) and cause-specific hazard ratios and 95% confidence intervals *

Event  B-13
 
   B-19
 
   B-23
 
  
 Surgery + XRT MF HR (95% CI) MF CMF HR (95% CI) CMF AC HR (95% CI) 
 (n = 369) (n = 362)  (n = 535) (n = 539)  (n = 983) (n = 972)  
Local-Regional   13.5  5.6 0.61 (.48–.78)  10.0  5.3 0.67 (.54–.83)  5.0  6.0 1.03 (.77–1.37) 
    IBTR § 15.3 2.6 0.48 (.30–.74) 8.2 4.8 0.67 (.46–.97) 3.1 5.9 2.10 (1.12–3.92) 
    Other 8.5 4.7 0.68 (.51–.91) 7.0 3.5 0.67 (.51–.88) 3.2 2.8 0.75 (.52–1.08) 
Distant 15.9 14.5 0.89 (.74–1.07) 13.8 10.1 0.83 (.70–.99) 9.8 9.3 0.96 (.78–1.19) 
Contralateral breast 6.1 4.5 0.84 (.65–1.09) 5.5 5.3 0.91 (.72–1.14) 2.6 5.6 1.09 (.72–1.66) 
Second primary 4.2 2.5 0.81 (.58–1.12) 4.1 3.4 0.99 (.75–1.3) 5.1 5.5 1.09 (.77–1.50) 
Deaths, NED 2.5 2.6 0.93 (.68–1.27) 2.9 1.7 0.71 (.49–1.03) 2.0 1.7 1.00 (.61–1.64) 
Unknown 1.1 0.8 0.72 (.35–1.47) — — — 0.3 0.3 1.00 (.25–3.99) 
Event  B-13
 
   B-19
 
   B-23
 
  
 Surgery + XRT MF HR (95% CI) MF CMF HR (95% CI) CMF AC HR (95% CI) 
 (n = 369) (n = 362)  (n = 535) (n = 539)  (n = 983) (n = 972)  
Local-Regional   13.5  5.6 0.61 (.48–.78)  10.0  5.3 0.67 (.54–.83)  5.0  6.0 1.03 (.77–1.37) 
    IBTR § 15.3 2.6 0.48 (.30–.74) 8.2 4.8 0.67 (.46–.97) 3.1 5.9 2.10 (1.12–3.92) 
    Other 8.5 4.7 0.68 (.51–.91) 7.0 3.5 0.67 (.51–.88) 3.2 2.8 0.75 (.52–1.08) 
Distant 15.9 14.5 0.89 (.74–1.07) 13.8 10.1 0.83 (.70–.99) 9.8 9.3 0.96 (.78–1.19) 
Contralateral breast 6.1 4.5 0.84 (.65–1.09) 5.5 5.3 0.91 (.72–1.14) 2.6 5.6 1.09 (.72–1.66) 
Second primary 4.2 2.5 0.81 (.58–1.12) 4.1 3.4 0.99 (.75–1.3) 5.1 5.5 1.09 (.77–1.50) 
Deaths, NED 2.5 2.6 0.93 (.68–1.27) 2.9 1.7 0.71 (.49–1.03) 2.0 1.7 1.00 (.61–1.64) 
Unknown 1.1 0.8 0.72 (.35–1.47) — — — 0.3 0.3 1.00 (.25–3.99) 
*

Hazard ratios (HR) were adjusted for clinical tumor size, age at study entry, and type of surgery. Surg. = surgery; MF = methotrexate, 5-fluorouracil; XRT = radiation therapy; CMF = cyclophosphamide, methotrexate, 5-fluorouracil; AC = doxorubicin, cyclophosphamide; IBTR = ipsilateral breast tumor recurrence; NED = no evidence of disease

Relative to all patients.

Cumulative incidence: percent at 10 yr of follow-up.

§

Relative to patients who underwent lumpectomy and breast irradiation: B-13, surg. only, 119 pts.; MF, 116 pts.; B-19, MF, 195 pts.; CMF, 194 pts.; B-23, CMF, 547 pts.; AC, 538 pts.

These studies were supported by Public Health Service grants (U10-CA-12027, U10-CA-69651, U10-CA-37377, and U10-CA-69974) from the National Cancer Institute and the Department of Health and Human Services.
We acknowledge the contributions made by the investigators who enrolled patients in the NSABP B-13, B-19, and B-23 trials. The names of those individuals are listed in the appendices that have been included in the initial reports of those studies ( 13 ) . The authors thank Tanya Spewock for editorial assistance, Mary Hof for preparation of the manuscript, and Jean Coyle for the graphic representations. Ronell Evans and John Scafidi are acknowledged for their assistance in data management.

R eferences

1
Fisher B, Redmond C, Dimitrov NV, Bowman D, Legault-Poisson S, Wickerham DL, et al. A randomized clinical trial evaluating sequential methotrexate and fluorouracil in the treatment of patients with node-negative breast cancer who have estrogen-receptor-negative tumors.
N Engl J Med
 
1989
;
320
:
473
–8.
2
Fisher B, Dignam J, Mamounas EP, Costantino JP, Wickerham DL, Redmond C, et al. Sequential methotrexate and fluorouracil for the treatment of node-negative breast cancer patients with estrogen receptor-negative tumors: eight-year results from National Surgical Adjuvant Breast and Bowel Project B-13 and first report of findings from NSABP B-19 comparing methotrexate and fluorouracil with conventional cyclophosphamide, methotrexate, and fluorouracil.
J Clin Oncol
 
1996
;
14
:
1982
–92.
3
Fisher B, Anderson S, Tan-Chiu E, Wolmark N, Wickerham DL, Fisher ER, 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
 
2001
;
19
:
931
–42.
4
Efron B. Forcing a sequential experiment to be balanced.
Biometrika
 
1971
;
58
:
403
–17.
5
Kaplan EL, Meier P. Nonparametric estimation from incomplete observations.
J Am Stat Assoc
 
1958
;
53
:
457
–81.
6
Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration.
Cancer Chemother Rep
 
1966
;
50
:
163
–70.
7
Cox DR. Regression models and life tables.
J R Stat Soc
 
1972
;
34
:
187
–200.
8
Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals.
Biometrika
 
1994
;
81
:
515
–26.
9
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.
10
Cleveland WS, Devlin SJ. Locally weighted regression: an approach to regression analysis by local fitting.
J Am Stat Assoc
 
1988
;
83
:
596
–610.
11
Gaynor JJ, Feuer EJ, Tan CC, Wu DH, Little CR, Straus DJ, et al. On the use of cause-specific failure and conditional failure probabilities: examples from clinical oncology data.
J Am Stat Assoc
 
1993
;
88
:
400
–9.
12
Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: Wiley;
1980
.
13
Kay R. Treatment effects in competing-risks analysis of prostate cancer data.
Biometrics
 
1986
;
42
:
311
–23.
14
Fisher B. Highlights from recent National Surgical Adjuvant Breast and Bowel Project studies in the treatment and prevention of breast cancer.
CA Cancer J Clin
 
1999
;
49
:
159
–77.
15
Bertino JR, Sawicki W, Lindquist C, Gupta VS. Schedule-dependent antitumor effects of methotrexate and 5-fluorouracil.
Cancer Res
 
1977
;
37
:
327
–8.
16
Gewirtz AM, Cadman E. Preliminary report on the efficacy of sequential methotrexate and 5-fluorouracil in advanced breast cancer.
Cancer
 
1981
;
47
:
2552
–5.
17
Pitman SW, Kowal CD, Papac RJ, Bertino JR. A sequential methotrexate-5-fluorouracil: a highly active drug combination in advanced squamous cell carcinoma of the head and neck.
Proc Am Soc Clin Oncol
 
1980
;
21
:
473
(abstr).
18
Shapiro CL, Gelman RS, Hayes DF, Osteen R, Obando A, Canellos GP, et al. Comparison of adjuvant chemotherapy with methotrexate and fluorouracil with and without cyclophosphamide in breast cancer patients with one to three positive axillary lymph nodes.
J Natl Cancer Inst
 
1993
;
85
:
812
–7.
19
Sutherland RL, Green MD, Hall RE, et al. Tamoxifen induces accumulation of MCF-7 human mammary carcinoma cells in the G0/G1 phase of the cell cycle.
Eur J Cancer Clin Oncol
 
1983
;
19
:
615
–21.
20
O’Brian CA, Liskamp RM, Solomon DH, Weinstein IB. Inhibition of protein kinase C by tamoxifen.
Cancer Res
 
1985
;
45
:
2462
–5.
21
Lam H-Y P. Tamoxifen is a calmodulin antagonist in the activation of cAMP phosphodiesterase.
Biochem Biophys Res Commun
 
1984
;
118
:
27
–32.
22
Pollak M, Costantino J, Polychronakos C, et al. Effect of tamoxifen on serum insulin-like growth factor - I levels of stage I breast cancer patients.
J Natl Cancer Inst
 
1990
;
82
:
1693
–7.
23
Berry J, Green BJ, Matheson DS. Modulation of natural killer cell activity by tamoxifen in stage I postmenopausal breast cancer.
Eur J Cancer Clin Oncol
 
1987
;
23
:
517
–20.
24
Mandeville R, Ghali SS, Chausseau JP. In vitro stimulation of human NK activity by estrogen antagonist (tamoxifen).
Eur J Cancer Clin Oncol
 
1984
;
20
:
983
–5.
25
Albain KS, Green SJ, Ravdin PM, Cobau CD, Levine EG, Ingle JN, et al. Adjuvant chemohormonal therapy for primary breast cancer should be sequential instead of concurrent: initial results from intergroup trial 0100 (SWOG-8814).
Proc Am Soc Clin Oncol
 
2002
;
21
:
143
(abstr).
26
Fisher B. From Halsted to prevention and beyond: advances in the management of breast cancer during the twentieth century.
Eur J Cancer
 
1999
;
35
:
1963
–73.
27
Fisher B, Jeong J-H, Bryant J, Anderson S, Dignam J, Fisher ER, et al. Treatment of lymph-node-negative, oestrogen-receptor-positive breast cancer: long-term findings from National Surgical Adjuvant Breast and Bowel Project randomised clinical trials.
Lancet
 
2004
;
364
:
858
–68.
28
Fisher B, Redmond C, Brown A, Wickerham DL, Wolmark N, Allegra J, et al. Influence of tumor estrogen and progesterone receptor levels on the response to tamoxifen and chemotherapy in primary breast cancer.
J Clin Oncol
 
1983
;
1
:
227
–41.
29
Early Breast Cancer Trialists’ Collaborative Group. Polychemotherapy for early breast cancer: an overview of the randomised trials.
Lancet
 
1998
;
352
:
930
–42.
30
Fisher B, Redmond C, Legault-Poisson S, Dimitrov NV, Brown AM, Wickerham DL, 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
–18.
31
Hutchins L, Green S, Ravdin P, Lew D, Martino S, Abeloff M, et al. CMF versus CAF with and without tamoxifen in high-risk node-negative breast cancer patients and a natural history follow-up study in low-risk node-negative patients: first results of Intergroup Trial Int 0102.
Proc Am Soc Clin Oncol
 
1998
;
17
:
1a
(abstr).
32
Martin M, Villar A, Sole-Calvo A, Gonzalez R, Massuti B, Lizon J, et al. Doxorubicin in combination with fluorouracil and cyclophosphamide (i. v. FAC regimen, day 1, 21) versus methotrexate in combination with fluorouracil and cyclophosphamide (i.v. CMF regimen, day 1, 21) as adjuvant chemotherapy for operable breast cancer: a study by the GEICAM group (Spanish Breast Cancer Research Group).
Ann Oncol
 
2003
;
14
:
833
–42.