Predictors of positive axillary lymph nodes after sentinel lymph node biopsy in breast cancer.

OBJECTIVE: The purpose of this study was to determine the factors that predict the presence of metastasis in nonsentinel lymph nodes (SLN) when the SLN is positive. METHODS: A prospective database was analyzed and included patients who underwent SLN biopsy for invasive breast cancer from July 1997 to August 2000 (n = 442). One hundred (22.6%) patients had one or more positive SLNs, and were analyzed to determine factors that predicted additional positive axillary nodes. RESULTS: Of the 100 patients with a positive SLN, 40 patients (40%) had additional metastasis in non-SLNs. The only significant variables that predicted non-SLN metastasis were tumor lymphovascular invasion (P = 0.004), extranodal extension (P < 0.001), and increasing size of the metastasis within the SLN (P = 0.011). In analyzing just those patients who had lymphovascular invasion, extranodal extension, and a SLN metastasis > 2mm, 92% were found to have additional positive nodes. CONCLUSIONS: In patients with invasive breast cancer and a positive sentinel lymph node, lymphovascular invasion, extranodal extension, and increasing size of the metastasis all significantly increase the frequency of additional positive nodes.

Selective analysis of the sentinel node in breast cancer.

BACKGROUND: This study was designed to determine the minimum number of sentinel nodes necessary to accurately stage patients with breast cancer. METHODS: Between August 1997 and February 2001, 509 consecutive patients were enrolled in a prospective sentinel node database. Nodes were characterized as either blue or hot (>2 times background), or both, and ranked based on the order harvested. Predictive value of the sentinel node based on these characteristics was evaluated to determine the minimum number necessary to stage the basin. RESULTS: In all, 990 sentinel nodes were harvested from 465 basins. Pathologic stage in 126 of 128 positive basins was predicted by the first or second node harvested. The remaining 2 patients were positive by immunohistochemistry only. The hottest node predicted the status in 114 of 128 basins. CONCLUSIONS: Although all nodes should be examined, these data suggest that limiting frozen section analysis to the first two sentinel nodes identified will not compromise the accuracy of staging and may provide a vehicle for resource savings.

Tamoxifen and chemotherapy for axillary node-negative, estrogen receptor-negative breast cancer: findings from National Surgical Adjuvant Breast and Bowel Project B-23.

PURPOSE: Uncertainty about the relative worth of doxorubicin/cyclophosphamide (AC) and cyclophosphamide/methotrexate/fluorouracil (CMF), as well as doubt about the propriety of giving tamoxifen (TAM) with chemotherapy to patients with estrogen receptor-negative tumors and negative axillary nodes, prompted the National Surgical Adjuvant Breast and Bowel Project to initiate the B-23 study. PATIENTS AND METHODS: Patients (n = 2,008) were randomly assigned to CMF plus placebo, CMF plus TAM, AC plus placebo, or AC plus TAM. Six cycles of CMF were given for 6 months; four cycles of AC were administered for 63 days. TAM was given daily for 5 years. Relapse-free survival (RFS), event-free survival (EFS), and survival (S) were determined by using life-table estimates. Tests for heterogeneity of outcome used log-rank statistics and Cox proportional hazards models to detect differences across all groups and according to chemotherapy and hormonal therapy status. RESULTS: No significant difference in RFS, EFS, or S was observed among the four groups through 5 years (P =.96,.8, and.8, respectively), for those aged < or = 49 years (P =.97,.5, and.9, respectively), or for those aged > or = 50 years (P =.7,.6, and.6, respectively). A comparison between all CMF- and all AC-treated patients demonstrated no significant differences in RFS (87% at 5 years in both groups, P =.9), EFS (83% and 82%, P =.6), or S (89% and 90%, P =.4). There were no significant differences in RFS, EFS, or S between CMF and AC in patients aged < or = 49 or > or = 50 years. No significant difference in any outcome was observed when chemotherapy-treated patients who received placebo were compared with those given TAM. RFS in both groups was 87% (P =.6), 87% in patients aged < or = 49 (P =.9), and 88% and 87%, respectively (P =.4), in those aged > or = 50 years. CONCLUSION: There was no significant difference in the outcome of patients who received AC or CMF. TAM with either regimen resulted in no significant advantage over that achieved from chemotherapy alone.

Variation in axillary node dissection influences the degree of nodal involvement in breast cancer patients.

BACKGROUND AND OBJECTIVES: The number of positive axillary lymph nodes predicts prognosis and is often important in determining adjuvant chemotherapy in breast cancer patients. This study was undertaken to determine if differences in the extent of axillary node dissection would alter the number of reported positive nodes. METHODS: The study population consisted of 302 patients with invasive breast cancer who underwent complete (level I/II/III) axillary lymph node dissection. Assuming that all patients had undergone a level I/II dissection, it was determined how frequently a patient's nodal category (0, 1-3, 4-9, >10 positive nodes) would have been altered if a level I or level I/II/III dissection were performed. RESULTS: Assuming that all 302 patients had undergone a level I/II dissection, performing only level I dissection would have resulted in a change in nodal category in 15.9% of all patients and 36.1% of patients with positive nodes. The corresponding changes for a level I/II/III dissection would have been 4.3% and 9.5%, respectively. CONCLUSIONS: Variations in the level of axillary node dissection for breast cancer can result in significant changes in the number of positive axillary nodes. This can potentially bias adjuvant chemotherapy recommendations if treatment decisions are based on this prognostic factor.

NCCN Practice Guidelines for Breast Cancer.

The therapeutic options for patients with noninvasive or invasive breast cancer are complex and varied. In many situations, the patient and physician have the responsibility to jointly explore and ultimately select the most appropriate option from among the available alternatives. With rare exception, the evaluation, treatment, and follow-up recommendations contained within these guidelines were based largely on the results of past and present clinical trials. However, there is not a single clinical situation in which the treatment of breast cancer has been optimized with respect to either maximizing cure or minimizing toxicity and disfigurement. Therefore, patient and physician participation in prospective clinical trials allows patients not only to receive state-of-the-art cancer treatment but also to contribute to the improvement of treatment of future patients.

Genetic induction and upregulation of cyclooxygenase (COX) and aromatase (CYP19): an extension of the dietary fat hypothesis of breast cancer.

A novel model of mammary carcinogenesis is proposed involving sequential induction and upregulation of cyclooxygenase and aromatase genes by essential fatty acids prominent in the US diet. The basic carcinogenic processes are: (1) constitutive prostaglandin biosynthesis and formation of mutagenic oxygen and nitrogen free radicals responsible for tumor initiation; (2) PGE-2-induced expression of aromatase and constitutive estrogen biosynthesis which sustains mitogenesis and tumor promotion; and (3) PGE-2-induced expression of vascular endothelial growth factor which stimulates angiogenesis and tumor metastasis.

Routine preoperative lymphoscintigraphy is not necessary prior to sentinel node biopsy for breast cancer.

BACKGROUND: This prospective study was performed to ascertain the added benefit of lymphoscintigraphy to a standard method of intraoperative lymphatic mapping and sentinel node biopsy for breast cancer. METHODS: Patients with invasive breast cancer were injected with 99mTc sulfur colloid prior to sentinel node biopsy; preoperative lymphoscintigraphy was then performed in half of the patient population. RESULTS: Sentinel node identification was possible in 45 of 50 patients (90%). All 14 patients (31%) with axillary nodal metastases had at least one histologically positive sentinel node (0% false negative rate). Lymphoscintigraphy revealed sentinel nodes in 17 of the 24 patients (70.8%) imaged. All 17 of these patients had one or more axillary sentinel nodes identified using intraoperative lymphatic mapping. In addition, 5 of 7 patients with a negative preoperative lymphoscintogram had an axillary sentinel lymph node(s) identified intraoperatively. None of the tumors showed drainage to the internal mammary lymph node chain by lymphoscintigraphy despite the fact that there were 5 patients with inner quadrant tumors. There was no significant advantage with respect to sentinel lymph node localization (91.7% versus 88.5%, P = not significant) or false negative rate (0%, both groups, P = not significant) in the group undergoing preoperative lymphoscintigraphy when compared with the patients in whom lymphoscintigraphy was not performed. CONCLUSIONS: Preoperative lymphoscintigraphy adds little additional information to intraoperative lymphatic mapping, and its routine use is not justified.

Prospective study of nonsteroidal anti-inflammatory drugs and breast cancer.

We conducted a prospective cohort study of Nonsteroidal anti-inflammatory drugs (NSAIDs) and breast cancer among 32,505 women in central Ohio, USA. After 5 years of follow-up, a total of 393 cases have been detected. The annual incidence of breast cancer per 100,00 women varied inversely with increasing intake of NSAIDs, declining from 323 among non-users to 183 among heavy users (p<0.01). Breast cancer rates decreased by about 50% with regular ibuprofen intake (p<0.01), and by about 40% with regular aspirin intake (p<0.05). The results suggest that specific NSAIDs may be effective chemopreventive agents against breast cancer.

Biodistribution and localization of radiolabeled NR-LU-10 Fab fragment in human breast cancer xenografts.

Radioimmunodetection, which takes advantage of tumor-specific or tumor-associated radio-labeled monoclonal antibodies or other biologic molecules to diagnose the extent of disease in cancer patients, has been of limited use in studies to date in patients with breast cancer. The difficulty is in finding an antibody that is both sensitive and specific enough to localize in breast tumors. This study undertook immunohistochemical and in vivo evaluation of tumor localization and biodistribution of NR-LU-10 Fab (antibody fragment) in breast tumors to determine its ability to bind selectively to malignant tissue. NR-LU-10 Fab recognizes a pancarcinoma glycoprotein antigen found on tumors of epithelial cell origin. NR-LU-10 Fab reacted with 6/6 (100%) breast cancer cell lines and 14/16 (87.5%) breast tumors with varying degrees of immunostaining intensities. Athymic mice bearing ZR-75-1 breast cancer xenografts were injected with 125I-labeled NR-LU-10 Fab (12 microg/5 microCi) and sacrificed at fixed time intervals. These studies demonstrated the highest tumor uptake of labeled Fab at 12 h postinjection (4.58+/-1.59% of injected dose/gram [% ID/g] of tissue); this gradually decreased to 0.13+/-0.05% ID/g of tissue by 72 h postinjection of the radiolabeled Fab. Biolocalization to normal tissues was as predicted for a Fab fragment; i.e., initially high in clearance organs (kidney), followed by rapid clearance over the 72-h test period. NR-LU-10 Fab displays adequate breast tumor localization with minimal biolocalization to normal tissues, thus supporting its potential use in radioimmunoscintigraphy and the RIGS system (radioimmunoguided surgery).

Keratinocyte growth factor (KGF) induces aromatase activity in cultured MCF-7 human breast cancer cells.

Estrogen is the major hormonal stimulus for growth of the hormonal-dependent type of breast cancer. The rate-limiting step in the conversion of androgens to estrogens in breast tumors is catalyzed by aromatase, one of a series of related P-450 enzymes involved in the production of steroid hormones. An interesting correlation has been found between KGF mRNA and aromatase mRNA expression in human breast tumors. Tumors that express aromatase mRNA exhibit strong KGF expression, while tumors that do not express aromatase are weak or negative for KGF expression. Thus, it is reasonable to theorize that a possible association between KGF and aromatase in controlling human breast tumor growth exists. The purpose of the current study was to establish whether there is any interaction between KGF, which is known to have epithelial-specific mitogenic activity on breast cancer cells in vitro, and the synthesis of estradiol within the hormone-dependent breast cancer epithelial cells. In the present study, we have demonstrated that KGF stimulates aromatase activity in human breast cancer cells (MCF-7) in a dose-dependent manner. Our data shows that recombinant human KGF, at a dose as low as 10 ng/ml, can significantly increase aromatase activity 2-fold over controls. In agreement with this observation, we also found that aromatase mRNA levels were increased after 10 ng/ml KGF treatment in MCF-7 cells. These results indicate that the stimulatory effect of KGF on aromatase activity may be mediated by alterations in aromatase mRNA levels or in the efficiency of the translation of the message in MCF-7 cells. In addition, our results have demonstrated that modulation of aromatase activity appears to correlate with the stimulation of proliferative activity by KGF in MCF-7 cells. These results are consistent with our previous observations that estradiol-17 beta stimulates KGF expression in human breast cancer stromal cells, leading to the speculation that breast malignant transformation is associated with a positive feedback stimulation, whereby estradiol-17 beta stimulates breast cancer stromal cell production of KGF, and KGF subsequently stimulates aromatase activity in breast cancer cells, consequently raising levels of estradiol-17 beta, in turn acts on breast stromal cells to yield more KGF. Such a positive feedback loop could play an important role in the loss of growth control in human breast cancer cells.

Estrogen-induced keratinocyte growth factor mRNA expression in normal and cancerous human breast cells.

The local recurrence rate of breast cancer has been reported to be unusually high at the surgical scar. Such breast cancer recurrence is believed to be triggered by the release of growth factors into the healing wound. Observations from an animal model have also demonstrated that KGF expression is dramatically induced by creation of full thickness wounds in mouse skin. Since KGF is an epithelial cell-specific mitogen in rat mammary epithelium, it is reasonable to speculate that KGF may be also involved in regulating human breast cancer cell growth. The purpose of the present study was to determine the effect of estradiol-17 on KGF gene expression in normal human breast stromal cells, as well as in human breast cancer stromal cells, and the mechanisms by which estradiol-17 regulates breast epithelial proliferation. Our results show that KGF expression was not effected by estradiol-17 treatment in normal human breast stromal cells. In contrast, KGF expression was stimulated by estradiol-17 in human breast cancer stromal cells. KGF mRNA levels have also been examined in normal human breast stromal cells and human breast cancer stromal cells. An interesting correlation was found between KGF expression and estradiol-17 regulation in these cell types. Normal human breast stromal cells which do not response to estradiol-17 have lower KGF mRNA level than the cancer cells which KGF expression is stimulated by estradiol-17. Our data also demonstrate that recombinant human KGF significantly stimulate normal human breast and human breast cancer epithelial cell proliferation in a dose-dependent manner. Since we have shown that estradiol-17 induces KGF mRNA expression in human breast cancer stromal cells, KGF may be involved at least in part in the stimulatory pathway that is initiated by estradiol-17 in human breast cancer epithelial cells.

Stress and immune responses after surgical treatment for regional breast cancer.

BACKGROUND: Adults who undergo chronic stress, such as the diagnosis and surgical treatment of breast cancer, often experience adjustment difficulties and important biologic effects. This stress can affect the immune system, possibly reducing the ability of individuals with cancer to resist disease progression and metastatic spread. We examined whether stress influences cellular immune responses in patients following breast cancer diagnosis and surgery. METHODS: We studied 116 patients recently treated surgically for invasive breast cancer. Before beginning their adjuvant therapy, all subjects completed a validated questionnaire assessing the stress of being cancer patients. A 60-mL blood sample taken from each patient was subjected to a panel of natural killer (NK) cell and T-lymphocyte assays. We then developed multiple regression models to test the contribution of psychologic stress in predicting immune function. All regression equations controlled for variables that might exert short- or long-term effects on these responses, and we also ruled out other potentially confounding variables. RESULTS: We found, reproducibly between and within assays, the following: 1) Stress level significantly predicted lower NK cell lysis, 2) stress level significantly predicted diminished response of NK cells to recombinant interferon gamma, and 3) stress level significantly predicted decreased proliferative response of peripheral blood lymphocytes to plant lectins and to a monoclonal antibody directed against the T-cell receptor. CONCLUSIONS: The data show that the physiologic effects of stress inhibit cellular immune responses that are relevant to cancer prognosis, including NK cell toxicity and T-cell responses. Additional, longitudinal studies are needed to determine the duration of these effects, their health consequences, and their biologic and/or behavioral mechanisms.

Androgens influence estrogen-induced responses in human breast carcinoma cells through cytochrome P450 aromatase.

The aromatase cytochrome P450 complex is responsible for the in vivo conversion of androgens to estrogens. Although breast cancer epithelial cells have been reported to have appreciable aromatase activity, its biologic significance remains uncertain. To address this, the effect of androgens on the expression of the estrogen-regulated gene pS2 in hormone-dependent human breast carcinoma cells in vitro was examined. Steroid-deprived MCF-7 cells were exposed to varying concentrations (1 nM, 10 nM, and 100 nM of androstenedione or testosterone for 2,4, and 6 days. Baseline aromatase activity was 4.9 (+/-3.1) fmol 3H2O/hour/microgram DNA [34.3 (+/-21.3) fmol/hr/10(6) cells] and was not influenced by the androgens. As an indication of estrogen biosynthesis, northern analysis was performed to quantitate pS2 mRNA expression. Although no significant pS2 induction was observed at 2 days, both 4 and 6 day exposure to 100 nM testosterone resulted in a 3-fold increase in pS2 mRNA expression. 5 alpha-dihydrotestosterone (5 alpha-DHT) failed to elicit a similar pS2 response. This testosterone-induced response was inhibited with the aromatase inhibitor 7 alpha (4'-amino) phenylthio-1,4-androstadiene-3,17-dione (7 alpha-APTADD) and with 10 microM tamoxifen. MCF-7 breast cancer cells possess endogenous aromatase activity at high enough levels to convert androgens to estrogens and elicit an estrogen-induced response. The expression of aromatase may offer a potential advantage to hormone-responsive cells, providing an additional autocrine growth pathway which may be exploited.

Seroma formation following axillary dissection for breast cancer: risk factors and lack of influence of bovine thrombin.

BACKGROUND: Seromas of the axillary space following breast surgery can lead to significant morbidity and delay in the initiation of adjuvant therapy. A prospective, randomized study was undertaken to evaluate the effect of bovine spray thrombin on seroma formation following either modified radical mastectomy (MRM) or lumpectomy with axillary dissection (LAD). In addition, risk factors for seroma formation were analyzed and identified. METHODS: A total of 101 patients were randomized to receive either bovine thrombin (20,000 units) (treatment group) or no thrombin (control group) applied to their axilla following either MRM or LAD. Drains were left in place until the preceding 24-hour drainage was < 40 milliliters. The number of days the drains were in place and wound complications (including seroma formation) were recorded. RESULTS: Forty-nine (n = 49) patients were assigned to the treatment gorup and 52 (n = 52) to the control group. MRM was performed on 60 patients (59%) and LAD oN 41 (41%). Eighteen of the 49 patients (37%) in the thrombin group developed a seroma in comparison to 21 of the 52 control patients (40%) (P = 0.71). Significant risk factors for seroma formation included increased age, patient weight, initial 72-hour wound drainage, and LAD. No statistically significant differences were observed between treatment and control groups with respect to time to drain removal, or the incidence of other wound complications. CONCLUSION: Although thrombin by itself appears to have no effect on subsequent seroma development following axillary dissection, the identification of predictive variables will be helpful in designing future trials aimed at reducing the incidence of this common complication of breast surgery.

Nonsteroidal antiinflammatory drugs and breast cancer.

We examined the association of nonsteroidal antiinflammatory drugs and breast cancer risk in a case-control study of 511 breast cancer patients and 1,534 population control subjects. The relative risk of breast cancer was reduced in women using these compounds at least 3 times per week for > or = 1 year [odds ratio (OR) = 0.66; 95% confidence interval (CI) = 0.52-0.83]. Odds ratios were similar for use of ibuprofen (0.57) or aspirin per se (0.69). Breast cancer risk declined with increasing exposure, and the greatest risk reduction (40%; odds ratio = 0.60; 95% CI = 0.40-0.91) occurred at the highest level of use (daily intake for > or = 5 years). These results indicate that nonsteroidal antiinflammatory drugs may have chemopreventive potential against the development of breast cancer.

Adjuvant CMFVP versus adjuvant CMFVP plus ovariectomy for premenopausal, node-positive, and estrogen receptor-positive breast cancer patients: a Southwest Oncology Group study.

PURPOSE: To determine whether the addition of surgical ovariectomy to standard chemotherapy prolongs disease-free survival (DFS) and overall survival in premenopausal patients with estrogen receptor (ER)-positive operable breast cancer with positive axillary nodes. PATIENTS AND METHODS: Three hundred fourteen premenopausal patients with ER-positive, node-positive breast cancer were enrolled between July 1979 and July 1989. Patients were stratified according to number of involved nodes and type of primary surgery and randomized to receive either of the following: (1) cyclophosphamide 60 mg/m2/d by mouth for 1 year, methotrexate 15 mg/m2 intravenously (i.v.) weekly for 1 year, fluorouracil (5-FU) 400 mg/m2 i.v. weekly for 1 year, vincristine .625 mg/m2 i.v. weekly for the first 10 weeks, and prednisone weeks 1 to 10 with doses decreasing from 30 mg/m2 to 2.5 mg/m2 (CMFVP); or (2) bilateral ovariectomy followed by CMFVP. RESULTS: The median follow-up time is 7.7 years and the maximum 13.2 years. Treatment arms are not significantly different with respect to either survival or DFS (one-sided log-rank, P = .55 and .70, respectively). The 7-year survival rate is 71% on the CMFVP arm and 73% on CMFVP plus ovariectomy. No significant differences were observed in node or receptor level subsets. CONCLUSION: We conclude that, in this study, the addition of ovariectomy did not improve results over chemotherapy alone in the treatment of premenopausal women with node-positive, ER-positive, operable breast cancer. Our sample size was too small to detect a small improvement. The death hazards ratio of CMFVP/CMFVP plus ovariectomy was 1.22 (95% confidence interval [CI], .79 to 1.89).

The natural history of breast cancer with more than 10 positive nodes.

BACKGROUND: Experimental protocols are being used increasingly to treat breast cancer with > 10 positive nodes. An appreciation of the natural history of this disease is crucial for choosing the optimal therapeutic approach. PATIENTS AND METHODS: We retrospectively reviewed the records of 141 patients who had breast cancer with > 10 positive nodes and received definitive therapy at our institution in the years 1969 through 1991. Because therapy evolved during this period, we compared the results from 1969 through 1981 to those from 1982 through 1991. RESULTS: Ninety-one patients (65%) were > or = 50 years of age. Fifty-four (38%) were estrogen receptor (ER) positive, the remainder were ER negative or ER status unknown. Fifty-seven (40%) had 10 to 15 positive nodes, 63 (45%) had 16 to 25, and 21 (15%) had > 25. The ratio of positive nodes to total nodes was < 50% in 22 patients, 50% to 75% in 49, and > 75% in 70. One hundred thirty-four patients (95%) underwent modified or radical mastectomy. Forty (28%) received adjuvant chemotherapy, including 16 (11%) of 58 patients treated prior to 1981. Eleven patients (8%) were treated with adjuvant radiation therapy. The median survival for all patients was 52 months, with an actuarial survival of 29% at 10 years. Patients treated after 1981 had significantly improved survival. They lived a median of 68 months postoperatively, as compared to 41 months among patients treated earlier. CONCLUSIONS: This is a high-risk group of patients, yet there is a small subset who can obtain a long survival with standard treatment modalities.

Adjuvant CMFVP versus tamoxifen versus concurrent CMFVP and tamoxifen for postmenopausal, node-positive, and estrogen receptor-positive breast cancer patients: a Southwest Oncology Group study.

PURPOSE: To compare chemohormonal therapy, chemotherapy alone, and hormonal therapy alone in postmenopausal patients with estrogen receptor (ER)-positive operable breast cancer and positive axillary nodes with respect to survival and disease-free survival (DFS). PATIENTS AND METHODS: Eight hundred ninety-two postmenopausal women with ER-positive, node-positive breast cancer were enrolled by the Southwest Oncology Group (SWOG) from July 1979 to March 1989 and 74 by the Eastern Cooperative Oncology Group (ECOG) between June 1987 and March 1989. Patients were stratified according to number of involved nodes and type of primary surgery and randomized to receive the following: (1) tamoxifen 10 mg twice daily by mouth for 1 year; (2) cyclophosphamide 60 mg/m2/d by mouth for 1 year, methotrexate 15 mg/m2 intravenously (IV) weekly for 1 year, fluorouracil (5-FU) 400 mg/m2 IV weekly for 1 year, vincristine .625 mg/m2 IV weekly for the first 10 weeks, and prednisone during weeks 1 to 10 with doses decreasing from 30 mg/m2 to 2.5 mg/m2 (CMFVP); or (3) the combination of tamoxifen and CMFVP. RESULTS: The median follow-up duration is 6.5 years, with a maximum of 12.8 years. Treatment arms are not significantly different with respect to either survival or DFS (log-rank, 2 df, P = .82 and .23, respectively). The 5-year survival rate is 77% for the tamoxifen arm, 78% for CMFVP, and 75% for the combination. No significant differences were observed in node or receptor level subsets. Severe or worse toxicity was experienced by 56% of patients on CMFVP and 61% on CMFVP plus tamoxifen, compared with 5% on tamoxifen alone. CONCLUSION: CMFVP chemotherapy, either alone or in combination with tamoxifen, has not been shown to be superior to tamoxifen alone in the treatment of postmenopausal women with node-positive, ER-positive, operable breast cancer.