Combined sequential approach in locally advanced breast cancer

Annals of Oncology 10. 305-310, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.

Original article Combined sequential approach in locally advanced breast cancer M. Zambetti, S. Oriana, P. Quattrone, P.Verderio, M. Terenziani, R. Zucali, P. Valagussa & G. Bonadonna Istituto Nazionale Tumori, Milano, Italy

Background: The interaction between primary and adjuvant chemotherapy is a crucial point in the treatment of locally advanced breast cancer. Objective: To evaluate the therapeutic efficacy of a sequential treatment with primary anthracyclines and adjuvant CMF in this patient subset. Design: Prospective cohort study. Patients: Eighty-eight breast cancer patients, stage T3b-T4 abc, N0-2> M o . Results: From February 1991 to July 1994, 88 consecutive patients with locally advanced breast cancer were treated at the Istituto Nazionale Tumori, Milano, with full-dose doxorubicin (75 mg/m 2 ) or epirubicin (120 mg/m 2 ) for three cycles followed by surgery, adjuvant chemotherapy with i.v. CMF for six cycles and local radiotherapy ± Tamoxifen. A high rate of

Introduction

The use of chemotherapy as first treatment modality for women with locally advanced breast cancer has had a major impact on the prognosis of the patients [1]. The most widely used therapeutic programs are generally comprehensive of anthracyclines and obtain high incidence of tumor regression, although the reported rate of complete remission is widely variable among different reports [2-5]. Local control is generally obtained by radical mastectomy and radiotherapy on chest wall, but in a selected subgroup of women, i.e., with tumor of limited size without diffuse edema or erythema, conservative surgery may be offered, even though this option can not yet be considered conventional [6-8]. The standard approach of sequential administration of systemic and loco-regional treatment has improved the long-term outcome of these patients, but the prognosis is still deadly in about two thirds of them atfiveyears [2, 3]. The sequential administration of anthracyclines and CMF allows for a high density schedule of the drugs that has a theoretical therapeutic edge over other schedules of administration [9]. The adjuvant treatment with a sequential administration of doxorubicin and CMF also demonstred a higher therapeutic efficacy than an alter-

objective responses (70%), but a low incidence of pathologic complete remission (2%), were observed following primary treatment with single-agent anthracyclines. Frequency of responses was not associated with tumor estrogen or progesterone receptors status, Mib-1 or grading. In 28 patients (32%) conservative surgery could be performed. At a median follow-up of 52 months, relapse free survival and overall survival are 52% and 62%, respectively. A multivariate analysis demonstrated a significant favorable prognosis in patients with limited nodal involvement at surgery and negative Mib-1 values. This drug sequence failed to significantly ameliorate the long term results in this unfavorable patient subset and more effective drug regimens and innovative therapeutic strategies are needed.

Key words: conservative surgery, locally advanced breast cancer, primary chemotherapy

nating schedule in a pivotal study performed in our institution [10]. We therefore started a study with the same drug sequence to evaluate its efficacy in locally advanced breast cancer patients. In addition, the study aimed at investigating the feasibility of conservative surgery in the same group of women. Patients and methods From February 1991 to July 1994 a total of 90 consecutive patients were treated at the Istituto Nazionale Tumori, Milano, and 88 were eligible and evaluable. Two cases were excluded because of metastatic disease in one and T 3a N o M o stage in the other (see criteria below). Locally advanced breast cancer was defined as T4 noninflammatory carcinomas and tumors > 5 cm with extension to the chest wall, unrespective to the axillary status, and without clinical and instrumental findings of metastatic disease. Women with bilateral synchronous cancer, serious cardiovascular or metabolic illnesses, pregnancy and previous neoplastic diseases were not eligible to this study. Pathological diagnosis was obtained by fine needle aspirates in 36 patients or surgical biopsy of the primary tumor in 51 women, while for one patient the only diagnostic procedure was the aspiration of clinically positive axillary nodes. In all patients, assessment of estrogen (ER) and progesterone receptors (PgR) by hystochemical methods, proliferation markers (Mib-1 determination) and tumor grade according to Scarffand Bloom, was planned. Pretreatment staging consisted of physical examination, bilateral

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

Summary

306 Table 1. Main patient characteristics at diagnosis. n (%) Total evaluable

88(100)

Age

No N, N2

ER-ICA negative PgR-ICA negative Grading 3 Mib-1 positive 1

49

31-65 55(62) 44 (50) 11(13) 38 (43) 21 (24) 18(20) 17(19) 48(55) 23 (26) 28/66(42) 36/67 (54) 30/70(43) 48/68(71)

Relapse-free (RFS) and overall survival (OS) were estimated by means of the product limit method [12] starting from the day of surgical intervention. The role of each of the prognostic variables (univariate analysis) and their joint effect (multivariate analysis) on RFS was investigated using a Cox regression model. In this model each of the regression coefficients is the logarithm of hazard ratio (HR), which is assumed constant in time. Under the null hypothesis that a variable has no prognostic role on RFS. HR is expected to be 1.00. The hypothesis of HR = 1.00 was tested by the Wald statistic. In the initial multivariate model, all the variables that were statistically significant at the level of 5% in the univariate analysis were included. Afinalmore parsimonious model was then obtained using a backward selection procedure which retained only the variables reaching the conventional significance of 5% level (final model). The variables not statistically significant at the conventional significance of 5% were excluded from the final model. The impact of each variable on clinical outcome in addition to that of the remaining variables was assessed by means of the Likelihood Ratio Test (LRT). In both univariate and multivariate analysis of independent dichotomous variables, the putative better category was considered as the reference.

According to Haagensen criteria [13]. Results

mammography, chest X-ray, bone scan, liver echography, kidney and liver function tests, cardiologic evaluation with clinical exam and ECG registration. Primary chemotherapy with i.v. bolus of doxorubicin (ADM) or epirubicin (EPI) at the dose of 75 mg/m2 and 120 mg/m2, respectively, was planned for three cycles every three weeks. The criteria for dose reduction because of iatrogenic toxicity are described elsewhere [10], Blood cell count was repeated at every cycle and kidney and liver function tests were assessed every three cycles. Cardiologic evaluation was planned before surgery and at the end of radiotherapy. At the end of primary chemotherapy, all patients underwent clinical and mammographic evaluation to assess response and plan surgery. Clinical and radiological complete response (CR) was defined as the disappearance of any tumor sign in the breast and in the axilla; partial response (PR) as a clinical reduction >50% in the product of the largest diameters. Minimal regression or disease stabilization were considered as no response. Any increase of Tor N size or the appearance of new lesions were judged as disease progression. After primary chemotherapy, surgical procedure (radical mastectomy or quadrantectomy) was empirically decided on the basis of the following clinical criteria: limited Tsize, no diffuse microcalcifications or multifocal disease and absence of extended skin edema qualified a patient for a conservative approach, while patients with extensive edema at diagnosis were not considered for conservative surgery. Adjuvant chemotherapy with intravenous CMFon days 1 and 8 every four weeks for six total cycles was planned in all patients. Irradiation of the residual breast or the chest wall was added in all cases at the end of the chemotherapy program. After conservative surgery the breast was irradiated with two tangential fields up to 50 Gy infiveweeks and with a boost of 10 Gy to the tumor bed. After mastectomy the chest wall was irradiated with electrons with a total dose of 45-50 Gy in four to five weeks. In the last 41 women Tamoxifen, 20 mg day forfiveyears, was started at the end of the entire treatment program, unrespective of ER and menopausal status. A pathological evaluation was performed based on the following prognostic factors: ER and PgR receptors by immunohistochemical methods, proliferation markers (Mib-1 determination), tumor grading according to Bloom and Richardson [11] were performed in the majority of patients, while other prognostic factors (p53. bcl-2, c-erb2, vimentin. etc.) were evaluable only in a limited fraction of women. Immunoreactivity of the Mib-1 was semiquantitatively evaluated with regard to the percentage of immunostained cells and scored as follows: negative = staining of up 10 "a of the cells: positive = staining of more than 10% of the cells.

Main patient characteristics are described in Table 1. Median age was 49 years (range 31-65) and 62% of women were premenopause. Tumor size was between 3 and 7 cm in 67% of cases. In 81% of patients axillary nodes were clinically positive and 50% of women presented unfavorable clinical signs according to the criteria of Haagensen and Stout [13]. ER-ICA and PgR-ICA determination was negative in 42% and 54% of cases, respectively, and 30 out of 70 patients presented a grade 3 tumor. Sampling at diagnosis was inadequate for assessment of each of the listed tumor characteristics in about 20%-25% of cases. Toxicity of drug treatment Primary chemotherapy with ADM (47 patients) or EPI (41 patients) was administered for three cycles with a median dose-intensity of 0.98 (0.71-1.14) for both drugs. Bone marrow toxicity was mild. WBC count less than 2500/mm 3 and neutrophil count less than 1000/mm 3 were recorded in 3% and 7% of cycles respectively. No platelets decrease below 100,000/mm 3 was observed. All patients experienced complete alopecia. Nausea and vomiting were infrequentely reported. Adjuvant CMF was administered with a dose-intensity of 0.93 (0.50-1.09). The bone marrow toxicity on WBC and neutrophils, according to the above-mentioned criteria, was respectively observed in 27% and 15% of all cycles. In 14% of treatment cycles, platelet count was under 100,000/mm3 (lowest value: 75,000/mm 3 ). Thirtyone patients (35%) experienced a transient and modest increase of liver function tests, essentially represented by pathological S-GOTand S-GPT > 1.5 UNL. In 16% of cycles, mucositis of grade 2-3 was observed and congiuntivitis was reported in 11% of cycles. Pathological ECG-signs were observed in five women:

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

Median Range Premenopausal Unfavorable presentation3 Tsize in cm 5J3.0 3.1-5.0 5.1-7.0 >7.0 Nodal status at palpation

Statistical analysis

307 Table 2. Objective response to primary chemotherapy with singleagent anthracyclines.

Mib-1 Positive Negative

70 69 72 67 76

Total evaluable Primary size at surgery ^ 2 cm Unfavorable presentation Responsive disease

Mastectomy, n (%)

Total, n (%)

28 (33) 16(57) 11 (39) 23 (82)

56 (67) 12(21) 30(53) 38 (68)

84(100) 28(33) 41 (49) 61 (73)

100

75 66 74 68

Conservative, n (%)

80 \

,

S? 60

75 67

40

75 62 0

alteration of ST-T trait was documented in two patients after epirubicin and in other two during CMF treatment; one patient had an episode of paroxysmal atrial tachycardia. All ECG alterations completely recovered within three to six months. Iatrogenic amenorrhea was recorded in 39 of 44 premenopausal women. In 12 patients menstrual cessation was reported during anthracycline therapy. Clinical response to primary chemotherapy Primary chemotherapy achieved objective remission in 62 patients (70%), including three cases with clinical complete response (3%). Disease progression while on primary chemotherapy was documented in three patients (locoregional in two; distant and locoregional in one). According to the univariate analysis, the likelihood of achieving an objective response was not significantly associated with any of the variable listed in Table 2. Surgery Eighty-four patients underwent surgery, that was not possible because of disease progression in three women and inadequate tumor reduction in a fourth patient. After primary chemotherapy, 28 patients (32%) showed a downstaging of the primary tumor to a size ^ 2 cm. Conservative surgery was offered to the majority of candidates and was performed in 33% of cases (Table 3). In all other patients, modified radical mastectomy was performed.

12

24

Months

36

48

60

55 43

35 24

15 11

Number of patients at nsk

OS RFS

84 84

82 73

73 56

Figure I. Survival after primary anthracyclines and surgery. — relapse free; overall survival.

axillary nodes were pathologically free of disease (11%), while in 20 (24%) more than 10 positive nodes were involved by tumor cells. The margins of resection were positive in 2 of the 28 patients undergoing conservative surgery, and in one of them a pathologically negative reexcision was performed. Assessment of immunohistochemical values after primary chemotherapy could be performed in the vast majority of the specimens and indicated a good to moderate level of concordance for all the three variables considered: ER-ICA 90%, PgR-ICA 88%, Mibl 81%. Since tumor grading was defined with two different methods before and after primary chemotherapy, hystological vs. nuclear, we selected not to perform any correlation analysis. Disease-free and overall survival

Among the 84 women submitted to surgery, at a median follow-up of 52 months, 36 patients relapsed and 1 woman developed a second cancer. A total of 27 women died, 25 because of breast cancer, 1 of a second malignancy and the remaining one because of fulminant hepatitis after surgery. The five-year relapse free survival is 52% and the corresponding overall survival is 62% (Figure 1). Pathological evaluation In the 36 women who showed disease relapse after surgery, the sites of involvement were the following: local Pathological complete remission was documented in 2 19%, regional 14%, loco-regional 3%, distant 47%, disof 84 patients, while persistent disease was detected in tant plus loco-regional 11%. Two patients (6%) develthe others (T ^ 2 cm: 26; T > 2 cm: 56). In 9 women the oped contralateral breast cancer. The pattern of relapse

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

Total Tumor size at palpation ^ 5 cm >5cm Menopausal status Premenopause Postmenopause Disease presentation Favorable Unfavorable ER-ICA Positive Negative PgR-ICA Positive Negative

Table 3. Type of surgery in relation to main tumor characteristics.

308 Table 4. Five-year relapse-free survival related to main characteristics %

Univariate analysis

Multivariate analysis

HR

65 42

0.01

2.49

0.002

3.5(1.5-7.6)

74 45

0.02

3.56

0.O2

3.5(1.2-10.1) 24

36 Months

67 48

0.02

2.66

0.O1

2.8(1.2-6.6)

Number of patients at nsk S3 nodes > 3 nodes

63 41

0.00

2.66

61 45

0.03

2.23

0.O2

2.3(1.1-4.7)

36 48

34

30 26

22 21

Figure 3. Relapse free survival according to degree of nodal involvement at histology. — 4 3 positive nodes; — > 3 positive nodes

Not entered into the model

Comments

Final model.

100

L

80

S 60

40 y,

o 24

36

48

Months Number of patients al nsk MIDI negative 20 M b i positive 58

1_ se 60

52

20 48

17 35

9 12

Figure 2. Relapse-free survival according to Mibl status at surgery. — negative; positive.

was similar in women undergoing radical mastectomy or breast sparing surgery, also as far as local and regional recurrences are concerned. Three and four patients in the two groups developed a local relapse only. A Cox univariate analysis showed that the extent of nodal status at histology and residual tumor size, ER-ICA, PgR-ICA and Mib-1, as assessed on surgical specimens, were able to significantly influence the fiveyear relapse-free survival (Table 4). The multivariate analysis resulted in a final model indicating that the extent of nodal involvement and the status of Mib-1 were the most important prognostic variables (Figures 2 and 3), followed by residual tumor size and estrogen receptor status (Table 4), while progesterone receptor status failed to enter the final model. Additional determinations such as c-erb-B2 by immunohistochemistry, p53 expression, bcl-2 status and histologic grading failed to significantly influence the five-year results, as did presentation at diagnosis, menopausal status and response to primary chemotherapy.

The current approach to the treatment of locally advanced breast cancer consists in the sequential application of systemic therapy and local-regional modalities. The compilation of results from several uncontrolled studies clearly showed that primary chemotherapy as an induction treatment profoundly ameliorated the relapsefree and overall survival of patients with locally advanced breast cancer [1] with respect to local-regional treatment only, or to the use of chemotherapy followed by irradiation [14]. In almost all trials published in the literature, doxorubicin-based induction chemotherapy was administered for four or three cycles, obtaining major objective response of the primary tumor and of enlarged regional lymph nodes in 60%-80% of patients [2-5]. Our study shows that primary chemotherapy with single agent anthracyclines at full dose can achieve a similar objective response rate (70%), regardless of whether doxorubicin or epirubicin were administered at the doses adopted. In addition, in spite of the fact that 50% of the patients were not candidate for surgery at presentation according to the criteria defined by Haagensen and Stout (13), 28 patients (32%) had their tumor down-staged to T0_i, 84 (95%) could undergo surgery, and 28 (33%) had a breast-sparing procedure. Local relapse had similar incidence in patients undergoing mastectomy and in those undergoing breast sparing surgery. Of note, only three patients progressed during primary chemotherapy. While the clinical goal of achieving a major reduction of the primary tumor and performing adequate surgery was obtained, the rate of clinical and pathological complete response was low (3% and 2%, respectively). These results may appear inferior to those of other groups reporting rates of 10%—20% clinical complete response, with as many as one third of them being confirmed as pathological complete response. However, the inherent difficulty of assessing a clinical complete response is well recognized in the literature, and results from different

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

Total Nodes 43 >3 Mib-1 Negative Positive Residual Tsize 4 2 cm > 2 cm ER-ICA Positive Negative PgR-ICA Positive Negative

HR (95% CI)

309 in women undergoing primary chemotherapy for operable breast cancer [22, 23]. Therefore, a reasonable goal for obtaining better long-term efficacy in addition to local control is that of defining chemotherapy programs that are capable of substantially increasing the rate of pathological complete response. The addition of new and more effective drugs, such as the taxanes paclitaxel and docetaxel, is promising, as indicated by a pilot study of paclitaxel and doxorubicin from our group [24], and supported by data of significant effects of paclitaxel on relapse-free and total survival in high-risk operable patients [25]. In addition, the frequency of complete remissions is inversely proportional to the initial tumor size, as clearly shown by recent studies in stage II breast cancer [22, 23]. Although there are no data available to support the hypothesis, it is again reasonable that more prolonged application of non cross resistant chemotherapy before surgery may obtain a higher rate of pathological complete response in the presence of a large tumor burden such as in women with locally advanced breast cancer. This hypothesis is now being tested in a pilot study ongoing in our Institute with the sequential administration of doxorubicin plus paclitaxel, vinorelbine plus fluorouracil and intermediate dose cyclophosphamide for 10 total cycles before surgery.

References 1

2.

3.

4.

5.

6.

7.

8.

9. 10.

11.

Hortobagyi GN, Buzdar AU, Strom EA et al. Primary chemotherapy for early and advanced breast cancer. Cancer Lett 1995; 90: 103-9. Hortobagyi GN. Ames FC, Buzdar MD et al. Management of stage III primary breast cancer with primary chemotherapy, surgery and radiation therapy. Cancer 1988; 62: 2507-16. Valagussa P, Zambetti M, Bignami Pet al. T 3 b-T 4 breast cancer: Factors affecting results in combined modality treatments. Clin Expl Metast 1983; 1: 191-202. PerlofT M, Lesnick GT, Korzun FC et al. Combination chemotherapy with mastectomy or radiotherapy for stage III breast carcinoma. A Cancer and Leukemia Group B study. J Clin Oncol 1988; 6: 261-9. Calais G, Descamps P, Chapet S et al. Primary chemotherapy and radiosurgical breast-conserving treatment for patients with locally advanced operable breast cancer. Int J Radiat Oncol Biol Phys 1993; 26: 37-42. Singletary SE, McNeese MD, Hortobagyi GN. Feasibility of breast-conservation surgery after induction chemotherapy for locally advanced breast carcinoma. Cancer 1992; 69: 2849-52. Touboul E, Buffat L, Lefranc JP et al. Possibility of conservative local treatment after combined chemotherapy and preoperative irradiation for locally advanced non-inflammatory breast cancer. Int J Radiat Oncol Biol Phy 1996: 34: 1019-28. Merajver SD, Weber BL, Cody R et al. Breast conservative and prolonged chemotherapy for locally advanced breast cancer: The University of Michigan experience. J Clin Oncol 1997; 15: 287381. Norton L. Conceptual basis for advances in the systemic drug therapy of breast cancer. Semin Oncol 1997; 24: SI 1-2. Bonadonna G, Zambetti M, Valagussa P. Sequential versus alternating doxorubicin and CMF regimens in breast cancer with more than three positive axillary nodes. Ten-year results. JAMA 1995; 273: 542-7. Bloom HJG. Richardson WV. Histological grading and prognosis in breast cancer. Br J Cancer 1957; 11: 359-77.

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

groups cannot be compared. A similar caution must also be applied when considering the different rates of pathological complete response because of the highly variable criteria used for defining the lack of residual tumor cells after induction chemotherapy. Differently from other reports [15-17], none of the clinical and biological variables taken into account in our study, including ER status, was able to predict for the achievement of an objective response to primary chemotherapy. In addition, the analysis of concordance between assessment of selected tumor variables at diagnosis and after primary chemotherapy, that was possible in 71% of our patients, failed to support the hypothesis indicated by other investigators that systemic treatment may cause phenotypic changes of the tumor population, possibly because of selection due to heterogeneous sensitivity to treatment within the tumor [18-21]. The five-year relapse-free (52%) and overall survival (62%) are in the range of the best multimodality programs starting with induction chemotherapy in patients with locally advanced breast cancer. Prognosis was mainly associated with classical variables assessed at surgery, such as extent of nodal involvement, Mib-1, residual tumor size and estrogen receptor status and the first two variables were the most important prognostic factors in a multivariate analysis. Presentation at diagnosis and clinical response to primary chemotherapy had no value as determinants of long-term efficacy. The even distribution of the patients meeting or not meeting the criteria for a 'poor' presentation according to Haagensen allows for the tentative conclusion that clinical criteria for defining locally advanced breast cancer as surgically operable or inoperable have little weight on the overall outcome of a multimodality treatment. The inability of the response to primary chemotherapy to predict for long-term results is not surprising, given the low rate of complete remission observed in the present study. Selected trials of primary chemotherapy in stage II breast cancer have recently indicated that pathological complete response is a strong determinant of long-term treatment outcome [22, 23]. The design of the overall treatment program applied in this study was prompted by our aim of mimicking, in this patient subset, the approach of sequential, dosedense administration of anthracyclines and CMF that proved significantly superior to alternating cycles of the same chemotherapy in women with operable breast cancer and metastasis to more than 3 axillary lymph-nodes [10]. With all the limitations inherent to a comparison with data from the literature, the hope for improving therapeutic results in locally advanced breast cancer found no support in the present uncontrolled study. In this study we introduced surgery after the initial treatment with anthracyclines, which has decreased the intensity of chemotherapy between anthracycline and CMF administration and may have had an influence on final treatment outcome. As already mentioned, the ability to reach a pathological complete response is a strong prognostic factor

310

22. 23. 24.

25.

progesterone receptors in patients receiving preoperative chemotherapy for locally advanced breast carcinoma. Ann Surg 1996; 62: 162-5. Bonadonna G, Valagussa P, Brambilla C et al. Primary chemotherapy in operable breast cancer: Eight-year experience at Milan Cancer Institute. J Clin Oncol 1998; 16: 93-100. Fisher B, Bryant J, Wolmark N et al. Effects of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 1998; 16: 2672-85. Moliterni A, Tarenzi E, Capri G et al. Pilot study of primary chemotherapy with doxorubicin plus paclitaxel in women with locally advanced or operable breast cancer. Semin Oncol 1997; 24 (Suppl 17): S17-14. Henderson IC, Berry D, Demetri G et al. Improved disease free and overall survival from the addition of sequential paclitaxel but not from the escalation of doxorubicin dose level in the adjuvant chemotherapy of patients with node-positive primary breast cancer. Am Soc Clin Oncol 1998; 390.

Received 11 November 1998; accepted 2 February 1999.

Correspondence to: M. Zambetti, MD Istituto NationaleTumori ViaVenezian I 20133 Milan Italy

Downloaded from https://academic.oup.com/annonc/article-abstract/10/3/305/155943 by guest on 14 January 2019

12. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53' 457-81. 13. Haagensen CD, Stout AP. Carcinoma of the breast - criteria of unoperability. Ann Surg 1943; 118: 859-66. 14. De Lena M, Zucali R, Viganotti G et al. Combined chemotherapy - radiotherapy approach in locally advanced (T3b-T4) breast cancer. Cancer Chemother Pharmacol 1978, 1. 53-9. 15. Billgren A, Rutqvist LE, Skoog L et al. Changes of proliferating fraction during neoadjuvant chemotherapy of primary breast cancer as a predictor of objective response Breast Cancer Res Treat 1994; 32: 63. 16. Makris A, Powles TT, Allred AC et al. Changes in hormone receptors and proliferation markers in Tamoxifen treated breast cancer patients and the relationship with response. Breast Cancer Res Treat 1998; 48: 11-20. 17. Ellis PA, Smith TE, Salter J et al. Preoperative neoadjuvant chemotherapy for early breast cancer induces apoptois. Proc Am Soc Clin Oncol 1998; 15: A112. 18. Bottini A, Berruti A, Bersipa A et al. Effect of neoadjuvant chemotherapy on Ki67 labelling index, c-erbB-2 expression and steroid hormone receptor status in human breast tumors. Anticancer Res 1996; 16: 3105-10. 19. Lo SS, Wang HC, Shyr YM et al. Can the hormonal receptor status of primary breast cancer be alterated by neoadjuvant chemotherapy? J Surg Oncol 1994; 57. 94-6. 20. Mall VM, Chumas J. Morphologic effects of neoadjuvant chemotherapy in locally advanced breast cancer. Path Res Pracl 1997; 193: 187-96. 21. Jain V, Landry M, Levine EA. The stability of estrogen and