Treatment of large and locally advanced breast cancers using neoadjuvant chemotherapy

Treatment Breast

of Large and Locally Advanced Cancers Using Neoadjuvant Chemotherapy

Amin Eltahir, FRCS (Ed), Steven D. Heys, MD, PhD, FRCS (Glas), Andrew W. Hutcheon, MD, FRCP, Tarun K. Sarkar, FRCR, Ian Smith, MBChB, Leslie G. Walker, PhD, Antoine K. Ah-See, ChM, Oleg Eremin, MD, FRACS, Aberdeen, Scotland

BACKGROUND: Neoadjuvant (primary) chemotherapy is being used increasingly in the treatment of patients with large and locally advanced breast cancer with the aim of reducing the size of the primary tumor and,eliminating micrometastatic disease. Response rates to, compliance with, and survival of patients following neoadjuvant chemotherapy have been variable. We report the results of a consecutive series of 77 patients with breast cancer who received neoadjuvant chemotherapy. METHODS: Seventy-seven patients with locally advanced breast cancers were treated with multimodality therapy comprising up to six cycles of chemotherapy (cyclophosphamide, vincristine, doxorubicin, and prednisolone), radiotherapy, and then surgery. The median follow-up was 54 months. Clinical response rates to therapy and overall survival have been documented. In addition, prognostic factors for survival were identified using the Cox proportional hazards model. RESULTS: The overall objective response rate of the primary tumor to chemotherapy alone was 87% (25% complete and 82% partial responses, UICC criteria). Following radiotherapy the response rate was 90% (52% complete and 38% partial responses). The overall 5-year survival for all patients was 0.48. However, the probability of survival at 5 years was 0.74 in those with a complete response, and 0.38 if there was a partial clinical response, but no patients who had either stasis of disease or progression survived for 5 years. Independent predictors of better survival that were identified were a complete histopathological response after chemotherapy and radiotherapy, a complete clinical response to chemotherapy, and five or six cycles of chemotherapy versus four or less.

From the Surgical Nutrition and Metabolism Unit and the De3artments of Surgery, Medical Oncology, and Clinical Oncology, Jniversity of Aberdeen, Aberdeen, United Kingdom. Requests for reprints should be addressed to S. D. Heys, MD, ‘hD, FRCS (Glas), Department of Surgery, Medical School Buildng, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, Scotland, United Kingdom. Manuscript submitted February 26, 1997 and accepted in reked form July 11, 1997.

0 1996 by Excerpta All rights reserved.

Medica,

Inc.

CONCLUSIONS: Neoadjuvant chemotherapy in patients with large and locally advanced breast cancers can result in satisfactory local control and overall survival rates, especially in patients with a complete clinical or histopathological response after treatment. Am J Surg. 1998;175: 127-132. 0 1998 by Excerpta Medica, Inc.

T

he treatment of patients with large and/or locally advanced breast cancers by mastectomy has been associated with a high risk of loco-regional recurrence and metastatic disease recurrence, resulting in a poor 5-year survival.“2 In order to minimize loco-regional recurrence and improve long-term prognosis, treatment of large and locally advanced breast cancers has involved the use of neoadjuvant (primary) chemotherapy. This form of therapy has resulted in substantial decreases in the size of the primary tumor, with approximately 70% of patients having a response (complete and partial, standard UICC criteria).’ U sin g a variety of chemotherapeutic regimens, response rates have ranged from 51% to 93%.4-” In contrast, histopathological response rates are markedly less, with approximately 10% of patients (ranging from 3% to 18%) having no residual tumor in the breast following chemotherapy.’ ‘-16 In an attempt to improve these response rates, and thus survival, different and more complicated approaches have been employed. For example, attempts have been made to synchronise tumor cell DNA synthesis prior to administration of chemotherapy.12 Although a high clinical response rate (over 90%) has been documented, this was associated with a substantial toxicity and difficulties in ensuring patient compliance. An alternative approach has been to administer a continuous infusion of S-fluorouracil for a period of 6 months, together with concurrent intravenous bolus injections of epirubicin and cisplatin at .&weekly intervals. In a pilot study of 50 patients (tumors ranging from 3 to 12 cm diameter) the complete clinical response rate was 66%.17 However, treatment delays and toxicities were common, frequently necessitating cessation of treatment (10%). In addition, 10% of patients had developed disease recurrence after only a median follow-up of 15 months. We have previously reported that the treatment of patients with bolus combination chemotherapy (doxorubicin, vincristine, cyclophosphamide, prednisolone) resulted in a response rate of 72%, and a 3-year survival of 65%.” The aim of the current study was to update our experience of these patients and include additional patients with large 0002-961 O/98/$1 PII sooo2-910(97)00279-l

9.00

127

TABLE

I Tumor TNM

T2 (n = 10) T3(n = 17) T4(n=50)

and/or locally advanced treated with neoadjuvant tween 1986 and 1992.

PATIENTS

Stage

NO

Staging Nl

N2

6 14 18

4 1 19

0 2 13

breast cancers who have been chemotherapy consecutively be-

AND METHODS

Patients Seventy-seven women with large or locally advanced breast cancers (T2 >4 cm, T3, T4 or N2, MO) were treated in the Professorial Surgical Unit, Aberdeen Royal Infirmary between 1986 and 1992. The diagnosis was confirmed by mammography, fine needle aspiration cytology, and TruCut needle biopsy, if required. All patients underwent staging investigations comprising serum urea, electrolytes and liver function tests, hemoglobin and full blood count, plain radiographs of chest, pelvis and lumbar spine, and isotope bone scans. Abdominal ultrasound was also performed if the patients’ liver function tests were abnormal. Treatment Protocol Patients received pulses of chemotherapy (detailed below) at Zl-day intervals. Of these patients, 40 received 3 days of L-arginine supplementation (30g per day for 3 days) given immediately prior to the chemotherapy as part of a feasibility study. This was followed by radiotherapy to the breast and lymph draining areas and then surgery (quadrantectomy or mastectomy) was performed. Axillary surgery was left to the discretion of the surgeon concerned. Chemotherapeutic regimen. All patients received four cycles of combination chemotherapy, given at Zl-day intervals, comprising cyclophosphamide 1000 mg/m*, vinc50 mg/m’ (all given as ristine 1.4 mg/m’, doxorubicin intravenous bolus injections), and prednisolone 40 mg/day orally for 5 days. In addition, 17 patients received a further 2 pulses of chemotherapy in view of the large tumor size at initial presentation. Radiotherapy. Three weeks following the completion of chemotherapy, patients received radiotherapy as follows: 5,000 cGY to the primary tumor and 4,500 cGY to the lymph-draining areas (axilla, supraclavicular nodes, internal thoracic chain), all given as 20 fractions over a 4-week period. Surgical treatment. Following chemotherapy and radiotherapy patients underwent surgery as follows: (a) If there was residual tumor detected clinically (>3 cm or patient preference), then a mastectomy was performed and the excised breast tissue examined histologically. If the tumour was ~3 cm, quadrantectomy was offered to the patient. (b) If there was no clinically detectable tumor and mammography showed no soft tissue mass or only residual calcification, then the patient was kept under regular clinical and mammographic surveillance. (c) If there was residual tumor detected on mammography, the patient underwent 128

THE AMERICAN

JOURNAL

OF

SURGERY@

VOLUME

175

and Size

cm

2-<5

Tumour 5-10

10 0 4

Size cm 0 17 38

s-10 cm 0 0 8

-

surgery (either quadrantectomy or mastectomy), depending on the size of the residual tumour). Treatment of the axilla was left to the discretion of the surgeon, as the protocol dictated that all patients would receive radiotherapy to the lymph-draining areas. Endocrine therapy. All patients were given tamoxifen, 20 mg/day, commencing after chemotherapy had been completed, and for a 5-year period. Assessment of Tumour Response to Therapy Clinical measurements (callipers) of the tumor were taken prior to treatment starting, 3 weeks after completion of chemotherapy, and finally, 6 weeks after completion of radiotherapy. The response was defined according to the standard UICC criteria. Partial response was a reduction of greater than 50% in the products of the two maximum perpendicular diameters, a complete response was no detectable tumor, and stasis was less than a 50% reduction or less than a 25% increase in the product of the maximum two perpendicular diameters.” Statistical Analysis The data for survival and local control of disease were examined using the Kaplan-Meier method and the log rank test where appropriate.” The combined and independent effects of the possible prognostic variables for survival were analyzed using the Cox proportional hazards model.z”~21 Stepwise procedures were used to identify the best model and to identify independent prognostic factors. Default settings were used for entry and removal of variables, ie, 0.05 for entry and 0.10 for removal.

RESULTS Patients Seventy-seven patients with locally advanced breast cancer were evaluated, with a follow-up period of up to 96 months (median 54). The patients were aged between 33 and 81 years (mean 56). Twentysix were premenopausal and 5 1 postmenopausal. The size of the primary tumors and TNM staging of the patients are shown in Table I. Response Rates After Chemotherapy Clinical assessment. All patients were treated using the protocol described above. Chemotherapy alone produced an overall clinical response rate of 87% (67 of 77). Of these, 24.7% (19 of 77) had a complete response and 62.3% of patients (48 of 77) had a partial response. Although 13% of patients (10 of 77) had no clinical response, all except 2 had stasis of the disease with no evidence of progression. The relationship between the initial tumor size and the clinical response to chemotherapy is shown in Table II. F:EBRUARY

1998

1 NEOADJUVANT TABLE

II Tumour

Size

and

Partial Response 2-15 cm 5-l 0 cm >lO cm Number djsease

8 36 4

Response

CHEMOTHERAPY

FOR

ADVANCED

BREAST

CANCEWELTAHIR

ET AL 1

to Chemotherapy Complete Response

Stasis or Progression

5 12 2

of cases showing part,a/ response, compiete or progression of disease (U/CC criteria).

1 7 2 response.

stasis

of 4”

L”

0”

Time (month)

Mammographic assessment. A mammographic complete response after chemotherapy was observed in only 12 patients (15.6%), a partial response with a residual soft tissue mass in 51 patients (66.2%), stasis of the disease was observed in 14 patients (18.2%), and tumor could not be visualized either prior to treatment or subsequently in 2 cases.

Figure 1. The effect of neoadjuvant trol of disease in patients undergoing

TABLE Ill Local Failure of Disease Control and Type Carried Out on the Breast Tumor Number Type

Clinical Response Rates After Chemotherapy and Radiotherapy Following radiotherapy, the overall clinical response rate had only increased to 89.6% (69 of 77). A notable feature was the increase in complete clinical response rate (5 1.9% of patients (40 of 77) and a decrease in partial response rate (37.7%, 29 patients); 8 patients (10.4%) showed stasis of the disease only. The mammographic response rate after radiotherapy was 87%, with a complete response in 29.9% (23 of 77 patients), a partial response in 57.1% (44 of 77 patients), and stasis of disease in 13% (10 of 77 patients). Histopathological Response Rates Patients with complete clinical response. Of the 40 patients who had a complete clinical response after completing chemotherapy and radiotherapy, 4 declined any surgical intervention (as they had a complete response) and were kept under clinical review (3 had no residual mammographic abnormality and 1 had microcalcificat-ion present on mammography). Of the other 36 patients with a complete clinical response, 7 also had a mammographic complete response. Careful histological examination of the excised breast tissue (2 had mastectomy and 5 quadrantectomy) showed no evidence of residual tumor. The remaining 29 patients with a complete clinical response had a residual soft tissue mass on mammography and histological examination of the removed breast tissue revealed residual tumor cells in all cases. Patients with a partial clinical response. All 29 patients who had a partial clinical response were also advised to undergo surgery; 5 patients declined any surgical intervention (their own wish). Although all these patients died of distant metastatic disease subsequently, 3 of them had progression of local disease prior to death. The remaining 24 patients proceeded to surgery (20 mastectomies, 4 quadrantectomies), and in all patients there was also evidence of a partial response on mammography. Histological examination of the removed breast tissue demonstrated residual malignant cells in all the specimens. Patients with static disease. In the remaining group of 8 patients who had stasis of disease as assessed clinically, 5 THE

AMERICAN

chemotherapy on local consurgery to the breast.

of Surgery

Breast conservation (n = 21) Mastectomy (n = 44)

of Surgery

of Patients with Disease Relapse

Local

4 8

patients underwent mastectomy and 3 had no surgery (these patients had inoperable locally advanced disease). Pathological assessment. A complete histological response, with no evidence of residual tumor, was found in 16 patients (20.8%). In a further 17 patients (22.106) there was evidence of only residual ductal carcinoma in situ, and in 9 patients (11.7%) the residual invasive cancer measured less than 1 cm in diameter. Residual tumor measuring 1 to 2 cm was present in 9 patients (11.7%), and residual tumour 2 to 3 cm in diameter was found in 14 patients (18.2%). No breast tissue was available for examination in 12 patients (15.6%) who did not undergo surgery. Local Disease Control The probability of local control of the disease in patients who underwent surgery is shown in Figure 1, being 0.78 at 20 and 0.74 at 60 months after diagnosis. Of the patients who did not undergo surgery (not included in this analysis), none of those wtth a complete clinical response have had local recurrence, but of those 5 with only a partial clinical response after chemotherapy and radiotherapy, 3 have had local progression of disease. The type of local treatment given to the breast in terms of local recurrence and progression of disease is shown in Table III. Overall Survival The overall 5-year probability of survival for all patients was 0.48. However, if survival was considered according to the clinical response to chemotherapy, a different pattern emerged (Figure 2). The patients with a complete clinical response had a 5-year survival probability of 0.74, and in those who had a partial clinical response this was 0.36. In contrast, none of the patients in the group with either stasis or progression of disease were alive at 5 years. Prognostic factors for survival. Table IV shows the factors examined for possible prognostic value for survival JOURNAL

OF

SURGERYe’

VOLUME

175

FEBRUARY

1998

129

EOADJUVANT

CHEMOTHERAPY

FOR

ADVANCED

BREAST

CANCEWELTAHIR

ET AL 1

1

COMMENTS

I

SDiPD

io

io

so

Time (month) Figure 2. The survival of patients according to the clinical response of their tumor after completion of chemotherapy. CR = complete response; PR = partial response; SD/PD = stasis and progression of disease. Log rank, P = 0.0001.

ool

0

4-o

50 Time

i

(month)

Figure 3. Patient survival according to the pathological response after completion of chemotherapy and radiotherapy. CR = complete pathological response; all others = residual tumour in mastectomy or breast tissue removed. Log rank, P = 0.003.

and the results of the univariate statistical analysis (log rank test). The results revealed that the following factors were associated with a better survival: complete clinical response after chemotherapy, a complete histopathological response (Figure 3), younger age, no clinical evidence of lymph node involvement, presence of detectable estrogen receptors, and absence of local recurrence of disease. However, statistical analysis using the Cox proportional hazards model (all the possible variables were entered into the model) revealed that a complete histopathological response, a better clinical response after chemotherapy, and the administration of more than four cycles of chemotherapy (as against four or less), were all independent predictors of better survival (Table V). Toxicity All patients developed alopecia, requiring the temporary use of a wig (grade IV), and all experienced various degrees of nausea and vomiting. Four patients had a mild peripheral neuropathy (grade l), thought to be secondary to vincristine, which resolved rapidly and completely. Nine patients had a l-week delay in the administration of their chemotherapy (low hemoglobin, 2; low white cell count [WCC], 3; infection, 4). Three patients required a 30% reduction in their dose of chemotherapy, due to a low WCC, but all other patients tolerated the full dose.

Neoadjuvant or primary medical chemotherapy has now been used in the treatment of patients with large and/or locally advanced breast cancers for over 20 years with varying degrees of tumor response and long-term beneiit.3-16 Response rates have also varied, but in our study reported here, the clinical response rate of 87% after chemotherapy (with 25% of patients having a complete response) is one of the highest reported in the literature. However, it is important to note that in some series the “clinical” response rate is lower because mammographic response rate has also been taken into account. Similarly, in our series of patients the mammographic complete response rate was slightly lower at 82%. The addition of radiotherapy after chemotherapy did not alter the overall response rate, but did increase the number of patients with clinical complete responses. However, higher clinical response rates have been documented in other studies. For example, the complicated regimen used by Sorace et alI3 resulted in a response rate of over 90%. Similarly, a complete clinical response rate of 66% (overall response rate was 98%) was reported by Smith et alI7 using a chemotherapy regimen requiring the continuous infusion of drugs for 6 months. In these studies, patient compliance was not good, and complications were common. Moreover, the effects of these more complicated regimens on disease recurrence and patient survival in the long term are not clear. In contrast, the regimen used in our study resulted in excellent patient compliance (no patient withdrew from the studies) and a low incidence of side effects, as already documented. Clinical response rates are believed to be important because some studies have shown that this correlated wirh patient survival.z2-“4 The patients reported in this study also demonstrated that in a univariate statistical analysis the clinical response after chemotherapy was an important determinant of survival, those patients with a complete response having a significantly increased 5-year probability of survival (0.74) when compared with those having a partial response only (0.36). All patients who had either stasis or progression of disease died within the 5-year period of assessment. If all patients are analyzed together, the overall 5-year survival was 0.48. Although this is better than some previous studies, others have reported higher survival rates following neoadjuvant chemot:herapy. However, interpretation and comparison with published studies is difficult because of the heterogeneous nature of the groups of patients being studied. It is the histological response that is probably more important in predicting patient survival. A complete histological response to chemotherapy, or an almost complete destruction of tumor cells, is associated with better longterm survival.25 In our study, histological examination of the residual tumor (mastectomy, or quadrantectomy specimens) was performed after the patient had received radio, therapy and chemorherapy. We found that a complete histological response occurred in 20.8% of all patients. However, no breast tissue was available for examination in 12 other patients, and this response rate represents, therefore, the lowest possible rate. This is in comparison with reported complete histopathological response rates of up to -I

130

THE AMERICAN

JOURNAL

OF

SURGERY@’

VOLUME

175

FEBRUARY

1998

1 NEOADJUVANT

TABLE

CHEMOTHERAPY

FOR

BREAST

CANCEWELTAHIR

ET AL

IV Prognostic

Prognostic

Factors,

Univariate

Analysis

factor

Log

Age Tumor size Clinical node status Menopausal status Estrogen receptor status Clinical response after chemotherapy Histopathological response Local recurrence Number of cycles of chemotherapy L-arginine

TABLE

ADVANCED

(Kaplan-Meier)

Rank

Degrees Freedom

38.37 6.70 25.14 32.25 2.8 9.03 6.55 0.02 0.78 0.86

3 1 2 3 3 3 1 1 1 1

of P <0.00005 0.009 <0.0001 <0.00005 0.0001 0.003 0.01 0.87 0.08 0.34

V Independent

Prognostic

Factors

for Survival

Range Histopathological response* No of cycles of chemotherapy+ Clinical response after chemotherapy*

l-2 l-2 1-3

* Residual tumour = 1, no residual tumour = 2. + More than four cycles of chemotherapy = 7, four or less cycles = 2. * Complete response = 1, p&/a/ response = 2: staas and progression b = regress,on coefficient: expb = proportional hazard: CI = confidence

Proportional expb

-0.9128 0.6933 0.8120

of dfsease interval.

10% following chemotherapy alone in patients with locally advanced breast cancers4 In the present study, those patients who had a complete histopathological response after chemotherapy and radiotherapy did have a significantly prolonged survival, when compared with those patients who had residual tumor (Figure 3), in a univariate analysis. Further multivariate statistical analysis was carried out to determine independent prognostic factors for survival in our patients. This revealed that the clinical response after cessation of chemotherapy, the histopathological response after completion of chemotherapy and radiotherapy, and the administration of more than four cycles of chemotherapy (as against less than four) were independent prognostic factors for survival (Table V). An important decision to be made following neoadjuvant chemotherapy is whether or not patients should undergo surgical resection of the residual tumor or the breast. It is difficult to predict whether or not viable tumor cells are still present following neoadjuvant chemotherapy. In our series of patients we found that if there was a complete clinical response in conjunction with a complete mammographic response (7 patients), there was no evidence of residual tumor. Therefore, it is these patients in whom it may be safe to observe the breast and not undertake any surgery. If there was a complete clinical response but a residual soft tissue mass on mammography (29 patients), all had residual tumor on histological examination. If there was a partial clinical response, then our data would indicate that these patients have residual tumor and should undergo surgery. Other workers are currently evaluating the role of magnetic resonance mammography in identifying patients who have residual tumor in the breast followTHE AMERICAN

(Cox’s b

Hazards) 95%

0.40 2.00 2.25

Cl for expb 0.21-0.75 1 .lO-3.61 0.99-5.17

P 0.004 0.02 0.055

= 3.

ing neoadjuvant chemotherapy. Preliminary results have been encouraging in small numbers of patients,z6,27 but further studies are required to confirm these early results. In summary, satisfactory loco-regional control was achieved using multimodality therapy in patients with large and locally advanced breast cancer. There were relatively few side effects and excellent patient compliance. The overall survival of patients who demonstrated a clinical response to chemotherapy was 0.74 at 5 years, and patients who had a complete histopathological response also had a better survival. However, these survival figures need to be improved, and further studies evaluating the role of new chemotherapeutic agents and regimens are urgently required.

REFERENCES 1. Haaaensen CD, Stout Al’. Carcinoma of the breast: II: criteria of operability. Ann Surg. 1943;118:859-870. 2. Arnold DJ, Lessnick OJ. Survival following mastectomy for stage III breast cancer. Am J Surg. 1979;137:362-366. 3. Heys SD, Eremin JM, Sarkar TK, et al. Role of multimodality therapy in the management of locally advanced carcinoma of the breast. .I Am Coil Surg. 1994;179:493-504. 4. Jacquillat CJ, Weii M, Baillet F, et al. Results of neoadjuvant chemotherapy and radiation in the breast-conserving treatment of 250 patients with all si-ages of infiltrative hreast cancer. Cnncer. 1990;66:119-129. 5. Morrow M, Braverman A, Thelmo W, et al. Multimodal therapy for locally advanced breast cancer. Arch Surg. 1986;121:1291-1296, 6. Hortobagyi G, Ames FC, Buzdar AU, et al. Management of stage III primary breast cancer with primary chemotherapy, surgery, and radiation therapy. Cancer. 1988;62:2507-2516. 7. Loprinzi C, Carbonc PP, Tormey DC, et al. Aggressive comJOURNAL

OF SURGERY’*

VOLUME

175

FEBRUARY

1998

131

EOADJUVANT

CHEMO

kined modality therapy for advanced local-regional breast carcinoma. J Clin Oncol. 1984;2:157-163. 8. Schwartz GF, Cantor RI, Biermann WA. Neoadjuvant chemotherapy before definitive treatment for stage III carcinoma of the breast. Arch Surg. 1987;122:1430-1434. 9. Hobar PC, Jones RC, Schouten J, et al. Multimodality treatment cd locally advanced breast carcinoma. Arch Surg. 1988;123:95 l955. 10. Luboinski G, Nagadowska, Pienkowski T. Preoperative chemotherapy in primarily inoperable cancer of the breast. Eur J Surg Oncol. 1991;17:603-607. 11. Piccart MJ, De Valeriola D, Paridaens R, et al. Six-year results of a multimodality treatment strategy for locally advanced breast cancer. Cancer. 1988;62:2501-2506. 12. Sorace RA, Bagley CS, Lichter AS, ec al. The management of nonmetasratic locally advanced breast cancer using primary induction chemotherapy with hormonal synchronization followed by radiation therapy with or without debulking surgery. World .I Surg.

1985;9:775-785. 13. Heys SD, Sarkar

TK, Ah-See AK, et al. Multomodality treatin the management of locally advanced breast cancer. J Roy C0l1 Surg Ed. 1993;38:9-15. 14. Bonadonna G, Vetonesi LJ, Brambilla C, et al. Primary chemotherapy to avold mastectomy in tumors with diameters of three centimetres or more. J Nat! Cancer Inst. 1990$2:1539-l 545. 15. Swain SM, Sorace RA, Bagley SA, et al. Neoadjuvant chemotherapy in the combined modality approach of locally advanced breast cancer. Cancer Res. 1987;47:3889-3894. 16. Gazet J-C, Coombes RC, Ford HT, et al. Assessment of the effect of pretreatment with neoadjuvant therapy on primary breast cancer. Br J Cancer. 1996;73:758-762. ment

132

THE

AMERICAN

JOURNAL

OF

SURGERY=

VOLUME

175

17. Smith IE. Walsh G, Jones A, er al. High complete remission rates with primary neoadjuvant infusional chemotherapy for large early breast cancer. ./ Clin Oncol. 1995;13:424-429. 18. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reportmg results of cancer treatment. Cancer. 1981;47:207-214. 19. Mantel N. Evaluation of survival data and two new rank statistics arising in Its consideration. Cancer Chemother Rep, 1966; 50:163-170. 20. Cox DK. Regression models and life tables (with discussion). J Roy Stat Sot GB. 1972;34:187-220, 21. Christensen E. M&variate survival analysis usmg Cox’s regression model. Hepnrology. 1987;7:1346-1358. 22. Scholl SM, Pierga JY, Asselain B, et al. Breast tumour response to primary chemotherapy predicts local and distant control as well as survival. EUTJ Cancer. 1995;31A( 12):1969-1975. 23. De Lana M, Zucali R, Viganotti G, et al. Combined chemotherapy-radiotherapy approach in locally advanced (T3b-T4) breast cancer. Cancer C&mother Pharrnacol. 1978;1:53-59. 24. Scholl SM. Asselain B, Palangie T, et al. Neoadjuvant chemotherapy in operable breast cancer. Eur 1 Cancer. 1991;27:16681671. 25. Feldman L, Hortobagyi G, Buzzdar A, et al. Pathological assessment of response to induction chemotherapy in breast cancer. Cancer Research. 1986;46:2578-2581. 26. Gilles R, Guinebretiere JM, Toussaint C, et al. Locally advanced breast cancer: contrast-enhanced subtraction MR imaging of response to preoperative chemotherapy. Radiolog?;. 1994;191:

633-638. 27. Mumtaz neoadjuvant 335-341.

FEBRUARY

H, Davidson T, Spittle M, et al. Breast surgery after treatment. Is it necessarv! Eur .I Surg Oncol. 1996;22:

1998