Prediction of supraclavicular lymph node metastasis in breast carcinoma

Int. J. Radiation Oncology Biol. Phys., Vol. 52, No. 3, pp. 614 – 619, 2002 Copyright © 2002 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/02/$–see front matter

PII S0360-3016(01)02680-3

CLINICAL INVESTIGATION

Breast

PREDICTION OF SUPRACLAVICULAR LYMPH NODE METASTASIS IN BREAST CARCINOMA SHIN-CHEH CHEN, M.D.,* MIIN-FU CHEN, M.D.,* TSANN-LONG HWANG, M.D.,* TZU-CHIEH CHAO, M.D., PH.D.,* YUNG-FENG LO, M.D.,* SWEI HSUEH, M.D.,† JOSEPH T.-C. CHANG, M.D.,‡ AND WEI-MAN LEUNG, M.D.‡ Departments of *Surgery, †Pathology, and ‡Radiation Oncology, Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Taoyuan, Taiwan Purpose: Supraclavicular lymph node metastasis in breast cancer patients has a poor prognosis, and aggressive local treatment has usually resulted in severe morbidity. The purpose of this study was to select high-risk neck metastasis patients for prophylactic radiotherapy. Methods: Between 1990 and 1998, 2658 consecutive invasive breast cancer patients underwent surgery and adjuvant therapy in the hospital. The median age was 47 years (range 22–92). The median follow-up period was 39 months. The following factors were analyzed: age, tumor size, tumor location, histologic type, histologic grade, estrogen and progesterone receptor status, DNA flow cytometry study results, number of positive axillary lymph nodes, use of chemotherapy, radiotherapy, and/or hormonal therapy, and level of involved axillary nodes. Results: Of the 2658 patients, 113 (4.3%) developed supraclavicular lymph node metastasis during this period. Young age (<40 years), tumor size >3 cm, high histologic grade, angiolymphatic invasion, negative estrogen receptor status, synthetic phase fraction >4%, >4 positive nodes, and level II or III involved nodes were all significant for predicting neck metastasis in the univariate analysis. Three predictive factors were significant after multivariate analysis: high histologic grade, >4 positive nodes, and axillary level II or III involved nodes. In patients with axillary level I involved nodes and <4 positive nodes, the incidence was 4.4%. If axillary level III was involved, the rate of supraclavicular lymph node metastasis was 15.1%. Conclusion: The incidence of supraclavicular lymph node metastasis was higher in the groups with >4 positive nodes and in those with axillary level II or III involved nodes. Selective use of comprehensive radiotherapy for these high-risk patients will achieve good locoregional control. © 2002 Elsevier Science Inc. Breast cancer, Supraclavicular lymph node metastasis, Axillary lymph node metastasis, Axillary lymph node level.

patients and only those with 1–3 positive nodes and tumors ⱕ5 cm (10, 11). Moreover, the benefits of aggressive local therapy such as RT must be balanced against the potential morbidities, including radiation pneumonitis (12), brachial plexus injury (13), lymphedema (14), and second tumor occurrence (15). To our knowledge, no report has studied the prediction of SLNM and then selected high-risk patients for additional systemic or local treatment. This retrospective study attempted to define the risk factors for SLNM and select high-risk patients for whom aggressive local treatment such as RT is indicated.

INTRODUCTION The incidence of supraclavicular lymph node metastasis (SLNM) after definite surgery for breast cancer is low, but SLNM is as common as axillary nodal failure in patients after complete axillary dissection (1, 2). SLNM in breast cancer is usually considered an almost invariable signal of micrometastasis with a grave prognosis (3, 4) and was thus classified as M1 disease in the 1988 AJCC, TNM staging manual (5). A wide variety of treatments are used for SLNM, including radiotherapy (RT), chemotherapy, and surgery (6 – 8). The lack of a consensus on treatment indicates that SLNM is a difficult problem, although RT is considered mandatory (9). A worldwide overview of postmastectomy RT showed a significant survival benefit; however, those trials included only a relatively small number of

METHODS AND MATERIALS Between January 1990 and December 1998, 2658 patients with a history of primary invasive breast cancer

Reprint requests to: Shin-Cheh Chen, M.D., Department of Surgery, Chang Gung Memorial Hospital, 5, Fu-Hsing Street, Kwei-Shan, Taoyuan, Taiwan. Tel: 886-3-3281200, Ext. 3219; Fax: 886-3-3285818; E-mail: [email protected] Acknowledgments—We thank the Chang Gung Cancer Center,

Chang Gung Memorial Hospital, and Terry Fox Foundation for their support. Received Feb 21, 2001, and in revised form Sep 17, 2001. Accepted for publication Sep 20, 2001. 614

Supraclavicular metastasis in breast cancer

without systemic metastasis were treated at the surgical department of our hospital. Patients with distant metastasis were excluded, and the systemic survey included chest X-ray, bone scan, and liver ultrasound scan. All the medical files were reviewed and updated annually by the data manager and supervised by one of us (S.C.C.). The median age was 47 years (range 22–92). The median follow-up time was 39 months. Most patients underwent modified radical mastectomy without dividing the pectoralis minor muscle (n ⫽ 2351), and 307 women received breast-conserving surgery. All the patients underwent surgery in the hospital using a relatively uniform technique of axillary dissection. The axillary lymph nodes were divided into three levels on the basis of the position related to the pectoralis minor muscle. Level I consisted of the specimens lateral to the muscle, level II comprised those located under the muscle, and level III consisted of those medial to the muscle in the apex of the axilla to the Halsted ligament. Six hundred sixty-one patients underwent dissection of all 3 levels because of a clinically large tumor size or enlarged lymph nodes were found at level I during surgery, and the others underwent at least level I and/or level II dissection. Adjuvant systemic therapy was given to 1978 patients. The chemotherapy regimens included cyclophosphamide 600 mg/m2, methotrexate 40 mg/m2, and 5-fluorouracil 600 mg/m2 for high-risk, node-negative, and 1–3 node-positive patients and cyclophosphamide 500 mg/m2, epirubicin 70 mg/m2, and 5-fluorouracil 500 mg/m2 for patients with ⱖ4 positive nodes. Hormonal therapy with tamoxifen was given to those patients with positive estrogen or progesterone receptors. RT was given to 553 patients, including those patients with breast tumors ⬎5 cm or ⱖ10 positive axillary nodes and the patients who received breast-conserving surgery. The RT was delivered by 6-MV photon beam for breastconserving treatment and 9 –12 MeV electrons to the chest wall area. The radiation schedule was 2.0 Gy per fraction and 5 fractions weekly. The dose of RT was 50 Gy to the entire breast or chest wall and then a final boost to the scar area with margins for another 10 Gy. The radiation field included the chest wall, internal mammary lymph node area, and supraclavicular fossa, but not the axillary fossa, as a single portal electron treatment technique. A customized Polyflex bolus was routinely used to prevent radiation pneumonitis. Boost doses of 10 Gy to the region of the primary tumor site were given to all the patients who received RT.

Histologic review The following histologic characteristics were reviewed: histologic type, lymphovascular invasion, and histologic grading according to the modified Scarff–Bloom–Richardson grading system (SBR grade), including ductal formation, nuclear pleomorphism, and mitotic count.

● S.-C. CHEN et al.

615

Estrogen and progesterone receptor assay Estrogen and progesterone receptors were determined by the dextran-coated charcoal method, using fresh tumor tissue. Patients were considered to be estrogen receptor positive if the cytosol protein measured ⬎10 fmol/mg and were considered negative if it was ⬍3 fmol/mg. These assays were not routinely used at our institution during the period covered by the study.

Flow cytometry study Cell samples for flow cytometric study were prepared from fresh, unfixed surgical specimens preserved in the transport medium. The tissue was then cut into small fragments, gently teased to release single cells, and washed through a 55– 60-␮m nylon mesh. The cell suspension was centrifuged (1200 rpm, 6 min 4°C), and then 1 mL of Roswell Park Memorial Institute (RPMI) medium, 1 mL of fetal calf serum, and 6 mL of 70% cold alcohol were added and mixed with a Vortex mixer. The extracted cell was washed with phosphate-buffered saline and centrifuged (1200 rpm for 6 min), and 10 ␮L of RNAse (1 mg/mL) and 1 mL of propidium iodine (0.5 mg/mL) were added. The samples were run on a Coulter XL Epics Flow Cytometer (Coulter, Miami, FL) with the multicycle-III analysis software. The DNA index, a value that expresses the amount of DNA relative to normal, was calculated as the ratio of the peak channel number of the tumor G0/G1 peak to the peak channel number of the normal content G0/G1 peak. By definition, the DNA index for a diploid population is 1.0. The DNA content was defined as aneuploid if two discrete G0/G1 peaks could be confirmed. Tetraploid histograms were similar to the diploid, except that they had ⬎10% of nuclei in the G2M peak. Nondiploid histograms included aneuploid and tetraploid histograms. The synthetic phase fraction was defined as the quotient between the estimated number of S-phase cells and the total number of cells included in the cell population. The rectangular box technique was used in calculating the number of S-phase cells. SLNM was defined as the first development of tumor over the ipsilateral supraclavicular fossa in the follow-up period or simultaneously with the development of local or distant metastasis. All the axillary, supraclavicular recurrences had been proved by fine needle aspiration cytology or histologically confirmed by excisional biopsy. The following factors were included in the analysis: age (ⱕ40 vs. ⬎40 years), tumor stage (T1, T2, T3, or T4), tumor size (ⱕ3 vs. ⬎3 cm), tumor location (medial or lateral site), histologic type, histologic grade, angiolymphatic invasion, estrogen and progesterone receptors (negative vs. positive), ploidy status by DNA flow cytometry (diploid or nondiploid), synthetic phase fraction (ⱕ4% vs. ⬎4%), number of positive axillary lymph nodes, use of chemotherapy, RT, and/or hormonal therapy, and the level of axillary lymph nodes involved.

616

I. J. Radiation Oncology

● Biology ● Physics

Table 1. Numbers and patterns of regional nodal failure

Volume 52, Number 3, 2002

Table 2. Risk factors of supraclavicular lymph node metastasis

Failure (n) Variable Axilla Supraclavicular Supraclavicular only Supraclavicular and chest wall relapse Supraclavicular and distant metastasis

28 (1.1) 113 (4.3) 68 (2.6) 16 (0.6) 29 (1.1)

Numbers in parentheses are percentages.

Statistical analysis The SLNM and recurrence-free survival rates were calculated using the Kaplan–Meier product–limit method. The significance was tested using the log–rank test. A forward stepwise procedure using the Cox proportional hazard regression model was used in multivariate analysis and included any variables found to be significant in the univariate analysis. RESULTS SLNM occurred in 113 patients (4.3%) as the first site of treatment failure. This included 68 isolated supraclavicular fossa failures, 16 supraclavicular and chest wall failures, and 29 supraclavicular failures with simultaneous distant metastasis (Table 1). The median time to SLNM development was 23 months (range 3–79). An analysis of the risk factors contributing to SLNM was performed, and 11 factors were identified that significantly increased the risk of SLNM on univariate analysis (Table 2). Tumor size was a significant factor in predicting SLNM. The incidence of SLNM for T1, T2, T3, and T4 tumors was 3.2%, 4.8%, 7.0%, and 10.0%, respectively. For a tumor size of ⬍3 cm or ⬎3 cm, the incidence was 3.4% and 6.1%, respectively. High histologic SBR grade, angiolymphatic invasion, negative estrogen receptor status, and synthetic phase fraction ⬎4% by DNA flow cytometry study significantly correlated with SLNM development. The status of axillary lymph node metastasis and the number of metastatic nodes also correlated well with SLNM. Only 1.9% of the node-negative patients and 10.0% of the patients with ⬎4 positive axillary nodes developed SLNM (p ⬍0.001). The levels of axillary nodes metastasis (level I, II, III, and Rotter’s) correlated significantly with the development of SLNM. Only 5.5% of the patients with positive level I axillary nodes had SLNM and 8.3%, 9.2%, and 15.1% of the patients with Rotter’s, level II, and level III positive nodes, respectively, had SLNM (p ⬍ 0.001). On multivariate analysis, the level of axillary lymph nodes involved, number of metastatic axillary nodes, and histologic grade remained significant (Table 3). Those patients with positive level II or III nodes developed more SLNM than did those with only positive level I nodes (relative risk ⫽ 4.53, p ⫽ 0.0462). The patients with highhistologic SBR grade (Grade 2 and 3) had a higher incidence of SLNM than did those with a lower histologic grade

Patients (n)

Age (yr) ⱕ40 707 ⬎40 1951 Tumor size (cm) ⱕ3 1821 ⬎3 837 Location Lateral 1702 Medial 956 Histologic type IDC 2289 Others 369 Histologic SBR grade I 563 II 889 III 554 Unknown 652 Angiolymphatic invasion Yes 518 No 1262 Unknown 878 Estrogen receptor Negative 1593 Positive 634 Unknown 431 Progestrone receptor Negative 1571 Positive 652 Unknown 435 Ploidy status Diploid 1351 Nondiploid 570 Unknown 737 Synthetic phase fraction (%) ⱕ4 621 ⬎4 1265 Unknown 772 Axillary lymph node metastasis Negative 1440 Positive ⱕ4 nodes 670 Positive ⬎4 nodes 548 Axillary metastasis by anatomic level Level I (⫹) 885 Level II (⫹) 228 Level III (⫹) 93 Rotter’s (⫹) 12 Chemotherapy Yes 1976 No 682 Hormonal therapy Yes 807 No 1851 Radiotherapy Yes 553 No 2105 T stage T1 915 T2 1489 T3 200 T4 54

Rate of SLNM (%)

p

5.5 3.8

0.032

3.4 6.1

0.001

3.8 5.0

0.156

4.8 3.2

0.077

0.9 4.8 5.4 4.6

0.003

6.4 2.3 4.6

0.002

5.1 2.5 3.8

0.010

5.2 2.5 3.7

0.006

3.8 4.6 4.9

0.693

1.8 5.3 4.5

0.000

1.9 4.5 10.0

0.000

5.5 9.2 15.1 8.3

0.000

4.9 2.5

0.001

2.7 4.9

0.011

6.0 3.8

0.008

3.2 4.8 7.0 10.0

0.000

Abbreviations: SLNM ⫽ supraclavicular lymph node metastasis; SBR grade ⫽ Scarff-Bloom-Richardson grade; IDC ⫽ infiltrating ductal carcinoma.

Supraclavicular metastasis in breast cancer

● S.-C. CHEN et al.

617

Table 3. Univariate and multivariate analysis of supraclavicular lymph node metastasis Multivariate analysis Characteristic

Univariate analysis p

Relative risk (95% CI)

p

Size ⬎3 cm High SBR grade Negative estrogen receptor status Synthetic phase fraction ⬎4% More than 4 positive nodes Positive level II, III lymph nodes Chemotherapy Hormonal therapy Radiotherapy

0.0015 0.0028 0.0112 0.0008 0.0001 0.0001 0.0011 0.0126 0.0090

1.89 (1.28–2.79) 5.97 (1.85–19.31) 2.10 (1.18–3.71) 3.15 (1.61–6.15) 4.94 (3.35–7.30) 4.53 (2.98–6.88) 2.72 (1.49–4.97) 0.53 (0.32–0.87) 1.79 (1.16–2.76)

NS 0.0186 NS NS 0.0134 0.0462 NS NS NS

Abbreviations: NS ⫽ not significant; SBR ⫽ Scarff-Bloom-Richardson.

(relative risk ⫽ 5.97, p ⫽ 0.0186). For positive axillary node patients, patients with ⬎4 positive nodes had a higher incidence of SLNM than did those with ⱕ4 positive nodes (relative risk ⫽ 4.94, p ⫽ 0.0134). In the 166 patients who had level II or III involved nodes and high SBR grade, 21 (12.7%) developed SLNM. In the 317 patients with level I involved nodes of only axillary and low SBR grade, 2 (0.6%) developed SLNM. In patients with positive axillary level I nodes, the incidence of SLNM in the subset of patients with ⬎4 positive nodes was 8.0%, significantly higher than among the patients with ⱕ4 positive nodes (p ⫽ 0.039) (Table 4). In positive level II patients, the incidence of SLNM in the subset of patients with ⬎4 positive nodes was higher than in those with ⱕ4 positive nodes; the incidence was 10.4% and 4.4%, respectively, but this was not a statistically significant difference. In the 90 patients who had level III involved nodes and ⬎4 positive nodes, 14 (18.6%) developed SLNM. DISCUSSION Regional lymphatic relapse is difficult to manage, and the symptoms of relapse are poorly controlled by any of the treatments. Approximately 50% of the patients with reTable 4. Incidence of supraclavicular lymph node metastasis by level and number of involved axillary nodes

Level I ⱕ4 ⬎4 Level II ⱕ4 ⬎4 Level III ⱕ4 ⬎4

Total (n)

Patients with SLNM (%)

610 275

27 (4.4) 22 (8.0)

45 183

2 (4.4) 19 (10.4)

3 90

0 (0) 14 (15.6)

p 0.039 0.385

SLNM ⫽ supraclavicular lymph node metastasis.

⬍0.001

gional lymphatic relapse will have another distant metastasis, and the life expectancy is limited (4). The incidence of axillary relapse in patients who had adequate axillary level I and II dissection is low (1). Axillary local control can be achieved with a modified radical mastectomy and the selective use of comprehensive RT in patients with negative apical biopsy (1). It may not even be necessary to use axillary radiation in the subsets of patients with negative or only 1–3 positive nodes treated with axillary dissection (2). The treatment of axillary relapse is relatively easy, and good local control can be achieved with axillary clearance (18). By contrast, SLNM occurs more frequently than axillary relapse, and the treatment of SLNM is more difficult and controversial (19 –21). Fentiman et al. (1) reported that SLNM patients had longer survival than those with multiple skin recurrence, and that 80% of the patients had experienced metastasis within 4 years. Kiricuta et al. (3) reported that the 5-year survival rate of SLNM was only 16%, similar to the rate for patients with distant metastasis. However, Asaga et al. (7) reported that patients with extended surgery of lymph nodes had significantly higher disease-free survival. Others reported that RT of SLNM would permit long-term control and longer overall survival (8, 10). The wide variety of treatments and different outcome indicate that SLNM is a complex problem; it is thus important to find the high-risk patients who will develop SLNM. Axillary nodal status was the most important factor relating to adjuvant therapy and correlated well with survival and regional nodal failure. In the study by Kiricuta et al. (3), 72% of the patients who developed SLNM had positive axillary nodes when the primary tumor was diagnosed. McKinna et al. (4) found that 22 of the 27 patients who had positive axillary nodes at presentation developed SLNM. The number of positive axillary nodes also correlated with the development of SLNM. Galper et al. (2) showed that 10% of 3 positive axillary node patients developed SLNM compared with only 2% of the node-negative breast cancer patients. In our study, the axillary node status, number of positive nodes, and level of axillary involvement all corre-

618

I. J. Radiation Oncology

● Biology ● Physics

lated well with SLNM in the univariate and multivariate analyses. Only 1.9% of the node-negative patients and 4.5% of the 1– 4-positive node patients developed SLNM; however, for the patients with ⬎4 positive axillary nodes, the incidence of SLNM was 10.0%. The classic three levels of dissection (complete axillary dissection) are very important not only for accurate staging but also for better locoregional control and to prolong overall survival (22, 23), although some controversy exists (24). In our study, the rate of SLNM in patients with positive level III nodes was 15.1%, and for patients with ⬎4 positive nodes in axillary level II, it was 10.4%. The patients with positive level II, III, or Rotter’s nodes had a significantly higher incidence of SLNM compared with the patients with only positive level I nodes (relative risk ⫽ 4.5, p ⫽ 0.0462). Moreover, the level of involved axillary nodes was as powerful predictor of SLNM as the number of positive nodes in multivariate analysis. The rate of SLNM in patients with ⬎4 positive nodes increased significantly from 8.0% to 10.4% and up to 15.6% for positive nodes located in axillary levels I, II, and III, respectively. The direct spread of involved nodes from the infraclavicular space (levels II and III) to the supraclavicular fossa explains why a patient with involved nodes at axillary level II or III is more likely to have SLNM (25). Tumor size is one of the most important predictors of axillary node metastasis (9). In our study, tumor size was a significant predicting factor of SLNM in univariate analysis but lost its predictive ability on multivariate analysis. The histologic SBR grade of breast carcinoma was related to axillary lymph node metastasis (26). In the present study, patients with a high-histologic SBR grade (Grades 2 and 3) had significantly more SLNM than did low-grade patients in univariate and multivariate analysis. However, the frequency of SLNM did not increase by histologic grade for the patients in the groups with positive level II and III nodes and ⬎4 positive nodes.

Volume 52, Number 3, 2002

The effect of adjuvant therapy (chemotherapy, hormonal therapy, and RT) on the development of SLNM was controversial in our study. On univariate analysis, the rate of SLNM decreased in the patients who received tamoxifen and increased in the patients who received RT or adjuvant chemotherapy. However, all lost their predictive power in the multivariate analysis. Furthermore, it was difficult to make conclusions about the effect of adjuvant therapy on SLNM in our study because the decision to use adjuvant therapy depended on various prognostic factors, as described in the “Methods and Materials” section. Although several reports have identified risk factors related to locoregional relapse (27, 28), only a few studies, with inconsistent results, have attempted to identify the risk factors for SLNM (29, 30). The incidence of SLNM was low in the node-negative breast cancer patients: only 1.9% in the present series. For those who had ⱕ4 positive nodes, only 4.4% developed SLNM, regardless of whether the positive nodes were in axillary level I or II or that the RT over the supraclavicular fossa was not suggested. The incidence of SLNM increased to 8% in the patients with ⬎4 positive nodes in axillary level I and to 10.4% in the patients with ⬎4 positive nodes at level II. In the consensus statement, RT is recommended for patients with a tumor size of ⬎5 cm or with ⱖ4 positive axillary lymph nodes to substantially reduce the rate of locoregional relapse (9). From the current results, RT is necessary for those patients with ⬎4 positive nodes in axillary level I and is mandatory for those patients with positive nodes at axillary level II or III. In conclusion, the results of our study have shown that the extent of axillary node involvement, as well as the number of positive nodes, is an independent factor in predicting SLNM. Patients with ⬎4 positive nodes in the axilla or involvement at level II or III are at high risk of SLNM development and should be selected for RT.

REFERENCES 1. Fentiman IS, Lavelle MA, Caplan D, et al. The significance of supraclavicular fossa node recurrence after radical mastectomy. Cancer 1986;57:908 –910. 2. Galper S, Recht A, Silver B, et al. Factors associated with regional nodal failure in patients with early stage breast cancer with 0 –3 positive axillary nodes following tangential irradiation alone. Int J Radiat Oncol Biol Phys 1999;45:1157– 1166. 3. Kiricuta IC, Willner J, Kolbl O, et al. The prognostic significance of the supraclavicular lymph node metastases in breast cancer patients. Int J Radiat Oncol Biol Phys 1993;28:387– 393. 4. McKinna F, Gothard L, Ashley S, et al. Lymphatic relapse in women with early breast cancer: A difficult management problem. Eur J Cancer 1999;35:1065–1069. 5. Sobin LH, Wittekind CH, editors. UICC TNM classification of malignant tumours. 5th ed. New York: John Wiley & Sons; 1997. 6. Abranham R, Nagy T, Goss PE, et al. High dose chemotherapy and autologous blood stem cell support in women with

7.

8.

9. 10.

breast carcinoma and isolated supraclavicular lymph node metastases. Cancer 2000;88:790 –795. Asaga T, Masuzuma C, Yoshida A. Evaluation of parasternal (Ps) and supraclavicular (Sc) node dissection combined with targeting chemotherapy (activated carbon particle adsorbing aclarubicin) in breast cancer with Ps node involvement– usefulness of extended surgery for improvement of prognosis. J Jpn Surg Soc 1994;95:528 –532. Ballo MT, Strom EA, Prost H, et al. Local-regional control of recurrent breast carcinoma after mastectomy: Does hyperfractionated accelerated radiotherapy improve local control? Int J Radiat Oncol Biol Phys 1999;44:105–112. Harris JR, Halpin-Murphy P, McNeese M, et al. Consensus statement on postmastectomy radiation therapy. Int J Radiat Oncol Biol Phys 1999;44:989 –990. Overgaard M, Hansen PS, Overgaard J, et al., for the Danish Breast Cancer Cooperative Group 82b Trial. Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 1997;337:949 –955.

Supraclavicular metastasis in breast cancer

11. Ragaz J, Jackson SM, Le N, et al. Adjuvant radiotherapy and chemotherapy in node-positive premenopausal women with breast cancer. N Engl J Med 1997;337:956 –962. 12. Lingos T, Recht A, Vicine F, et al. Radiation pneumonitis in breast cancer patients treated with conservative surgery and radiation therapy. Int J Radiat Oncol Biol Phys 1991;21:355– 369. 13. Pierce S, Recht A, Lingos T, et al. Long-term radiation complications following conservative surgery (CS), and radiation therapy (RT) in patients with early stage breast cancer. Int J Radiat Oncol Biol Phys 1992;23:915–923. 14. Larson D, Weinstein, M, Goldberg I, et al. Edema of the arm as a function of the extent of axillary surgery in patients with stage I and II breast cancer treated with primary radiotherapy. Int J Radiat Oncol Biol Phys 1986;12:1575–1582. 15. Neugut A, Robinson E, Lee W, et al. Lung cancer after radiation therapy for breast cancer. Cancer 1993;71:3054 – 3057. 16. Aristei C, Marsella A, Chionne F, et al. Regional node failure in patients with four or more positive lymph nodes submitted to conservative surgery followed by radiotherapy to the breast. Am J Clin Oncol (CCT) 2000;23:217–221. 17. Bijker N, Rutgers EJT, Peterse JL, et al. Low risk of locoregional recurrence of primary breast carcinoma after treatment with a modification of the Halsted radical mastectomy and selective use of radiotherapy. Cancer 1999;85:1773–1781. 18. Grabau DA, Jensen MB, Blichert-Toft M, et al. The importance of surgery and accurate axillary staging for survival in breast cancer. Eur J Surg Oncol 1998;24:499 –507. 19. Halverson KJ, Perez CA, Kuske RR, et al. Isolated localregional recurrence of breast cancer following mastectomy: radiotherapeutic management. Int J Radiat Oncol Biol Phys 1990;19:851– 858. 20. Fisher BJ, Perera FE, Cooke AL, et al. Extracapsular axillary node extension in patients receiving adjuvant systemic therapy: An indication for radiotherapy? Int J Radiat Oncol Biol Phys 1997;38:551–559. 21. Willner J, Kiricuta IC, Kolbl O. Locoregional recurrence of

22.

23.

24.

25.

26.

27.

28.

29.

30.

● S.-C. CHEN et al.

619

breast cancer following mastectomy: Always a fatal event? Results of univariate and multivariate analysis. Int J Radiat Oncol Biol Phys 1997;37:853– 863. Canavese G, Catturich A, Vecchio C, et al. Prognostic role of lymph-node level involvement in patients undergoing axillary dissection for breast cancer. Eur J Surg Oncol 1998;24:104 – 109. Zurrida S, Morabito A, Galimberti V, et al. Importance of the level of axillary involvement in relation to traditional variables in the prognosis of breast cancer. Int J Oncol 1999;15:475– 480. Haffty BG, Ward B, Pathare P, et al. Reappraisal of the role of axillary lymph node dissection in the conservative treatment of breast cancer. J Clin Oncol 1997;15:691–700. Ellis H, Colborn GL, Skandalakis JE. Surgical embryology and anatomy of the breast and its related anatomic structures. Surg Clin North Am 1993;73:611– 632. Barth A, Craig PH, Silverstein MJ. Predictors of axillary lymph node metastases in patients with T1 breast carcinoma. Cancer 1997;79:1918 –1922. Pisansky TM, Ingle JN, Schaid DJ, et al. Patterns of tumor relapse following mastectomy and adjuvant systemic therapy in patients with axillary lymph node-positive breast cancer. Cancer 1993;72:1247–1260. Fowble B, Solin LJ, Schultz DJ, et al. Frequency, sites of relapse, and outcome of regional node failures following conservative surgery, and radiation for early breast cancer. Int J Radiat Oncol Biol Phys 1989;17:703–710. Diab SG, Hilsenbeck SG, Moor CD, et al. Radiation therapy and survival in breast cancer patients with 10 or more positive axillary lymph nodes treated with mastectomy. J Clin Oncol 1998;16:1655–1660. Recht A, Gray R, Davidson NE, et al. Locoregional failure 10 years after mastectomy, and adjuvant chemotherapy with or without tamoxifen without irradiation: Experience of the Eastern Cooperative Oncology Group. J Clin Oncol 1999;17: 1689 –1700.