Patients

We reviewed the records of all 598 patients with LABC who were treated on three prospective trials of combined-modality therapy for LABC at M.D. Anderson Cancer Center between 1974 and 1991. Results of these trials have been updated through May 1999. We identified 70 patients who presented with ipsilateral supraclavicular metastases but no evidence of other distant metastases.

Before chemotherapy, each patient underwent a staging work-up that consisted of a clinical history, physical examination, bilateral mammography, chest roentgenography, nuclear bone scan, nuclear liver scan (in more recent years, computed tomography or sonography of the liver, or both), and biopsies of the breast to document invasive disease. Patients with distant metastases other than ipsilateral supraclavicular metastases were excluded. In accordance with institutional policy, written consent was obtained from each patient after the research study was explained. Each patient was evaluated in a multidisciplinary planning clinic before therapy was initiated. The clinical team consisted of a radiation oncologist, a surgical oncologist, and a medical oncologist and was supported by a breast radiologist.

Treatment Plan

In the first study, which was initiated in 1974 and for which accrual was completed in 1985, the chemotherapy regimen was fluorouracil, doxorubicin, and cyclophosphamide (FAC).⁵ The treatment schema and details of the chemotherapy regimen are shown in Fig 1. After three cycles of FAC, all patients received local therapy, which consisted of a total mastectomy with axillary dissection (levels 1 and 2), radiotherapy, or both. Radiation was delivered to a total dose of 50 to 60 Gy to the breast or chest wall, internal mammary lymph nodes, and the supraclavicular fossa. Patients with either no response to induction chemotherapy or a complete clinical response to chemotherapy were treated with irradiation exclusively. Systemic chemotherapy was interrupted during local therapy and was resumed after local reactions had subsided. During the early years of the study, patients completed irradiation before systemic therapy was resumed; after 1978, irradiation was administered after all systemic chemotherapy was completed. Chemotherapy with FAC was continued until patients had received 450 mg of doxorubicin per square meter of body-surface area; subsequent maintenance therapy consisted of cyclophosphamide, methotrexate, and fluorouracil. Originally, chemotherapy was continued for a total of 2 years. After 1980, maintenance therapy with the cyclophosphamide, methotrexate, and fluorouracil combination was eliminated. After completion of three cycles of neoadjuvant chemotherapy, or earlier if there was clinical evidence of no response to systemic therapy, patients were re-evaluated in the multidisciplinary planning clinic to determine the local therapies that would be used and their sequence. At the same time, mammography of the affected breast was repeated to evaluate the response to the neoadjuvant chemotherapy. Patients with significant clinical primary tumor shrinkage after the first or second chemotherapy cycle underwent breast ultrasonography and placement of metallic markers to facilitate subsequent intraoperative localization and specimen mammography.

In the second study, which was initiated in 1985 and for which accrual was completed in 1989, the role of alternate noncross-resistant drugs was evaluated. The treatment schema and details of the chemotherapy regimen are shown in Fig 2. For their initial therapy, patients received a combination of vincristine, prednisone, doxorubicin, and cyclophosphamide (VACP), which was administered in escalating doses to tolerance. After completion of three cycles of chemotherapy, patients with objective responses underwent surgery. If the pathologic specimen had no tumor or if the amount of residual disease was less than 1 cm³, patients received five more cycles of VACP. Patients with objective responses but with an amount of residual disease greater than or equal to 1 cm³ were randomly assigned to receive five more cycles of VACP or five cycles of an alternate noncross-resistant regimen of vinblastine, methotrexate, fluorouracil and folinic acid. After completion of all chemotherapy, patients received radiotherapy. Patients with a minimal response or no response to the first three cycles of chemotherapy were treated with five cycles of vinblastine, methotrexate, fluorouracil, and folinic acid after completion of surgical therapy.

In the third study ( Fig 3), which was initiated in 1989 and completed in 1991, the feasibility of breast conservation was evaluated prospectively. All patients received FAC at the same doses and on a similar schedule as in the initial study with two exceptions: doxorubicin was administered as a continuous infusion during a 72-hour period to reduce the risk of cardiac dysfunction, and local therapy was offered after four cycles of FAC instead of after three cycles. At the conclusion of all therapy, patients older than 50 years who had estrogen receptor–positive tumors received tamoxifen for 5 years.

Patient Follow-Up

During the first 2 years of follow-up, patients underwent history and physical examinations, complete blood counts, and liver function tests every 4 months. Any abnormalities in liver function tests were followed up with sonography, computed tomography, or a bone scan as warranted. A chest x-ray was obtained every 6 months. Mammograms of the contralateral breast and, in the case of breast conservation therapy, mammograms of the treated breast were obtained every 6 months. After 2 years, these studies were performed at yearly intervals.

Data Analysis

Clinical responses to induction chemotherapy were classified according to the following five categories: (1) complete response, total resolution of the breast tumor and adenopathy as assessed by physical and radiologic examination; (2) partial response, 50% or greater reduction in the product of the two largest perpendicular dimensions of the breast mass and adenopathy; (3) minor response, less than 50% reduction in the product of the two largest perpendicular dimensions of the breast mass and adenopathy; (4) stable disease, no change in clinical status; and (5) progressive disease. Response outcomes were compared between groups with the χ² test. Multivariate analysis was performed by a logistic regression model.⁹ The Kaplan-Meier method was used to calculate overall survival from the date of diagnosis, and disease-free survival was calculated from the date of complete response, whether this was achieved with chemotherapy, surgery, or radiotherapy.¹⁰ To determine whether the overall and disease-free survival for patients with regional stage IV LABC differed significantly from that of patients with stage III LABC, we calculated overall and disease-free survival for the 239 patients with stage IIIB LABC who were treated on the three prospective trials of combined-modality therapy. The log-rank statistic was used for univariate comparisons of survival end points.¹¹ Two-tailed results are reported, and P ≤ .05 was considered statistically significant. Data were analyzed using the Statistica (Statsoft, Tulsa, OK) or SAS (SAS Institute, Cary, NC) software.

Patient Characteristics

Patient characteristics are listed in Table 1. The median follow-up time from the initiation of therapy was 11.6 years (range, 4.8 to 22.6 years). The median patient age was 49 years (range, 24 to 78 years). Estrogen receptor status was available in 39 patients. Progesterone receptor assays were performed seldom during the early trials, so progesterone receptor status was known in only 11 patients. Forty-five patients were treated on the first protocol, 13 on the second, and 12 on the third.

Response Rates

Response rates are listed in Table 2. The overall response rate (complete response plus partial response) to neoadjuvant chemotherapy was 89%. Twenty-one percent of the patients experienced clinical complete response, 67% had a partial response, 6% had a minimal response, and 4% had no change; 1% of patients experienced disease progression during chemotherapy.

Local-Regional Control

Local-regional failure was defined as recurrence in the ipsilateral chest wall, skin, or regional lymph nodes. With a median follow-up duration of 11.6 years (range, 4.8 to 22.6 years), the local-regional control rate was 81%. No statistically significant differences in the pattern of failure were observed among the three studies.

Disease-Free and Overall Survival

With a median actuarial follow-up duration of 11.6 years to date (range, 4.8 to 22.6 years), 51 patients have died. The median duration of survival was 3.5 years; the probabilities of survival to 5 and 10 years were 41% and 31%, respectively. There was no statistically significant difference between the overall survival curve for the 239 patients with stage IIIB disease and that for 70 patients with regional stage IV disease ( Fig 4). Five patients died of other causes; two patients died of congestive heart failure, one of acute leukemia, and two of unknown causes. There were no breast cancer-related deaths. No known prognostic factors were associated with significant differences in disease-free survival.

Analyses of the probability of freedom from recurrence were performed on all 70 patients. The median duration of disease-free survival was 1.9 years. The 5- and 10-year disease-free survival rates were 34% and 32%, respectively. There was no statistically significant difference between the disease-free survival curve for the 239 patients with stage IIIB disease and that for the 70 patients with regional stage IV disease ( Fig 5).

We compared the overall survival of patients with regional stage IV breast cancer with 1,531 patients with distant metastatic breast cancer who were treated with doxorubicin and cyclophosphamide at our institution.¹² The clinical outcome of the patients with regional stage IV breast cancer was significantly superior to that of patients with chemotherapy-naïve metastatic breast cancer (Fig 6). We also compared the overall survival of patients with stage IV breast cancer with patients who had oligometastatic disease (one or two sites) ( Fig 6). The patients with oligometastatic disease also experienced worse outcomes than patients who presented with regional stage IV breast cancer. ( Fig 7).

Side Effects

Myelosuppression was universal, but no treatment-related deaths occurred. The majority of patients reported nausea and vomiting. Alopecia was universal and rapidly reversible. Surgical complications were within the expected range. Skin changes after radiotherapy included minimal to moderate erythema, desquamation, and fibrosis.^5,6,7