Efficacy of a Primary Chemotherapy Regimen Combining Vinorelbine, Epirubicin, and Methotrexate (VEM) as Neoadjuvant Treatment in 89 Patients with Operable Breast Cancer

Abstract

Introduction

Primary chemotherapy has become the standard treatment of inflammatory and locally advanced breast cancer (LABC) and has more recently been extended to the management of patients with operable disease, eligible for mastectomy, mainly in order to increase the rate of breast conservation. Preoperative (neoadjuvant) chemotherapy followed by the appropriate local treatment (either radiotherapy or surgery or both) has, therefore, become an established means of improving the outcome for patients whose disease is unlikely to be controlled by other means. The functions of chemotherapy are to produce a major objective regression of the tumor that will permit management with a more conservative surgical resection (and allow breast conservation, if possible) while also treating clinically undetectable micrometastatic disease. In addition, exposure of the tumor to chemotherapy as initial therapy enables the clinician to test the responsiveness of the malignant cells and provides an early opportunity to change the agents if the tumor appears resistant [1–3].

The appropriate selection of drugs for neoadjuvant treatment must reflect the need to use agents likely to achieve rapid control of the underlying disease. The most active groups of cytotoxic compounds in this setting will include anthracyclines, active antimetabolites, and newer classes of agents, such as second generation vinca alkaloids or taxanes, which have demonstrated high response rates in advanced disease [3–7]. Previous experience with other anthracycline-based combinations has resulted in responses that subsequently permitted breast conservation in 85% of patients [8], and it was the intention of this study to establish the activity of a novel combination that might achieve comparable results with less toxicity.

Methods

Neoadjuvant chemotherapy was administered to patients with either noninflammatory bulky tumors (>30 mm) for which extended resection would be required or those with tumors located in the central area of the breast near the nipple. All patients were required to have measurable disease not previously treated by chemotherapy or hormonal agents. Other inclusion criteria included age 18-70 years, performance status <2, life expectancy >3 months, and adequate hematological, hepatic, and renal functions. Patients were excluded if they had evidence of metastatic disease or active second malignancy and also if they had either impaired cardiac function (left ventricular ejection fraction [LVEF] <50%) or a history of ischemic myocardial disease or congestive cardiac failure.

All patients gave written consent according to the standard procedures of the treating hospital. The study protocol was designed according to the principles of the Helsinki guidelines and approved by the Ethics Committee of Centre Jean Perrin, Clermont Ferrand, France, in December 1990.

The initial assessment of patients entering the study included clinical examination, bilateral mammography, bilateral breast ultrasound, and cytological or pathological diagnosis by core biopsy. When invasive adenocarcinoma was demonstrated, the tumor was evaluated for Scarff Bloom Richardson (SBR) grading, hormone receptors were assayed by radioimmunology, and DNA analysis was done using flow cytometry (EPICS XL, Coulter; Miami, FL). The absence of distant metastases was confirmed by chest x-ray, liver ultrasound, and radionuclide bone scan. Laboratory assessment consisted of full blood count, electrolytes, serum creatinine, alkaline phosphatase, gamma glutamyltransferase, and the tumor markers carcinoembryonic antigen and CA15.3.

Treatment was administered according to the following schedule: vinorelbine (Navelbine^®) 25 mg/m², epirubicin (Farmorubicin^®) 35 mg/m², and methotrexate (Ledertrexate^®) 20 mg/m², given on day 1 and day 8 of a 28-day cycle to a total of six cycles. Patients were clinically assessed and monitored for hematological and biochemical tolerance before each cycle.

After induction chemotherapy, responding patients underwent appropriate surgery according to the size of their residual tumor. Radiotherapy was also given to patients who underwent a conservative surgical procedure or who were found to have either axillary node involvement or extracapsular invasion. Patients who were found to have extensive residual disease were permitted to proceed to adjuvant therapy adapted to the prognostic factors established at initial staging.

Response to treatment was assessed halfway through induction chemotherapy and at the end of cytotoxic treatment before surgery. The main methods utilized were clinical ultrasound and mammographic bidimensional measurement of lesions, and definitions of response were according to the World Health Organization (WHO) classification [9]. Complete responses were defined as the disappearance of all known lesions; partial responses were defined as a reduction in each lesion by at least 50%; stable disease was defined as a decrease of less than 50%, but increase of not greater than 25% with no new lesions; and progressive disease represented an increase of greater than 25% or the appearance of new lesions. Complete and partial responses were confirmed by the same techniques after a period of 1 month. After surgical resection, a separate evaluation of pathological response was performed in which microscopic evaluation of at least 15 sections from the breast specimen and axillary dissection were examined. The pathological responses were classified as follows [10]: class 1) disappearance of all tumor (pathological complete response (pCR); class 2) presence of in situ carcinoma but no invasive tumor in the breast, no tumor in the lymph nodes; class 3) presence of invasive carcinoma with alteration of stroma or cells, and class 4) few or no modifications of the tumoral appearance. A central review of all pathology and radiology was performed. Survival curves (disease-free survival and overall survival) were calculated by the Kaplan Meier method and log-rank test, and the multivariate analysis was done using the Cox Model.

Results

Between October 1991 and April 1996, 89 patients were included in the study, 87 of whom were evaluable for response (two patients were lost to follow-up). The distribution of clinical involvement showed that 73 had tumors >30 mm, and 16 had tumors situated near the nipple. The median age was 52 years (range 31-72 years), and 41 (46%) of these women were premenopausal. All patients had WHO performance status of 0. Details of the extent of the disease are shown in Table 1. The median clinical diameter of the tumors at presentation was 42 mm (range 15-120 mm).

Eighty-four (94%) patients received all the planned induction therapy, while in five patients, treatment was discontinued after three cycles because of fever and neutropenia. No patients required hospitalization for infection, and no toxic deaths were recorded. A total of 519 cycles of treatment were administered with a median of 5.9 cycles per patient. The dose intensity calculated in mg/m²/week was 11.75 (94% of theoretical dose) for vinorelbine, 17.15 (96%) for epirubicin, and 9.7 (97%) for methotrexate. Hematological toxicity is shown in Table 2 and was most significantly related to neutropenia, which was recorded as WHO grade 3-4 in 28% of cycles affecting 22 patients; no anemia or thrombocytopenia greater than grade 1 was observed, and neither red cell nor platelet transfusions were required. Nonhematological toxicity consisted of grade 3 alopecia, which occurred in 80% of patients despite the use of cooling caps. There were also less frequent treatment events, consisting of six episodes of stomatitis, four cases of venous toxicity, and six patients who reported back pain for which no explanation was found; two patients experienced headaches and two reported skin reactions. Nausea, vomiting, and diarrhea were mild and infrequently reported. No cardiac events were described, but consistent monitoring of LVEF was not undertaken. Three of 49 premenopausal patients (6%) underwent menopause in response to chemotherapy.

Clinical evaluation of the efficacy of treatment indicated objective responses in 78 of the 87 evaluable patients (90%), with 24 (28%) complete responses and 54 (62%) partial responses; in nine patients (10%) there was stable disease. The greater sensitivity provided by other methods of assessment resulted in lower overall response rates, with 71% of responses in the 87 evaluable patients confirmed by ultrasound and 64% by mammography. Assessment of the response during treatment showed that there was a small but significant increase in the overall response rate after six cycles compared with four cycles by all assessment criteria (clinical p < 0.01, ultrasound p < 0.05, mammography p < 0.06), with an additional 10% of responses achieved at the end of the scheduled treatment (Table 3). Histological material was received for all 87 patients for assessment of response (Table 4) and revealed pCR in seven patients (8%) and in situ carcinoma only in an additional five (6%), for a total overall response rate of 14% with noninvasive tumor, confirmed by pathological criteria. Twenty-four patients had tumor persistence but evidence of alteration, while 53 patients showed some unmodified tumor after therapy. A multivariate analysis (Cox model) of factors affecting the achievement of pCR failed to show any effect for age, menopausal status, receptor status, or tumor/node/metastasis (TNM) score, but identified a significant adverse effect for grade III SBR score (p < 0.03).

When the 59 patients who underwent axillary surgery were analyzed by the presence or absence of palpable lymph nodes, pathological evidence of node involvement was found in only 2 out of 15 without clinical nodal disease (13.3% of N0), but in 22 out of 44 with palpable nodes (50% of N1 or N2).

Postoperative treatment was administered to 26 patients, one of whom received adjuvant chemotherapy, four who had more than 10 positive nodes were autografted, and 21 postmenopausal patients received tamoxifen.

The study was analyzed in December, 2000, with a median follow-up of 100 months (range 39-108 months). The overall and disease-free survivals are shown in Figure 1, which demonstrate a survival at 5 years of 90% and a disease-free survival at 5 years of 78.2%. Of the 13 patients who had relapsed, four had local recurrence and one had a contralateral recurrence, while an additional eight (four with local and metastatic recurrence) had developed distant metastases, five of whom died as a result of their disease.

Figure 1

Overall survival (OS) and disease-free survival (DFS) (n = 89). Median follow-up was 86 months (range 39-108) or 7.2 years. Actuarial overall survival at 5 years was 90%. Actuarial DFS at 5 years was 78.2%.

Open in new tabDownload slide

Discussion

The goal of neoadjuvant therapy is to provide the means for good locoregional control of disease that would otherwise be difficult to manage. Initial treatment should also generate long intervals of freedom from disease and improved survival in the same way that the late application of adjuvant chemotherapy is designed to prevent recurrence. The means by which these intentions are achieved involves the “downstaging” of the tumor, which then permits conservative surgery and the production of a better cosmetic result for the patient. However, the opportunity to examine the effects of chemotherapy on the tumor also makes important prognostic information available to the oncologist. The precise criteria by which the assessment of residual disease in surgically resected specimens is made are not internationally agreed, but the recent approach used for inflammatory breast cancer can be adapted for other clinical situations [10]. When the results of published studies are compared, it is important to make a clear identification of responding patients, distinguishing between those achieving true pCR, with no evidence of tumor, and those combined as a group that included those with residual carcinoma in situ who could be termed “pathological evidence of response.” Patients in whom pCRs are achieved have demonstrated that their disease is highly susceptible to chemotherapy and benefit in terms of longer disease-free survival, while those in whom unmodified tumors remain after chemotherapy might well benefit from subsequent adjuvant treatment. The impact of chemotherapy on survival and disease-free survival given for stage I or II breast cancer is similar whether the treatment is administered before or after surgical resection, but the additional information relating to the response of the disease available in those patients who have received preoperative chemotherapy may be helpful in predicting the efficacy of subsequent adjuvant treatment [3]. The benefits in terms of disease control of giving preoperative chemotherapy are at least as great as those of adjuvant chemotherapy, while the patients additionally benefit from a greater chance of receiving a breast-sparing surgical procedure.

The selection of cytotoxic drugs that are suitable for use in neoadjuvant therapy depends, in part, on the evidence for activity against advanced disease. However, the priority to achieve activity without serious toxicity is even greater in patients who have early-stage disease and in whom it is intended that their clinical condition after treatment be sufficient for them to tolerate surgical resection of the tumor. The substitution of an active antimetabolite, such as methotrexate, for an alkylating agent (e.g., cyclophosphamide) might be expected to diminish the risks of therapy-related menopause and also produce less late effects, such as the induction of second malignancies. The VEM schedule only induced menopause in 6% of patients at risk in contrast to reported rates of 68% for CMF [11].

Anthracycline-based therapy (FAC or AC) is the gold standard neoadjuvant chemotherapy. In a recent comparative study, AC achieved a clinical response rate of 66%, with 58% of patients able to undergo conservative surgical procedures, and a pCR rate of 9%, results which were less positive than the doxorubicin/paclitaxel combination with which it was compared [12].

In our previous study of AVCF/AVCFM in 142 patients, we were able to achieve a clinical response in 85% of patients but, following the experience of other investigators [13], those who had major tumor reductions received radiotherapy alone and, therefore, the pCR rate could not be established. Of the 41 patients who received radiotherapy alone after chemotherapy, 15% later had a local relapse and, as a result, our subsequent trial was designed to incorporate surgical resection of residual disease [8].

An alternative experience of investigators affiliated with our group involved the neoadjuvant use of a combination of THP-doxorubicin, vinorelbine, cyclophosphamide, and fluorouracil given on a 3-week schedule [14]. Patients were selected for this approach on the basis of poor prognostic factors at presentation (grade III disease, negative hormone receptors) and younger age. Hematological toxicity with this combination was severe, involving grade 3-4 neutropenia in 81% of cycles despite some growth factor support, anemia in 25%, and thrombocytopenia in 20%; clinically significant episodes of infection requiring hospital admission were frequent. So, despite the fact that this schedule produced a 22% rate of pCR, there are reservations concerning the safety of this approach to treatment in an unselected patient population [14].

The goal of this trial was to evaluate the efficacy of the VEM regimen, especially by demonstrating a high level of conservative surgery and a worthwhile pathological response rate, as had been suggested by the preliminary results [15]. In terms of toxicity, the regimen was confirmed to be safe in that there were no significant infections and no toxic deaths, although the rate of grade 3-4 leukopenia was 28%. Other toxicities were mild, and the fractionated administration of all three drugs allowed full-dose administration without the problems experienced when trying to give the total anthracycline dose on day 1 [8]. The response rates of greater than 80% in this population of stage II and IIIA patients are comparable with those we obtained with our previous multidrug schedule of AVCF [8] and are reflected in an appropriately high level of conservative surgical procedures, achieved in 87% of the patients in this study. The opportunity for pathological examination of residual disease revealed that, of the 24 patients (27%) who achieved clinical complete responses, 12 patients (14%) were in pCR (class 1 and 2), and so complete clearance of disease was documented in 33% of this group. The equivalent experience in the National Surgical Adjuvant Breast Project B-18 study of preoperative AC revealed a 32.8% clinical complete response rate, with true pCR obtained in 26% of the group of responders [3], and is clearly a comparable level of efficacy (Table 5).

The outcome of this study now represents a mature experience of this approach to neoadjuvant chemotherapy with half of the patient population passing the 5-year point. The long-term achievements of preoperative chemotherapy can be identified in this patient population in whom adjuvant chemotherapy or dose intensification was only administered to a small subset. Relapses were documented in only 14.1% during the follow-up period, a rate of disease recurrence that is consistent with the efficacy of adjuvant therapy [15].

We conclude that VEM is a well-tolerated and active schedule for neoadjuvant therapy and could be justifiably compared with FAC in a phase III trial.

Acknowledgment

Dr. Stephen Johnson (Taunton and Somerset Hospital, United Kingdom) assisted in the preparation of the final text.

References