The Oncologist, Vol. 11, No. 6, 563-573, June 2006; doi:10.1634/theoncologist.11-6-563 © 2006 AlphaMed Press
Multidisciplinary Therapy of Locally Far-Advanced or Inflammatory Breast Cancer with Fixed Perioperative Sequence of Epirubicin, Vinorelbine, and Fluorouracil Chemotherapy, Surgery, and Radiotherapy: Long-Term Resultsa First Department of Medical Oncology, b 4th Department of Surgery, c 2nd Department of Radiation Oncology, and d Pathology Department, St. Savas Anticancer Hospital, Athens, Greece; e Department of Biochemistry and Molecular Biology, Faculty of Biology, University of Athens, Athens, Greece Key Words. Locally advanced • Inflammatory • FEN • Surgery • Radiotherapy Correspondence: Alexandros Ardavanis, M.D., St. Savas Anticancer Hospital, 171 Alexandras Avenue, 115 22 Athens, Greece. Telephone: 0030-210-6409231; Fax: 0030-210-6420146; e-mail: ardavanis{at}yahoo.com Received December 14, 2005; accepted for publication April 14, 2006.
Background. Based on phase II data in advanced breast cancer (BC), the fluorouracil, epirubicin, and vinorelbine (FEN) combination was assessed as perioperative chemotherapy, integrated in a multidisciplinary treatment for locally advanced BC. Patients and Methods. Patients with newly diagnosed inoperable (stage IIIB or inflammatory) BC. Multimodality treatment protocol consisted of four preoperative courses of fluorouracil (600 mg/m2 day 1), epirubicin (75 mg/m2 day 1), and vinorelbine (25 mg/m2 day1andday8), all i.v. every 21 days, followed by radical or conservative surgery according to clinical response and four postoperative identical chemotherapy courses aimed to eradicate micrometastatic disease. Locoregional radiotherapy was offered to all patients after the completion of chemotherapy followed by hormonotherapy according to hormone receptor status. The primary end points of the study were: (a) clinical and pathological response, (b) downstaging and conversion to operable disease, and (c) recurrence-free survival (RFS) and overall survival (OS). Results. Forty-eight women, one stage IIIA, 27 (56.2%) stage IIIB, two stage IIIC (4.1%), and 12 (25%) with inflammatory BC, aged 3475 years (median, 52), were accrued. Thirty-eight and 34 patients completed the planned pre- and postoperative chemotherapy, respectively. Totals of 175 and 135 cycles were administered pre- and postoperatively, respectively. Toxicity of both phases, mainly hematologic, was in general acceptable without treatment-related death. Venous reactions were a frequent problem. All but three tumors were converted to operable, 31.3% with breast conservation. The clinical response rate (RR) was 77.7% (22.2% complete) and pathological RR was 73.3% (complete, 20% in both primary and axilla). After a median follow-up of 72 months, 62.5% and 16.7% of patients remain relapse free at 3 and 5 years, respectively, while 83% and 58.3% were alive 3 and 5 years, respectively, after the start of chemotherapy. Median RFS and OS have not yet been reached, and are currently 37+ and 62+ months, respectively. Conclusion. This fixed number of FEN perioperative courses schedule followed by radiotherapy is safe and highly active in inducing both local and distant control of locally far-advanced BC. This strategy is at least not inferior to other established regimens or strategies for locally far-advanced BC, while the integration of taxanes or new targeted agents may help show its true value for this challenging clinical entity.
The term locally advanced breast cancer (LABC) encompasses a heterogeneous group of breast neoplasms; in the last revision of the American Joint Committee on Cancer (AJCC) staging system [1], all of stage III disease is considered locally advanced, as well as a subset of stage IIB (T3N0). Yet, in this last revision, the involvement of supra-clavicular lymph nodes also was classified as N3 rather than M1; therefore, LABC has four subgroups, of which IIIB and IIIC (T4N23) are typically considered inoperable. These cancers vary widely in biological characteristics and clinical behavior, ranging from locally aggressive and systemically "indolent" to de novo generalized disease. The prognosis for women with LABC depends on tumor size, extent of lymph node involvement, and the presence or absence of an inflammatory component [2]. According to the 19731998 Surveillance, Epidemiology, and End Results data, the 3- and 5-year relative survival rates for women with stage III breast cancer are 70% and 55%, respectively, while the median survival time is 4.9 years [3]. Inflammatory breast cancer (IBC) is a distinct clinicopathologic entity from LABC, with aggressive behavior; IBCs often present highly angiogenic characteristics, are more likely to be high grade, aneuploid, and hormone-receptor negative, and have a high S-phase fraction and p53 mutations. However, prognostic factors for IBC and treatment are similar to those for LABC, although prognosis is clearly worse in the former [2, 4]. Multidisciplinary therapy is the established treatment for patients with LABC and IBC. Primary or neoadjuvant chemotherapy followed by locoregional therapy, either surgery and/or radiotherapy, and postoperative systemic chemotherapy is the widely accepted sequence to optimally control LABC [2, 5]. Upfront or primary chemotherapy integrated into a multimodality program has become the standard of care in LABC and IBC the past two decades [69]. Responses to primary chemotherapy, both clinical and pathological, are strongly associated with long-term outcome, while the conversion of an initially inoperable tumor to operable or, even more, to conservatively operable is a major goal, usually feasible with neoadjuvant chemotherapy [2, 511]. Since the Early Breast Cancer Trialists Group established the superiority of anthracycline-based chemotherapy regimens [12], they have been considered as standard of care for neoadjuvant chemotherapy for LABC [2, 4]. Epirubicin, a semisynthetic derivative of doxorubicin, has been extensively investigated in patients with breast cancer and has demonstrated efficacy both in metastatic disease and the adjuvant setting [1316]. Although the utility of anthracyclines is somewhat restricted by their cardiotoxicity, epirubicin clearly has less cardiotoxic and myelotoxic potential than doxorubicin at equimolar doses, allowing the safe administration of cumulative doses between 950 and 1,000 mg/m2 [17]. The semisynthetic vinca alkaloid vinorelbine (Navelbine®; GlaxoSmithKline, Philadelphia) is among the most active drugs in advanced breast cancer, yielding response rates (RRs) of 35%53% in first-line therapy [1823] at the cost of acceptable toxicity. Further, the anti-metabolite 5-fluorouracil (5-FU) has long been a standard component of active chemotherapy regimens (CEF, FEC, etc.) for both advanced and early breast cancer [24]. A series of epirubicin/vinorelbine-based doublets or triplets with a third agent in advanced breast cancer have been reported so far. After the pioneering work of Spielmann and coworkers with the doxorubicinvinorelbine combination [2527], not reproduced however in the phase III setting [28], an interesting first-line activity of the epirubicinvinorelbine doublet has emerged in phase II [2932]: in a recently published phase III trial, the addition of vinorelbine to epirubicin (90 mg/m2) monotherapy led to a significantly greater complete RR (RR) and progression-free survival time but not overall survival (OS) time [33]. Moreover, other investigators have shown that combinations of vinorelbine with 5-FU in doublets or triplets with a third agent have significant activity in metastatic and LABC, although in some of them 5-FU was administered in protracted i.v. infusions and not bolus [3441]. In view of the above, we conducted a phase II trial of the 5-FU, epirubicin, and vinorelbine combination (FEN) in advanced breast cancer [42, 43]; subsequently, based on the encouraging efficacy and tolerance of FEN, we integrated it as upfront chemotherapy into a multidisciplinary approach of a fixed schedule of administration (four preoperative and four postoperative courses) with surgery and radiotherapy in inoperable (stage IIIB or IIIC or IBC) breast cancer patients. Clinical response, pathological response, and conversion to operability were the primary end points, while recurrence-free survival (RFS) and OS were the secondary end points. The rationale of this protocol was: (a) to maximally debulk the tumor by full adjuvant-cumulative levels of epirubicin (300 mg/m2), assisted by vinorelbine and 5-FU; (b) to attempt to eradicate systemic disease with four identical courses in cases of at least a partial objective response, improving therefore both survival variables; and (c) to enhance local control with both appropriate surgery and radiotherapy. The short-term results (response, operability) as well as the mature long-term RFS and OS results are presented here.
Patient Selection Women characterized as stage IIIB (T4N02M0) disease or IBC (T4d, any N, M0) after full staging procedures were eligible. Other criteria for entry into the study included age 1875 years, good performance status (Eastern Cooperative Oncology Group [ECOG] score 2), and adequate cardiac, renal, and hepatic function as well as normal marrow reserves. Patients with creatinine clearance <50 ml/minute, WBC <3.5 x 109/l, neutrophil count <1.5 x 109/l, or platelet count <100 x 109/l within the 2 weeks preceding the start of the study were excluded. Patients were also excluded if distant disease was detected within a month before or after initial diagnosis. Patients with other malignant tumor or tumor history, except for nonmelanoma skin cancer or radically excised in situ carcinoma of the uterine cervix, were also excluded. Furthermore, patients with severe chronic obstructive lung disease, patients who were pregnant, or patients with active infections or other serious underlying medical or mental conditions, which would impair their ability to safely receive protocol treatment, could not participate. The study was conducted according to the Declaration of Helsinki and the guidelines for Good Clinical Practice. The local ethics committees approved the protocol, and informed consent was obtained from all patients prior to study entry.
Initial Evaluation of Patients All patients had to undergo an electrocardiogram (ECG) and left ventricular ejection fraction (LVEF) evaluation either by multigated acquisition (MUGA) scan or echography before study entry; ECGs were repeated before each course and at the end of treatment. Furthermore, during the follow-up period, full clinical and laboratory cardiologic evaluations were planned every 34 months for at least 3 years.
Treatment Protocol
Patients had to receive 600 mg/m2 of 5-FU on day 1, as a bolus i.v. injection; 75 mg/m2 epirubicin, diluted in 50 ml of normal saline as a 5- to 10-minute i.v. infusion on day 1; and 25 mg/m2 vinorelbine, diluted in 50 ml of normal saline as a 5-minute i.v. infusion on days 1 and 8. Treatment was recycled every 21 days for a total of four courses; appropriate surgery (conservative or radical) was performed after full assessment of response and restaging within 3 weeks from the last chemotherapy session; postoperative chemotherapy with four courses of the same regimen was given if there was at least a partial clinical response and evidence of pathological remission, starting within 3 weeks from the operation. Nonresponders were withdrawn from postoperative FEN chemotherapy and treated appropriately; they had to receive at least appropriate radiotherapy and they were used in the calculations for survival. Radiotherapy was administered after the completion of chemotherapy, according to the operation performed and the microscopic findings; in cases of radical surgery, a total dose of 5060 Gy in 1.82.0 Gy daily fractions was delivered to the chest wall and the ipsilateral axilla/supraclavicular fossa. In cases of conservative operation, a total dose of 5060 Gy in 1.82.0 Gy daily fractions was delivered to the breast and the ipsilateral axilla/supraclavicular fossa with a 10 Gy boost in the tumor bed. Hormonal manipulation was administered according to the hormonal status of the tumor before and after exposure to the induction chemotherapy (tamoxifen with or without a luteinizing hormone-releasing hormone analogue, starting after the completion of radiotherapy). Patients with progressive disease (PD) at any phase of the protocol therapy were withdrawn and offered alternative treatment accordingly; if they had received at least two preoperative cycles they were included in the toxicity and response evaluations as well as in the survival analysis.
Dose Modifications for Adverse Events
Assessment of Response to Preoperative Chemotherapy Pathological response in the surgically resected specimen was based on exhaustive microscopic examination of multiple sections from the breast and axillary lymph nodes; a pathological response was considered complete (pCR) if no residual or in situ tumor could be detected. Invasive tumor of maximal diameter less than or equal to that found mammographically and or clinically, tumor cell foci amid fat necrosis and fibrosis with inflammatory cell infiltration, and tumor cell vacuolization were all considered as a pathological partial response (pPR), according to the literature available at that time referring to radiation- or chemotherapy-induced microscopic changes of the tumor and breast tissue [4446]. Pathological "No Change" or PD was considered when clinical response was of the above types together with the presence of infiltrative tumor without evidence of necrosis in microscopy. All patients were considered to be assessable for response and toxicity by intent-to-treat analysis if they had received at least two preoperative chemotherapy cycles. Patients were analyzed after stratification into two groups, those with stage IIIB and those with IBC. Toxicity and RRs were analyzed by descriptive methods. RFS and OS were calculated according to standard definitions from enrollment and start of preoperative chemotherapy, respectively.
Statistics
Between October 1996 and September 1998, 48 patients with LABC, aged 3475 years (median, 52 years), were accrued. Of these, one with stage IIIA (T3N2M0) and two with stage IV (T4N2M1, due to involved supraclavicular lymph nodes), according to the tumor, nodes, metastasis (TNM) classification valid at that time (1988), were erroneously enrolled. All these patients were, however, taken into account in the present study, because with the last modification of the AJCC staging system, the last two are classified as stage IIIC [47], while stage IIIA is not quite different from IIIB in prognosis and operability. Although IBC patients have a distinct behavior from other stage III patients, no different analysis was performed for this subgroup because of its rather small number (n = 12).
The characteristics of all patients enrolled are summarized in Table 1
All but three and two patients were operated and irradiated, respectively, being therefore evaluable for survival (n = 48). Eight of nine patients with initially unknown hormonal receptor status according to FNA biopsy were operated on, three of those eight were found to be ER/PgR positive, with a caveat that these results were obtained after chemotherapy.
Toxicity The postoperative phase of chemotherapy was, as expected, less well tolerated, although again toxicities occurred at acceptable levels. Of note, venous toxicity complicating 28.2% of the postoperative chemotherapy patients was the main reason for its discontinuation in 5 of 10 patients who did not complete postoperative chemotherapy. The mean relative dose intensity in the postoperative phase was maintained at somewhat lower levels (90% for all drugs), leading to a median postoperative cumulative dose of epirubicin of 270 mg/m2 and a total dose of around 570 mg/m2. Grade III or IV radiation dermatitis was also an important problem in seven patients; however, in all cases, the syndrome regressed without major complications. Despite the relatively high cumulative epirubicin dose (600 mg/m2 planned, 570 mg/m2 administered), no significant short- or long-term cardiac toxicity was detected during the 72-month median follow-up period, with the exception of two patients: one 68-year-old patient without a prior cardiac history and one 58-year-old hypertensive patient, both with left breast cancer and both irradiated, developed congestive heart failure (CHF) 12 and 28 months, respectively, after completion of chemotherapy; of note, the first patient received a 40 mg/m2 cumulative dose of mitoxantrone in the intervening period, for recurrent disease, in another oncology center. No clinically evident neurotoxicity was recorded. Constipation, which was recorded in some of our patients, could not be clearly attributed to either vinorelbine or the setrones used as antiemetics.
All toxicity results are summarized in Tables 2
Response and Survival Forty-five patients were evaluable for response to the induction chemotherapy phase and 48 patients were evaluable for survival (three patients were lost to follow-up after 38, 43, and 46 months). In the 45 patients evaluable for response, there were 35 (77.7 %; 95% confidence interval [CI], 74.282.2) clinical responses (10 complete [22.2 %] and 25 partial [55.5%]) and 33 (73.3%; 95% CI, 70.276.4) pathological responses (9 pCR [20%] and 24 pPR [53.3%]). pCR in the breast was achieved in 11 patients (24.4%) and in the axillary lymph nodes in 10 patients (22.2%), while absence of tumor in both the breast and the axillary nodes occurred in nine patients (20%). When analyzed separately, in IBC, both clinical and pathological RRs were 41.6% (5 of 12 patients) with only one (8.3%) CR. The clinical RR in stage IIIB/IIIC patients was 86.1% (31 of 36), with 22.2% (8 of 36) CRs; however, the pathological RR was 72.2% (26 of 36) with 16.6% (6 of 36) CRs. All but three patients were operated on, 15 (31.3%) initially T4abN12, with conservative surgery. All but two patients were irradiated.
After a median follow-up of 72 months (1296), 30 (62.5%) and 8 (16.7%) patients remained relapse free at 3 and 5 years, while 83% (40 of 48) and 58.3% (28 of 48) were alive 3 and 5 years, respectively, after the start of chemotherapy. Five patients remained alive and free of relapse 6896 months after accrual; therefore, the RFS and OS durations have not yet been reached. The median RFS time is 37+ months (mean ± standard error [SE], 41.5 ± 4.22; 95% CI, 33.249.75). The median OS time is 62+ months (mean ± SE, 59.5 ± 3.66; 95% CI, 52.366.7). Patients with a pCR had a significantly higher probability of longer RFS (p = .023), but only a trend to longer OS (p = .087). Response and survival data are shown in Tables 4
One young patient presented with an obviously (different histology and receptor status) new primary in the other breast (T1N0M0) 5 years after the start of chemotherapy; the patient was successfully operated on, and she was alive 8 years after enrollment into the study, having been classified in this analysis as a recurrence-free long-term survivor. The amount of preoperative chemotherapy seemed to affect clinical response (four cycles led to more cCRs and cPRs; p = .015), pathological response (four cycles led to more pCRs and pPRs; p = .002), and the operability and type of surgery (p = .002); however, the full preoperative chemotherapy program did not have any significant impact on type of recurrence and the probability of recurrence and death (p = .78 and p = .6, respectively). Postoperative chemotherapy was related to the type of recurrence (more distant failures in patients without postoperative chemotherapy; p = .001); however, similarly to preoperative chemotherapy, the postoperative chemotherapy did not have any significant influence on the probability of recurrence and death (both p = .12). Clinical response was related to pathological response (cCR and cPR to pCR and pPR; p < .001) and type of recurrence (cPR to both local and distant recurrence; p = .007). However, neither cCR nor cPR was significantly correlated with the rate of recurrence (p = .29) and clinical response was marginally associated with a lower probability of death (p = .057). Pathological response was related to type of surgery (more conservative operations in pCR and pPR patients; p = .001) and type of recurrence (patients with a pPR had more local recurrences; p < .001) and marginally related to the probability of recurrence (less recurrences in pCR and pPR patients; p = .059); no relationship with the probability of death was detected (p = .15). Patients with stage IIIB or IIIC had more pathological responses than IBC patients (p = .005) as well as a higher probability of conservative surgery than IBC patients (p = .008); however, no difference in the probability of recurrence and death was found (p = .135 and p = 1.0, respectively) between stage IIIB/IIIC and IBC. Surgery and type of surgery were not significantly related either to type of recurrence or probability of recurrence and death.
Although the probability of recurrence and death was not significantly improved by any of the above variables studied, further analysis of the median RFS and median OS times showed that both clinical and pathological responses (CR and PR) were strongly correlated with both survival variables. Clinically responding patients had longer RFS and OS times (p = .001 and p = .004, respectively, Kruskal-Wallis or log-rank test) than nonresponders; further, pathologically responding patients also had longer RFS and OS times (both p < .001, Kruskal-Wallis and log-rank test) than nonresponders. RFS was favorably affected by pCR (Fig. 3
LABC remains a major clinical problem despite progress accomplished in the field. Despite the use of multimodal treatment, patients with LABC ultimately die from generalized disease. Every effort to improve prognosis of this entity is therefore welcome. In the present study, we investigated the value of a fixed perioperative chemotherapy and postoperative radiotherapy treatment plan aimed at maximally debulking the tumor locally and eradicating systemic disease in LABC and IBC patients. We used an anthracycline-based, FEC-like regimen with vinorelbine instead of cyclophosphamide, previously tested in metastatic disease [42, 43]. The present study was carried out in a mixed population of locally far-advanced breast cancer and IBC patients. Because the prognosis of those two subgroups is not quite different from that of patients with systemic disease, the results of the treatment used should be compared with the latter rather than LABC. The overall toxicity of both preoperative and postoperative FEN followed by radiotherapy was as expected, acceptable. Perioperative morbidity was not greater than with standard therapies. Of note, with the exception of two patients (one of whom subsequently received mitoxantrone and developed CHF), despite the relatively high total dose of epirubicin administered in most patients, no significant short- or long-term morbidity attributable to the initial manipulations (chemotherapy and radiotherapy) was recorded during the median 72 months of follow-up time. Of note, however, is that, with the exception of the three patients lost to follow-up and the above cited case, no patient was allowed to receive cardiotoxic chemotherapy at relapse. Nevertheless, subclinical myocardial damage might not have been detected by the methods used in the long-term assessment of cardiac function in this study. Overall, FEN was proven to be a safe regimen. The efficacy of preoperative FEN as evidenced by the 77.7% clinical RR with a 22.2% cCR rate and 20% pCR rate in both the breast and axilla is among the highest reported in the literature [2, 4, 5, 8] and similar to that of another recent study using epirubicin and docetaxel in LABC [48]. Further, although a discrepancy between clinical and pathological response is reported in many studies, this was not true in our patients (77.7% and 68.7% overall clinical and pathological RRs, respectively). Whether this may be attributed to the different criteria of response used or to other factors cannot be answered here. All but three of our patients (93.7%) became operable after induction FEN. The high percentage of radically operated patients should be attributed to the far-advanced stage at presentation, although in almost one third of the patients, conservative surgery was judged feasible. To our knowledge, this conversion-to-operability rate is among the highest reported in the literature [611, 4852], given the advanced stage of almost all cases and the presence of an inflammatory component in 25% of them. The short-term local control rate was influenced by the amount of preoperative chemotherapy administered, while postoperative chemotherapy seemed to enhance long-term systemic disease control. The survival outcomes of this study are comparable with those of previous reports. Hortobagyi et al. [10] from the M.D. Anderson Cancer Center, in their pioneering work used three cycles of preoperative doxorubicin-based chemotherapy, radical surgery, radiotherapy, and additional postoperative chemotherapy in noninflammatory stage IIB-regional IV (IIIC by the latest classification) LABC. They reported median DFS and OS durations of 30 and 48 months, respectively, for stage IIIB and regional IV patients; 5-year DFS and OS rates were 33% and 84%, respectively, declining to 30% and 44% at 10 years. Similar results were obtained in two subsequent studies assessing four cycles of anthracycline-based preoperative chemotherapy, with a different schedule of administration (72-hour continuous infusion of doxorubicin) or a noncrossresistant postoperative regimen [49, 50]. The 16.7% and 58.3% 5-year DFS and OS rates in our study are clearly inferior to the above MDACC results; however, the median DFS and OS durations of our patients have not yet been reached, being at the time of the last analysis 37 and 62 months, respectively; therefore, they are at least equivalent to the MDACC results. Further, it should be noted that a high proportion (25%) of our patients presented with an inflammatory component, by definition having a more dismal prognosis, in contrast with the relatively high proportion (80%) of patients with hormone-dependent tumors. Several other investigators have reported RR and survival outcomes similar to those reported in this study [610, 51, 5356].
It is well known that clinical and pathological response both predict for better outcomes [2, 55, 56]. In accordance with the literature, both clinical and pathological response in this study were associated with DFS and OS. Although there is, in general, considerable difficulty in accurately assessing response, pCR is associated with a favorable outcome [55, 56]. This was partially true in our patients, in whom a significant difference was observed in DFS but not in OS between the pCR and pPR/nonresponding subgroups; indeed, although <20% and >50% of complete and partial/nonresponders, respectively, were alive at 5 years, the difference was not significant (Figs. 3 To our knowledge, this is the first study exploring the efficacy of an anthracycline plus vinorelbine-based regimen administered in a fixed perioperative schedule followed by radiotherapy in locally far-advanced breast cancer or IBC. The mature results of this approach suggest a high activity of this approach in the local control of the disease. Compared with the relatively poor remission and survival rates of therapy in stage IV breast cancer, to the outcome of which, in our opinion, stage IIIB and IIIC should be seen, the results of multimodal FEN appear clearly superior; however, the survival outcomes of our patients are not superior if compared with those reported in the entire LABC population (stage IIBIIIC). We conclude that this perioperative strategy shows at least noninferiority in terms of downstaging rate, DFS, and OS versus other established regimens or strategies for locally far-advanced breast cancer reported in the literature. Perhaps the integration of taxanes and newer biologic agents like trastuzumab, already widely used in c-erb-B-2-overexpressing tumors, using similar perioperative scheduling, may help show the true value of the model of fixed perioperative systemic and local therapy in treating this challenging disease entity more effectively.
The authors indicate no potential conflicts of interest.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||