Sequential administration of dose-dense epirubicin/cyclophosphamide followed by docetaxel/capecitabine for patients with HER2-negative and locally advanced or node-positive breast cancer
- First Online:
- Cite this article as:
- Nieto, Y., Aramendía, J.M., Espinós, J. et al. Cancer Chemother Pharmacol (2010) 65: 457. doi:10.1007/s00280-009-1049-y
- 111 Views
Capecitabine is effective against metastatic breast cancer (MBC). We hypothesized that sequential treatment with dose-dense epirubicin/cyclophosphamide (EC) and docetaxel/capecitabine would be active and tolerable in the adjuvant/neoadjuvant setting.
In this prospective phase II clinical trial patients with HER2-negative and node-positive or locally advanced tumors were eligible to receive four cycles of EC (100/600 mg/m2) every 2 weeks with G-CSF on days 3–10, followed by four cycles of docetaxel/capecitabine (75/1,000 mg/m2 b.i.d., days 1–14) every 3 weeks.
Fifty-five patients were enrolled with median age of 49, and 80% had hormone receptor-positive disease. The median tumor size was 2.5 cm, with a median of two axillary nodes involved. Seventy-five percent of the first 20 patients had grade 2/3 hand-foot syndrome (HFS). Dose reduction of capecitabine to 800 mg/m2 reduced the grade 2/3 HFS incidence to 31% in the remaining patients. No grade 4/5 toxicities were observed. All 20 patients treated preoperatively responded, with 5 (25%) pathologic complete responses and 3 additional pT0N1 tumors. At a median follow-up of 48 (range 28–60) months, the event-free and overall survival rates are 91 and 98%, respectively.
Sequential treatment with dose-dense EC followed by docetaxel/capecitabine, using a lower capecitabine dose than that approved for MBC, has an acceptable toxicity profile and encouraging activity when used as neoadjuvant or adjuvant treatment of breast cancer.
KeywordsPhase II trialCapecitabineDocetaxelAdjuvantNeoadjuvantBreast cancer
Recent advances in the treatment of node-positive breast cancer include the development of dose-dense anthracyline schemas  and combinations of anthracyclines and taxanes. Several trials have shown the benefit of both sequential and concurrent combinations of anthracyclines and paclitaxel or docetaxel [2–6].
Capecitabine is an effective agent in metastatic breast cancer (MBC). O’Shaughnessy and colleagues have shown the superiority of docetaxel/capecitabine over docetaxel alone in patients with metastatic disease . The synergy between these two agents seems to result from docetaxel-induced upregulation in tumor cells of thymidine phosphorylase (TP) [8, 9], the enzyme responsible for the intracellular activation of capecitabine from its intermediate metabolite 5′-deoxy-5-fluorouridine (5′-FUDR) to the active form 5-fluorouracil (5-FU). Numerous studies have demonstrated that higher TP content in tumor cells has a favorable effect on patient outcomes after treatment with capecitabine [10–13].
In view of its effectiveness in metastatic disease, capecitabine holds substantial promise in the treatment of nonmetastatic breast cancer, a setting where it has not been studied extensively. Prior treatment with doxorubicin/cyclophosphamide (AC) or epirubicin/cyclophosphamide (EC), but not with a 5-FU-containing regimen, has been shown to upregulate the expression of TP in breast cancer cells . We hypothesized that sequential treatment with TP-upregulating dose-dense epirubicin/cyclophosphamide (EC) followed by docetaxel/capecitabine would be active and well tolerated in the adjuvant or neoadjuvant settings. Given the observed antagonism between the anti-HER2 antibody trastuzumab and 5-FU (and, by extension, capecitabine) in preclinical studies [15, 16], we restricted enrollment to patients with HER2-negative disease, who do not benefit from trastuzumab.
We report here a prospective phase II study of four cycles of dose-dense EC followed by four cycles of docetaxel/capecitabine as adjuvant or neoadjuvant treatment of patients with HER2-negative, node-positive or locally advanced breast cancer.
Patients and methods
This trial was conducted at the Clinica Universitaria de Navarra, Spain, between 2005 and 2006. The study protocol was approved by the Ethics Committee of our institution. All patients gave informed consent prior to enrollment. Eligible patients were 18–70 years of age and had histologically proven breast cancer, stage IIA to IIIC, with either involvement of axillary nodes or locally advanced disease that was deemed suitable for preoperative chemotherapy; performance status 0–2; normal end-organ function (creatinine clearance >50 ml/min, left ventricle ejection fraction >50%, glutamic oxaloacetic transaminase (GOT, AST)/glutamic pyruvic transaminase (GPT, ALT)/bilirubin <2× upper limit of normal, and normal peripheral blood counts); and HER2-negative tumors (0 or 1+ by immunohistochemistry or negative by fluorescence in situ hybridization). Exclusion criteria were pregnancy, presence of metastases, and previous chemotherapy.
The study endpoints were: (1) to define the hematologic and nonhematologic toxicities of the proposed sequential regimen as adjuvant or neoadjuvant treatment, (2) to assess the event-free survival (EFS) and overall survival (OS) of the study population after this treatment, and (3) to determine the clinical response rate and pathological complete response (pCR) rate in the subset of patients treated preoperatively.
Surgery of the primary tumor consisted of modified radical mastectomy or lumpectomy, with or without sentinel lymph node biopsy. Adjuvant chemotherapy was to start within 6 weeks following surgery. Pretreatment workup included a complete blood count with differential, chemical profile (creatinine, urea, lactate dehydrogenase, bilirubin, alkaline phosphatase, transaminases), chest X-ray, liver ultrasonography or abdominal computed tomography, bone scan, tumor markers (CEA and CA 27.29), and an echocardiography or radionuclide ventriculography.
Patients received four cycles of EC, with epirubicin (100 mg/m2) and cyclophosphamide (600 mg/m2) administered intravenously (IV) every 14 days. Granulocyte colony-stimulating factor was administered subcutaneously at 5 μg/kg daily from day (d) 3 to 10. Starting 2 weeks after the fourth cycle of E100C, patients were given four cycles of docetaxel/capecitabine, with docetaxel (Taxotere®) (75 mg/m2 IV, d1 of each cycle) and capecitabine (Xeloda®) at 1,000 mg/m2 twice a day (b.i.d.) orally (PO), d1–14, administered every 21 days. During the four cycles of docetaxel/capecitabine, patients received levofloxacin at 500 mg daily PO, on d5–18 of each cycle, and pyridoxine 50 mg PO three times daily.
Patients receiving treatment preoperatively were required to undergo magnetic resonance imaging (MRI) and breast ultrasonography with axillary assessment prior to starting chemotherapy. These tests were repeated upon completion of treatment to evaluate clinical response. Response Evaluation Criteria in Solid Tumors (RECIST) criteria were used for clinical response evaluation . Surgery was performed 4–6 weeks after the last cycle. Pathological response was evaluated in the surgical specimens from the breast and axilla. A pCR required the eradication of viable invasive ductal carcinoma cells at both sites, with or without presence of ductal carcinoma in situ in the breast pathological specimen.
In the case of incomplete hematologic recovery from the previous cycle (absolute neutrophil count <1,000/mm3 and or platelets <75,000/mm3), treatment was delayed until recovery. After each episode of grade 3–4 nonhematological toxicity treatment was withheld until resolution of toxicity to ≤grade 1, and the doses of epirubicin and cyclophosphamide (if after EC) or of capecitabine (if after docetaxel/capecitabine) were reduced by 20%. The dose of docetaxel was not reduced for nonhematological toxicity. After each episode of neutropenic fever the doses of epirubicin and cyclophosphamide or docetaxel were reduced by 20%. The dose of capecitabine was not reduced for hematological toxicity.
Post-chemotherapy and post-surgical treatment
Since all patients in the study had axillary involvement or locally advanced tumors, the need for locoregional radiotherapy was assessed in all cases. Following completion of chemotherapy, premenopausal patients with hormone receptor-positive tumors received adjuvant hormonal therapy with tamoxifen for 5 years and a luteinizing hormone-releasing hormone analogue for 2 years; postmenopausal patients were treated with an aromatase inhibitor for 5 years. Patients were monitored every 6 months for 5 years upon treatment completion, and on a yearly basis thereafter.
Previous reports have established the separate feasibility of the two components of the treatment sequence, epirubicin/cyclophosphamide  and docetaxel/capecitabine , at the doses used in this study. In this trial, we intended to study the tolerability and activity of their sequential administration in the nonmetastatic setting. Toxicity was graded according to the Common Toxicity Criteria of the National Cancer Institute .
The trial followed a two-stage design. Twenty patients were accrued in the first stage of the study. If five or more patients experienced grade 3–4 toxicity, the lower limit of the 95% confidence interval of incidence of grade 3–4 toxicity would be greater than 5%. The dose of capecitabine (if nonhematological toxicities) and/or docetaxel (if hematological toxicities) would then be decreased by 20% for all patients.
If fewer than five patients among the first 20 experienced grade 3–4 toxicity then another 20 patients would be enrolled at the same dose. If the cumulative number of grade 3–4 toxic events among the 40 patients was 8 or more, the conclusion, with 95% probability, would be that the grade 3–4 toxicity rate exceeds 5%. If the cumulative number of grade 3–4 toxic events among the 40 patients was 15 or more, the conclusion would be that the grade 3–4 toxicity rate exceeds 20%, with 95% probability. This two-stage design has greater than 90% power to detect a 20% rate of grade 3–4 toxicity, with a type I error rate of 5% .
Event-free survival (EFS) was estimated from the first chemotherapy day until tumor progression, relapse, or death from any cause. Overall survival (OS) was estimated from the start of chemotherapy until death. The comparisons between the toxicity grades encountered in both trial stages employed the chi-square test.
Patient Demographics (N = 55)
Age, median (range)
49 (27–73) years
Tumor size, median (range)
2.5 (1–13) cm
No. + nodes: median (range)
Nodal ratio: median (range)
Ki67, % + cells: median (range)
Toxicity (Table 2)
Grade 2–3 toxicities
1st stage (N = 20)
2nd stage (N = 35)
None of the other observed toxicities necessitated a dose reduction, and their incidences did not differ significantly between the first and second stages of the trial. Grade 2 nail changes were seen in 30 and 23% of patients in the first and second stages, respectively (P = 0.5), and grade 1 nail changes were observed in 30 and 11% of patients, respectively. Stomatitis was mild throughout the study: the incidences were 10% for grade 2 and 5% for grade 1 in the first stage and 11% for grade 2 and 6% for grade 1 in the second stage. No patients in this study experienced peripheral neuropathy or a drop in the left ventricle ejection fraction. Five patients experienced neutropenic fever with grade 3 neutropenia after EC and three after docetaxel/capecitabine. The incidences of neutropenic fever in the first and second stages were 20 and 14%, respectively (P = 0.6). No grade 3–4 anemia or thrombocytopenia was observed. No grade 4–5 toxicities of any sort were seen throughout the trial.
None of the surgeries had to be delayed for patients treated preoperatively. There were no post-surgical complications in those cases.
All 20 patients treated preoperatively responded clinically as determined by MRI, ten of them with a CR. Fifteen patients underwent breast-conserving surgery and five patients had a modified radical mastectomy. Five patients (2 with stage IIIA and 3 with stage IIB tumors) experienced a pCR (25%, 95% confidence interval, 6–44%). Three additional patients had pT0N1 tumors in their surgical specimens.
Following completion of chemotherapy, 43 patients with hormone receptor-positive tumors were prescribed hormone treatment for 5 years. Locoregional radiotherapy was administered to 54 (98%) patients without unexpected side effects.
The five patients whose tumor relapsed did not differ significantly from the other 50 patients in age (P = 0.2) or histological grade (P = 0.9). However, their tumors seemed to be larger (median 5.5 vs. 2.4 cm, P = 0.1), presented with more involved axillary nodes (median 6 vs. 2, P = 0.00001) and a higher axillary nodal ratio, defined as the quotient between the no. positive nodes and the no. dissected nodes (0.55 vs. 0.105, P = 0.00001), and showed a higher proliferative fraction (70 vs. 15% Ki67-positive cells, P = 0.01). Two of the five relapsing patients were treated neoadjuvantly, with a clinical CR and a partial response, respectively. The pathological stages in their surgical specimens were pT1N0 and pT2N2, respectively.
The sequential combination of dose-dense EC followed by docetaxel/capecitabine in patients with node-positive or locally advanced tumors was well tolerated and showed very encouraging activity in this feasibility study. This new sequential schema incorporates capecitabine, an effective agent in the metastatic setting, into the adjuvant or neoadjuvant treatment of breast cancer.
We elected to add capecitabine to a sequential framework of anthracyclines followed by docetaxel, because emerging data suggest that sequential treatment with these agents might be superior to their concurrent administration. Francis et al. compared both sequential and concurrent doxorubicin and docetaxel to a control arm of doxorubicin with or without cyclophosphamide, followed in all cases by cyclophosphamide/methotrexate/5-FU, and reported improved EFS in the sequential, but not the concurrent, experimental arm compared to control treatment . Results of other randomized neoadjuvant trials are in keeping with these observations. Sequential doxorubicin–docetaxel resulted in higher response and pCR rates than doxorubicin alone in several studies [22–24]. In contrast, concurrent doxorubicin/docetaxel was not shown to be more active than doxorubicin/cyclophosphamide . Finally, the use of a sequential arm of doxorubicin/cyclophosphamide followed by docetaxel resulted in higher response and pCR rates than a dose-dense regimen of concurrent doxorubicin/docetaxel .
The combination of docetaxel and capecitabine is attractive because of their different mechanisms of action and largely nonoverlapping side effects, namely, myelotoxicity and cutaneous/mucosal toxicity, respectively. In metastatic disease, docetaxel/capecitabine achieved superior RR, EFS and OS compared to docetaxel alone . The synergy of the combination may be due to docetaxel-induced tumor upregulation of TP [8, 9], the enzyme responsible for the intracellular activation of capecitabine from 5′-FUDR to 5-FU. Preclinical observations have shown that transfecting cells with TP sensitizes them to 5′-FUDR[27–29] and that this compound is more active in TP-overexpressing tumor xenografts . Furthermore, TP expression in tumor cells is a favorable predictive factor after treatment with 5′-FUDR in patients with breast cancer [10–12, 31, 32]. In a recent study, TP expression in metastatic tumor samples correlated directly with time to progression after treatment with docetaxel/capecitabine . Further supporting the hypothesis of docetaxel-induced tumor sensitization to capecitabine, concurrent treatment of xenografts with docetaxel and capecitabine resulted in synergistic activity, whereas the combination of docetaxel with either 5-FU or uracil plus tegafur, neither of which require activation by TP, showed only additive activity .
The therapeutic schema tested in this trial pursues additional upregulation of TP by EC administered prior to docetaxel/capecitabine. Toi et al.  have described increased expression of TP in breast cancer specimens after preoperative chemotherapy with AC or EC but not after using 5-FU-containing regimens (i.e., FAC or FEC). It is possible that 5-FU selectively kills or suppresses TP-overexpressing cells or that, at high concentrations, 5-FU, a pyrimidine substrate, downregulates the expression of TP. Additional considerations about toxicity prompted us to use EC instead of FEC for the initial dose-dense part of the therapeutic sequence. In one adjuvant study, biweekly FEC was associated with excessive pericardial, pleural and pulmonary toxicity , whereas dose-dense EC can be administered without difficulty . Thus, EC was a better fit for our strategy of administering dose-dense TP-upregulating chemotherapy followed by TP-dependent chemotherapy.
In the original phase I evaluation of docetaxel/capecitabine, the dose of capecitabine considered to be feasible was identified as 1,250 mg/m2 b.i.d. for 14 days, combined with 75 mg/m2 of docetaxel, every 3 weeks . When compared to docetaxel in MBC, these doses of docetaxel/capecitabine were associated with a manageable 24% incidence of grade 3 HFS . Subsequent reports of dose reductions of capecitabine to 950 mg/m2 suggested that the toxicity of the regimen could be ameliorated without compromising its efficacy . In contrast to the latter observations, our starting capecitabine dose of 1,000 mg/m2 caused excessive HFS. A subsequent protocol-mandated dose reduction to 800 mg/m2 improved the tolerability of docetaxel/capecitabine. The paradoxical observation of greater toxic effects from capecitabine in our nonmetastatic patients than in prior reports from the metastatic setting may be explained by the upregulation of TP by EC. Since TP is preferentially expressed in tumor cells [37, 38], the combined effect of TP upregulation and subsequent dose reduction of capecitabine should, at least in theory, improve its therapeutic index.
There is growing interest in the use of capecitabine in the adjuvant setting. Joensuu and colleagues  recently reported an interim safety analysis in a subset of 600 patients enrolled in a randomized trial of docetaxel/capecitabine at a similar dose and schedule but in the inverse therapeutic sequence to ours. These authors compared in 1,500 patients with node-negative and node-positive disease the use of three cycles of docetaxel (60 mg/m2)/capecitabine (900 mg/m2 b.i.d., d1–15) followed by three cycles of cyclophosphamide (600 mg/m2)/epirubicin (75 mg/m2)/capecitabine (900 mg/m2 b.i.d., d1–15), to a control arm that received three cycles of docetaxel alone (75 mg/m2) followed by three cycles of FEC, all cycles administered every 3 weeks. The toxicity profile (9.6% incidence of grade 3–4 HFS) of docetaxel/capecitabine seen in this patient subset was similar to that we observed in the second stage of our study.
The preoperative activity of our regimen, evidenced by a 25% pCR rate in the small subset of patients with LABC, is encouraging, particularly in view of the fact that our patients had tumor phenotypes (HER2-negative and largely hormone receptor-positive) associated with lower pCR rates after preoperative chemotherapy [40–42]. Our use of triweekly docetaxel is supported by reports of single-agent docetaxel having greater activity when given every 3 weeks as compared to weekly, in both the adjuvant  and metastatic settings . Lee and collaborators compared neoadjuvant docetaxel/capecitabine (at the same doses as in the first stage of our trial) with AC, both given every 3 weeks . This trial enrolled 209 patients, mostly with stage II tumors. The pCR rate in the breast and axilla was higher after docetaxel/capecitabine than after AC (15 vs. 7%). Other authors have tested preoperatively a variant schedule in which docetaxel/capecitabine was administered monthly, with weekly doses of docetaxel at 36 mg/m2 (days 1, 8 and 15) and capecitabine at 500–625 mg/m2 bid for 14–17 days from day 5 of each cycle. Layman et al.  treated 26 patients, most of whom had T1–T2 and hormone receptor-negative tumors, with the variant docetaxel/capecitabine schedule. The pCR rate in their trial was only 7.6% after four cycles of docetaxel/capecitabine, but it increased to 26.9% after an additional four courses of dose-dense AC. Natoli et al.  employed a similar sequence to ours, with four cycles of dose-dense EC (90/600 mg/m2) followed by four cycles of variant docetaxel/capecitabine in 44 patients, most of them with T2 tumors. The pCR rate in their trial was 17% (8% among 24 patients with HER2-negative tumors).
The randomized GeparQuattro trial compared the preoperative use of four cycles EC (90/600 mg/m2) every 3 weeks followed by either four cycles of docetaxel alone at 100 mg/m2, four cycles of concomitant docetaxel/capecitabine (60/900 mg/m2 every 12 h × 15 days) every 3 weeks, or the sequential use of four cycles of docetaxel (75 mg/m2) followed by four cycles of capecitabine (900 mg/m2 every 12 h) [48, 49]. This study enrolled 1,510 patients with operable tumors. Trastuzumab was added to chemotherapy for the 30% of patients with HER2-positive tumors. The first interim analysis of toxicity showed a 27.6% incidence of grade 3–4 HFS in the concomitant docetaxel/capecitabine arm, which appears higher than that observed in the second phase of our study (14%), perhaps due to the use of a slightly higher dose of capecitabine. The first preliminary analysis of responses showed no differences between the three arms, with pCR rates of around 21% in all three. Important differences between this study and ours were the inclusion in GeparQuattro of patients with smaller tumors and with HER2-positive disease, the latter group treated with concurrent trastuzumab-chemotherapy. Both groups of patients typically achieve higher pCR rates than the population of patients with larger and HER2-negative tumors we targeted in our trial [40–42].
Trials testing docetaxel/capecitabine as neoadjuvant treatment
EC → Doc
Doc: 75 mg/m2 IV, d1
21% in all three arms
EC → Doc/Cap
Cap: 900 mg/m2 PO BID, d1–14,
EC → Doc → Cap
Every 3 weeks
(+Trastuzumab if HER2+)
Doc: 75 mg/m2 IV, d1
Cap: 1,000 mg/m2 PO BID, d1-14
Every 3 weeks
EC → Doc/Cap
Doc: 36 mg/m2 IV, d1, 8 and 15
Cap: 1,250 mg/m2 PO BID, d5–18
Every 4 weeks
Doc/Cap → AC
Doc: 36 mg/m2 IV, d1, 8 and 15
Cap: 1,000 mg/m2 PO BID, d5–21
Every 4 weeks
EC → Doc/Cap
Doc: 75 mg/m2 IV, d1
Cap: 800-1,000 mg/m2 PO BID, d1-14 every 3 weeks
In conclusion, sequential treatment with dose-dense EC followed by docetaxel/capecitabine, using a lower capecitabine dose than that approved for MBC, has an acceptable toxicity profile and encouraging activity when used as neoadjuvant or adjuvant treatment of breast cancer.