S-1 combined with docetaxel following doxorubicin plus cyclophosphamide as neoadjuvant therapy in breast cancer: phase II trial

Women with previously untreated clinical stage II or III breast cancer were eligible for this neoadjuvant trial if the following eligibility criteria were met: i) pathologically confirmed invasive ductal or lobular carcinoma from a core biopsy specimen; ii) HER2-negativity of 0 or 1+ by immunohistochemistry, or HER2 non-amplification by fluorescent in situ hybridization; iii) age ≥18 years; iv) Eastern Oncology Cooperative Group performance status of ≤1; v) adequate cardiac function (left ventricular ejection fraction > 50%); vi) adequate bone marrow (neutrophils ≥1.5 × 10³/μl, platelets ≥100 × 10³/μl, Hb ≥10.0 g/dl), renal function (serum creatinine ≤1.5 times the upper normal limit or creatinine clearance ≥ 50 ml/min by Cockroft formula), and liver function (serum bilirubin ≤1.5 times the upper normal limit, aspartate aminotransferase/alanine aminotransferase ≤2.5 times the upper normal limit) within 2 weeks before starting the therapy; and vii) no previous chemo-, radio- or hormone therapy. Patients were excluded if they met the following criteria: i) T4d/inflammatory breast cancer; ii) potentially or currently pregnant or lactating; iii) taking medications that alter the pharmacokinetics of S-1 (for example, allopurinol, phenytoin); or iv) inability to swallow S-1 tablets or malabsorptive gastrointestinal condition.

All patients provided written informed consent and this study was approved by the Institutional Review Board of Severance Hospital, Seoul, Korea.

This was a single-arm, single-center, phase II study of S-1 combined with docetaxel (SD) following doxorubicin plus cyclophosphamide (AC) as neoadjuvant chemotherapy in patients with stage II-III breast cancer. Neoadjuvant chemotherapy schedule was as follows: For the initial AC treatment, doxorubicin (60 mg/m² i.v. on day 1) and cyclophosphamide (600 mg/m² i.v. on day 1) were administered every 3 weeks for four cycles. Upon completion of AC treatment, if the diseases had not progressed by physical and radiological examinations and toxicities were acceptable, SD treatment with S-1 (30 mg/m² orally b.i.d. on days 1–14) and docetaxel (75 mg/m² i.v. day 1) was given every 3 weeks for four cycles.

Modifications of doses and dosing schedules were as follows: If the neutrophil count was ≥1.5 × 10³/μl and the platelet count was ≥100 × 10³/μl, we would begin the next cycle for both AC and SD. If these values were not reached, AC or SD was delayed by 1 week. At the end of the first week of delay, if the neutrophil count was 1.0–1.5 × 10³/μl and the platelet count was 75–100 × 10³/μl, the next doses of doxorubicin/cyclophosphamide and S-1/docetaxel were reduced by 1 step to 50/500 mg/m² and 25/60 mg/m², respectively. If the neutrophil count was ≤1.0 × 10³/μl or the platelet count was ≤75 × 10³/μl, therapy was delayed for one more week. After the second week of delay, if the neutrophil count was ≥1.5 × 10³/μl and the platelet count was ≥100 × 10³/μl, AC and SD were reduced by 1 step; however, if these values were not reached, the trial would be stopped. Granulocyte colony-stimulating factor (G-CSF) was administered in the case of febrile neutropenia. Prophylactic G-CSF usage was permitted from the second cycle. Regarding non-hematological toxicity, if grade 3–4 S-1 specific toxicities such as diarrhea, stomatitis, or epigastric pain were observed, S-1 was omitted and restarted at a 1-step lower dose when toxicities were recovered to grade 1 or baseline; however, docetaxel administration was continued. If grade 3–4 peripheral neuropathy was observed, docetaxel was omitted and restarted at a 1-step lower dose when toxicities recovered to grade 1 or baseline. If grade 3–4 abnormalities of serum bilirubin, alkaline phosphatase, aspartate aminotransferase, or alanine aminotransferase increase were observed, docetaxel and S-1 were delayed and restarted at a 1-step lower dose when toxicities recovered to grade 1 or baseline. If the other grade 3–4 non-hematological toxicities were observed, AC or SD could be delayed and reduced according to the investigator’s discretion.

The type of breast surgery after neoadjuvant chemotherapy was determined by using a multidisciplinary team approach considering the clinical and radiological response to neoadjuvant chemotherapy, presence of multifocal or multicentric tumor, and patient preference. All patients with initial axillary node-positive disease underwent standard level I/II axillary lymph node dissection with or without explorative sentinel lymph node biopsy. Patients with Initially cytology-proven supraclavicular lymph node metastasis underwent supraclavicular lymph node dissection. After surgery, all patients received adjuvant radiotherapy covering the whole breast and regional nodes including supraclavicular area. Patients with hormone receptor-positive tumors received adjuvant endocrine therapy. However, postoperative adjuvant chemotherapy was not scheduled. If tumors were not resectable, patients were taken off the clinical trial.

The primary endpoint was the pCR rate, defined as the absence of invasive carcinoma in the breast and no involvement of carcinoma to axillary nodes (AXLNs) (i.e., pathological stage T0N0 or TisN0). Secondary endpoints were the clinical response rate, safety, pCR rate in the breast regardless of AXLN, and BCS rate.

Tumor measurement by mammography, ultrasonography, and/or MRI was scheduled at baseline, at the completion of four cycles of AC, and before surgery. The breast and regional lymph nodes were examined by palpation every cycle. A PET-CT scan was performed at baseline and at the completion of chemotherapy to exclude distant metastases. Pathological responses of the breast tumor and infiltration of regional lymph nodes were assessed and staged according to the TNM system. The clinical response was defined according to the Response Evaluation Criteria for Solid Tumors (version 1.1) [12]. Toxicity was graded based on the National Cancer Institute’s Common Terminology Criteria for Adverse Events version 3.0. For pathological assessment, excised tissues from the breast were cut to 5-mm thickness and those from the regional lymph nodes to 2-mm thickness. Patients were considered to have had BCS if the final surgical procedure was tumorectomy, segmentectomy, or quadrantectomy.

The null hypothesis is that the pCR rate is 11%. This pCR rate was estimated lower than 14.3% obtained by docetaxel following by AC in GEPARDUO study in which only operable T2-3 and N0-2 tumors were enrolled but our study enrolled more advanced stages with pathologically confirmed axillary node involvement (T1-4c and N1-3) including inoperable, stage III tumors [13]. The alternative hypothesis is that addition of S-1 in combination with docetaxel will increase the pCR rate to 25%. With a two-sided significance level of 0.05 and a power of 0.8 to reject the null hypothesis, the minimum sample size was determined to be 44. Assuming a 10% of drop-out rate, the final sample size was calculated to be 49 patients. An optimal cut-off point for Ki67 was determined as 30% using the minimum P value approach in the analysis of predictors of pCR.

From July 2009 to September 2011, 49 patients were enrolled in this study at Severance Hospital, Yonsei University Health System, Seoul, Korea. All 49 patients received at least one dose of study treatment and were included in the evaluation of pCR and toxicity [intention to treat (ITT) population]. Baseline characteristics of the 49 patients are summarized in Table 1. All patients were positive for involvement of axillary lymph node and five patients also had supraclavicular lymph node (SCL) metastasis. Involvement of axillary and supraclavicular node was confirmed by fine needle aspiration and histological examination.

Forty-one patients (83.7%) completed four cycles of AC and four cycles of SD chemotherapy and also underwent surgery. Among eight patients (16.3%) who could not complete the scheduled chemotherapy, one was lost to follow-up during AC cycles and seven were taken off the study during SD cycles for the following reasons: two withdrew consent after 1 cycle; three exhibited toxicity after 1, 2, or 3 cycles; one had disease progression after 2 cycles; one died from septic shock after 2 cycles. Six of the seven patients who dropped out during SD cycles received surgery.

Efficacy outcomes are summarized in Table 2. The primary endpoint, the rate of pCR in both breast and AXLN, was 22.5% (n = 11) in the ITT population. The rate of pCR in the breast and in AXLN was 30.6% (n = 15) and 40.8% (n = 20), respectively. In the per protocol population, which excluded patients who were lost to follow-up (n = 1) or withdrew consent (n = 2), the rate of pCR in the breast and AXLN was 23.9% (11/46). The rate of BCS was 42.9% (21/49). The pCR rate in the breast and AXLN was 30.8% (4/13) for the triple-negative disease and 19.4% (7/36) for hormone receptor-positive disease.

Clinical response was observed in 67.4% of patients (6 complete and 27 partial responders among 49 patients). Fourteen patients (28.6%) showed stable disease and one patient (2.0%) showed progressive disease. According to the site of the tumor, the clinical response rates were 67.4% in the breast (7 complete and 26 partial responders), 55.1% in the AXLN (27 complete responders), and 40.0% in the SCL (2 complete responders out of 5 with SCL metastasis). Subsequent clinical response to SD chemotherapy was evaluated in 31 patients who had serial sonographs. The median tumor size in the breast was 26 mm, 13 mm, and 10 mm at baseline, completion of AC, and completion of SD chemotherapy, respectively. Compared to tumor size before SD chemotherapy, subsequent clinical response rates in the breast to SD chemotherapy were: complete response, 6.9% (2/29); partial response, 31.0% (9/29); stable disease, 62.1% (18/29). Two complete responses to AC chemotherapy were maintained after subsequent SD chemotherapy.

In an exploratory finding, tumors with high Ki-67 (≥30%) or small initial size of tumor (≤2 cm) showed a trend toward a higher pCR rate than those without but they did not reach statistical significance (Table 3).

All grades of hematologic (by cycle) and non-hematologic (by patient) toxicities are listed in Table 4. Toxicities of AC chemotherapy were mainly grade 1–2 nausea, vomiting and anorexia as expected. In the SD cycles, the dominating grade 3–4 toxicity was myelosuppression and epigastric pain. Grade 3–4 hematological toxicities were neutropenia (8.5%) and febrile neutropenia (2.9%). After two cycles of SD, one neutropenic patient experienced infection that led to death. S-1 specific toxicities were as follows: epigastric pain, 20.4% at grade 1–2 and 12.2% at grade 3; stomatitis, 28.6% at grade 1–2 and 4.1% at grade 3; diarrhea, 8.2% at grade 1–2 and 2.0% at grade 3; hand-foot syndrome, 8.2% at grade 1–2 and 2.0% at grade 3.

Since AC chemotherapy is a routine treatment and SD chemotherapy is experimental, dose-intensity analysis focused on SD chemotherapy. In total, 172 cycles of SD chemotherapy were delivered. Docetaxel was delayed in 7 cycles (4.1%), mainly due to neutropenia. At least one dose of S-1 was omitted in 27 cycles (15.7%) due to stomatitis (6 cycles), epigastric pain (4 cycles), and other toxicities (17 cycles). Of 48 patients who received at least one cycle of SD chemotherapy, 12 patients (25%) required dose reduction of both docetaxel and S-1 due to infection (n = 2), neutropenia (n = 2), asthenia (n = 2), and other toxicities (n = 6). Two patients (4.2%) required dose reduction in only docetaxel and 10 patients (20.8%) in only S-1, mainly due to epigastric pain (n = 7). Five of these 10 patients discontinued S-1 completely after 1 (n = 1), 2 (n = 2), or 3 cycles (n = 2). The mean relative dose intensity (RDI) of docetaxel and S-1 was 91.4% and 79.6%, respectively.