Phase I/II study of S-1 combined with weekly docetaxel in patients with metastatic gastric carcinoma

We designed a phase I/II trial of S-1 combined with weekly docetaxel to determine the maximum tolerated dose (MTD) and recommended dose (RD) and to evaluate the efficacy and toxicity in metastatic gastric carcinoma (MGC). Patients with measurable disease received S-1 orally b.i.d. on days 1–14 and docetaxel intravenously on days 1 and 8 every 3 weeks. In phase I (n=30), each cohort received escalating doses of S-1 (30–45 mg m−2 b.i.d.) and docetaxel (25–40 mg m−2); MTD was 45 mg m−2 b.i.d. S-1/35 mg m−2 docetaxel and RD was 40 mg m−2 b.i.d. S-1/35 mg m−2 docetaxel. Dose-limiting toxicities included grade 3 elevated liver enzymes, gastric perforation, grade 3 diarrhoea/fatigue, febrile neutropenia with grade 3 anorexia/fatigue, and neutropenic infection with grade 3 stomatitis/anorexia. In phase II (n=52), the overall response rate was 66.7% (95% confidence interval (CI): 53.8–79.6%) and the median time to progression and overall survival were 6.5 months (95% CI: 4.9–8.1) and 13.7 months (95% CI: 9.9–17.5), respectively. The most common grade 3/4 toxicity was neutropenia (29.4%), and febrile neutropenia/neutropenic infection occurred in 19.6% of patients. Non-haematological toxicities were generally mild. There was one treatment-related death due to pneumonitis. S-1 combined with weekly docetaxel is active in MGC with moderate toxicities.

Despite its decreasing incidence over the past few decades, gastric cancer remains one of the major cause of death due to cancer worldwide (Parkin et al, 2005). Survival benefits have been demonstrated by systemic chemotherapy in patients with locally advanced or metastatic gastric carcinoma (MGC) (Pyrhonen et al, 1995;Glimelius et al, 1997); however, the results of most combination regimens have been unsatisfactory, with median survival times of 6 -9 months (Webb et al, 1997;Vanhoefer et al, 2000;Ohtsu et al, 2003). To date, the most commonly used combination chemotherapies have been based on fluorouracil (5-FU) and/or cisplatin, with 5-FU/cisplatin (FP) and epirubicin/ cisplatin/5-FU being currently regarded as reference treatments. Although a recent phase III trial showed that patients treated with docetaxel combined with FP had superior survival to patients treated with FP alone, the former regimen also had severe toxicities, thereby limiting its application (Van Cutsem et al, 2006). In addition, cisplatin-based chemotherapy is frequently associated with an unfavourable toxicity profile, including severe emesis, neurotoxicity, and nephrotoxicity. Continuous intravenous infusion of 5-FU also results in inconvenience to patients and catheter-related complications. Therefore, there is a need to develop active, but less-toxic, chemotherapy regimens, which include new active compounds.
Docetaxel, which inhibits microtubule depolymerisation, has been widely used in the treatment of MGC, with response rates of 16 -24% when used as a single agent in phase II trials (Sulkes et al, 1994;Bang et al, 2002). Compared with the 3-weekly regimen, docetaxel administered once weekly has a favourable safety profile, including myelosuppression (Schuette et al, 2005;Bria et al, 2006;Camps et al, 2006).
Due to the activity of S-1 and docetaxel in MGC, and their synergistic activity in gastric cancer cell lines and xenografts (Takahashi et al, 2005;Wada et al, 2006), we designed a phase I/II trial of S-1 combined with weekly docetaxel in patients with MGC to determine the maximum tolerated dose (MTD) and recommended dose (RD) of these agents when used together, and to evaluate their efficacy and toxicity.

Eligibility
Patients were eligible for this trial if they were over 18 years of age with histologically proven MGC, unidimensionally measurable disease, an ECOG (Eastern Cooperative Oncology Group) performance status of 0 -2, and possessed adequate baseline haematological function (ANC (absolute neutrophil count) X1.5 Â 10 9 l À1 , platelet count X100 Â 10 9 l À1 ), hepatic function (serum aspartate aminotransferase and alanine aminotransferase p2.5 times ULN (upper limit of normal) and serum bilirubin pULN), and renal function (serum creatinine pULN). Patients in the phase II part had received no prior chemotherapy, including adjuvant chemotherapy, whereas patients in the phase I part were permitted up to two previous chemotherapy regimens, except for prior taxane or S-1, and were required to have discontinued chemotherapy for at least 4 weeks before participating in this study.
Patients were excluded if they had a history of other malignancies within the previous 3 years, had severe comorbid conditions, or lacked the ability to comply with the requirements of the protocol. Patients receiving drugs with potential interactions with S-1 (e.g., flucytosine, allopurinol, and phenytoin) were also excluded. All patients provided written informed consent, and the protocol was approved by our Institutional Review Board.

Pretreatment evaluations
Baseline evaluations included medical history, physical examination, ECOG performance status, complete blood cell count, serum chemistries and electrolytes, urinalysis, urine pregnancy test (for women), 24-h urine creatinine clearance, chest X-ray, computed tomography, electrocardiogram, and recording of concomitant medications.

Treatment and study design
During each 3-week cycle, patients received oral S-1 twice daily (within 1 h after morning and evening meals) on days 1 -14 and a 1 h intravenous infusion of docetaxel on days 1 and 8. All the patients received oral dexamethasone (4 mg twice daily for 4 doses, starting 12 h before docetaxel) and parenteral pheniramine maleate (45.5 mg) prophylactically. Prophylactic administration of granulocyte colony-stimulating factor and antiemetics was not allowed; however, secondary prophylaxis or therapy with antiemetics in subsequent cycles was allowed. Treatment was continued in the absence of disease progression or unacceptable toxicity for a maximum of 12 cycles.
During phase I, each cohort of at least three patients was treated with escalating doses of S-1/docetaxel: 30 mg m À2 b.i.d./25 mg m À2 (level 1), 35 mg m À2 b.i.d./25 mg m À2 (level 2), 35 mg m À2 b.i.d./ 30 mg m À2 (level 3), 40 mg m À2 b.i.d./30 mg m À2 (level 4), 40 mg m À2 b.i.d./35 mg m À2 (level 5), 45 mg m À2 b.i.d./35 mg m À2 (level 6), and 45 mg m À2 b.i.d./40 mg m À2 (level 7). Dose escalation was continued until at least one-third of the patients in a given cohort showed dose-limiting toxicity (DLT) during the first cycle. Before escalating to the next dose level, all three patients should have received at least one treatment cycle. If none of the first three treated patients developed DLT during the first cycle at a specific dose level, dose escalation was continued. If one out of the first three treated patients developed DLT at any dose level, three additional patients were entered at the same dose level; if only one in six patients at a given level experienced a DLT, dose escalation was continued. The MTD was defined as the dose level at which one-third or more of patients experienced a DLT. The RD for the subsequent phase II study was defined as the dose level preceding the attainment of the MTD.
Phase II was performed using the RD determined during phase I.

Dose modifications
The next chemotherapy cycle was delayed if patients had ANC o1.5 Â 10 9 l À1 , platelet count o100 Â 10 9 l À1 , or any grade 41 non-haematological toxicity, excluding alopecia. Docetaxel dose was reduced by 20% in subsequent cycles if patients experienced grade 3/4 neutropenia associated with infection or fever (X38.31C as single temperature or X38.01C for 1 h), grade 4 thrombocytopenia, grade 3 non-haematological toxicity, grade 2/3 neurological toxicity, or recurrent fluid retention. Docetaxel treatment on day 8 was delayed to day 10 for ANC o1.5 Â 10 9 l À1 , platelet count o75 Â 10 9 l À1 , or grade X2 non-haematological toxicity on the day of scheduled treatment. On day 10, docetaxel dose was reduced by 50% for ANC of 0.5 Â 10 9 l À1 -1.0 Â 10 9 l À1 or platelet count of 50 Â 10 9 l À1 -75 Â 10 9 l À1 , and omitted for other toxicities. S-1 dose was reduced by 20% in subsequent cycles if patients experienced grade 3/4 neutropenia associated with infection or fever, grade 4 thrombocytopenia, a second occurrence of a grade 2 nonhaematological toxicity, or any grade 3 non-haematological toxicity. If patients experienced a grade 4 non-haematological toxicity, docetaxel and S-1 were definitively interrupted or continued at doses 50% less than the starting dose.

Evaluation during chemotherapy
During phase I, complete blood cell count and serum chemistries were monitored twice weekly and once weekly, respectively, during the first cycle, except for the first week and on days 1 and 8 of each subsequent cycle. During phase II, complete blood cell count and serum chemistries were monitored weekly during the first two cycles and on days 1 and 8 of each subsequent cycle.

Assessment of efficacy and toxicity
Computed tomography scans were performed every two cycles to evaluate the tumour response, which was assessed according to the Response Evaluation Criteria in Solid Tumors (Therasse et al, 2000). Objective responses were confirmed by a second evaluation 4 -6 weeks later. Time to progression (TTP) was calculated from the date of first chemotherapy cycle to the date of disease progression, and overall survival (OS) was calculated from the date of first chemotherapy cycle to either the date of death due to any cause or the date of the last follow-up visit. Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria (version 2.0).

Phase II study: statistical planning and analysis
The primary end point for the phase II part of the study was the objective response rate, and the secondary endpoints were TTP, OS, and safety. A two-stage optimal design proposed by Simon was used to determine the sample size of phase II (Simon, 1989).    Twelve patients (23.5%) had stable disease and five (9.8%) had progressive disease. All objective responses were confirmed by follow-up computed tomography at least 4 weeks after the initial documentation of response. The median duration of response was 6.3 months (range: 1.6 to 13.8 þ ). The median follow-up time was 13.1 months (range: 8.5 -18.0), during which the median TTP was 6.5 months (95% CI: 4.9 -8.1 months) ( Figure 1) and the median OS was 13.7 months (95% CI: 9.9 -17.5) (Figure 2). The one-year survival rate was 58.5% (95% CI: 44.2 -72.8%).
Toxicity Fifty-one patients were assessable for toxicity. Table 3 summarises chemotherapy toxicities per patient. The most common grade 3/4 haematological toxicities were neutropenia (29.4% of patients) and leukopenia (29.4%). Grade 3/4 febrile neutropenia and grade 3 infection with neutropenia each occurred in five patients (9.8%). All these patients were successfully treated with antibiotics and granulocyte colony-stimulating factor. Grade 3 anaemia occurred in three patients (5.9%), but no patient experienced grade 3/4 thrombocytopenia.

DISCUSSION
We have shown here that S-1 combined with weekly docetaxel is a highly active first-line chemotherapy regimen for MGC. The overall response rate of 66.7%, median TTP of 6.5 months, and median OS of 13.7 months are comparable to results of trials using older (Webb et al, 1997;Vanhoefer et al, 2000;Ohtsu et al, 2003) or newly developed chemotherapeutic agents (Roth et al, 2000;Ridwelski et al, 2001;Al-Batran et al, 2004;Bouche et al, 2004;Park et al, 2004;Chun et al, 2005;Kim et al, 2005;Moehler et al, 2005;Giordano et al, 2006;Kang et al, 2006;Orditura et al, 2006;Van Cutsem et al, 2006), including oxaliplatin, irinotecan, capecitabine, and other docetaxel-containing regimens. With the latter agents, response rates ranged from 9 to 60%, median TTP from 1.9 to 6.9 months, and median OS from 5.7 to 12.0 months (Roth et al, 2000;Ridwelski et al, 2001;Al-Batran et al, 2004;Bouche et al, 2004;Park et al, 2004;Chun et al, 2005;Kim et al, 2005;Moehler et al, 2005;Giordano et al, 2006;Kang et al, 2006;Orditura et al, 2006;Van Cutsem et al, 2006). Recently, phase III trials of S-1 alone and/or S-1 combined with cisplatin showed that S-1 alone had non-inferior OS compared with infusional 5-FU (11.4 vs 10.8 months, non-inferiority Po0.001), and then S-1 combined with cisplatin had superior OS to S-1 alone (13.0 vs 11.0 months, P ¼ 0.0366) (Boku et al, 2007;Narahara et al, 2007). Considering the potential of the S-1-based combination regimen, the combination of S-1 and the new cytotoxic agent in the present study deserves comparison in further phase III trials.
During phase I of our trial, we determined that the RD and treatment schedule for phase II was 40 mg m À2 b.i.d. S-1 on days 1 -14 and 35 mg m À2 docetaxel on days 1 and 8 of each 3-week cycle. Two recent trials of S-1 combined with docetaxel used different doses and schedules; the first used 40 mg m À2 S-1 b.i.d. on days 1 -14 and 40 mg m À2 docetaxel on day 1 of each 3-week cycle , and the second used 40 mg m À2 b.i.d. S-1 on days 1 -14 and 40 mg m À2 docetaxel on day 1 of each 4-week cycle (Yamaguchi et al, 2006). Despite these variations, the DLTs in phase I of these trials included neutropenia and complicated neutropenia, similar to our trial (Yoshida et al, 2004;Yamaguchi et al, 2006). Although a direct comparison between other phase II trials and our trial is difficult, one previous trial, which had the same planned dose intensity of S-1 and a different dose intensity of docetaxel as our trial, showed similar efficacy, with an overall response rate of 56.3% (compared with 66.7% in our trial), a median TTP of 7.3 months (compared with 6.5 months), and a median OS of 14.3 months (compared with 13.7) . About 10% of patients in this trial, however, had locally advanced disease, whereas all of our patients had metastatic disease . Moreover, the patients in this trial received a median of four cycles (in a 3-week cycle) and had a median response duration of only 5.1 months . This discrepancy between the relatively low number of median treatment cycles per patient and long median TTP may not be usual treatment outcomes. In our study, patients received a median eight cycles of chemotherapy and had a median response duration of 6.3 months. Another previous trial, which used a lower dose intensity (40 mg m À2 b.i.d. S-1 on days 1 -14 and 40 mg m À2 docetaxel on day 1 of each 4-week cycle) compared with our trial, showed inferior efficacy, with an overall response rate of 46% and a median progression-free survival of 4.1 months (Yamaguchi et al, 2006). It may be attributed to the weekly regimen that a higher dose of docetaxel could be determined as RD in our trial than in previous two trials; docetaxel administered once weekly has lower toxicity with comparable efficacy compared with the 3weekly regimen (Schuette et al, 2005;Bria et al, 2006;Camps et al, 2006).
The regimen used in the present trial resulted in a generally low incidence of grade 3/4 haematological toxicities, including neutropenia (29.4%), and non-haematological toxicities. Febrile neutropenia and infection with neutropenia, however, occurred in a relatively high proportion of patients (19.6%). This may be related to the relatively high incidence of grade 1/2 stomatitis (80.4%) and diarrhoea (72.5%), in that disruption of the mucosal barrier may make patients susceptible to infection. This high incidence of grade 1/2 non-haematological toxicities may be associated with the long treatment duration, a median of eight cycles per patient. Due to the frequency of febrile neutropenia or infection with neutropenia and the actual dose intensity of S-1 and docetaxel in phase II of our trial, we propose that the RD of this regimen be lowered to 35 mg m À2 b.i.d. S-1 on days 1 -14 and/or 30 mg m À2 docetaxel on days 1 and 8 of each 3-week cycle, or prophylactic granulocyte colony-stimulating factor be used.
Notably, this regimen resulted in a low incidence of nausea and vomiting despite the absence of primary prophylactic antiemetics. Grade 2 nausea and vomiting occurred in 31.4 and 17.6% of patients, respectively, and grade 3 nausea developed in only one patient (2.0%). During the entire treatment period, only 19 patients (36.5%) in 49 cycles (12.7%) were given a serotonin antagonist for secondary prophylaxis during subsequent cycles or therapy. The current regimen compares favourably with cisplatin, irinotecan, oxaliplatin, or infusional 5-FU-containing regimens, in which the incidence of grade 3/4 nausea/vomiting ranged from 4.9 to 26% with prophylactic antiemetics (Webb et al, 1997;Roth et al, 2000;Vanhoefer et al, 2000;Ridwelski et al, 2001;Ohtsu et al, 2003;Al-Batran et al, 2004;Bouche et al, 2004;Moehler et al, 2005;Kang et al, 2006;Van Cutsem et al, 2006). This regimen also resulted in a very low incidence of grade 2/3 hand-foot syndrome (5.9%), which is quite troublesome and has been reported to occur in 12.9% to more than 50% of patients treated with capecitabine, another oral fluoropyrimidine, combined with docetaxel (Park et al, 2004;Chun et al, 2005;Kim et al, 2005;Giordano et al, 2006).
Taken together, we conclude that S-1 combined with weekly docetaxel is a highly active outpatient regimen in MGC with moderate toxicity. Further studies with appropriate dose modifications are warranted.