Introduction

Gastric cancer is the fifth most common malignancy and the third leading cause of cancer mortality worldwide [1]. Although surgical resection offers the only chance of long-term cure, the majority of patients with gastric or gastroesophageal junction (GEJ) cancer present with locally advanced or metastatic disease at diagnosis, and they are candidates to receive systemic chemotherapy [2, 3]. Among chemotherapy agents, cisplatin (CDDP) and oxaliplatin (l-OHP), docetaxel (D), the fluoropyrimidines 5-fluorouracil (5FU) and capecitabine, and the anthracyclines epirubicin (EPI) and doxorubicin (DOX) are the most active drugs, which are conventionally administered in advanced gastric or GEJ cancer [4]. The oral agent S-1 plus CDDP achieved a better median progression-free survival (PFS) and better overall survival (OS) compared with CDDP alone in a randomized phase III trial, and this combination regimen became the standard treatment for advanced gastric cancer in Japan [5]. In addition, the randomized TOGA trial demonstrated that patients with positive human epidermal receptor 2 (HER-2) tumors may benefit from chemotherapy combined with the monoclonal antibody trastuzumab in terms of PFS and OS [6]. The substitution of l-OHP for CDDP guarantees a better toxicity profile, and EOF or EOX and DOF or DOX regimens are effective and largely used treatments for patients with advanced oesophagogastric cancer [7]. These chemotherapy regimens are conventionally administered until the occurrence of progressive disease or unacceptable toxicity and achieve about 50 % objective response rate, 6- to 7-month median PFS and <12-month median OS. Triplet drug combinations usually achieve better results than those obtained with single agents or with doublet drug combinations of l-OHP and 5FU [7–9]. However, despite the administration of triple drug combination chemotherapy, the prognosis of patients with advanced gastric or GEJ cancer remains very poor and more effective treatment options are required. In this setting, a sequential strategy might offer the opportunity to administer full doses of most active drugs as first line with acceptable toxicity. In a sequential treatment, drugs are given as planned, even in the presence of response with no signs of resistance at the end of the first part of the sequence [10, 11]. The administration of a non-cross-resistant agent with a different mechanism of agent by a sequential treatment might allow drug resistance to be circumvented [10]. Nevertheless, in other solid tumors such as lung cancer, the prolonged chemotherapy administration does not seem more effective than 3–4 cycles of a specific regimen [12]. The sequential chemotherapy has achieved promising results in non-small cell lung cancer, colorectal and breast cancer [13–15]. In the past, Cascinu et al. reported promising results in terms of overall response rate and PFS in advanced gastric cancer patients treated with a multidrug combination regimen followed by single-agent D [17]. However, to date, no efficacy and toxicity data exist in advanced gastroesophageal cancer about a sequential strategy, which could combine a decreased anthracycline exposure with the possibility to receive both anthracyclines and taxanes. The primary objective of this study was to determine the activity and safety profile of the sequential treatment with EOF/DOF in patients with advanced gastric or GEJ cancer.

Patients and methods

Eligibility criteria

The study enrolled patients with histologically proven advanced adenocarcinoma of the stomach or GEJ, who had not previously received chemotherapy for advanced disease. Patients who had been treated with adjuvant 5-FU or capecitabine-based chemotherapy were eligible provided that they had remained disease free at least 12 months after the completion of adjuvant therapy. The other eligibility criteria included age more than 18 years, Eastern Cooperative Oncology Group performance status of 0–2, bidimensionally measurable disease, a life expectancy of at least 3 months, adequate hematological parameters (an absolute neutrophil count of ≥1.5 × 109/l and a platelet count of ≥100 × 109/l), creatinine and total bilirubin levels <1.25 × the upper normal limit, aspartate and alanine aminotransferase <3.0 × the upper normal limit, and absence of a second primary tumor other than non-melanoma skin cancer or in situ cervical carcinoma. Patients with operable metastatic disease were excluded from the study, as were those with severe cardiac dysfunction, chronic diarrhea or uncontrolled sites of infection. Patients with HER-2-positive tumors were also excluded. This study was approved by the local ethical and scientific committee, and all of the patients gave their written informed consent.

Patient evaluation

The pre-treatment evaluation, performed within 2 weeks before study entry, included a detailed history and physical examination, a complete blood cell count with differential and platelet counts, whole-blood chemistry and computed tomography (CT) scans and/or magnetic resonance imaging (MRI) of the chest and abdomen. During treatment, a complete blood cell count with differential and platelet counts was performed every 10 days. In addition, the patients were clinically assessed every 3 weeks, and routine biochemical tests were performed. Heart function by echocardiography was evaluated before the start and after the completion of EOF and DOF treatment regimens, respectively. Treatment response by means of CT scan and/or MRI was evaluated every four 3-weekly cycles or sooner if clinically indicated. Tumor response was assessed using the RECIST criteria [16].

Treatment delivery

Treatment consisted of intravenous (i.v.) Epi 50 mg/m2 combined with 6-h i.v. l-OHP 130 mg/m2 on day 1, and continuous infusion (c.i.) 5FU 750 mg/m2 days 1–5 (EOF regimen), every 3 weeks for a maximum of 4 cycles. After EOF completion, patients received i.v. D 70 mg/m2 combined with 6-h i.v. l-OHP 130 mg/m2 on day 1, and c.i. 5FU 750 mg/m2 days 1–5 (DOF regimen), every 3 weeks for a maximum of 4 cycles. If patients developed progressive disease before EOF completion, they were transferred to the DOF regimen.

After the completion of the sequential treatment strategy, the patients who achieved complete or partial response or stable disease continued the maintenance treatment with c.i. 5FU 750 mg/m2 days 1–5 every 3 weeks until progressive disease or unacceptable toxicity.

Toxicity

Toxicity was assessed using the common toxicity criteria of the National Cancer Institute (NCI), version 3.0. Treatment was delayed if, on the planned day of treatment, the neutrophil count was <1500/mm3, the platelet count was <100,000/mm3, or the patient had persistent diarrhea or stomatitis >grade 1. Any patient who required more than 2 weeks for recovery from adverse reactions was excluded from the study. In the event of grade 4 hematological or any other severe (≥grade 3) organ toxicity in individual patients, the doses of chemotherapy drugs were reduced by 25 % for subsequent courses. The l-OHP dose was also reduced by 25 % in case of persistent (≥14 days) paresthesia or temporary (7–14 days) painful paresthesias or functional impairment. In case of persistent (>14 days) painful paresthesia or functional impairment, l-HOP was omitted from the subsequent cycles until recovery.

In order to prevent nausea and vomiting, i.v. hydroxytryptamine-3 antagonists plus dexamethasone 8 mg i.v. were administered before the chemotherapy infusion. Oral loperamide 2 mg every 2 h and oral rehydration were prescribed in the case of delayed diarrhea. No cytokine prophylactic treatment was recommended. The other concomitant medications were primarily used to palliate pain.

Statistical considerations

The primary endpoint was 6-month disease control rate (DCR). Most studies investigating effective combination chemotherapy suggest that approximately 60 % of advanced gastric cancer patients are progression free 6 months after the start of treatment [7]. The hypothesis for the current study was that, using the sequential EOF/DOF strategy, at least 80 % of patients would be progression free 6 months after the start of chemotherapy. It was calculated that a total of 43 patients would have to be recruited to yield an 80 % probability to correctly select the treatment when it is superior by absolute difference of 20 % in 6-month DCR (Simon’s optimal design) [18]. PFS was calculated as the time from the first chemotherapy infusion to disease progression or death. Secondary end points included safety, OS (measured from the date of start of treatment to the date of death) and response rate. Kaplan–Meier method was used to determine PFS and OS. Statistical analyses were conducted by STATA IC 2012 software.

Results

Patient characteristics

Forty-five patients were enrolled from September 2010 to April 2014. All patients were radiologically evaluated at our Medical Oncology Unit. Baseline characteristics of the patients are presented in Table 1. The median age was 69 years (range 34–83). In all, 82.2 % of patients had a performance status of 0 or 1, 62.2 % had liver metastases, and 80.0 % had multiple metastatic sites.

Table 1 Patient characteristics
Full size table

Primary tumor was gastric in 26 patients and GEJ in 19 patients. Seven patients (15.5 %) had received prior 5-FU-based adjuvant chemotherapy (3 patients had received prior l-OHP).

Treatment efficacy

All patients received at least one treatment cycle and were evaluable for response and toxicity: The median of the cycles of each treatment was 4.

In an intent-to-treat analysis, after the completion of the EOF/DOF sequential treatment, documented complete response (CR) was observed in one patient (2.2 %) and partial response (PR) in 22 patients (48.8 %), with an overall response rate of 51.1 % (95 % CI 35.7–66.2 %). Nineteen patients (42.2 %) had stable disease (SD), and 3 (6.6 %) had progressive disease (PD).

Ten patients improved their outcome in the switch from EOF to DOF: One patient with PR achieved CR, seven with SD achieved PR, and two with PD achieved SD (Table 2).

Table 2 Best response to treatment
Full size table

All the responses were confirmed by a blinded radiologic review committee.

All the enrolled patients were included in PFS analysis: Two patients who progressed after EOF and then responded to sequential DOF were included in PFS analysis as non-progressing subjects after sequential EOF/DOF. The median PFS of the patients was 9.5 months (95 % CI 8.0–11.9 months): Forty-two patients (93.3 %) were progression free after about 6 months from the onset of sequential chemotherapy. After the completion of EOF/DOF, 41 of these 42 patients were treated with c.i. 5FU 750 mg/m2 days 1–5 every 3 weeks as maintenance for a median of 4.2 months (range 3–12 months). One patient was lost to follow-up after 7 months from the onset of sequential EOF/DOF for treatment-unrelated reasons. A total of 28 patients received a further treatment line at documentation of progressive disease: 16 patients received irinotecan-based chemotherapy, 8 days alone, and 4 patients received CDDP-based chemotherapy.

At a median follow-up time of 16.7 months, 37 patients were deceased: The median OS was 15.8 months (95 % CI 13.6–18.9 months). PFS and OS of the 45 enrolled patients are illustrated in Fig. 1.

Fig. 1
figure 1

Kaplan–Meier curves for PFS and OS of 45 advanced gastric or GEJ cancer patients

Full size image

The 14 patients who achieved PR after EOF and then switched to DOF exhibited 11.2-month median PFS and 16.9-month median OS.

The median PFS was 9.5 months, and the median OS was 15.2 months in patients with gastric cancer; the median PFS was 8.4 months, and the median OS was 14.3 months in patients with GEJ cancer.

Toxicity

A total of 180 EOF cycles and 178 DOF cycles were administered. Toxicity during sequential EOF/DOF was acceptable and easily manageable with adequate supportive care (Table 3). One patient interrupted DOF treatment after 2 cycles for toxicity-unrelated reasons. Similar proportions of hematological and non-hematological adverse events were observed with either EOF than with DOF. Fifteen of 45 enrolled patients (33.3 %) experienced at least one episode of grade 3 neutropenia; two patients had two episodes of grade 3 neutropenic fever during DOF, and other two patients experienced grade 4 neutropenia during DOF regimen. Among non-hematological toxicities, grade 4 was not observed: Grade 3 stomatitis was reported in six patients (13.3 %). Grade 3 neurotoxicity was documented in one patient after 8 cycles of sequential EOF/DOF. No moderate to severe cardiotoxicity was observed. A total of 19 EOF cycles (10.5 %) and 23 DOF cycles (12.9 %) were delayed for at least 1 week because of toxicity, and dose reductions were required in five patients during EOF and in seven patients during DOF regimen.

Table 3 Number of patients with the most frequent treatment-related adverse events EOF/DOF
Full size table

Discussion

The results of the current study indicate the feasibility and efficacy of a first-line treatment with a sequential strategy that includes EOF regimen followed by DOF regimen in the treatment of advanced gastric or GEJ cancer patients. Albeit with the limitations of a phase II study, 93.3 % 6-month DCR, 9.5-month median PFS and 15.8-month median OS compare well with the best ones reported in this disease with most active combination chemotherapies. Interestingly, 14 patients who achieved a response after EOF and who switch to DOF achieved a favorable median PFS of 11.2 months and median OS of 16.9 months. In clinical practice, EPI- or D-based triple drug regimens are conventionally administered until the occurrence of progressive disease and usually achieve <9-month PFS and <12-month OS [4, 7]. It should be noted that in the current study, the interpretation of its primary endpoint might be difficult as well as the comparison with other studies since the PFS analysis excluded patients who progressed after EOF with consequent potential over interpretation of its results. Nevertheless, if we try to exclude from the PFS analysis five patients who progressed after EOF, the value of median PFS remains unchanged.

Encouraging results in terms of PFS and OS were observed in the current study both in gastric cancer patients and in patients with GEJ tumors. It is known that the incidence of GEJ cancer is progressively increasing worldwide, and its prognosis has been usually worse compared with gastric cancer, with median PFS <6–7 months and OS not longer than 10 months [19, 20]. In this setting, our results suggest that a sequential treatment strategy that includes the most active agents may achieve a high and similar efficacy in patients with gastric cancer and in patients with GEJ cancer. Moreover, our findings seemed similar or even better than those reported in other studies, which investigated D-based chemotherapy administered as second-line treatment upon documented disease progression in advanced gastric cancer [21, 22]. Nevertheless, an advantage of a sequential strategy in these aggressive tumors may be the possibility to administer full doses of all most active agents such as EPI, l-OHP, 5FU and D in first-line treatment. It should be also considered that l-OHP and 5FU were administered for all the 8 planned cycles, suggesting the key role of the combination of l-OHP and 5FU/FA in the treatment of advanced oesophagogastric cancer. Nevertheless, we acknowledge that the small sample size and the mono-institutional nature of the current study are limitations, which should be considered for a correct interpretation of data, and thus, caution should be exercised before drawing firm conclusions. The actual benefit of a sequential treatment with EOF/DOF should have to be proven in a randomized comparison before this treatment strategy is preferred to conventional chemotherapeutic approach.

Other studies performed in advanced gastric cancer reported interesting results with a sequential strategy, which included a number of different chemotherapeutic agents. CDDP, leucovorin and 5FU followed by D showed an effective palliative option with a more favorable toxicity profile than the simultaneous use of the same drugs in patients with metastatic gastric cancer [23]. A randomized phase II study demonstrated that the sequential administration of 5FU/FA combined with D or irinotecan achieves similar efficacy to triple drug combination, with less toxicity [24]. A sequential strategy became standard in neoadjuvant treatment in breast cancer, and interesting results were described by a sequential alternating treatment in advanced NSCLC, as well as in a recent randomized phase II trial in advanced pancreatic cancer [25–27]. In the search to improve patient outcomes with targeted therapy in advanced gastric cancer, some innovative agents were recently investigated, such as the antiangiogenic agent bevacizumab, the anti-EGFR panitumumab and the oral mammalian target of rapamycin everolimus [28–30]. However, to date, no significant improvements in PFS and OS have been obtained with new agents combined with chemotherapy, and further research is required about treatment of patients with advanced gastric or GEJ cancer. In this scenario, a sequential treatment strategy might have a role in order to strike a balance between prolonged clinical response and mild toxicity.

An interesting point in the current study is that the sequential EOF/DOF we applied was well tolerated, with a low percentage of cycle delays and dose reductions of drugs. The proportion of patients suffering from grade 3 myelotoxicity, grade 3 nausea and vomiting and diarrhea appeared similar to that usually reported with the conventional EOF, ECF, EOX or DCF, DCF and DCX [7]. The use of EPI and D for only 4 cycles, respectively, may have contributed to the low incidence of these agents-related adverse events. L-OHP-related neurotoxicity was mild and reached grade 3 in only one patient. The proportions of adverse events were in line with those reported in other studies in which l-OHP has been substituted for CDDP, suggesting that l-OHP caused lower incidences of hematological and non-hematological toxicity than CDDP, but was associated with slightly higher incidences of neuropathy [7, 9]. Furthermore, a prolonged l-OHP infusion time is usually adopted at our oncology institution, and as reported in a previous study, we believe it is the main reason for the low incidence rate of moderate to severe neurotoxicity in the current study [31].

As far as 5FU administration is concerned, we adopted 5-day c.i. 5FU since previous reports and our encouraging experience suggested activity, tolerability and patient’s convenience for this treatment approach [32–34]. 5FU is administered as 200 mg/m2 every day by protracted continuous infusion in most studies about EOF, ECF, DCF and DOF regimens, but we believe that a continuous 5FU infusion for several months by an external infusion pump may be discomfortable for some patients. In this setting, capecitabine, a 5FU prodrug, is often substituted for c.i. 5FU. However, patients who present with locally advanced GEJ or gastric cancer or that exhibit severe peritoneal metastasis or massive ascites may have severe difficulties in oral intake, and therefore, the assumption of approximately 4–6 capecitabine pills every day may be discomfortable [35].

In summary, the conventional triple drug regimens EOF or EOX, or DOF or DOX actually remain the standard platform to consider for further clinical studies in oesophagogastric cancer in western countries. The results of the current study suggest that a sequential treatment strategy with EOF and DOF is feasible and may achieve encouraging results in terms of PFS and OS in well-selected patients with advanced gastric or GEJ cancer. The EOF/DOF sequential treatment might deserve further investigation.