Irinotecan plus S-1 versus S-1 in patients with previously treated recurrent or metastatic esophageal cancer (ESWN 01): a prospective randomized, multicenter, open-labeled phase 3 trial

Esophageal cancer is the sixth most common type of malignancy in the world [1]. Esophageal squamous cell carcinoma (ESCC) is the predominant histological type of esophageal cancer in Eastern Europe and Asia. The use of platinum plus paclitaxel or fluorouracil-based chemotherapy are the preferred regimens for the first-line treatment of esophageal cancer, but despite such, the overall survival (OS) of these patients remains less than 1 year [2].

The clinical benefits of irinotecan [3], docetaxel [4], and ramucirumab [5] as second-line regimens for advanced esophageal adenocarcinoma have been validated in phase 3 trials, but the second-line treatment options for advanced ESCC patients remain investigational. The benefit of second-line targeted treatment in patients with advanced ESCC is still uncertain, and only a few treatment options are available [3]. The Cancer Oesophagus Gefitinib (COG) trial [4], which included 50 patients with ESCC in the gefitinib arm, is the largest randomized phase 3 trial performed for esophageal cancer patients whose disease progressed after chemotherapy. In that study, the gefitinib arm demonstrated a modest antitumor efficacy with an overall response rate (ORR) of 2.7% and a disease control rate (DCR) of 24%, and the observed median progression-free survival (PFS) and OS were 1.57 months and 3.73 months, respectively. Although no OS benefit from gefitinib over placebo was observed, a small benefit in PFS (1.57 vs. 1.17 months, P = 0.020) was achieved. By comparison, in another phase 2 trial, when pretreated ESCC patients having epidermal growth factor receptor (EGFR) overexpression or amplification were treated with icotinib, an EGFR-tyrosine kinase inhibitors (TKI), they demonstrated an ORR of 16.7% and a median PFS of 52 days [5]. It is noteworthy that despite the increased response rate in selected patients observed in the latter trial, the median PFS was still disappointing. Recently, the effect of immune checkpoint inhibitors on ESCC have also been evaluated in clinical trials, with ORR ranging between 17% and 33% and median PFS between 1.5 and 3.6 months [6–8]. However, only a small number of patients (n = 18–65) with ESCC were included in these phase 1/2 trials, and no standard treatments were established. As such, chemotherapy still remains the cornerstone in the treatment of advanced ESCC.

The combination of docetaxel with platinum has demonstrated a response rate of 25%–34.2% in metastatic ESCC patients whose disease progressed after first-line chemotherapy on platinum plus fluorouracil [9–11]. S-1, a fluoropyrimidine derivative, has been reported in a phase 2 trial as a second-line chemotherapy for patients with unresectable and recurrent ESCC, resulting in a median OS of 330 days [12]. Irinotecan, a topoisomerase I inhibitor, has demonstrated moderate activity as a single agent in metastatic gastro-esophageal cancer [13]. In our previous retrospective single institution study, we have observed a response rate of 29.6% in recurrent or metastatic ESCC patients treated with irinotecan plus a fluorouracil derivative as a second-line or third-line chemotherapy [14]. However, since the treatment options for recurrent or metastatic ESCC is still limited and the reported results have not shown significant improvement in patients’ survival, in this open-labeled phase 3 ESWN 01 trial, we intended to explore the benefit of adding irinotecan to S-1 as a second- or third-line chemotherapy for recurrent or metastatic ESCC.

To the best of our knowledge, this is the first prospective, randomized, controlled study comparing doublet with single-agent chemotherapy in patients with advanced ESCC progressing after systemic cytotoxic therapies. Patients treated with irinotecan plus S-1 demonstrated a significantly longer PFS and higher response rate than did those treated with S-1 monotherapy. The treatments were well tolerated in both groups. The toxicity profile observed in this study was generally consistent with that observed in other tumor types [15, 16].

A total of 15 centers across China registered to participate in this study, but only 148 patients could be screened. Patient accrual was difficult partly due to the open-label design of the trial. The most frequent reason for refusal to enroll, as reported from the participating centers, was the unwillingness of the patients to receive single-agent treatment. There were also 14 cases (3 in the irinotecan plus S-1 group and 11 in the S-1 monotherapy group) of early withdrawal of consent after randomization.

Despite the promising breakthrough in immunotherapy, cytotoxic chemotherapy still remains the mainstream in the management of recurrent or metastatic ESCC patients. However, no randomized phase 2 or 3 trials have been reported for patients with recurrent or metastatic ESCC who had resistance to platinum- or taxane-based chemotherapy. Currently, S-1 or fluorouracil, platinum, taxanes, and irinotecan are the accepted drugs for these patients [17]. The combination of irinotecan with fluorouracil or capecitabine has demonstrated activity in patients with metastatic esophageal cancer progressing after platinum-based first-line chemotherapy [15, 16]. However, these phase 2 trials included both ESCC and esophageal adenocarcinoma patients, and the number of patients enrolled was limited. In this ESWN 01 trial, which enrolled only ESCC patients, the results showed that treatment with irinotecan plus S-1 significantly prolonged PFS as compared with S-1 alone, and the observed treatment outcomes can be regarded clinically meaningful. The ORR and DCR were also higher in the irinotecan plus S-1 group than in the S-1 monotherapy group (ORR: 24.6% vs. 9.7% [P = 0.002]; DCR: 57.3% vs. 35.4% [P < 0.001]). Both the ORR and median PFS of the irinotecan plus S-1 regimen in the present study were numerically slightly inferior to the findings observed in our previous retrospective study [14], nevertheless, the efficacy of the doublet regimen from these two studies were comparable even though the study design and patient population were different. The addition of irinotecan did not significantly prolong the median OS in the present study (irinotecan plus S-1 group [7.1 months] vs. S-1 monotherapy group [6.2 months], P = 0.141), despite an observed superior 6-month OS rate in the irinotecan plus S-1 group (61.9% vs. 50.2%, respectively). This could have been partially attributed due to a substantial number of patients in the S-1 monotherapy group who received irinotecan or irinotecan-based chemotherapy in subsequent lines of treatment after failure with S-1 monotherapy. Another reason could have been the decreased willingness of the patients in the irinotecan plus S-1 group to undergo further chemotherapy when their diseases progressed beyond the study treatment for the fear of severe adverse events. In this prospective randomized setting, our results substantiate the role of the irinotecan plus S-1 regimen as a possible important alternative regimen for patients with recurrent or metastatic ESCC. Our future direction could be exploring the combination of irinotecan plus S-1 chemotherapy with immune checkpoint inhibitors or other targeted treatments.

The regimens used in both groups of the present study appeared to be well tolerated. Irinotecan and S-1 have distinct toxicity profiles. In this study, more patients developed grade 3–4 neutropenia, leucopenia, and nausea in the irinotecan plus S-1 group. Despite a higher rate of adverse events in the irinotecan plus S-1 group, only a few patients discontinued the treatment due to the adverse events. Several phase 2 trials evaluated the efficacy of irinotecan-based chemotherapy in esophageal or esophago-gastric cancer patients in which the dosage schedules of irinotecan were 180 mg/m² repeated every 2 weeks and 250 mg/m² repeated every 3 weeks respectively, and the reported rates of grade 3–4 neutropenia ranged between 26.4% and 31% [15, 16], whereas the rate of common adverse events in the irinotecan plus S-1 group of the present study was much lower, at 14.8%. Similarly, the rate of grade 3–4 diarrhea was lower in the irinotecan plus S-1 group (3.3%) than those reported in the two above-mentioned trials (7.9%–15.0%) [15, 16]. A possible explanation for this difference in observed adverse events could have been due to the lower dose of irinotecan (160 mg/m²) used in our study.

The present study had several limitations. First, the randomized controlled trial was prematurely stopped due to the stagnation in subject recruitment. As a result, the number of randomized patients was smaller than proposed in the protocol. Second, the open-labeled design of this trial had led to a certain number of patients who were randomized (n = 123) but did not receive the study treatment (n = 11) in the S-1 monotherapy group. This could have partly affected the evaluation of the efficacy of S-1.

The addition of irinotecan to S-1 monotherapy in treating patients with recurrent or metastatic ESCC in the second- or third-line setting may prolong PFS and increase ORR, and can therefore be considered an alternative option as to S-1 monotherapy in this patient population.