Treatment of stage IV gastric cancer with induction chemotherapy using S-1 and cisplatin followed by curative resection in selected patients

Patients with unresectable stage IV gastric cancer who were treated at the Saitama Medical Center, Jichi Medical University, Japan from January 2008 to March 2010 were retrospectively reviewed. Inclusion criteria for the study were: (1) gastric cancer diagnosed by upper gastrointestinal endoscopy and histologically confirmed as adenocarcinoma; (2) unresectable features detected on thoracic, abdominal, or pelvic multidetector computed tomography (CT) scan; (3) ability to maintain an oral liquid or solid intake without vomiting; (4) Eastern Cooperative Oncology Group Performance Status (ECOG PS) 0 or 1; and (5) white blood cell count ≥ 3,000/ml, platelet count ≥100,000/ml, serum creatinine ≤ 1.2 mg/dl, aspartate aminotransferase (AST) < 70 IU/l, and alanine aminotransferase (ALT) < 70 IU/l. Unresectable features were defined as: invasion of the primary lesion into adjacent organs (T4), ascites or peritoneal nodules (P1), para-aortic lymph node metastasis (N3), bulky lymph node metastasis or invasion into the area around the celiac trunk (bulky N2), or hepatic metastasis (H1). Exploratory laparoscopy was recommended for diagnosis of peritoneal dissemination, but was not mandatory. Peritoneal dissemination was diagnosed by imaging by a specialist in radiological interpretation. Exclusion criteria were: (1) prior history of gastric cancer treatment; and (2) known distant metastasis to sites other than the lymph nodes, liver or peritoneum. Patients were not routinely investigated for distant metastasis, but those with symptoms suggestive of metastasis underwent appropriate investigations. The gross and histological tumor types, depth of tumor invasion, and lymph node metastasis (TNM) stage were described according to the Japanese Classification of Gastric Carcinoma (JCGC; 13th edition) of the Japanese Gastric Cancer Association[16]. This study was approved by the Research Ethics Committee at Jichi Medical University.

All patients received induction chemotherapy using S-1 and cisplatin according to the protocol described in the SPIRITS trial[8]. In each cycle, oral S-1 (80 mg/m²) was administered for 3 weeks followed by a 2-week drug holiday. Intravenous cisplatin (60 mg/m²) was administered on day 8 after adequate premedication and hydration.

Complete blood cell count, and serum creatinine, total bilirubin, AST and ALT levels were measured before each cycle and regularly during each cycle. Chemotherapy toxicity was evaluated according to the National Cancer Institute Common Toxicity Criteria for Adverse Events (NCI CTCAE) version 3.0. If grade 4 leukopenia or grade 3 or 4 thrombocytopenia were observed, the S-1 and cisplatin doses were reduced by 1 dose level. If the serum creatinine level exceeded 1.5 mg/dl, cisplatin was discontinued and S-1 monotherapy was continued.

Response was assessed after every two cycles of chemotherapy. Measurable tumors were evaluated using the Response Evaluation Criteria in Solid Tumors (RECIST)[17]. The best overall response was evaluated and the response was not confirmed for 4 weeks. The primary gastric lesion was assessed using endoscopy, gastroduodenal barium contrast study, or CT scan according to the JCGC criteria.

If the unresectable features were resolved at the time of reassessment, surgery was performed 4 to 6 weeks after completion of S-1 administration. Informed consent was obtained from all patients. If surgical exploration in these patients did not reveal unresectable features, R0 resection was performed. Patients underwent total or distal gastrectomy according to the location of the gastric tumor with D2 lymph node dissection, plus extended para-aortic lymph node dissection (PAND) if they were para-aortic lymph node-positive (cN3) at presentation. If surgical exploration revealed unresectable features, exploratory laparotomy or bypass surgery was performed. Postoperative follow-up included blood testing every 3 months and abdominal ultrasonography every 6 months. Patients also underwent yearly abdominal CT scans and upper gastrointestinal endoscopy.

Response to induction chemotherapy, surgical procedure, extent of lymph node dissection, intraoperative blood loss, postoperative complications, pathological examination, and final disease stage were recorded. The therapeutic response was graded according to histological features. Survival was defined as the time from the start of induction chemotherapy until death or the last follow-up. Survival data were updated to December 2013.

Patients who underwent R0 resection received S-1 alone as postoperative adjuvant chemotherapy for 1 year. The remaining patients continued to receive S-1 and cisplatin therapy until disease progression was evident. When postoperative recurrence was found or the tumor was refractory to S-1 and cisplatin, irinotecan monotherapy (100 mg/m² intravenously on days 1, 8 and 15) was administered as second-line treatment and paclitaxel monotherapy (80 mg/m² intravenously on days 1, 8 and 15) was administered as third-line treatment, and continued until an adverse effect occurred or further disease progression was confirmed. Twenty-nine patients received second-line treatment and 12 received third-line treatment.

Overall survival rates for the whole group and for the R0 resection group were analyzed using the Kaplan-Meier method. Prognostic factors for survival were analyzed with univariate and multivariate analyses using the Cox proportional hazards model. All reported P- values were 2-sided, and P < 0.05 was considered to be statistically significant. Statistical analyses were performed using SPSS version 22.0 (SPSS, Chicago, IL, USA).

Of the 88 patients with stage IV gastric cancer who were treated at our institute during the study period, 59 (45 male and 14 female) met the inclusion criteria. The median age of patients was 65 years (40 to 74 years). ECOG PS was 0 in 46 patients and 1 in 13 patients. Thirty patients had differentiated adenocarcinoma and 29 had undifferentiated adenocarcinoma. Preoperative laparoscopy was performed in 10 patients, but the results did not influence subsequent management. The frequencies of features that deemed the tumor to be unresectable are shown in Table 1.

According to the RECIST criteria, 35 of the 59 patients had metastatic lesions detected on abdominal CT scan before induction chemotherapy (lymph node metastasis in 28 patients and liver metastasis in 7). Of these 35 patients, 13 (37.1%) responded to chemotherapy (9 patients had a partial response and 4 had a complete response). Overall, patients received 1 to 8 cycles (median 4 cycles) of induction chemotherapy with S-1 and cisplatin. Patients who underwent surgery received 2 to 10 cycles (median 2 cycles) of induction chemotherapy (Table 2).

According to the JCGC criteria, 30 of the 59 patients (50.8%) responded to chemotherapy. The primary lesion showed a partial response in 28 patients and a complete response in 2. Table 3 shows the adverse effects of induction chemotherapy treatment. Nineteen patients (32.2%) had grade 3 adverse effects, and no patients had grade 4 effects. There were no treatment-related deaths.

Sixteen patients were deemed eligible for surgery because of resolution of the factors that had made their cancer unresectable. Of these sixteen patients, four presented with cN3, four with bulky N2, six with cP1, and two with cT4 cancer. Of the six patients with cP1, three were found to have residual peritoneal seeding at the time of surgery, which made them ineligible for R0 resection. Two of these three underwent bypass surgery only and one underwent exploratory laparotomy only. The remaining 13 patients underwent R0 resection, resulting in an R0 resection rate of 22% (13/59). The details of surgery are shown in Table 4. We found no correlation between histological response and clinical outcome.

At the time of analysis in December 2013, 34 of the 59 patients had died, and the median follow-up period of the remaining 25 patients was 26 months (range 1 to 63 months). Kaplan-Meier analysis showed an overall median survival time of 13 months and a 3-year survival rate of 18.2% (Figure 1). Multivariate analyses identified negative para-aortic lymph nodes (hazard ratio (HR) 0.086, 95% confidence interval (CI) 0.020 to 0.383, P = 0.002) and undergoing R0 resection (HR 0.285, 95% CI 0.131 to 0.623, P < 0.001) as independent prognostic factors for survival (Table 5). Age, sex, T stage, bulky N2, peritoneal metastasis, distant metastasis, response according to the JCGC criteria, and second-line chemotherapy were not significantly associated with survival.

Among the patients who underwent R0 resection, six developed recurrence (three with para-aortic lymph node, and three with peritoneal dissemination). All patients with recurrence received chemotherapy but they all died: one each with lung metastasis at 15 months, brain metastasis at 12 months, liver metastasis at 31 months, lymph node recurrence at 36 months, peritoneal dissemination at 24 months, and peritoneal dissemination at 28 months) (Table 2). Kaplan-Meier analysis of patients who underwent R0 resection showed a mean survival time of 53 months and a 3-year survival rate of 53.8% (Figure 2).