Abstract
Feasibility and efficacy of S-1 and cisplatin followed by surgery was evaluated, and factors contributing to survival benefit were analyzed. Patients and Methods: In total, 120 consecutive patients with highly advanced gastric cancer were treated with S-1 (80 mg/m2 for 21 consecutive days) and cisplatin (50 mg/m2 on day 8). Results: The response rate was 62.5% overall, and 75.7% for these with metastatic lymph nodes. Grade 3/4 adverse events were less than 10%. The median survival time was 41.9 months among 93 patients whose primary lesion was resected. Liver metastasis, R2 resection, poor performance status and lack of response were identified as independent risk factors by a multivariate analysis. Conclusion: Preoperative chemotherapy with S-1 and cisplatin was effective. The results show the need for different approaches in the treatment of patients with metastases and these without.
The only curative treatment for gastric adenocarcinoma is R0 resection, arguably accompanied by D2 lymph node dissection according to the Guidelines of the Japanese Gastric Cancer Association (JGCA) (1). Local control is considered an essential component of the treatment for gastric carcinoma, and extended lymphadenectomy can accomplish this task safely in experienced hands (2). The prognosis for stage III/IV advanced gastric cancer (AGC) remains unsatisfactory, however, and further improvement in surgical technique is unlikely to lead to notable progress in the outcome (3-4). Hence the development of an effective multimodal strategy has been sought after by various study groups.
A recent meta-analysis performed in the West showed that combination regimens achieve better survival outcomes than those with 5-fluorouracil (5-FU) monotherapy and that regimens containing 5-FU, anthracyclines and cisplatin (CDDP) are the most effective (5). A perioperative chemotherapy using this triplet was actually found to improve survival of potentially curable AGC significantly when compared with treatment by surgery alone (6). Downstaging and eradication of micrometastases through the preoperative chemotherapy component may have had a particularly important role in such a strategy. In Japan, postoperative adjuvant chemotherapy by S-1, a dehydropyrimidine dehydrogenase-inhibiting 5-FU derivative, was found to improve survival of patients with curatively resected stage II/III gastric cancer. Addition of adequate preoperative therapy to this strategy may enhance the survival of patients with resectable AGC, while downstaging through preoperative chemotherapy may provide patients with more advanced cancer some chance for cure. One candidate for use is intensive preoperative chemotherapy under such circumstances would be a combination of S-1 and CDDP, which led to a response rate of 54% and median survival time of 13 months among patients with unresectable gastric cancer in a phase III trial (7). This combination was also shown to be feasible as a preoperative induction therapy in a case series involving a smaller number of patients (8).
In the current study, 120 consecutive AGC patients who were treated with S-1/CDDP therapy prior to surgery were retrospectively analyzed to assess the efficacy and safety of this combination as a preoperative therapy and to identify the subset of patients who may benefit from this strategy.
Patients and Methods
Patients. The files for one hundred and twenty consecutive patients with primary AGC who were treated preoperatively by a combination of S-1 and CDDP between October 2000 and December 2005 were retrieved from the prospective database of Niigata Cancer Center according to the following criteria: histologically confirmed adenocarcinoma of the stomach; clinically diagnosed as locally advanced T3/T4-stage disease or metastatic disease; evaluable lesions on computed tomography (CT) scan, at upper gastrointestinal series and/or upper digestive endoscopies; age less than 75 years; ECOG performance status between 0 and 2; no prior chemotherapy or radiotherapy; sufficient organ functions represented by leukocyte count of more than 3,000/mm3, platelets more than 10×104/mm3, GOT/GPT less than 2 times the upper limit of normal range (ULN), total bilirubin less than 2.0 mg/dl, BUN and creatinine less than the ULN; no serious co-morbidities; no concurrent active malignancy; no serious psychosomatic disorder; and provision of written informed consent. Staging laparoscopy was performed only for patients with linitis plastica or those with macroscopically type 3 cancer with preoperatively estimated diameter of >8 cm. Cytological examination of the peritoneal washes was performed at the time of staging laparoscopy, but the result was used only as a reference, and detection of cancer cells in this examination did not preclude patients from receiving preoperative chemotherapy followed by surgery.
Treatment schedule. All patients received systemic chemotherapy consisting of S-1 and CDDP. S-1 was orally administered at a dose of 80 mg/m2 for 21 consecutive days, followed by 14 days of rest. With the intent to deliver the treatment on an outpatient basis, the dose of CDDP was modified from the original version by Koizumi et al. in which 60 mg/m2 had been administered. CDDP was administered intravenously on day 8 at a dose of 50 mg/m2. The treatment was repeated every 5 weeks. Patient status was evaluated after each course of the treatment. Toxicity was assessed using the National Cancer Institute-Common Toxicity Criteria version 3.0. The response of measurable lesions was evaluated according to the RECIST criteria. The primary lesion, when not considered as measurable by the RECIST criteria, was assessed according to the Japan Gastric Cancer Association (JGCA) clinical criteria for response assessment of chemotherapy and radiotherapy. The assessment was based on shrinkage and morphological change of the primary tumor as evaluated by barium contrast study and/or endoscopic examinations (9).
Patients with locally advanced cancer were treated by chemotherapy until primary cancer or massive nodal metastases responded and resection with curative intent was deemed possible. Patients with metastatic cancer (those with hepatic or peritoneal metastases) were treated until metastatic lesions achieved complete response by CT or became co-resectable. Patients who remained with clear evidence of unresectable disease and those who did not respond to the chemotherapy were discouraged from receiving surgery. Surgery with intent to cure was performed at least 3 weeks after the final cycle. Most patients were treated by S-1 monotherapy as an adjuvant therapy after surgery. Treatments after R2 resection or at detection of recurrent disease were decided at the discretion of each physician.
Study design and statistical analysis. Median survival time (MST) was calculated from the initiation of chemotherapy to death or the day when the patient was last interviewed. Survival curves were calculated by the Kaplan-Meier method and compared by the log-rank test. Univariate and multivariate analyses using Cox's proportional hazards model was performed to identify independent prognostic factors. All statistical calculations were performed using statistical analysis system (SAS) version 8.2 (IBM, North Caroline, USA) and a value of p<0.05 was considered as statistically significant.
Results
Patient demographics. Characteristics of the 120 patients are shown in Table I. There were 75 men and 45 women with a median age of 61 years (range 29 to 83 years). There were 44 patients with linitis plastic type cancer (36.7%). Non-curative factors included liver metastasis in 8 patients, peritoneal dissemination in 30, involvement of abdominal para-aortic lymph nodes in 34 and locally advanced and potentially unresectable gastric cancer in 12. The pretreatment clinical stage (c-stage) was diagnosed according to the classification of JGCA, which was based on the findings of CT, upper GI series, endoscopy, and staging laparoscopy. Preoperative stages were decided according to the JGCA staging system (c-stage II: 1 case; c-stage III: 33 cases; c-stage IV:86 cases). Distribution of the c-stage IV factors was as follows: metastasis to the paraaortic nodes in 34 cases, cT4N2 in 12 cases, hepatic metastasis in 8 cases, peritoneal dissemination in 30 cases, and other distant metastasis in 4 cases.
Proportion of the treatment performed at outpatient clinic. The median number of administered courses was 3 (range: 1-7), and the proportion of care given in the outpatient setting was 86%. Forty-seven patients who underwent staging laparoscopy were admitted for the procedure and given the first course of chemotherapy during the same stay in the hospital. Of the 73 remaining patients, 22 managed to receive chemotherapy entirely on an outpatient clinic basis. However, the rest of the patients needed admission for hydration and antiemetic therapy during administration of CDDP.
Surgery. After chemotherapy, 27 patients failed to be treated by surgery, mostly because of persistence of metastatic disease through imaging studies. The remaining 93 patients underwent surgery and gastrectomy was performed in all patients. The overall resection rate was 77.5%. The surgical procedure was total gastrectomy in 57 patients and distal gastrectomy in 36 patients. R0 resection was possible in 68 patients (73.1% of all patients who underwent surgery), of whom 25 received combined resection of the involved adjacent organs and 14 underwent extended lymph node dissection including of the para-aortic lymph nodes. Of those who underwent surgery, there were 59 males and 34 females, with a median age of 61 years (range: 29 to 77 years) (Table II). The median hospital stay was 18 days. The median duration of surgery was 195 minutes and the median blood loss was 225 ml. The distribution of postoperative c-stage was as follows; 26 patients in c-stage I/II, 26 in c-stage III, and 41 in c-stage IV. R0 resection was successfully performed in 68 (73.1%) patients. Downstaging was obtained in 32 (34.4%) patients.
Clinical response to chemotherapy. The objective response of the evaluable lesions is shown in Table III. The overall response rate (ORR) was 62.5% (95% confidence interval (CI): 53.8-71.2%). There were 75 responders (one complete response (CR) and 74 partial responses (PR)), when the response to the primary lesion was disregarded. Response rate for regional/para-aortic lymph nodes, primary gastric tumor (assessed based on JGCA clinical criteria for response assessment of chemotherapy and radiotherapy), liver metastases and peritoneal metastases was 75.7% (56/74), 61.7% (74/120), 28.6% (2/7) and 23.8% (5/21), respectively. Twenty-five other patients (42.4%) had stable disease (SD), and only 2 patients had progressive disease (PD). Pathologic CR of the metastatic lymph nodes, including para-aortic lymph nodes, was confirmed after surgery in 4 patients. Of the 75 responders, residual tumor was completely resected in 51 (68.0%). Out of the 47 patients who underwent staging laparoscopy, 31 were found to have peritoneal metastasis; of these, complete remission of the peritoneal disease was confirmed at surgery in 9 (29.0%).
Patient characteristics (N=120).
Patient characteristics of resected cases.
Response.
Toxicity.
Cumulative probability of overall survival as estimated by the Kaplan-Meier method in 120 patients. The median survival time was 23.0 months, and a 4-year survival rate was 30.6%.
Toxicity. The adverse reactions during 308 cycles of S-1/CDDP regimen were evaluated according to NCI-CTC grade (Table IV). The most frequent toxicities of S-1/CDDP were myelosuppression and gastrointestinal symptoms. The incidence of notable adverse events were 55.8% for neutropenia, 43.3% for leukocytopenia, 65.0% for anemia 65.0%, 40.8% for thrombocytopenia, 67.5% for anorexia, 56.7% for nausea, respectively. However the incidence of grade 3/4 toxicity was infrequent: neutropenia 7.5%, leucopenia 0.8%, anemia 6.7%, thrombocytopenia 6.7%, anorexia 5.8% and nausea 2.5%. The preoperative chemotherapy was generally well tolerated. There was no surgical mortality, and postoperative surgical morbidity was remarkably low at 17.2%.
Survival and analysis of prognostic factors. The median survival time of patients overall was 23.0 months, with a 4-year survival rate of 30.6% (Figure 1). The median survival time of patients who went on to receive surgery was 41.9 months (95% CI: 31.9-51.9 months) and the 3-year survival rate was 51.2% (95% CI: 37.4-64.9%) (Figure 2). There was a statistically significant difference in survival between these patients and these who failed to receive gastrectomy.
For all patients, response to chemotherapy, location of the tumor, resectability of the primary lesion, liver metastasis, and peritoneal metastasis were predictive of the overall survival (Table V). In the multivariate analysis, response to chemotherapy, peritoneal metastasis and hepatic metastasis were the only independently prognostic factors (Table VI). For the patients who were treated by gastrectomy, curability of surgery, response to the chemotherapy, hepatic metastasis, peritoneal metastasis, the extent of lymph node dissection, N category, and performance status were identified as significant prognostic determinants (Table VII). Of these, hepatic metastasis, curability of surgery, performance status and response to the chemotherapy were identified as independent prognostic factors (Table VIII).
Discussion
Gastric carcinoma remains a major health problem worldwide, primarily because it is often diagnosed at an advanced stage. In addition, it often relapses even after a potentially curative resection, and multimodal treatments have been sought after by various study groups to combat residual micrometastases. One of the consequences is that postoperative chemoradiation was found to significantly improve outcome of curatively resected patients and has become a standard of care in North America (10-11). There is a suspicion, however, that radiation as a local therapy may have compensated for poor local control due to suboptimal surgery, and the Japanese surgeons remained confident that extended nodal dissection precludes the need for adjuvant treatment focused around the gastric bed. However, the Japanese experts felt promise when S-1, a novel oral fluoropyrimidine derivative, became available. This drug achieved a response rate of more than 40% when used as a single agent (12-13), while the response rate rose to 50-75% when used in combination with CDDP (12), irinotecan (14-15), docetaxel (17-18), and paclitaxel (19-20). Their expectations were met when an interim analysis of a pivotal phase III study revealed that postoperative adjuvant chemotherapy with single agent S-1 significantly improved survival of stage II-III gastric cancer patients when compared with surgery alone (21).
Cumulative probability of survival in the gastrectomy group (93 patients) and the non-gastrectomy group (27 patients) as estimated by the Kaplan-Meier method. A 3-year survival rate of the former and the latter was 51.2% and 10.8%, respectively. There was a statistically significant difference.
Gastrectomy causes various gastrointestinal symptoms and nutritional deficits, and additive toxicity through postoperative chemotherapy could be a substantial burden for the patients. More than 10% of the Japanese patients in the aforementioned phase III trial had actually failed to continue with oral S-1 at three month postoperatively. Post-gastrectomy deterioration of compliance regarding chemotherapy was also observed in the British MAGIC trial, in which 88% of patients received preoperative chemotherapy whereas only 55% tolerated the same therapy postoperatively (6). Thus, there is a rationale for delivery of a somewhat toxic but effective chemotherapy preoperatively, and neoadjuvant chemotherapy is a promising option for resectable AGC. In addition, the indication for preoperative chemotherapy could be extended to include AGC with synchronous metastases, provided the metastatic lesions are co-resectable or become resectable after chemotherapy. Although chemotherapy is the standard of care for metastatic gastric cancer, it does not cure the disease. One can argue therefore that surgery remains an option as a part of multimodal therapy for patients with resectable metastases. When such is the case, preoperative chemotherapy provides useful information as regards drug sensitivity and biology of cancer, besides the potential to downstage the disease, and prevents futile surgery for cancer that is destined for rapid progression. In either of the settings, efficacy along with safety of preoperative chemotherapy and its influence on surgery that follows need to be addressed.
Univariate analysis of 120 patients with primary AGC who were treated preoperatively by a combination of S-1 and CDDP (log-rank test).
Multivariate analysis of 120 patients with primary AGC who were treated preoperatively by a combination of S-1 and CDDP. Cox's proportional hazard model (SAS ver. 8.2, score method).
Chemotherapeutic regimens with high response rates are required to achieve downstaging along with eradication of micrometastases whilst preventing disease progression. In due course, a combination of S-1 and CDDP has become acknowledged in Japan as a candidate for neoadjuvant chemotherapy owing to its remarkable response rate, in excess of 70%, and this is the regimen with which the authors chose to treat AGC patients preoperatively. A combination of S-1 and CDDP was first established by Koizumi et al., by which 60 mg/m2 of CDDP was to be administered on day 8 of a 5-week course. Administration at this dosage was feared to cause nausea and potential damage to renal function, and patients usually had to be admitted for a few days for continuous intravenous infusion along with extensive use of antiemetics. To lower the risk of organ dysfunction prior to surgery and in an attempt to deliver all the drugs on an outpatient basis, we modified the dose of CDDP to 50 mg/m2. Consequently, CDDP was delivered entirely on the outpatient basis in 22 out of 73 patients, but admission was still necessary for all remaining patients. Response rate for the nodal metastases was satisfactory at over 70%, but those for other metastatic lesions were substantially lower. Given that the number of beds available for preoperative chemotherapy is limited, establishment of a modified regimen with further dose reduction, perhaps through an increase in the number of intravenous deliveries per cycle to preserve the dose intensity, may be warranted.
Response to the chemotherapy is undoubtedly a valuable parameter in deciding whether or not to proceed to surgery for metastatic cancer. When chemotherapy is performed in the neoadjuvant setting, however, cancer is usually resectable before the treatment. The response will then have to be evaluated even more cautiously and diligently to avoid delay in surgery when the cancer is not responding to the chemotherapy. When the tumor is not accompanied by distant metastasis or bulky lymphadenopathy, the primary lesion would be the only target for evaluating response. In addition, shrinkage of the primary may allow surgeons to avoid total gastrectomy in cases of advanced cancer of the distal or mid-portion of the stomach. Thus, the Authors insisted on assessing response in the primary lesion according to the Japanese criteria, although these lesions are not considered as measurable by the RECIST criteria. Marked response to the primary was observed in 61.7% of the patients.
Univariate analysis of 93 patients underwent gastrectomy after chemotherapy (log-rank test).
Multivariate analysis of 93 patients underwent gastrectomy after chemotherapy. Cox's proportional hazard model. (SAS ver. 8.2, score method).
In the current population of advanced/metastatic cancer, the R0 resection rate among those who eventually underwent surgery was unexpectedly high at 73.1%. The MST was 23 months overall and 42 months among those who underwent surgery. The combination of S-1 and CDDP thus provided promising survival data with a favorable toxicity profile with no treatment-related deaths. Multivariate analysis of all patients identified peritoneal metastasis and hepatic metastasis as independent prognostic factors in all patients. Of patients with metastatic cancer, only those with hepatic metastasis that responded to chemotherapy went on to receive surgery. Nevertheless, hepatic metastasis remained an independent prognostic factor among those who underwent surgery. These results confirm that the outcome of patients with metastatic cancer is quite different from those with locally advanced cancer (those who undergo so-called neoadjuvant chemotherapy). In future, these two groups of patients should thus be treated by different strategies and analyzed independently.
Conclusion
S-1/CDDP at a reduced dose was safe and feasible when given preoperatively, without notable influence over the surgical morbidity. It remained effective against the primary tumor and nodal metastases. The survival benefit of cytoreductive surgery in metastatic cancer that responds to such chemotherapy needs to be addressed by a randomized trial, while another trial is needed to confirm its benefit in the neoadjuvant setting for locally advanced cancer.
Acknowledgements
We thank Professor Y. Kodera of the Department of Surgery II, Nagoya University Graduate School of Medicine for his critical review of this manuscript. This work was supported in part by as Grant-in-Aid from the Ministry of Health, Labour, and Welfare (Tokyo).
- Received July 27, 2009.
- Revision received October 6, 2009.
- Accepted October 12, 2009.
- Copyright© 2009 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- Feasibility Study of Triplet Combination Chemotherapy of Paclitaxel, Cisplatin and S-1 for Advanced Gastric Cancer
- Phase I Study of Neoadjuvant Chemoradiotherapy Consisting of S-1 and Cisplatin for Patients with Resectable Advanced Gastric Cancer (KOGC-01)
- Improvement of Biodistribution with PEGylated Liposomes Containing Docetaxel with Degradable Starch Microspheres for Hepatic Arterial Infusion in the Treatment of Liver Metastases: A Study in CC-531 Liver Tumor-bearing WAG RIJ Rats