Gastric adenocarcinomas, even in the absence of distant metastases, have a poor prognosis which is particularly dismal when tumors are located in the cardia, in the event of locoregional lymph node involvement and/or bulky tumors. Postoperative adjuvant chemotherapy has never clearly demonstrated its efficacy on survival. Besides ongoing trials using new and more active regimens, preoperative chemotherapy has been used for unresectable cancer due to loco-regional extension and when locally advanced cancer is potentially resectable but with poor prognosis such as bulkiness, when tumors are located in the cardia and when there is tumor in the coeliac area at CAT-scan with suspected metastatic lymph nodes. In case of unresectable tumor at initial surgery five publications have reported the ability of chemotherapy to reduce the tumor volume and to allow subsequent resection of the gastric tumor in 40% to 60% of the cases. In these cases there is a clear survival advantage as the median survival reported in 2 of these studies was 12 and 18 months compared to the 4 to 6 months median survival reported in historical studies in case of unresectable cancer [17, 18]. In case of locally advanced gastric tumors some Japanese case reports have demonstrated the ability of preoperative chemotherapy to concentrate in the tumor tissue and to downstage the tumors. Four North American and European studies have demonstrated that preoperative chemotherapy is feasible, and will probably increase the resection rate. J. Ajani has reported 2 studies in which tolerance was acceptable: a major response (R) observed in 24% and 31%, the resectability rates were 72% and 77% and the median survival 15 and 16 months, respectively. Our experience is based on 30 patients treated with a combination of continuous i.v. 5-FU and CDDP. Fifteen had a tumor of the cardia, 15/30 had enlarged lymph nodes and 7/30 a linitis plastica (diffuse type). After a mean number of 3 cycles, 27/30 patients were evaluable for response. One patient achieved a CR and 14 a PR (OR rate 56%, 95% CI: 38% to 74%) but only one of those with linitis plastica responded. Twenty-eight patients underwent surgery and 23 had a macroscopically complete resection (82%). Resectability rate was higher after OR (13/15) than in nonresponding patients (4/12). Toxicity was acceptable, however grade 4 leucopenia in 5 patients and one toxicity-related death were observed. There was no increase in postoperative complications. Nine patients received postoperative chemotherapy and 3 patients with positive margins received postoperative external radiotherapy.(ABSTRACT TRUNCATED AT 400 WORDS)