From 1990 to 1997 we observed, in our department, 267 patients affected by gastric cancer. In the first four years of our experience (1990-93) we resected 87 patients out of 136 gastric cancers observed (63.9%): 56 pts. (64.4%) were classified as ASA I-II, 21 (24.1%) as ASA III, 10 (11.5%) as ASA IV. In 2 cases (2.3%), operated in emergency, a DO-1 lymphectomy was performed, with a mean of 4 nodes resected; 67 pts. (77.0%) had a D2, with a mean of 36.5 nodes resected; in 18 pts. (20.7%) we performed a D3, with a mean of 64.3 nodes resected. Post-operative technical complications were 13 (14.9%). We observed 5 post-operatory deaths (5.7%), 3 due to technical complications. Absolutely and relatively curative resections have been 62 (71.3%). In the second period (1994-97) we resected 89/131 patients (67.9%): in this group 50 pts. (56.2%) were classified as ASA I-II, 24 (26.9%) as ASA III, 15 (16.9%) as ASA IV. With the exception of 3 patients (3.4%) who were operated in emergency (D0-1 procedures, with a mean number of 2.3 nodes resected) we adopted D2 lymphadenectomy plus hepatic peduncle as the procedure of choice, performed in 86 pts. (96.6%), with a mean of 34.1 nodes resected. Post-operative technical complications were 5 (5.6%). We observed 5 post-operatory deaths (5.6%), 1 due to technical problems. Absolutely and relatively curative resections have been 76 (85.4%). Morbidity and mortality due to technical complications in the second period are lower than observed in the first period, without any difference in the curability rate. D2 lymphectomy seems to be an effective procedure, safe even in high anesthesiological risk patients. Increasing experience and standardization of the technique reduced risk of surgical complications and mortality.