Following the researches performed on lymphatic spread of gastric carcinoma, our study aimed to add useful informations to establish the extension of exeresis with radical intent on the basis of recent studies on sentinel node mapping. The study includes 120 consecutive cases of gastric cancer operated between 1996 and 2000. Sixty-nine (57.5%) males and 51 (42.5%) females, with age ranging between 29 and 84 years, were evaluated. Cancer location was in 17 patients upper third, in 35 middle third, in 59 distal third and in 9 cases gastric stump. Early cases (T1-2) were 65 and advanced cases (T3-4) were 55. Sixty-eight STG and 52 TG were performed with 21 D1, 73 D2 and 26 D3 lymphoadenectomy. Overall 3501 lymph-nodes were examined. Eighty-two cases (68%) showed lymph-nodal metastases, while 38 (32%) were negative. The lymphatic spread of gastric cancer resulted related to the depth of wall invasion (T): 25% of T1, 75% of T2, 85% of T3 and 95% of T4 were node positive (N+). The correlation between T and N+ was also significative concerning the diffusion to the first, second and third level stations, increasing with the depth of wall invasion. The possibility of extension to one single lymphnodal station was inversely related to the T. When the first level stations were negative the second and third were never positive. This study seems to confirm the complex but systematic and predictable lymphatic gastric flow. The skip metastases, rare but described in other retrospective studies, were never observed. Our study shows that sentinel node mapping could be reliable also for gastric cancer surgery, allowing tailored and less aggressive resections. Meanwhile, from this analysis it appears that nodal stations 7-8 should be included in the first level lymphoadenectomy, widening D1 exeresis considered until now sufficient in less advanced cases. In conclusion, studies of lymphatic spread and sentinel node mapping seems to require a different therapeutic approach to gastric cancer concerning D1 lymphoadenectomy, tailoring the exeresis on the basis of sentinel node mapping in negative T1-2 cases, but systematically widening the lymphoadenectomy from the perigastric only, to 7 and 8 stations.