Background/aims: It is unclear whether gastric cancer prognosis is improved by extended lymph node dissection more than by lymph node dissection limited to the contiguous N1 perigastric lymph nodes.

Methodology: Four hundred and thirty-eight patients treated by curative gastrectomy were evaluated. Outcomes of D1/D1.5 lymphadenectomy, limited lymph node dissection and of D2/D2.5 lymphadenectomy, extended lymph node dissection and histopathological prognostic factors as in the 1993 TNM staging classification supplement were analyzed.

Results: Estimated overall 5-year survival was 54.9%. Five-year survival was 58.4% in the limited lymph node dissection group and 54% in the extended lymph node dissection (P n.s.). Stage I 5-year survival was 59% after D2.5 lymph node dissection, 58% after D1.5 and 50% after D2 dissection (P n.s.). Stage II 5-year survival was 86% in D2.5 group and 56% in D1.5 group (P = 0.041). Stage IIIa survival was 61% in the D2.5 group and 22% in the D1.5 group (P = 0.001). Stage IIIb 5-year survival was 42% after D2.5 resection and 0% in D1.5 group (P = 0.001). In the pT3 group 5-year survival was 72% after D2.5 dissection and 33% after D2 dissection (P = 0.001). In the positive N1 lymph nodes group 5-year survival was better after extended lymph node dissection than after limited lymph node dissection. In pN2a patients 5-year survival was 57% after D2.5 resection and 0% after D2 resection (P < 0.001). In pN2b and pN2c patients extended lymph node dissection did not statistically improve survival.

Conclusions: Even if no statistical differences were found in overall survival, prognosis was improved by extended lymph node dissection in stage II and III, particularly in T2 and T3 subgroups and in N1 and N2a subgroups. When large numbers of positive nodes were found, improved survival was dependent upon resection of extragastric nodes distal to the uppermost echelon of positive nodes.