PURPOSE: The important prognostic factors for gastric cancer are the depth of invasion by the primary tumor and the lymph node metastasis. The 5th edition of the Union Internationale Contrala Cancrums (UICC) TNM classification, which is based on the number of metastatic lymph nodes, has proved to be a reliable and objective method for predicting the prognosis of patients suffering with gastric cancer. However, its value for the prognosis of treating patients with serosal invasive (T3) gastric cancer, it is still being debated. So, we retrospectively studied the prognostic factors for T3 gastric cancer patients and we also evaluated the staging method according to the number of metastatic lymph nodes and the metastatic lymph node ratio. METHODS: This retrospective study was based on the medical records of 369 patients who underwent curative resection for serosal invasive (pT3) gastric cancer from 1992 to 2000. The patients were divided into four groups according to the number of metastatic lymph nodes and the clinicopathologic factors were evaluated by comparative study and the patients were then, classified into 4 groups by the metastatic lymph node ratio (<0.1, 0.1~0.3, 0.3~0.5, >0.5). We evaluated the prognostic factors and performed a survival analysis by using the Kaplan-Meier method and the Cox proportional hazard model. RESULTS: Among the four groups, the significant different factors were tumor size, the Borrmann type, the type of gastrectomy, the histologic type, and lymph node dissection. According to the univariate survival analysis, the tumor size, Borrmann type, lymph node stage, and the metastatic lymph node ratio significantly affected the prognosis. Yet, when comparing each survival rate, there was not significant difference between the pT3pN0 and pT3pN1 calassification. When we classified the metastatic lymph node ratio into 4 categories, each group then showed a significantly different survival rate. By conducting a multivariate analysis, only the metastatic lymph node ratio was an independent prognostic factor for serosal invasive gastric cancer (P=0.028). CONCLUSION: For evaluating patients with serosal invasive gastric cancer, there have been some problems with using the lymph node staging, so the metastatic lymph node ratio is a more reliable prognostic factor as it provides information about the extent of lymph node dissection and the degree of lymph node metastasis.