BACKGROUND: Although the results of gastric cancer treatment have markedly improved, this disease remains the most common cause of cancer death in Korea.METHODS: Clinicopathologic characteristics were analyzed for 10 783 consecutive patients who underwent operation for gastric cancer at the Department of Surgery, Seoul National University Hospital, from 1970 to 1996. We also evaluated survival and prognostic factors for 9262 consecutive patients operated from 1981 to 1996. The clinicopathologic variables for evaluating prognostic values were classified as patient-, tumor-, and treatment-related factors. The prognostic significance of treatment modality [surgery alone, surgery + chemotherapy, surgery + immunotherapy + chemotherapy (immunochemosurgery)] was evaluated in patients with stage III gastric cancer (according to the International Union Against Cancer TNM classification of 1987). For the assessment of lymph node metastasis, both the number of involved lymph nodes and the ratio of involved to resected lymph nodes were analyzed, as a quantitative system.RESULTS: The mean age of the 10 783 patients was 53.5 years and the male-to-female ratio was 2.07 : 1. Resection was performed in 9058 patients (84.0% resection rate). The 5-year survival rates were 55.9% for all patients and 64.8% for patients who received curative resection. Age, sex, preoperative hemoglobin and albumin levels, type of operation, curability of operation, tumor location, Borrmann type, tumor size, histologic differentiation, Lauren's classification, perineural invasion, lymphatic invasion, vascular invasion, depth of invasion, number of involved lymph nodes, ratio of involved to resected lymph nodes, and distant metastasis had prognostic significance on univariate analysis. Radical lymph node dissection, with more than 25 resected lymph nodes improved survival in patients with stage II and IIIa disease. As postoperative adjuvant therapy, immunochemotherapy was most effective in patients with stage III disease. Patients with identical numbers of lymph nodes -either the number of involved lymph nodes or the number of resected lymph nodes- were divided according to their ratios of involved-to-resected lymph nodes. In each numeric group, there were significant survival differences according to the ratio of involved-to-resected lymph nodes. However, patients who had the same involved-to-resected lymph node ratio did not show significant differences in survival rate according to either the number of involved or the number of resected lymph nodes. On multivariate analysis, curability of operation, depth of invasion, and ratio of involved to resected lymph nodes were independent significant prognostic factors.CONCLUSIONS: Curative resection, depth of invasion, and lymph node metastasis were the most significant prognostic factors in gastric cancer. With regard to the status of lymph node metastasis, the ratio of involved to resected lymph nodes had a more precise and comprehensive prognostic value than only the number of involved or resected lymph nodes. Early detection and curative resection with radical lymph node dissection, followed by immunochemotherapy, particularly in patients with stage III gastric cancer should be the standard treatment in principle, for patients with gastric cancer.