BACKGROUND/AIMS: We examined whether a new system for categorization of the number or level of lymph node metastases would improve survival prediction in patients with gastric cancer. METHODOLOGY: Node-positive gastric cancer patients (361) were retrospectively studied. Cox proportional-hazards model analysis of the numbers of metastatic lymph nodes (numbers 1 to 11) showed the first peak of the hazard ratio for death at a threshold value of 4 lymph node metastases. We calculated cumulative 5-year survival rates and examined prognostic factors by multivariate analysis according to the numbers and levels of lymph node metastases. The effects of other prognostic factors were also examined. RESULTS: We divided the number of lymph node metastases into two groups (1-3 vs. > or =4), and classified the level of lymph node metastasis as level I or level II as defined by the Japanese Classification of Gastric Carcinoma. The 5-year survival rate differed significantly according to the number of lymph node metastases (1-3 vs. > or = 4) (p<0.0001). On multivariate regression analysis, the level of lymph node metastasis (level I or level II) and the number of lymph node metastases (1-3 or > or = 4) were independent predictors of survival. CONCLUSIONS: Categorization of the number of metastatic lymph nodes (1-3 or > or = 4) was easy and provided an accurate predictor of survival in patients with gastric cancer.
BACKGROUND/AIMS: We examined whether a new system for categorization of the number or level of lymph node metastases would improve survival prediction in patients with gastric cancer. METHODOLOGY: Node-positive gastric cancerpatients (361) were retrospectively studied. Cox proportional-hazards model analysis of the numbers of metastatic lymph nodes (numbers 1 to 11) showed the first peak of the hazard ratio for death at a threshold value of 4 lymph node metastases. We calculated cumulative 5-year survival rates and examined prognostic factors by multivariate analysis according to the numbers and levels of lymph node metastases. The effects of other prognostic factors were also examined. RESULTS: We divided the number of lymph node metastases into two groups (1-3 vs. > or =4), and classified the level of lymph node metastasis as level I or level II as defined by the Japanese Classification of Gastric Carcinoma. The 5-year survival rate differed significantly according to the number of lymph node metastases (1-3 vs. > or = 4) (p<0.0001). On multivariate regression analysis, the level of lymph node metastasis (level I or level II) and the number of lymph node metastases (1-3 or > or = 4) were independent predictors of survival. CONCLUSIONS: Categorization of the number of metastatic lymph nodes (1-3 or > or = 4) was easy and provided an accurate predictor of survival in patients with gastric cancer.