Multivariate Analysis of Prognostic Factors in Gastric Cancer

ABSTRACT

BACKGROUND: Worldwide stomach cancer death rates have been steadily declining in most countries since World War II, but in Korea the incidence of stomach cancer is still high and is a major concern for public health. Surgical resection is the best method of treatment and offers a chance for cure, but the overall results of surgical treatment have been dismal. METHODS: The authors performed a retrospective study of 958 patients treated for adenocarcinomas of the stomach with curative intent over the period from March 1986 through December 1994 at Kyung Hee University Hospital. The study examined the prognostic influence of 2 host factors (age and sex of the patients), 7 tumor factors (location of tumor, number of lesions, size of tumor, depth of invasion, extent of lymph-node involvement, number of metastatic lymph nodes, and TNM stage), and 4 treatment factors (type of gastrectomy, combined resection, extent of lymph-node dissection, and adjuvant chemotherapy) by univariate and multivariate analysis. RESULTS: In univariate analysis, 9 prognostic factors - the age of the patients, the size of lesions, the depth of invasion, the extent of lymph-node involvement, the number of metastatic lymph nodes, the TNM stage, the type of gastrectomy, the extent of lymph-node dissection, and combined resection - were significantly correlated with survival. In multivariate analysis using the Cox proportional hazard model, 5 independent prognostic factors significantly correlated with survival the age of the patients (RR = 1.480), the depth of invasion (RR = 1.287), the number of metastatic lymph nodes (RR = 2.700), the TNM stage (RR = 1.592), and the extent of lymph-node dissection (RR = 1.390). CONCLUSIONS: The authors can make a prognosis for stomach cancer more precisely on the basis of these independent prognostic factors, and an extended D2 operation (D2 alpha operation) may be a valuable means to treat gastric cancer.