Prognostic value of tumor size in gastric cancer: an analysis of 2,379 patients.

Tumor size has been included into the staging systems of many solid tumors, such as lung and breast. However, tumor size is not integrated in the staging of gastric cancer, and its prognostic value for gastric cancer needs to be reappraised. A total of 2,379 patients who received radical resection for histopathologically confirmed gastric adenocarcinoma were enrolled in the present study. Tumor size, originally presented as continuous variable, was categorized into small gastric cancer (SGC) group and large gastric cancer (LGC) group using an optimal cutoff point determined by Cox proportional hazards model. The associations between tumor size and other clinicopathological factors were checked using Chi-square test. Survival of gastric cancer patients was estimated by using univariate Kaplan-Meier method, and the survival difference was checked by using the log-rank test. The significant clinicopathological factors were included into the Cox proportional hazards model to determine the independent prognostic factors, and their hazard ratios were calculated. With the optimal cutoff point of 4 cm, tumor size was categorized into SGC group (≤ 4 cm) and LGC group (>4 cm). Tumor size closely correlated with age, tumor location, macroscopic type, Lauren classification, and lymphatic vessel invasion. Moreover, tumor size was also significantly associated with depth of tumor invasion and status of regional lymph nodes. The 5-year survival rate was 68.7 % for SGC group which was much higher than 40.2 % for LGC group. Univariate analysis showed that SGC had a better survival than LGC, mainly for patients with IIA, IIB, and IIIA stage. Multivariate analysis revealed that tumor size as well as age, tumor location, macroscopic type, Lauren classification, lymphatic vessel invasion, depth of tumor invasion, and status of regional lymph nodes were independent prognostic factors for gastric cancer. Tumor size is a reliable prognostic factor for patients with gastric cancer, and the measurement of tumor size would be helpful to the staging and management of gastric cancer.