Abstract
BACKGROUND: To investigate influencing factors of the metastatic lymph nodes ratio (MLR) and whether it is related to survival in patients with gastric adenocarcinoma. METHODS: We retrospectively evaluated the clinical features of 121 patients with gastric adenocarcinoma enrolled in our hospital between 2000 and 2007. The receiver operating characteristic (ROC) curve was used to determine the cutoff of the MLR, and CK20 immunohistochemical staining was used to detect micrometastasis of the lymph nodes. RESULTS: The areas under the ROC curve of MLR used to predict the death of 3-year and 5-year postoperative patients were 0.826 +/- 0.053 and 0.896 +/- 0.046. Thus MLR = 30.95% and MLR = 3.15% were designated as cutoffs. The MLR was then classified into three groups: MLR1 (MLR<3.15%); MLR2(3.15% < or = MLR < or = 30.95%); and MLR3 (MLR>30.95%). We found that patients with a higher MLR demonstrated a much poorer survival period after radical operation than those patients with a lower MLR (P = 0.000). The COX model showed that MLR was an independent prognostic factor (P = 0.000). The MLR could also discriminate between subsets of patients with different 5-year survival periods within the same N stage (P < 0.05). The MLR has been shown to be 34.7% (242/697) by HE staining and 43.5% (303/697) by CK staining (P = 0.001). The clinicopathological characteristics of lymph vessel invasion and the depth of invasion could significantly affect the MLR. CONCLUSION: MLR is an independent prognostic factor in gastric cancer. The combined ROC curve with MLR is an effective strategy to produce a curve to predict the 3-year and 5-year survival rates.