Abstract
BACKGROUND: Predicting lymph node metastasis (LNM) accurately is vital to design optimal treatment strategies preoperatively for gastric cancer (GC) patients. However, conventional tumor biomarkers and imaging techniques are not sufficient to predict LNM before surgery. miR-126 has been reported to play important roles in tumor metastasis which may represent a novel tumor biomarker. OBJECTIVE: To assess the utility of the combination of serum miR-126 and multi-detector computed tomography (MDCT) in predicting LNM preoperatively in GC. METHODS: Quantitative real-time polymerase chain reaction was performed to examine the serum miR-126 expression levels in 338 GC patients. MDCT was also performed. The cut-off value of preoperative serum miR-126 level for LNM was determined by receiver characteristic curve (ROC) analysis. Logistic regression analysis was used to determine independent predictors for LNM. RESULTS: The serum miR-126 levels of GC patients with LNM were significantly lower compared with those without LNM (p< 0.05). In addition, the later the N stage was, the lower the serum miR-126 level was in GC patients (p< 0.05). With a cut-off value of 63.4, the sensitivity, specificity, accuracy, positive predictive value and negative predictive value of serum miR-126 for predicating LNM were 83.2%, 79.0%, 81.9%, 90.4% and 66.4%, respectively. The combination of serum miR-126 level and MDCT increased the accuracy of MDCT prediction for LNM from 69.2% to 86.7%. Serum miR-126 was an independent predictor for LNM. CONCLUSIONS: These results indicate for the first time that the combination of serum miR-126 and MDCT is useful for the prediction of LNM in GC.