Identification of a 5-gene signature for clinical and prognostic prediction in gastric cancer patients upon microarray data

In this study, we aimed to investigate the clinical and prognostic value of a 5-gene expression signature model for gastric cancer patients upon microarray data. A total of 158 gastric cancer patients were selected, with 33 cases used for microarray analysis as training set and 125 cases for validation real-time quantitative polymerase chain reaction analysis as test set. Unsupervised clustering algorithms and supervised clustering algorithm were used to identify differentially expressed genes. Gene ontology analyses were used to determine functional prediction of gene biomarkers and receiver operating characteristic analyses to verify the specificity and sensitivity of the evaluation. Moreover, the correlation between clinicopathological characteristics and 5-gene expression was evaluated. The results showed that there were poor disease progression and clinic prognosis in the patients with proximal gastric cancer compared with distal gastric cancer, including differentiation grade (P = 0.001), depth of tumor invasion (P < 0.01), lymph node metastasis (P < 0.01), UICC stage (P < 0.01), and survival status (P < 0.01). Furthermore, a 5-gene signature, including cordon-bleu protein-like 1, damage-specific DNA binding protein 1, BCL2-like 13, nuclear receptor coactivator-6, and F-box leucine-rich protein 11, was identified based on the combination of multiple bioinformatics algorithm. The expression of 5-gene signature (HR = 2.35; 95 % confidence interval 1.24–5.06; P = 0.026) and UICC stage (HR = 5.35; 95 % confidence interval 1.36–19.15; P = 0.032) was independent prognostic factors for overall survival in the survival multivariate analysis. Over-expression of the 5-gene signature in patients with proximal gastric cancer is strongly associated with disease progression and poor prognosis, suggesting that might be potential prognostic predictors in gastric cancer.