Abstract
Colorectal cancer (CRC) is a major cause of global cancer mortality. Gene expression profiles can help predict prognosis of patients with CRC. In most of previous studies, disease recurrence was analyzed as the survival endpoint. Thus we aim to build a robust gene signature for prediction of overall survival (OS) in patients with CRC. Fresh frozen CRC tissues from 64 patients were analyzed using Affymetrix HG-U133plus 2.0 gene arrays. By performing univariate survival analysis, 6487 genes were found to be associated with the OS in our cohort. KEGG analysis revealed that these genes were mainly involved in pathways such as endocytosis, axon guidance, spliceosome, Wnt signalling and ubiquitin mediated proteolysis. A seven-gene signature was further selected by a robust likelihood-based survival modelling approach. The prognostic model of seven-gene signature (NHLRC3, ZDHHC21, PRR14L, CCBL1, PTPRB, PNPO, and PPIP5K2) was constructed and weighted by regression coefficient, which divided patients into high- and low-risk groups. The OS for patients in high-risk group was significantly poorer compared with patients in low-risk group. Moreover, all seven genes were found to be differentially expressed in CRC tissues as compared with adjacent normal tissues, indicating their potential role in CRC initiation and progression. This seven-gene signature was further validated as an independent prognostic marker for OS prediction in patients with CRC in other two independent cohorts. In short, we developed a robust seven-gene signature that can predict the OS for CRC patients, providing new insights into identification of CRC patients with high risk of mortality.