Original Article
Development and validation of an individualized gene expression-based signature to predict overall survival in metastatic colorectal cancer
Abstract
Background: Metastatic colorectal cancer (mCRC) is a heterogeneous disease. Predictive biomarkers are in great demand to optimize patient selection at high risk for death and to provide a novel insight into potential targeted therapy.
Methods: The present study retrospectively analyzed the gene expression profiles of tumor tissue samples from 4 public CRC cohorts, including 1 RNA-Seq data set from The Cancer Genome Atlas (TCGA) CRC cohort and 3 microarray data sets from GEO. Prognostic analysis was performed to test the predictive value of prognostic gene signature.
Results: Of 192 patients, 108 patients (56.3%) were men and median age was 65 years. A prognostic gene signature that consisted of 15 unique genes was generated in the discovery cohort. In the meta-validation cohorts, the signature significantly classified patients into high-risk and low-risk groups with regard to overall survival (OS) in mCRC patients with advanced stage disease and remained as an independent prognostic marker in multivariable analysis (1.57; 95% CI: 1.16–2.11; P=0.003) after adjusting for clinical parameters and molecular types. Gene Set Enrichment Analysis showed that several biological processes, including angiogenesis (P<0.001), epithelial mesenchymal transit (P<0.001) and inflammatory response (P=0.001), were enriched among this prognostic gene signature.
Conclusions: The proposed prognostic gene signature is a promising prognostic tool to estimate OS in mCRC. Prospective larger studies to examine the clinical utility of the biomarkers to guide individualized treatment of mCRC are warranted.
Methods: The present study retrospectively analyzed the gene expression profiles of tumor tissue samples from 4 public CRC cohorts, including 1 RNA-Seq data set from The Cancer Genome Atlas (TCGA) CRC cohort and 3 microarray data sets from GEO. Prognostic analysis was performed to test the predictive value of prognostic gene signature.
Results: Of 192 patients, 108 patients (56.3%) were men and median age was 65 years. A prognostic gene signature that consisted of 15 unique genes was generated in the discovery cohort. In the meta-validation cohorts, the signature significantly classified patients into high-risk and low-risk groups with regard to overall survival (OS) in mCRC patients with advanced stage disease and remained as an independent prognostic marker in multivariable analysis (1.57; 95% CI: 1.16–2.11; P=0.003) after adjusting for clinical parameters and molecular types. Gene Set Enrichment Analysis showed that several biological processes, including angiogenesis (P<0.001), epithelial mesenchymal transit (P<0.001) and inflammatory response (P=0.001), were enriched among this prognostic gene signature.
Conclusions: The proposed prognostic gene signature is a promising prognostic tool to estimate OS in mCRC. Prospective larger studies to examine the clinical utility of the biomarkers to guide individualized treatment of mCRC are warranted.
