Identification and validation of a blood-based 18-gene expression signature in colorectal cancer

Clin Cancer Res. 2013 Jun 1;19(11):3039-49. doi: 10.1158/1078-0432.CCR-12-3851. Epub 2013 Mar 27.

Abstract

Purpose: The early detection of colorectal cancer (CRC) is crucial for successful treatment and patient survival. However, compliance with current screening methods remains poor. This study aimed to identify an accurate blood-based gene expression signature for CRC detection.

Experimental design: Gene expression in peripheral blood samples from 216 patients with CRC tumors and 187 controls was investigated in the study. We first conducted a microarray analysis to select candidate genes that were significantly differentially expressed between patients with cancer and controls. A quantitative reverse transcription PCR assay was then used to evaluate the expression of selected genes. A gene expression signature was identified using a training set (n = 200) and then validated using an independent test set (n = 160).

Results: We identified an 18-gene signature that discriminated the patients with CRC from controls with 92% accuracy, 91% sensitivity, and 92% specificity. The signature performance was further validated in the independent test set with 86% accuracy, 84% sensitivity, and 88% specificity. The area under the receiver operating characteristics curve was 0.94. The signature was shown to be enriched in genes related to immune functions.

Conclusions: This study identified an 18-gene signature that accurately discriminated patients with CRC from controls in peripheral blood samples. Our results prompt the further development of blood-based gene expression biomarkers for the diagnosis and early detection of CRC.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Biomarkers, Tumor / blood
  • Biomarkers, Tumor / genetics*
  • Colorectal Neoplasms / blood
  • Colorectal Neoplasms / diagnosis*
  • Colorectal Neoplasms / genetics*
  • Female
  • Gene Expression Profiling*
  • Humans
  • Male
  • Middle Aged
  • Neoplasm Staging
  • ROC Curve
  • Reproducibility of Results
  • Transcriptome*

Substances

  • Biomarkers, Tumor