ColoGuidePro: a prognostic 7-gene expression signature for stage III colorectal cancer patients

Clin Cancer Res. 2012 Nov 1;18(21):6001-10. doi: 10.1158/1078-0432.CCR-11-3302. Epub 2012 Sep 18.


Purpose: Improved prognostic stratification of patients with stage II and III colorectal cancer is warranted for postoperative clinical decision making. This study was conducted to develop a clinically feasible and robust prognostic classifier for these patients independent of adjuvant treatment.

Experimental design: Global gene expression profiles from altogether 387 stage II and III colorectal cancer tissue samples from three independent patient series were included in the study. ColoGuidePro, a seven-gene prognostic classifier, was developed from a selected Norwegian learning series (n = 95; no adjuvant treatment) using lasso-penalized multivariate survival modeling with cross-validation.

Results: The expression signature significantly stratified patients in a consecutive Norwegian test series, in which patients were treated according to current standards [HR, 2.9 (1.1-7.5); P = 0.03; n = 77] and an external validation series [HR, 3.7 (2.0-6.8); P < 0.001; n = 215] according to survival. ColoGuidePro was also an independent predictor of prognosis in multivariate models including tumor stage in both series (HR, ≥ 3.1; P ≤ 0.03). In the validation series, which consisted of patients from other populations (United States and Australia), 5-year relapse-free survival was significantly predicted for stage III patients only (P < 0.001; n = 107). Here, prognostic stratification was independent of adjuvant treatment (P = 0.001).

Conclusions: We present ColoGuidePro, a prognostic classifier developed for patients with stage II and III colorectal cancer. The test is suitable for transfer to clinical use and has best prognostic prediction potential for stage III patients.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Aged, 80 and over
  • Colorectal Neoplasms / diagnosis*
  • Colorectal Neoplasms / genetics*
  • Colorectal Neoplasms / mortality
  • Computational Biology
  • Female
  • Gene Expression Profiling / methods*
  • Humans
  • Male
  • Middle Aged
  • Neoplasm Staging
  • Prognosis
  • Transcriptome*
  • Workflow