Metastasis-Associated Gene Expression Changes Predict Poor Outcomes in Patients with Dukes Stage B and C Colorectal Cancer

Abstract

PURPOSE: Colorectal cancer prognosis is currently predicted from pathologic staging, providing limited discrimination for Dukes stage B and C disease. Additional markers for outcome are required to help guide therapy selection for individual patients. EXPERIMENTAL DESIGN: A multisite single-platform microarray study was done on 553 colorectal cancers. Gene expression changes were identified between stage A and D tumors (three training sets) and assessed as a prognosis signature in stage B and C tumors (independent test and external validation sets). RESULTS: One hundred twenty-eight genes showed reproducible expression changes between three sets of stage A and D cancers. Using consistent genes, stage B and C cancers clustered into two groups resembling early-stage and metastatic tumors. A Prediction Analysis of Microarray algorithm was developed to classify individual intermediate-stage cancers into stage A-like/good prognosis or stage D-like/poor prognosis types. For stage B patients, the treatment adjusted hazard ratio for 6-year recurrence in individuals with stage D-like cancers was 10.3 (95% confidence interval, 1.3-80.0; P = 0.011). For stage C patients, the adjusted hazard ratio was 2.9 (95% confidence interval, 1.1-7.6; P = 0.016). Similar results were obtained for an external set of stage B and C patients. The prognosis signature was enriched for downregulated immune response genes and upregulated cell signaling and extracellular matrix genes. Accordingly, sparse tumor infiltration with mononuclear chronic inflammatory cells was associated with poor outcome in independent patients. CONCLUSIONS: Metastasis-associated gene expression changes can be used to refine traditional outcome prediction, providing a rational approach for tailoring treatments to subsets of patients. (Clin Cancer Res 2009;15(24):7642-51).

Fig 1

Unsupervised clustering for stage B and stage C colorectal cancers based on metastasis-associated genes. Clustering divides 95 stage B and 93 stage C cases from this study (A–B) and 83 stage B and 73 stage C cases from expO (C–D) into groups with early-stage and metastatic profiles. Samples are arranged along the x-axis and genes along the y-axis. Orange represents increased and blue decreased expression relative to the mean- and sample-centered scaled expression. Genes are grouped into those found to be down-regulated (blue) and up-regulated (orange) in metastatic cancers as indicated by the color bars.

Fig 2

Comparison of disease-free survival among stage B and C patients when grouped by (A) pathological staging, (B) the 128-gene PAM classifier, (C) and both pathological staging and the PAM classifier. For single-sample PAM classification, prior (expected) six-year recurrence probabilities were set as 30% for all cases. Class predictions with a >90% probability were scored.

Fig 3

External prognosis classifier validation (Danish dataset). (A) Unsupervised clustering using metastasis-associated genes represented on HG-U133A GeneChip arrays. Samples are arranged along the x-axis and genes along the y-axis. Each square represents the expression level of a given gene in an individual sample. Orange represents increased expression and blue represents decreased expression relative to the mean- and sample-centered scaled expression of the gene across the samples. Genes are grouped into those found to be down-regulated (blue) and up-regulated (orange) in metastatic cancers as compared to early-stage cancers as represented by the color bars. The two main groups resulting from clustering show early-stage and metastatic profiles as indicated. (B) Survival curves generated using PAM classification show a significant difference in outcome. Class predictions with a >90% probability were scored. (C) Survival curves using Dukes’ staging criteria show a significant difference in outcome. (D) Survival curves grouped by both Dukes’ stage and molecular signature show that both stage B and C patients can be further subdivided into good and poor prognosis groups.