Gene expression profiling associated with risk of recurrence in patients with colorectal cancer

Abstract

3602 Background: Early identification of patients at increased risk for recurrence remains a central issue in the treatment of colorectal cancer. Altered expression levels in cancer-related genes have emerged as potential markers in determining risk of recurrence. However, due to the large and complex variety of pathways that influence colon cancer progression, a set of molecular markers has not been established as a basis for clinical prognosis. The use of microarray technology addresses this problem by taking a comprehensive approach to examining gene expression rather than a restricted pathway approach. We analyzed gene expression profiles of 38 colorectal cancer patients using microarrays to determine differential expression patterns of recurrence versus recurrence-free patients. Methods: 38 patients receiving adjuvant chemotherapy were selected, of which 14 experienced recurrence and 24 did not. RNA was extracted from fresh frozen tumor tissue and microarray analysis was conducted using Affymetrix U133A genechip containing 22,000 probe sets. Cluster analysis was performed to differentiate between recurrence and recurrence-free patients. In addition, significance analysis of microarrays (SAM analysis) was conducted to establish an expression profile based on recurrence. Genes with a greater than 1.5 fold difference in expression from the standard deviation were considered significant, with a false discovery rate of less than 10%. Results: By SAM analysis, 87 genes were found significantly differentially expressed in recurrence vs. non-recurrence patients. The 87 genes do not all associate functionally but rather are involved in a variety of cancer-related pathways, including angiogenesis, cell proliferation, apoptosis, cell cycle regulation, and cell adhesion. Additionally, clustering showed a mild correlation between recurrence and recurrence-free patients. Conclusions: These results demonstrate the ability of expression profiling using microarrays to predict recurrence in colorectal cancer. Expression profiling may be used in combination with traditional pathologic staging to more accurately characterize disease status and hence individualize treatment at the genetic level. Author Disclosure Employment or Leadership Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Chiron, Genentech, Response Genetics Lilly Oncology, Pfizer, Roche, Sanofi NCI, NIH