Research Abstracts Online
January 2009 - March 2010
University of Minnesota Twin Cities
Department of Laboratory Medicine and Pathlogy
PI: Stephen C. Schmechel
Identification and Validation of Prognostic Prostate Cancer Biomarkers
Standard pre-operative prognostic factors such as prostate specific antigen (PSA) are important for the management of prostate cancer but they provide only limited information on extent and aggressiveness of disease. Understanding extent and aggressiveness is crucial for initial treatment planning, as these factors dictate whether local control is likely to be possible, and are important predictors of ultimate clinical outcome. Treatment of prostate cancer involves invasive techniques that may not be necessary in a subset of patients. New biomarkers are needed in order to accurately differentiate aggressive from non-aggressive prostate cancer tissue.
This researcher has isolated a set of genes correlated to high aggressiveness and a set of genes correlated to non-aggressiveness. An ordered gene list (OGL) analysis method was used to rank genes whose expression data is consistently associated with outcome, and involved assigning a quality score to each gene as a predictor of aggressive vs. non-aggressive tumor biology. A voting algorithm was then used to weight genes based on their quality score and expression values. The result was that an eight-gene model optimally distinguished aggressive from non-aggressive tumors, and was validated by imposing on other datasets. Further analysis will compile a panel of genes to be subjected to validation by immunohistochemistry of prostate tumor tissue microarrays.