PDGFR-alpha expression was analyzed by immunohistochemistry, and expression was scored separately for epithelial cells, stroma fibroblasts and perivascular cells. In general, PDGFR-alpha expression was frequently seen in perivascular cells and fibroblasts, but not in epithelial cells. Fibroblast expression was up-regulated in tumors as compared to normal tissue. PDGFR-alpha

expression was higher in colon cancer fibroblasts than in rectal cancer fibroblasts. PDGFR-alpha expression in primary tumor CAFs was correlated with more advanced N stage. Several associations were observed between PDGFR-alpha expression in lymph find more node metastases and survival. Increased expression of PDGFR-alpha in lymph node fibroblasts was associated with worse survival in the whole patient cohort. High PDGFR-alpha expression in fibroblasts or pericytes in lymph nodes was associated with increased recurrence risk in curatively treated patients. The associations between

survival and stromal PDGFR-alpha lymph node expression were also significant in a multivariate analysis. Interestingly, also high expression of PDGFR-alpha in fibroblasts of normal mucosa was associated with worse over-all survival. These findings thus highlight the prognostic potential of tumor stroma and specifically demonstrate novel prognostic significance of stromal PDGFR-alpha in CRC. The associations between PDGFR-alpha status of normal mucosa and survival also points to the importance of “host factors” in tumor progression. Poster No. 58 Serum Levels PD0332991 concentration of Dermcidin Increase with Progression of Mammary Carcinogenesis OSBPL9 Heather Ann Brauer 1,2 , Tanya E. Libby1, Yutaka Yasui3, Anne McTiernan1, Henry J. Thompson4, Paul D. Lampe1,2 1 Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 2 Molecular and Cellular Biology Program, University of Washington, Seattle, WA, USA, 3 Department of Public Health Sciences, University of Alberta, Edmonton, AB, Canada,

4 Cancer Prevention Laboratory, Colorado State University, Fort Collins, CO, USA Early detection and prognostic profiling of cancers has the potential to increase lifespan and quality of life. The “field effect” hypothesis that motivated this investigation suggests that there are cellular changes that occur both within and around tumor cells that could be detectable in serum. These changes may be detectable before the disease is histologically identifiable using the current testing methods. This valuable information could potentially come from serum where early stages of tumorigenesis lead to changes in the serum peptidome. An experiment testing this idea was carried out using a rat model of mammary carcinoma. Samples were collected at different stages of progression and abundant proteins depleted to determine if MALDI-TOF mass spectrometry could provide a proteomic profile that could identify disease.