Tuesday, November 14, 2017
3:30 pm - 4:30 pm
701 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104
Title: Spatial statistics approach to develop novel protein cancer biomarkers
Abstract: Quantitative Immunofluorescence (QIF) is used for immunohistochemistry (IHC) quantification of proteins that serve as prognostic biomarkers or targets for cancer therapies. Traditionally, only the mean signal intensity (MSI) of the protein expression across cancer cells is considered for developing protein cancer biomarkers. We propose a new approach for developing IHC biomarkers using the information on spatial distribution of cellular signal intensity (CSI) of protein expression across the cancer cell population. We view protein QIF expression levels as marks for the marked point process of cancer cells in the tumor tissue and develop spatial index predictors of clinical outcomes based on nonparametric spatial statistics describing the relationship between marks (protein QIF expression) and points (cancer cell locations). The utility of the new spatial index IHC biomarkers is investigated and compared to the standard MSI predictors using the protein expressions in tissue microarrays (TMAs) incorporating tumor tissues from 2,000+ breast cancer patients. The new approach provides new insight into the standard breast cancer protein biomarkers and identifies novel protein biomarkers that do not have a prognostic value if only the mean signal intensity is considered.