The Biostatistics Core (new in this application) provides partial support for four biostatisticians embedded in specific projects or cores. The areas of support include study design, data analysis, and joint development, with the Bioinformatics Core, of methods for data processing, quality control, data management and data retrieval. The existence of this Core assures a uniform plan of protocol design, data handling and statistical analysis. Furthermore, it assures that appropriate resources are available to all investigators. The goal of Project 1 is to identify and test a set of markers that predict subsequent tumor events in women with DCIS. Biostatistical support includes sample size planning, investigation of cut-points to dichotomize potential biomarkers and investigation, using logistic regression and stratified Cox proportional hazards, of their ability to predict tumor events and survival. Useful biomarkers found in this project will be added to the Markov model predicting outcome based on tumor characteristics, mode of discovery, initial treatment and biomarkers. Project 2 is concerned with developing a model to predict response to chemotherapy in breast cancer cell lines and using that model to select breast cancer subtype specific drugs. The model is based on defining and testing patterns of gene expression and copy number that are correlated with drug response in cell lines and in tumors. Support for Project 3, which is concerned with developing targeted therapeutic agents, includes experimental design to define a small number (one or two) candidate agents that restrict tumor growth in xenograft mice. Several endpoints summarizing tumor growth over time will be evaluated statistically. Project 4 is focused on discovering agents that interfere with telomerase activity. Biostatistical support includes evaluating measures of response in cells and finding patterns in cell biomarkers that predict response to proposed anti-telomerase agents.
Showing the most recent 10 out of 339 publications
