Because of the longitudinal follow-up on the biomarker of PSA and large sample size, we would like to perform the following aims:
• Develop prediction models for prostate cancer development and also the metastasis by considering PSA and other demographic/clinical/social economic data using the PLCO;
• Develop the models and toolbox/website to analyze newly diagnosis patients using PLCO newly diagnosed cancer patients. We can perform data cleaning, process, harmonization, summary statistics, and further clustering analysis to identify subtypes. In particular, we can consider prediction of (overall and cancer-specific) mortality for overall prostate cancer patients, and also the subcohort with metastatic prostate cancer patients only.
• Develop the models and toolbox/website for evaluation of the treatment information and treatment decision making that is of most importance and clinical impact.

Nicholas Zaorsky, Penn State College of Medicine
Alicia McDonald, Penn State College of Medicine