Parametric and Nonparametric Analysis of Breast Cancer for Censored and Uncensored Data

• Activity Number: 168
• Type: Contributed
• Date/Time: Monday, August 3, 2009 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Chris P. Tsokos and Chunling Cong*+
• Companies: University of South Florida and University of South Florida
• Keywords: parametric analysis ; nonparametric analysis ; breast cancer ; censored date ; uncensored data

Abstract

In the present study, we utilized breast cancer data from Fyles et al.2004,to conduct analysis for survival time after breast-conserving surgery, and the underlying distributions are identified for two groups of both censored and uncensored data: women who received tamoxifen and those who received tamoxifen plus breast irradiation. We developed parametric and nonparametric survival analysis for both data sets and results were compared. Decision tree technique was also used to homogenize the data, and survival analysis was performed for the subject data sets. In addition, we developed a statistical model in which the attributable variables are treatment, tumor size, etc.; and the response variable is survival time. This model has been statistically evaluated and it gives a good estimate of the survival time after breast-conserving surgery given information of attributable variables.