A Flexible Cure Rate Model in Population-Based Cancer Survival Analysis

• Activity Number: 657
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Myron Chang* and Yang Li
• Companies: University of Florida and University of Florida
• Keywords: mixture model ; promotion time model ; relative survival function

Abstract

The cure rate is a useful measure of the effectiveness of anti-cancer treatments. When a clinical trial targets on elder people or a clinical trial has long-term follow-up, patients may die for reasons other than the cancer of interest, and the cause of death is often not available or incorrectly identified. Many authors consider the cure rate as the proportion of patients whose survival experience is equivalent to that in the general population with matched age, gender, and other potential covariates, and perform the so-called population-based analysis. The mixture model and the non-mixture model are the two major types of models commonly used in population-based studies. It may be difficult for practitioners to decide which model should be used in their data analysis. We extend the models proposed by Yin and Ibrahim (2005) to propose a class of models that include the mixture model and the non-mixture model as special cases for population-based studies. With an additional parameter, the proposed models offer great flexibility in model selection. The final model, determined through the data fitting, could be a model close to the mixture model, a model close to the non-mixture mode