Cox regression model with change-point(s) problems and their application in cancer drug development

本提案旨在开发和评估癌症临床试验实践中需要的新的统计方法，这将使：1)确定预测性生物标志物的截止点或建立癌症治疗选择的基因标记分类器；2)更好地估计处理效果；3)改进临床试验的计划和分析；4)研究其在卫生经济学中的意义。这些方法将在具有变化点问题的Cox回归模型框架中实现。

This proposal is for developing and evaluating new statistical methods needed in cancer clinical trial practice, which will enable: 1) identifying cutoff points in predictive biomarkers or build gene signature classifier for cancer treatment selection; 2) better estimating treatment effect; 3) improved planning and analysis of clinical trials, and 4) study its implication in health economics. These methods will be implemented in the frame work of Cox regression models with change-point(s) problems.