It becomes more common that clinical validation study for certain diagnostic devices involves time-event data. While survival analysis is always an interesting topic in statistics, new issues faced in time to event data analysis of risk prediction and companion diagnostic devices is practically challenging. How to address the time dimension and censoring has been an active research area in diagnostic medicine. In this talk, we are going to explore some recent advance and challenges and issues of survival analysis such as competing risks, cure rate in risk predication and companion diagnostic devices of personalized medicine. In our presentation, we will discuss the definition and statistical issues of competing risk and “cured”patients. We will also propose a statistical method in addressing competing risk and cured rate using Baysian approach. Simulations and examples will be presented to illustrate our method.