Cancer Survival: Analysis and Reporting
摘要
Time-to-event endpoints such as overall survival, metastasis-free survival, and ocular survival are commonly of interest in clinical ophthalmic oncology research studies. With time-to-event endpoints, subjects are considered censored if they were not observed to have the event during the study period. This type of endpoint shares the important feature of consisting of two main parts: the time until the event or censoring and whether the event occurred during the study period. As such, they require specialized statistical techniques known as survival analysis, including the Kaplan–Meier method and Cox proportional hazards regression analysis. This chapter will introduce readers to the reasons why special statistical methods are needed for time-to-event endpoints and will describe the basic approaches to the analysis and reporting of survival data.