Personalized Risk Prediction in Clinical Oncology Research: Applications and Practical Issues Using Survival Trees and Random Forests

摘要: ABSTRACTA crucial component of making individualized treatment decisions is to accurately predict each patient’s disease risk. In clinical oncology, risks are often measured through time-to-event data, such as overall survival and progression/recurrence-free survival, subject censoring. Risk prediction models based on recursive partitioning methods becoming increasingly popular largely due their ability handle nonlinear relationships, higher-order interactions, and/or high-dimensional covariates. The most versions the Classification Regression Tree (CART) algorithm, which builds a simple interpretable tree structured model. With aim increasing accuracy, random forest algorithm averages multiple CART trees, creating flexible risk used in oncology commonly use both traditional demographic tumor pathological factors well gene...