All Times EDT
Keywords: Subgroup identification, Personalized medicine, Censored survival data, Nonparametric, restricted mean survival time, Gradient tree boosting
In randomized clinical trials with survival data, there has been increasing interest in subgroup identification based on baseline genomic, proteomic markers or clinical characteristics. Some of the existing methods directly search for subgroups that benefit substantially from the experimental treatment by modeling outcomes or treatment effect. But these methods do not connect an overall expected clinical outcome to be maximized with the identified subgroups. When the goal is to find an optimal treatment for a given patient rather than finding the right patient for a given treatment, methods under individualized treatment regime framework estimate an individualized treatment rule, if followed by all patients, would lead to the best expected clinical outcome, which is measured by a value function. Connecting the concept of value function to subgroup identification, we propose a nonparametric method that searches for subgroup membership scores by maximizing a value function that directly reflects the subgroup-treatment interaction effect based on restricted mean survival time. A gradient tree boosting algorithm is proposed to search for the individual subgroup membership scores. We conduct simulation studies to evaluate the performance of the proposed method and an application to an AIDS clinical trial is performed for illustration.