Nonparametric Testing Methods for Treatment-Biomarker Interaction Based on Local Partial-Likelihood

• Activity Number: 128
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Yicong Liu*+ and Wenyu Jiang and Bingshu E. Chen
• Companies: Queen's University and Queen's University and NCIC Clinical Trials Group and Department of Community Health and Epidemiology, Queen's U
• Keywords: bootstrap ; clinical trials ; treatment-covariate interaction ; nonparametric estimation ; survival analysis

Abstract

A fair amount of research has been done on the interactions between treatment and biomarkers. Particularly, Fan et al. (2006) assumed the treatment effect is an unknown function of a putative biomarker, and give the local partial likelihood estimation (LPLE) of this function. However, no methods were developed for assessing whether the treatment effect is a function of the biomarker (interaction exists) or just a constant (no interactions). Motivated by this, we propose a local partial likelihood bootstrap (LPLB) test in this paper based on the idea of LPLE. By evaluating different null hypothesis, the LPLB test can be used to assess whether the treatment benefit varies among patient subsets of different biomarker values, or further, whether there are treatment effects of any types in any subsets of patients. Simulation shows that the LPLB test is very powerful when testing in both cases. Furthermore, it can also be used to study the interactions between a biomarker and a continuous and/or multidimensional covariate. Application to real data is given to show how it works in practice. Fan et al. (2006) Local partial-likelihood estimation for lifetime data. The Annals of Statistics.