ABSTRACT
This chapter aims to establish the relationships between the treatment effects along each endpoint, rather than to summarize the treatment effect on a single endpoint. For example, based on summary data from multiple clinical trials, the hazard ratios (HRs) for comparing two treatments on overall survival (OS) can be regressed on the HRs of treatment groups for progression-free survival. Meta-analytic regression equations take many different forms in the published literature, depending which endpoints are evaluated, whether transformations are needed, what is the most appropriate statistical model to be implemented, and how weights for each study are to be calculated. The analysis for validation of progression-free survival as a surrogate endpoint for OS in malignant mesothelioma was presented in a paper by Wang et al. The data sets consisted of single arm trials and HRs were constructed comparing risk groups. In other words, the meta-regression equation shows predicted relationship between HRs for progression-free survival and OS, based on studies included.
