Dynamic treatment regimens and biomarker-driven strategies: Bridging the two worlds

Biostatisticians have contributed a myriad of statistical methods for designing and analyzing clinical trials with the goal to develop the right treatments for the right patients at the right time. In biostatistical precision medicine research, there are two schools of thoughts that have been thriving relatively in parallel. On one hand, researchers focus on using molecularly targeted information to study treatment effects in marker defined subpopulations with the goal to improve patients’ clinical benefits with prognostic or predictive biomarkers. On the other hand, researchers focus on using intermediate patient outcomes, collected during the clinical care process, to select an optimized treatment pathway among all the sequential treatment options. Namely, the two strategies described are biomarker-driven designs and dynamic treatment regimes. In this chapter, we describe the basic characteristics of both strategies to illustrate their differences and similarities. Our goal is to provide readers a balanced introduction of biostatistical methods driving the precision medicine research. Our effort lies on making meaningful connections between the two strategies through understanding the different aspects they target.