PK/PD Modeling and simulation: Applications in discovery through development

Technology has provided us with excellent tools to measure drug effect and drug/metabolite concentration vs. time profiles, even during discovery and early development phases. We now have high sensitivity LC/MS-MS assays, biomarkers, and good old-fashioned pharmacologic endpoints, plus software that allows the use of sophisticated PBPK and nonlinear mixed effect PK/PD models on every desktop. It appears that in silico PBPK modeling is superior to in vivo allometric scaling approaches, thus making PK/PD relationships transportable across species. Given this scenario, when is it acceptable to measure only PK or PD? Not very often, if we consider the cost of delaying knowledge of key exposure-response relationships at each step of the discovery and development continuum. The key is maximum utilization of these tools early, which requires strong study designs aimed at explicit questions. Some recent examples from the speaker's work will be used to illustrate these recommendations.