The old paradigm for First In Human (FIH) studies involved increasing dose to find maximum tolerated dose. That is largely outdated, due to targeted therapy based on our presumed knowledge of a biological pathway, with the intention of producing a desirable clinical effect by interrupting the pathway. A critical early question is “To what degree are we interrupting the pathway?”, so that we can decide whether to stop development for insufficient target activation, or what dosage regimen to pursue. Development of a mathematical/statistical relationship between drug concentration and degree of interruption enables the insight to make those key decisions. When the pathway exists in normal subjects it may be possible to measure the degree of interruption by a pharmacodynamic biomarker (PD), and model the PD response as a function of the pharmacokinetics (PK). Developing usable PK/PD models may often involve working closely with the analytical laboratory to employ constrained nonlinear regression to derive the PD response, with pharmacologists to employ appropriate pharmacology models, and pharmacokineticists. This talk will focus on some of the statistical issues in developing the PK/PD models, and applications for dose selection or stopping a program for futility.