Translation from preclinical pharmacology to human dosing is often described as an exposure-scaling problem, yet systemic pharmacokinetics, tissue exposure, target engagement, downstream pharmacology, and clinical decision utility are not interchangeable evidence layers. A compound can achieve predicted plasma exposure without reaching the relevant tissue compartment, demonstrate measurable target occupancy without sufficient functional inhibition, or produce coherent pharmacodynamic changes without resolving safety or efficacy uncertainty. This state-of-the-art review examines these separations across free-drug and tissue-exposure concepts, target-engagement measurement, species scaling, pharmacokinetic–pharmacodynamic bridging, biomarker translation, and human dose projection. The evidence supports an explicitly conditional view of translation in which each transfer between exposure, access, engagement, function, and dose constitutes an inferential boundary requiring its own validation. We therefore develop a proposed falsification-oriented translational architecture: exposure predictions should be tested against target-site access, engagement should be distinguished from functional consequence, biomarkers should be interpreted according to context of use and timing, and dose hypotheses should remain revisable after human pharmacokinetic, pharmacodynamic, and safety observations. The resulting approach treats translational development as progressive reconciliation of partially independent evidence rather than confirmation of a single preclinical prediction.
IJPRAS
