Do Brain Extracellular Fluid Concentrations Predict PET ‐Inferred Receptor Occupancy in Human CNS ? A Mechanistic Model‐Based Analysis of Dopamine D2 Receptor Ligands

We previously developed a mechanistic model framework integrating physiologically based pharmacokinetic model (LeiCNS‐PK3.0), binding kinetics, and receptor dynamics, to predict regional human central nervous system (CNS) receptor occupancy (RO), using brain extracellular fluid (brain ECF ) concentrations. Positron emission tomography (PET)‐RO is a non‐invasive biomarker to estimate, and validate human CNS RO. However, for highly lipophilic drug buprenorphine brain ECF concentrations‐based RO predictions deviated substantially from PET‐RO. This mismatch and the lipophilicity driven brain cell membrane (brain M ) accumulation motivated us to investigate relevance of brain ECF vs. brain M concentrations for PET‐RO. We identified human studies reporting plasma PK and CNS PET‐RO for D2 receptor ligands spanning lipophilicities. Population PK parameters, derived via model fitting, served as inputs to the mechanistic framework to simulate brain ECF and brain M ‐based regional RO and compared with PET‐RO. We systematically assessed the impact of lipophilicity, receptor dynamics, endogenous ligands, and schizophrenia‐related physiological changes on simulations accuracy. Brain M concentrations showed better agreement with PET‐RO than brain ECF for five of eight D2 ligands. However, neither improvement nor brain M partitioning showed a clear trend with logP; substitution yielded modest improvement. Receptor internalization reduced simulated RO and enhanced PET agreement (remoxipride/haloperidol), independent of lipophilicity. Endogenous dopamine and schizophrenia‐related changes had negligible effects. This study demonstrated that brain M concentrations offer improved PET‐RO alignment for a subset of compounds, independent of lipophilicity suggesting role of other drug‐specific factors in brain M accumulation or lack thereof. The integrated mechanistic framework offers potential to enhance CNS drug treatment, subject to further development via qMRI, hMRI, MRS, and controlled clinical studies.