Model-based sequential human PET study design for Optimal PK/RO assessment
Marcella Petrone, Stefano Zamuner, Vincenzo Di Iorio, Roger Gunn, Marc Laruelle and Roberto Gomeni
Clinical Pharmacokinetic Modelling and Simulation- GlaxoSmithKline
Objective: Positron emission tomography (PET) is a non invasive imaging techniques to investigate ligand-receptor binding in the living brain and to assess the plasma concentration/receptor occupancy (PK/RO) time course. Although information are available from preclinical species regarding distribution of receptors into target organ, affinity of labelled compound to receptor (kon/koff), a conspicuous level of uncertainty is present when translating these information from animal to human to allow the study design of a PET experiment to volunteers.
The purpose of this work is to demonstrate the added value of an adaptive-optimal PK/RO study, with respect to traditional PET study design in which scans are mostly scheduled at the time of maximum plasma concentration and at 24 hours post dose.
Methods: A traditional PET study design consisted in 12-16 volunteers dosed at a maximum of 3 dose levels and PET scans recorded at approximately the time of Cmax and 24hours post dose.
The model-based stepwise PET study design consisted of few sequential cohorts of 4 subjects each. The PET scan of the first cohort were scheduled based on optimal study design using information available on kon and koff estimates from preclinical RO studies and on population PK information derived from FTIH studies.
A simulation approach was used to assess the advantage of adaptive design when human kon and koff are consistently different from the initial preclinical values.
In each subsequent simulated cohort an optimal design was applied based on information gathered from previous cohorts. PET scan times were adjusted at each step to allow the most informative assessment of PK/RO relationship.
Optimisations for adaptive-optimal design PET study were obtained using WinPOPT® 1.1 software. Comparison between traditional and stepwise-optimal design were made on D-optimality criterion.
Results: Simulations show that a model-based stepwise-optimal approach to design a PET study is endowed with added value respect to traditional PET study design. The use of the proposed sequential model design increased the efficiency of the study for the PK/RO assessment especially when the human receptor kinetic model cannot be properly inferred from animal estimates.
Conclusion: Model-based sequential or stepwise-optimal study design allows a more rationale and data-driven decision making, especially in those studies were intensive sampling strategies are not allowed for operational or ethical reasons.