Identification of a pharmacodynamic (PD) biomarker, which is predictive of the efficacy outcome, is of ultimate interest in drug development. The objectives of the current analyses are to develop the pharmacokinetic (PK)/PD model for biomarkers (thymidine kinase 1 [TK1] in serum and phosphor-retinoblastoma protein [pRb] and Ki67 in skin tissues) related to cyclin-dependent kinase (CDK) 4/6 inhibition by palbociclib and to explore the relationship of the biomarker response with the efficacy end point (progression-free survival). The data used for analysis consisted of extensive sampling of palbociclib PK and longitudinal rich sampling for the PD biomarkers TK1, pRb, and Ki67 in 26 patients. A 2-compartment model was used to describe the PK of palbociclib. A precursor-dependent indirect response PD model was developed to describe the pRb time course, whereas a similar PD model with an additional transit compartment to model the delayed effect on Ki67 and TK1 response was used to describe the Ki67 and TK1 time course. Palbociclib effect on biomarkers was modeled as a maximum inhibition model. A Cox proportional hazard model was used to assess the relationship of progression-free survival with the biomarker response. The PK/PD models adequately described the observed PK of palbociclib and the longitudinal change of pRb, Ki67, and TK1. Palbociclib exposure significantly correlated with the reduction of all 3 biomarkers, and the estimated concentration to achieve 50% inhibition of the synthesis rate values were 45.2, 42.4, 50.2 ng/mL, respectively, for pRb, Ki67, and TK1. The exploratory biomarker-response analyses showed that a longer PFS was associated with lower baseline TK1 and simulated minimum TK1. Such results may warrant further confirmation from future large-scale study. Clinical Trial Registration: NCT02499146.
Keywords: CDK4/6; modeling, PFS, pharmacodynamics, thymidine kinase.
© 2021, The American College of Clinical Pharmacology.