Axitinib is a second-generation small-molecule vascular endothelial growth factor receptor inhibitor. An axitinib steady-state area under the plasma concentration-time curve (AUCSS ) >300 ng/mL/hr is associated with superior progression-free and overall survival. This study sought to characterize the physiological and molecular characteristics driving variability in axitinib AUCSS using physiologically based pharmacokinetic modeling to identify exposure biomarkers for this drug. The capacity to predict subjects likely to fail to achieve an axitinib AUCSS >300 ng/mL/hr was evaluated as a secondary outcome. A full physiologically based pharmacokinetic model incorporating mechanistic absorption was developed and verified for axitinib in accordance with the US Food and Drug Administration Guidance using Simcyp (Version 17.1). This model was used to simulate axitinib exposure over 7 days with twice-daily dosing (5 mg) in a cohort of 1000 virtual cancer patients. Multiple linear regression modeling was used to identify patient characteristics associated with differences in axitinib exposure. A multivariable linear regression model incorporating hepatic cytochrome P450 (CYP) 3A4 abundance, albumin concentration, hepatic CYP1A2 abundance, hepatic CYP2C19 abundance, and intestinal CYP2C19 abundance provided robust prediction of axitinib AUCSS (R2 = 0.890; P < .001). By accounting for these variables, it was possible to identify subjects who would fail to achieve an effective axitinib AUCSS with a specificity of 88.7% and a sensitivity of 92.6%. Variability in axitinib AUCSS is primarily driven by differences in hepatic CYP3A4 abundance and albumin concentration. Consideration of these 2 characteristic is likely to be sufficient to individualize axitinib dosing.
Keywords: axitinib; between-subject variability; biomarkers; physiologically based pharmacokinetic modeling; precision dosing.
© 2019, The American College of Clinical Pharmacology.