Mechanistic modelling identifies and addresses the risks of empiric concentration-guided sorafenib dosing

The primary objective of this study is to evaluate the capacity of concentration-guided sorafenib dosing protocols to increase the proportion of patients that achieve a sorafenib maximal concentration (C_max) within the range 4.78 to 5.78 µg/mL. A full physiologically based pharmacoki-netic model was built and validated using Simcyp® (version 19.1). The model was used to simulate sorafenib exposure in 1000 Sim-Cancer subjects over 14 days. The capacity of concentration-guided sorafenib dose adjustment, with/without model-informed dose selection (MIDS), to achieve a so-rafenib C_max within the range 4.78 to 5.78 µg/mL was evaluated in 500 Sim-Cancer subjects. A multivariable linear regression model incorporating hepatic cytochrome P450 (CYP) 3A4 abundance, albumin concentration, body mass index, body surface area, sex and weight provided robust prediction of steady-state sorafenib C_max (R² = 0.883; p < 0.001). These covariates identified subjects at risk of failing to achieve a sorafenib C_max ≥ 4.78 µg/mL with 95.0% specificity and 95.2% sensitivity. Concentration-guided sorafenib dosing with MIDS achieved a sorafenib C_max within the range 4.78 to 5.78 µg/mL for 38 of 52 patients who failed to achieve a C_max ≥ 4.78 µg/mL with standard dosing. In a simulation setting, concentration-guided dosing with MIDS was the quickest and most effective approach to achieve a sorafenib C_max within a designated range.