The Drug Penetration Problem A drug must reach its target site - the tumor - at sufficient concentration to be effective. Measuring drug levels in blood plasma is straightforward, but plasma concentrations do not necessarily reflect what is happening inside tumor tissue. For EGFR-targeted therapies in lung cancer, the question of how much drug actually reaches mutated cancer cells in the lung is clinically critical.
Osimertinib and EGFR Osimertinib (Tagrisso) is a third-generation EGFR tyrosine kinase inhibitor (TKI) approved for NSCLC with EGFR mutations, including the T790M resistance mutation that causes failure of earlier-generation TKIs. Its standard dose is 80 mg daily. Understanding how osimertinib distributes within the body - particularly to lung tumor tissue - would help explain its clinical efficacy and guide dosing decisions.
PBPK Modeling Approach Van der Gaag and colleagues developed a whole-body physiologically based pharmacokinetic (PBPK) model for osimertinib that incorporates the drug's known binding kinetics to EGFR receptors in tumor cells. They then verified the model against real patient data obtained using PET imaging with a radiolabeled form of the drug.
Physiological Basis PBPK models represent the body as interconnected compartments representing major organs - lungs, liver, kidney, blood, muscle, fat, brain, and tumor - with drug movement between compartments governed by organ-specific blood flows, tissue volumes, and physiological parameters. This mechanistic structure allows predictions to be extrapolated across patient populations and dosing conditions.
EGFR Binding Kinetics A key innovation was incorporating nonlinear EGFR binding kinetics into the tumor compartment. Rather than treating drug-tissue binding as a simple equilibrium constant, the model explicitly represents the association and dissociation rates of osimertinib with mutant EGFR, capturing the drug's covalent-like irreversible binding characteristics.
PET Verification Data The model was verified against dynamic PET imaging data from 4 NSCLC patients who received a microdose of [11C]C-osimertinib (2.7 micrograms, 280 MBq). This radiolabeled tracer distributes identically to therapeutic osimertinib but at doses far below pharmacological effect, making it safe for research use while providing tissue-specific drug distribution measurements.
PET Verification Model predictions were compared to PET-measured drug concentrations in each organ at 60 minutes post-dose. Most organ predictions fell within a 2-fold range of observed values - an acceptable standard for PBPK model verification. The lung prediction at 60 minutes was slightly outside the 3-fold range, indicating an area for model refinement.
Plasma Pharmacokinetics The model predicted a maximum venous plasma concentration (Cmax,vein) of 201 nM compared to the observed 168 nM, a prediction error of 45.5%. The predicted half-life was 52.4 hours versus the observed 48 hours - a close match for a population-level PBPK prediction.
EGFR Saturation A clinically important prediction was that at steady-state therapeutic dosing (80 mg daily), EGFR receptors in the tumor are predicted to be 88% saturated on average (SD 5.9%). Near-complete target engagement at standard dose supports the clinical efficacy of osimertinib and suggests that dose reductions would substantially reduce target coverage.
Microdose Principle At microgram doses - thousands of times below the therapeutic milligram range - a drug does not achieve sufficient concentration to exert pharmacological effects. However, its distribution and elimination pharmacokinetics are governed by the same physiological mechanisms as at therapeutic doses, making microdose PET an ethically acceptable way to study drug distribution in patients.
Dynamic PET Imaging Dynamic PET acquires sequential images over time (rather than a single snapshot), allowing researchers to track how drug concentration changes in each organ minute by minute after injection. This temporal resolution is essential for fitting PBPK model parameters - a single time-point measurement would be insufficient to distinguish between competing distribution mechanisms.
Clinical Research Translation The combination of microdose PET verification with PBPK modeling represents a powerful paradigm for drug development: models can be built and refined using PET data from small numbers of patients and then scaled to predict therapeutic behavior across diverse patient populations without requiring invasive tissue sampling.
Personalized Dosing The PBPK model could be used to predict whether individual patients - particularly those with unusual body composition, renal/hepatic impairment, or drug-drug interactions affecting osimertinib metabolism - are achieving adequate tumor drug concentrations at standard doses. This could guide personalized dose adjustments.
Acquired Resistance Context Osimertinib resistance typically develops within 1-2 years of treatment, often through secondary EGFR mutations or bypass pathway activation. Suboptimal drug penetration to specific tumor compartments could contribute to the emergence of resistant subclones. The model could help explore whether higher doses in sanctuary sites (brain, specific metastases) would delay resistance.
Model Expansion Future development should expand the verified model to include the standard 80 mg therapeutic dose (beyond the microdose range verified here), tumor heterogeneity (variable EGFR mutation density), and combination therapy scenarios where osimertinib is paired with other agents that may affect its distribution.