Background: Variability in exposure and response to cancer medicines contributes to differences in treatment efficacy and toxicity. A precision-medicine framework that integrates biomarkers of drug disposition with pharmacodynamic determinants of response may better predict who will achieve effective drug exposure, who will respond, and who is at greatest risk of toxicity.
Methods: Across a program of translational studies, tissue-derived extracellular vesicles (EVs) are used as non-invasive liquid biopsies of otherwise inaccessible tissues. Drug-metabolising enzymes and transport pathways are quantified to characterise determinants of pharmacokinetic exposure, while drug targets and downstream pharmacodynamic pathways are assessed to characterise biological sensitivity and treatment response. These dynamic molecular phenotypes are integrated with genotype, pharmacokinetic data and clinical measures of efficacy and toxicity. Physiologically based pharmacokinetic modelling and AI-enabled multivariable approaches are used to translate these data into individualised predictions of exposure and response.
Results: Across exemplar translational studies, tissue-derived EV molecular profiles reflected corresponding tissue biology and captured clinically relevant interindividual variation in pathways governing medicine disposition and response. EV-derived biomarkers identified dynamic changes in drug-metabolising capacity associated with physiological and pathological states, and incorporation of these molecular phenotypes into pharmacokinetic models translated this variability into predicted differences in medicine clearance and systemic exposure. Collectively, these studies provide proof-of-concept that EVs can provide a non-invasive, dynamic measure of functional phenotype beyond genotype alone. This platform is now being extended to pharmacodynamic targets and response pathways, enabling integration of determinants of both medicine exposure and biological response.
Conclusions: Integrating dynamic biomarkers of drug exposure with pharmacodynamic measures of drug target engagement and biological response offers a pathway beyond conventional genotype- or dose-based precision medicine. Combining EV-derived phenotypes, pharmacokinetics, pharmacodynamics, genomics and clinical outcomes may enable models that predict both the dose a patient needs and the likelihood that their cancer will respond.