The value of target-site physiologically-based pharmacokinetics for high-affinity small molecules: a mechanistic foundation for precision dosing in oncology and hematology

Abstract

For decades, oncology dose selection has been guided by the maximum tolerated dose (MTD) and plasma pharmacokinetics (PK), reflecting assumptions appropriate for classical cytotoxic chemotherapies. However, the advent of high-affinity, targeted therapies, including kinase inhibitors, epigenetic modulators, and radioligands challenges this paradigm. These agents achieve robust target engagement at doses far below the MTD, and systemic plasma concentrations often fail to reflect pharmacologically relevant exposure at tumor or hematologic sites. Physiologically-based pharmacokinetic (PBPK) modeling, extended to incorporate target-site dynamics, offers a mechanistic framework linking dose, systemic exposure, and local pharmacology. By integrating tissue physiology, drug properties, and target interactions, target-site PBPK provides insights into heterogeneous tumor penetration, intracellular distribution, and variable target occupancy that plasma PK alone cannot capture. Clinical examples, such as PSMA-targeted radioligands and tyrosine kinase inhibitors, illustrate how these models can inform rational dose selection, optimize ligand design, and guide individualized therapy. As oncology moves toward mechanism-driven, biology-aligned development, target-site PBPK represents a pivotal tool for translating preclinical insights into patient-specific dosing strategies and for redefining the standard of precision pharmacology.

Introduction

For decades, dose selection in oncology was driven by the maximum tolerated dose (MTD) and plasma pharmacokinetics (PK). This framework suited classical cytotoxic chemotherapies as systemic exposure was closely linked to toxicity, while efficacy was often assumed to increase with dose within a certain therapeutic window. For example, higher exposures of alkylating therapies, are associated with increased tumor cell kill, but also with greater toxicity [1]. Importantly, this paradigm was not universally applicable, even among traditional therapies. Several agents, including methotrexate and asparaginase, already rely on exposure- or schedule-dependent pharmacology rather than simple dose escalation [2].

Today, however, oncology and hematology increasingly rely on high-affinity, highly selective agents, encompassing kinase inhibitors, epigenetic modulators, and targeted radioligands among them [3–5]. These modern therapeutics no longer conform to the old assumptions. For many targeted agents, escalation to the MTD is biologically inappropriate, as effective target engagement and downstream pathway inhibition are often achieved at doses well below those associated with dose-limiting toxicity. These agents are engineered to bind specific molecular targets with enhanced potency and frequently with slow dissociation rates. Consequently, pharmacological activity depends far more on achieving sufficient target-site exposure than on maintaining elevated plasma concentrations.

Because the pharmacologically relevant concentration resides at the tumor or hematologic niche, often intracellular, stromal, or otherwise protected, plasma profiles provide limited insight into real therapeutic exposure. While plasma PK remains highly valuable in clinical practice, particularly as a surrogate for systemic exposure and a predictor of off-target toxicity, it does not necessarily reflect pharmacological relevant concentrations at the site of action. Recognizing this gap, the FDA’s Project OPTIMUS advocates for a paradigm shift toward dose optimization based on pharmacological activity, exposure-response relationships, safety and efficacy, rather than relying primarily on toxicity. It urges drug developers to evaluate multiple dose levels early in development and identify doses that achieve adequate target engagement early rather than defaulting to the MTD [6]. This shift requires mechanistic insight that quantify spatial drug distribution and target engagement. These data capture intra-tumoral variability and enable model-informed optimization of dose and scheduling, particularly for drugs with narrow therapeutic windows, uneven tumor penetration, or complex intracellular kinetics [5, 6].

Why plasma PK and conventional ADME are insufficient

Traditional absorption, distribution, metabolism and elimination (ADME) assessments focus on plasma concentrations and their decline over time, but for high-affinity small molecules these measurements often do not reflect what happens within tumors or hematologic tissues. Solid tumors exhibit striking heterogeneity in perfusion, vascular permeability, stromal density, transporter expression, and pH [7, 8]. Even within a single patient, lesions may receive markedly different drug levels despite identical plasma exposure, yet these discrepancies remain invisible in conventional PK curves [9, 10].

In hematological malignancies, malignant cells reside in bone marrow, lymph nodes, spleen, and occasionally the central nervous system, with microenvironments that can influence drug penetration, regardless of circulating levels. Increasing the dose may, therefore, intensify systemic toxicity without improving exposure at these protected sites. Complicating matters further, many high-affinity small molecules act intracellularly, where physicochemical properties dictate transport, sequestration may prolong local retention, and target occupancy depends on local, not systemic, concentrations. Plasma PK does not always capture these dynamics [11, 12].

The contribution of PBPK and target-site PBPK

Physiologically-based pharmacokinetic (PBPK) modeling provides a structured, mechanistic framework that links dose and systemic exposure to drug concentrations within tissues and cells by integrating physiological parameters with drug-specific properties. When extended to include local tissue processes and explicit target interactions, PBPK evolves into target-site PBPK, designed to simulate concentrations where pharmacology actually occurs, providing a more informative basis for PK-efficacy analyses than plasma-based approaches alone.

In contrast to conventional PBPK models, which often rely on tissue partition coefficient representing average organ-level distribution, we propose that target-site PBPK explicitly incorporates spatial, cellular, and binding-level processes. This includes receptor binding kinetics, intracellular sequestration, and local microenvironmental factors, thereby enabling the exploration of heterogeneity within tissues rather than assuming homogeneous distribution. Target-site PBPK integrates physiological parameters (e.g., organ volumes, blood flows, transporter expression and activity, tissue composition) with drug-specific properties (e.g., protein binding, lipophilicity, ionization, metabolism and transporter affinity and target binding) to mechanistically describe tissue penetration, intracellular distribution and binding kinetics (Fig. 1). This framework reveals the causes of successful or inadequate target occupancy, insights that plasma PK cannot reveal.

Fig. 1.

Determinants and measurement of target-site PK integrated into PBPK for individualized dosing. Physicochemical properties, imaging data, pharmacodynamic biomarkers, and tumor biopsy–derived measurements feed into target-site PBPK models. These mechanistic models integrate processes such as receptor binding, protein binding, intracellular trafficking, and lysosomal trapping to simulate tissue- and cell-level drug exposure. The resulting predictions support translation to individualized dosing strategies that account for patient-specific variability in target-site pharmacokinetics

Clinical applications: PSMA radioligands and tyrosine kinase inhibitors as representative

Prostate-specific membrane antigen (PSMA)-targeted small molecules illustrate how predictable plasma PK can coexist with highly variable tumor uptake [13–15]. Off-target uptake in kidneys and salivary glands reflect physiological PSMA expression, shaping both biodistribution and toxicity. Target-site PBPK integrates physiological parameters to predict tumor-to-organ ratios, identifies underexposed lesions, and informs ligand optimization through adjustments in linker chemistry, or affinity. Because radioligands are imaged by positron-emission tomography (PET) for diagnosis and treatment purposes, they offer a unique opportunity to validate and refine PBPK predictions directly, at a macroscopic (tissue) level. When appropriately validated, these diagnostic PET-image based models may enable patient-specific therapy optimization.

Furthermore, tyrosine kinase inhibitors (TKIs) provide another example of variable target site binding. Although structurally small, TKIs may vary dramatically in tissue penetration, due to differences in protein binding, transporter interactions, intracellular dynamics and target affinity. Many possess extremely high protein binding, leaving only a small free fraction available to enter tissues; others undergo pH-dependent lysosomal trapping or face efflux across key barriers, such as the blood-brain barrier [12, 16, 17]. As a result, similar plasma concentrations between patients can yield very different target-site exposures. Moreover, even when target-site concentration are similar, differences in target affinity, binding kinetics, and downstream signaling mean that equivalent exposure does not guarantee equivalent efficacy. Target-site PBPK offers mechanistic explanations for such discrepancies and can support dose decisions in context of ranging from inadequate central nervous system penetration to altered α1-acid glycoprotein levels or perfusion deficits induced by prior therapy [16, 17].

Microscopic and macroscopic evaluation of target-site PK for individualized dosing

Target-site PK can be evaluated at both macroscopic and microscopic scales, each contributing complementary insights for individualized dosing. At the macroscopic level, PET imaging of radioligands provide quantitative measurements of tumor uptake across all lesions. These data allow calibration of PBPK parameters, characterization of lesion-level heterogeneity, and identification of underexposed tumors. PET-based PK assessment therefore provides a practical framework to validate and refine target-site PBPK predictions in individual patients.

Microscopic methods provide spatial and cellular insight that elucidate heterogeneity in tissue penetration and target engagement. Although microscopic PK data have not yet been routinely integrated into PBPK models, several studies demonstrate that such integration is technically feasible [18–21]. For example, a correlation between PET uptake and microscopic PSMA expression patterns within PSMA expressing tumors are investigated, illustrating alignment between macro- and micro-scale pharmacology [22]. For TKIs, mass spectrometry imaging has revealed heterogeneous drug distributions at a cellular level, demonstrating substantial spatial variability that cannot be inferred from plasma PK alone [21, 23, 24].

Quantitative studies have demonstrated substantial intra-tumoral variability in drug exposure, with reported coefficients of variation exceeding 50–100% within a single lesion depending on the compound and tumor type [19, 20, 24]. Such variability had been linked to differences in vascularization, stromal barriers, and cellular density, underscoring the limitations of average tissue concentrations, and may contribute to incomplete responses and the emerge or drug resistance [9, 21, 25–29]. We therefore propose to incorporate variable drug distributions, rather than single mean values, for tissue exposure within PBPK frameworks, enabling simulation of the fraction of tumor regions achieving adequate drug levels under different dosing strategies [25–27].

Together, macroscopic imaging and emerging microscopic PK techniques provide a more complete view of target-site exposure. Incorporating these datasets—whether PET-derived uptake patterns, biopsy quantification, or spatial expression mapping—enhances PBPK model accuracy and supports individualized dosing based on organ function, tumor phenotype, transporter expression, and predicted underexposure at critical sites.

Translational relevance

Target-site PBPK is emerging as a key bridge between preclinical discovery and clinical decision-making. It supports rational dose selection, mechanism-informed ligand optimization, strategic biomarker and imaging development, and evidence-based personalization of therapy. Its importance is especially evident for high-affinity small molecules, whose success depends on target-site exposure rather than plasma levels alone. Tumor heterogeneity, transporter activity, protein binding, and microenvironmental factors can all uncouple systemic exposure from local pharmacology. Target-site PBPK brings these determinants together in a coherent, mechanistic framework that highlights barriers to drug action and guides rational dosing and personalized therapy.

Despite its promise, the clinical implementation of target-site PBPK for precision dosing remains challenging. Direct measurement of target-site concentrations is often not feasible in routine care and substantial intra- and inter-patient heterogeneity further complicates interpretation. A practical path forward lies in embedding non-invasive imaging with model-informed precision dosing frameworks, enabling target-site PBPK models to integrate plasma PK, imaging-derived biomarkers, and patient-specific covariates. In early-phase development, inertly labeled PET tracers can provide quantitative, non-invasive measurements of drug distribution and target engagement at microdose levels, enabling direct characterization of target-site pharmacokinetics. Excised tumor tissues provide an opportunity for microscopic PK assessment. In this framework, plasma concentrations anchor systemic exposure, while (PET) imaging biomarker data can inform on local tissue distribution patterns, allowing Bayesian updating of individual model parameters to guide and individualize dosing.

Such approaches could enable individualized dose selection aimed at achieving sufficient target-site exposure while minimizing toxicity, even when direct measurement is not possible. However, prospective validation is required to demonstrate that model-informed strategies improve clinical outcomes compared with standard dosing approaches. As oncology drug development shifts toward pharmacology-driven dose optimization, target-site PBPK provides a framework to mechanistically link systemic exposure to local drug action.

Author contributions

All authors contributed to the conception and design. The first draft of the manuscript was written by Suzanne van der Gaag and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interest

The authors have no relevant financial or non-financial interests to disclose.