Robert B. de Mooij, Parth J. Upadhyay, Raju Prasad Sharma, Wenyi Wang, Sieto Bosgra
Genmab
Objectives:
In drug development, accurate prediction of clinical pharmacokinetics (PK) and pharmacologically active doses is imperative to decision-making at all stages. Employing a physiologically based PK (PBPK) model facilitates the examination of drug distribution across various organs where pharmacological effects occur. Cao’s minimal PBPK model [1] has demonstrated utility, particularly in early preclinical phases when specific drug candidate data are limited. In oncology, integrating a tumor compartment enables the prediction of target engagement at the site of pharmacological activity. Various methods exist for characterizing tumor compartments, ranging in complexity. A conventional approach, a single compartment with a reflection component (1C), fails to address exposure gradients and the heterogeneous distribution of target or effector cells. This analysis quantitatively compares clinical dose projections based on this single tumor compartment model against two more complex alternatives: a three-compartment model incorporating stroma and tumor interstitial sub compartments (3C) [2], and a Krogh cylinder model (KC) [3] that incorporates more physiological parameters governing permeability and diffusion.
Methods:
The three tumor characterizations will be incorporated into the minimal PBPK model and evaluated across diverse scenarios, including:
- Monoclonal and bispecific antibody administration.
- Homogeneous distribution of the target or a target gradient reflecting a higher density of effector cells closer to the vasculature.
- Various metrics for efficacy, including achieving 30% or 95% receptor occupancy (RO) at peak or trough, either averaged across the entire tumor or as a minimum within sub-compartments, and the peak trimers per effector cell for bispecific antibodies.
A sensitivity analysis was conducted to assess the influence of tumor perfusion/reflection, target density, effector cell counts and distribution and target expression outside of the tumor on dose projections.
Results:
For scenarios considering average target engagement across the tumor, the complex 3C and KC models generally projected doses within twice that of the 1C model. The largest discrepancy was noted when a minimum peak RO across the tumor was required. A homogeneous distribution of the target resulted in an anticipated dose to achieve a 30% peak RO in the 3C model that was up to 4.3-fold higher than in the 1C model under the conditions evaluated. The heterogeneous distribution partially mitigated this, leading to a 2.7-fold higher prediction compared to the 1C model. The discrepancies between the models appeared to be sensitive to alterations in tumor permeability, intercompartmental flows and target-mediated drug disposition (TMDD) characteristics.
Conclusions:
The findings propose that the dose estimations from the more complex 3C and KC models are largely consistent with those from the simplistic 1C model in most instances. Nevertheless, under certain conditions, the selection of the tumor model could significantly influence the projections of pharmacologically active doses, primarily conditional on the chosen efficacy metric.
References:
[1] Cao, Y., Balthasar, J. P., & Jusko, W. J. (2013). Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies. Journal of Pharmacokinetics and Pharmacodynamics, 40(5), 597–607. https://doi.org/10.1007/S10928-013-9332-2
[2] Wang, J., Chen, T., Ruszaj, D. M., Mager, D. E., & Straubinger, R. M. (2024). Integrated PK/PD Modeling Relates Smoothened Inhibitor Biomarkers to The Heterogeneous Intratumor Disposition of Cetuximab in Pancreatic Cancer Tumor Models. Journal of Pharmaceutical Sciences, 113(1), 72–84. https://doi.org/10.1016/J.XPHS.2023.10.019
[3] Chang, H. P., Le, H. K., & Shah, D. K. (2023). Pharmacokinetics and Pharmacodynamics of Antibody-Drug Conjugates Administered via Subcutaneous and Intratumoral Routes. Pharmaceutics, 15(4). https://doi.org/10.3390/PHARMACEUTICS15041132
Reference: PAGE 32 (2024) Abstr 10966 [www.page-meeting.org/?abstract=10966]
Poster: Methodology - Model Evaluation