Seoghwa Song1
1Chungnam National University
Objective Biological drugs have a crucial role in the treatment of various diseases, yet their pharmacokinetics (PK) has been considered highly complex due to interactions with specific biological targets. The Target-Mediated Drug Disposition (TMDD) model effectively has been captured these drug-target interactions, accounting for nonlinear PK and dose-dependent elimination kinetics. In contrast, conventional compartment models have provided advantages with simplified equations and high prediction performances, even though limited reflection of PK characteristics of biologics. Since comparison of characteristics of models should be essential for selecting the most predictive approach for dose optimization, particularly in ensuring long-term drug accumulation, efficacy, and safety, the main aim of this study was to evaluate the TMDD model and compartment models to refine clinical dosing strategies and improve therapeutic outcomes. Methods This study was conducted in four stages which consisted of steps for validation of observational data, PK model development and evaluation, model-based simulations, and analysis of simulation outcomes. Data from five real biologics (Anakinra (Drug A), recombinant human interleukin-7 hybrid Fc (Drug B), human interleukin-1 receptor antagonist hybrid Fc (Drug C), fully humanized recombinant anti-PD-L1 monoclonal antibody (Drug D), and PEGylated filgrastim (Drug E)) were analyzed under single-dose and multiple-dose conditions. Dose proportionality assessments were performed to determine PK linearity. Conventional compartment models, modified compartment models incorporating Michaelis-Menten kinetics, and TMDD models were developed. Predictive performance was assessed using the corrected Akaike Information Criterion (AICc) and Objective Function Value (OFV). Visual evaluations were performed through Visual Predictive Check (VPC) and Goodness-of-Fit (GOF) plots. Model validation was carried out in consequence of superposition-based Non-Compartmental Analysis (NCA) of observational data to calculate PK parameters and internal validation using training/testing split datasets to evaluate model performance. Lastly, Model-based simulations were conducted to estimate long-term PK exposure metrics, including AUCt, Cmax, and Ctrough. Software tools used included NONMEM 7.5.0, Berkeley Madonna 10.6.1, PSN 5.3.1, Pirana 3.0.0, and R 4.3.3 Results According to dose proportionality analysis based on observation, Drugs A and D were confirmed by following linear PK, whereas Drugs B, C, and E were displayed nonlinear PK kinetics. Based on those reference phenomenon, the conventional compartment and TMDD models yielded comparable AUCt estimates for Drug A and D having linear PK drugs. however, TMDD model was shown best prediction performance in cases of Drug B, C and E as non-linear PK drugs. While modified compartment model incorporating Michelis Menten kinetics were generated by significantly different estimates for all cases of drugs. In addition, TMDD models obviously provided superior predictive performance for drugs having nonlinear PK and effectively captured nonlinearity in higher doses, exhibiting the largest deviations from Superposition NCA predictions what it calculated by linear PK assumption. These findings indicated that model performances would be dependent on PK linearity of drugs. Conventional compartment models performed well for biologics having linear kinetics in dose range, whereas TMDD models should be more suitable for biologics with nonlinear PK, reflecting target-mediated clearance mechanisms. Conclusion The development and evaluation of conventional compartment, modified compartment models, and the TMDD model were conducted using five biological drugs having linear or non-linear PK characteristics. For drugs exhibiting linearity in the data, the conventional compartment model provided sufficient predictive performance, while for drugs showing non-linearity, the TMDD model demonstrated higher predictive performance. Taking account into linearity of PK could be changed in dose range and its characteristic, it would be suggested that confirmation of the dose linearity should be dominant step to figure out model having best model performances in biologics durg as well.
[1] L. Gibiansky et al., [Target-mediated drug disposition model: approximations, identifiability of model parameters and applications to the population pharmacokinetic-pharmacodynamic modeling of biologics], Expert Opinion on Drug Metabolism & Toxicology, vol. 5 (7) (2009) [2] Guohua An, [Concept of Pharmacologic Target-Mediated Drug Disposition (TMDD) in Large-Molecule and Small-Molecule Compounds], J Clin Pharmacol, 60(2), 149-163 (2020). [3] DR Mould et al., [Basic Concepts in Population Modeling, Simulation, and Model-Based Drug Development-Part 2: Introduction to Pharmacokinetic Modeling Method], Pharmacometrics & Systems Pharmacology (2013).
Reference: PAGE 33 (2025) Abstr 11681 [www.page-meeting.org/?abstract=11681]
Poster: Methodology - Model Evaluation