Physiologically-based pharmacokinetic (PBPK) modeling, combined with in vitro and in vivo extrapolation (IVIVE) approaches, have shown a major impact in drug development, with applications spanning from early lead identification to late clinical development. PBPK modeling also has the potential to accelerate small molecule (SM) drug discovery, reducing the need for in vivo animal studies and optimizing design cycle times. Building PBPK models in the early discovery space, however, is time consuming due to cumbersome data gathering, complex model interfaces, the need for learn and confirm cycles and limited in vitro data availability. Due to these factors, simpler yet limited approaches dominate the early discovery space (e.g., correlations, assumption-rich equations, etc.).
High throughput (HT)-PBPK modeling approaches, on the other hand, allow fast and seamless PBPK simulations leveraging the power of Machine Learning to fill in gaps in measured compound properties, thereby bringing PBPK-based ADME insights into the early discovery space. With HT-PBPK, teams can predict PK, PD and ADME properties based only on sparse in vitro data and in silico predicted inputs. This presentation will highlight the use of HT-PBPK modeling in early discovery projects and the implementation of this technology as part of the toolbox for early drug design and optimization.
Andrés Olivares-Morales, Ph.D.
DMPK/PD Project Leader & Translational Modeling and Simulation (Principal) Scientist