Pavel Balazki (1), Felix Mil (1), Rudolf Engelke (1), Anastasiia Kostiv (1), Stephan Schaller (1)
ESQlabs GmbH, Saterland, Germany
Objectives:
The increasing importance of Physiologically-Based Pharmacokinetic (PBPK) and Quantitative Systems Pharmacology (QSP) models for regulatory submissions in drug development necessitates user-friendly and quality-controlled tools and workflows. Our goal is to develop a robust framework in R that facilitates PBPK/QSP modeling and simulation (M&S) projects with models developed with the open-source Open Systems Pharmacology (OSP) (1) modeling software, promoting reproducibility, transparency, and automation in model development while minimizing error probability coming from maintaining complex R code.
Methods:
An open-source R package {esqlabsR} (2) has been developed that utilizes R packages from the OSP ecosystem: ospsuite-r, TLF, and parameter.identification. The Quarto framework is used for generation of Markdown/PDF reports, and R Shiny package for the graphical user interface (GUI). All packages adhere to high transparency and quality control standards in software development and are availble open-source.
Results:
The framework offers a user-friendly workflow for working with models developed with PK-Sim and MoBi. Key supported steps include:
- Model development in PK-Sim and MoBi
- Modifying species, individual characteristics, and populations in R
- Configuring simulation scenarios with species/individual/disease-specific parameters and application protocols
- Performing sensitivity analysis
- Running parameter identification routines
- Generating markdown and PDF reports
All workflows are built around the definition of simulation scenarios from a single simulation file. This approach ensures equal and up-to-date model structure for all scenarios and minimizes the risk reporting results for an outdated model version. The scenarios are defined as a set of parameter values applied to the model structure, whereby different hierarchy levels of parametrization are available:i) global parametrization, ii) species-specific, iii) individual-specific, iv) population-specific, v) disease state, vi) application protocol. The final parameter set is automatically supplied with the generated simulation report, ensuring full transparency and reproducibility.
The specified scenarios can be re-used in the subsequent steps of performing sensitivity analyses, parameter estimations, and generation of statistical and graphical results analyses.
The different workflow steps are configured in Excel files, ensuring project reproducibility, minimizing coding errors, and facilitating use by modelers with limited coding experience. A R Shiny application provides a convenient GUI for the execution of the entire workflow and generation of markdown report templates.
Conclusions:
We present a novel framework in R that simplifies user interaction with PBPK/QSP models developed in OSP software. This framework prioritizes quality control and user-friendliness, facilitating the creation of regulatory-compliant reports for pharmaceutical submissions. The {esqlabsR} package and the R Shiny application are freely available on GitHub, and users are encouraged to share feedback and participate in its development.
In the future, more workflows will be supported by the framework, such as pre-defined workflows for drug-drug-interaction simulations, pediatric scaling, or high-throughput PBPK modeling. The GUI app will be continuously developed to offer more model analyses functionalities and to allow generation of new model structure from pre-defined model blocks.
References:
[1] Open Systems Pharmacology Community. Open Systems Pharmacology [Internet]. 2018 [cited 2018 Feb 4]. Available from: www.open-systems-pharmacology.org
[2] esqLABS GmbH. esqlabsR utilities package [Internet]. [cited 2024 Mar 11]. Available from: https://esqlabs.github.io/esqlabsR/index.html
Reference: PAGE 32 (2024) Abstr 11027 [www.page-meeting.org/?abstract=11027]
Poster: Methodology - Other topics