Developing a shiny app to expedite the simulation of PKPD profiles of mAbS and exploration study designs
Aida N Kawuma(1,2), Rohan M Benecke(3), Maxwell T Chirehwa(4)
(1)Division of Clinical Pharmacology, Department of Medicine, University of Cape Town, South Africa. (2)Infectious Disease Institute, College of Health Sciences, Makerere University, Kampala, Uganda. (3)Division of Clinical Pharmacology, Department of Medicine, Faculty of Medicine and Health Sciences, Stellenbosch University, South Africa. (4)Clinical Pharmacology Modelling and Simulation, GSK medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
Objectives:
Monoclonal antibodies (mAbs) are an essential and expanding class of therapeutic agents [1]. MAbs possess a high degree of target selectivity, with many exhibiting nonlinear distribution and elimination, influenced by binding to their target [1]. There is an increasing interest in using target engagement (TE) (defined here as percent change in free target from baseline) to inform dose selection for first-in-human and Phase 2 studies. Due to limited data from clinical studies to develop TMDD models, pharmacologists combine data from several sources to simulate likely PKPD profiles and determine percent target engagement associated with each dose level or dosing regimen. Tools like an R shiny application can save time, and financial resources, minimize programming errors and increase reproducibility. The objective of this exercise were to develop an R Shiny application that can be used to (1) simulate and visualize PKPD profiles of mAbs to support the design of PKPD studies and (2) explore possible range of initial estimates for PKPD modelling.
Methods:
A shiny application was developed in R software and simulations are performed using mrgsolve package [2]. The golem R package was used to guide application development and definition of the structure of the backend directories[3]. A moduler programming approach was implemented to allow reuse of application components and future expansion on the application. Plots were generated using ggplot2 package [4] on the application interface and ggplot package [4] upon generation of a pdf or html report.
Results:
The user interface (UI) is built to accept PK and PD parameters together and dosing regimens. The app supports both IV and extravascular dosing. The user is able to build and edit a data frame of dosing regimens. The application provides flexibility in select quantities to be plotted and summaries such as AUC, Cmax, Cmin, and %TE. The server side of the app generates the data frame with different regimes and performs the simulations. An option to generate a report has been included as well as downloading simulated data for further manipulation. Using the app, one can generate the following plots which include; PK, PD, and TE vs Dose plots. Simulations can be performed with or without variability and the application provides an Initial estimate of different model parameters.
Conclusions:
This app is a useful tool for modellers/ pharmacologists to support the design of clinical trials and to visualize the influence of different parameters on the PK or PD profiles following dosing of a monoclonal antibody. By comparing simulated PK data with experimental data, researchers can use with realistic initial parameter estimates during model development.
References:
[1] An, G. (2020), Concept of Pharmacologic Target-Mediated Drug Disposition in Large-Molecule and Small-Molecule Compounds. The Journal of Clinical Pharmacology, 60: 149-163. https://doi.org/10.1002/jcph.1545
[2] Baron K (2023). mrgsolve: Simulate from ODE-Based Models. R package version 1.0.9, https://github.com/metrumresearchgroup/mrgsolve.
[3] Fay C, Guyader V, Rochette S, Girard C (2023). golem: A Framework for Robust Shiny Applications. R package version 0.4.1, https://github.com/ThinkR-open/golem
[4] Wickham H (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org
