(W0930-05-31) User-Friendly Web Applications for Pharmacokinetic Simulations and Predictions

Purpose: The purpose of the presented work is to develop user-friendly web applications that integrate and simplify pharmacokinetic (PK) analysis processes. These tools aim to reduce the need for complex modeling software and lower the barrier for non-experts in obtaining pharmacokinetic analysis. They significantly reduce the time required to complete these analyses, enabling DMPK scientists to more quickly inform program teams of the optimal study designs and compound assessments. Ultimately, this can potentially accelerate candidate drug identification.

Methods: Oral PK Dose Selection Tool: A tool was developed to predict the PK profile using one or two compartment PK models. This tool allows users to select doses to achieve targeted concentrations (Cmax, Cmin, or Cavg) based on the predicted PK profile. Optional inputs include potency and fraction unbound. The tool was built under the R Shiny framework. Human Clearance Prediction Tool: This tool incorporates multiple methods of human clearance prediction, including allometry scaling (multiple and single species) and IVIVC (In vitro - In vivo Correlation) using microsomes and hepatocytes. The rule of exponent for allometry is used to recommend an interspecies scaling approach. The tool was built under the R Shiny framework.

Results: For the Oral PK Dose Selection Tool, it allowed for dose prediction to achieve specified targeted exposure and visualized the predicted PK profile in context with potency. The tool also included an additional simulation option for dose adjustment with real-time updates of PK parameters.
For the Human Clearance Prediction Tool, it provided multiple methods for predicting human clearance in a single tool for easy comparison. Outputs include IVIVC using microsomes and hepatocytes and allometry scaling results, and all results are complemented with interactive visualizations. For demonstrated input, the tool recommended allometry scaling with MLP correction based on the rule of exponent for allometry. Single species allometry scaling was also provided as an option.

Conclusion: User-friendly web applications for pharmacokinetic simulations and predictions were successfully built to support quick PK simulation and prediction. The human clearance prediction tool integrates multiple prediction methods (IVIVC using microsomes and hepatocytes, and allometry scaling) in a single platform that allowed for quick prioritization of compounds for further characterization during lead optimization. The oral PK dose selection tool was effectively used for dose selection in toxicity and efficacy studies during drug discovery programs.

References: 1.Drug Metab Dispos,2008 Jul;36(7):1194-7
2. Xenobiotica , Volume 26, 1996 - Issue 9, Pages 887-895

Acknowledgements: The authors would like to express their sincere gratitude to colleagues in Pharmaceutical Candidate Optimization.


Figure 1. For the Oral PK Dose Selection Tool, The predicted time-concentration curve is shown along with potency to visualize the coverage of free potency at the predicted dose. PK parameters, potency, and projected dose are displayed in table.


Figure2 . Human Clearance Prediction Tool, a) The IVIVC interactive visualizer for human hepatic blood clearance with fu,b and fu,mic correction. b) Regression visualization of allometry scaling with MLP correction c) Predicted human plasma clearance value from allometry scaling. Recommended method and value based on the rule of exponent for allometry are highlighted by the tool.