Model-Based Individualized Treatment of Chemotherapeutics: Bayesian Population Modeling and Dose Optimization

Area of Science:

• Pharmacology
• Computational Biology
• Genetics

Background:

• 6-Mercaptopurine (6-MP) is crucial for treating pediatric cancers, autoimmune diseases, and inflammatory bowel disease.
• Individual responses to 6-MP vary significantly due to genetic differences in the thiopurine methyltransferase (TPMT) enzyme, impacting treatment efficacy.
• Current methods for optimizing 6-MP dosage and predicting patient response are insufficient.

Purpose of the Study:

• To develop a Bayesian population modeling approach for human 6-MP metabolism.
• To create a patient-specific pharmacological model for optimizing 6-MP dose scheduling.
• To integrate multi-level biological data for improved prediction of therapeutic targets.

Main Methods:

• Utilized a global sensitivity analysis and model reduction to minimize parameters in data-scarce clinical settings.
• Employed optimal experimental design (D-optimality) for accurate estimation of sensitive pharmacokinetic parameters.
• Implemented a model predictive control algorithm for optimizing 6-MP dose scheduling to maintain therapeutic 6-thioguanine nucleotide (6-TGN) levels.

Main Results:

• Developed a robust Bayesian model for 6-MP pharmacology.
• Successfully optimized dose scheduling to maintain 6-TGN within the therapeutic window.
• Demonstrated the critical role of integrating gene expression, enzyme phenotype, and drug phenotype data in reducing prediction uncertainty.

Conclusions:

• The developed model and control strategy enable individualized 6-MP dosing based on patient-specific drug metabolism.
• This approach moves beyond the standard-dose-for-all strategy, offering a path to more precise and effective patient care.
• The findings highlight the potential for clinical application in personalizing 6-MP therapy.