Numerical Verification of Tucuxi, a Promising Bayesian Adaptation Tool for Model-Informed Precision Dosing.

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Computational Biology and Bioinformatics
• Clinical Pharmacology

Background:

• Model-Informed Precision Dosing (MIPD) software aids clinical decisions for therapeutic drug concentrations.
• Accurate prediction of pharmacokinetic parameters is crucial for optimizing drug therapy.

Purpose of the Study:

• To evaluate the predictive accuracy of Tucuxi, a novel MIPD software.
• To compare Tucuxi's predictions against NONMEM, a widely accepted Bayesian pharmacokinetic modeling tool.

Main Methods:

• Developed a panel of pharmacokinetic models simulating various administration routes and dosing regimens.
• Created virtual patient populations (n=4000) with simulated sparse and rich sampling profiles.
• Compared Tucuxi and NONMEM predictions for concentrations, Cmin, Cmax, and AUC0-24h using relative differences, MPE, and RMSE.
• Applied bioequivalence criteria to AUC0-24h, Cmin, and Cmax for further comparison.

Main Results:

• Tucuxi demonstrated high predictive accuracy, closely matching NONMEM predictions across all simulated scenarios.
• A median of 99.8% of predicted concentrations at sampling times had relative errors <0.1%.
• Mean Prediction Error (MPE) and relative Root Mean Square Error (RMSE) were minimal (0% and 0.82%, respectively).
• Bioequivalence criteria were met for all models, with median values of 100% for AUC0-24h, Cmin, and Cmax.

Conclusions:

• Tucuxi exhibits excellent predictive accuracy comparable to the gold-standard NONMEM.
• The software's reliability suggests strong potential for adoption in clinical practice to guide precision dosing.
• Tucuxi can effectively support achieving therapeutic concentration targets in patients.