TITE-PK

TITE-PK implements a Bayesian time-to-event pharmacokinetic model to integrate safety data across sequential phase I dose-escalation trials with multiple dosing schedules.

Key Features:

• Sequential trial design: Tailored for sequential phase I trials exploring different dosing schedules (e.g., daily or weekly) and leverages data from completed and ongoing escalation phases to inform dose decisions.
• Time-to-event pharmacokinetic modeling: Models time-to-event toxicity outcomes within a PK framework to account for timing of events and drug exposure.
• Bayesian framework: Integrates prior knowledge and continuously updates posterior probabilities as new patient-level data accrue.
• Pharmacokinetics integration: Uses PK principles to combine and harmonize information across multiple dosing schedules and exposures.
• Comparative performance: In simulation studies, outperformed the bridging continual reassessment method and Bayesian logistic regression models with meta-analytic priors in recommending safe doses while minimizing overly toxic recommendations across most scenarios.
• Efficiency: Achieves improved operating characteristics without requiring substantially larger patient numbers.

Scientific Applications:

• Early-phase dose-escalation trials: Integrates safety and PK data across multiple dosing schedules to inform dose selection in phase I trials.
• Safety assessment across regimens: Supports selection of dose levels that balance toxicity risk across different dosing schedules.

Methodology:

Bayesian time-to-event pharmacokinetic modeling that synthesizes data across schedules using PK principles, incorporates prior information with continuous updating, and is evaluated via simulation studies comparing performance to the bridging continual reassessment method and Bayesian logistic regression with meta-analytic priors; implemented in R and Stan.