Limited Sampling and Dose Prediction Model for Cisplatin Using a Bayesian Approach

Abstract

A dose prediction method for cisplatin based on the Bayesian principle and with use of limited blood sampling is described. A reference population was used to establish plasma total platinum (Pt) pharmacokinetics. the limited sampling method, using two assays values, was validated with a second population. In the following step a cisplatin dosing method to achieve a target concentration of Pt was evaluated. This method was validated using the limited blood sampling in two circumstances: (i) during a first infusion of cisplatin, the estimated amount of cisplatin to be infused during the days four and five was 95.0 ± 12.8% of the actual amount (12 patients); (ii) during any other infusion, the estimated amount of cisplatin to be infused during the ensuing period was 104.8 ± 13.0% of the actual amount (17 patients). the variability of this method to predict individual pharmacokinetics and individual cisplatin dosage appeared acceptable for most clinical use.