PI-010 - LIMITED SAMPLING STRATEGIES FOR NONCOMPARTMENTAL ANALYSES OF ENCORAFENIB PHARMACOKINETIC DATA

T. Lewis¹, E. Hahn², L. Wollenberg³, M. Reddy³, J. Piscitelli⁴; ¹University of California San Diego, San Diego, CA, USA, ²Pfizer Inc., Boulder, CO, USA, ³Pfizer, Boulder, CO, USA, ⁴Pfizer, Inc., San Diego, CA, USA.

Background: Encorafenib, a BRAF inhibitor, is approved for multiple cancers. As encorafenib is evaluated in new patient populations its pharmacokinetics (PK) must be captured; however, PK sampling in cancer patients should be minimized where possible due to ethical concerns. To address this, limited sampling approaches have been utilized to alleviate unnecessary specimen collection. The objective of this analysis was to examine different limited sampling strategies for encorafenib.

Methods: Data from 31 healthy adult participants who completed an encorafenib food effect study (100-mg single dose administered in the fasted state) were used for this analysis. Plasma concentrations were collected from 0 to 72 hours. Noncompartmental analysis was performed (in Phoenix WinNonlin v8.3) using a variety of sampling schemes to estimate individual PK parameters, notably AUCinf. Limited sampling data were evaluated through direct comparison to the participants PK profile obtained from intensive sampling. Bias and precision were assessed through three different statistical criteria: mean percentage error (MPE) (-5 to 5%), mean absolute percent error (MAPE) ( < 10%), and root mean square percent error (RMSPE) ( < 15%).

Results: Bias and precision statistics for AUCinf estimation for four sampling schemes (S1 to S4) are provided in Table 1. All schemas met the pre-specified criteria for MPE, MAPE, and RMSPE.

Conclusion: Only minimal differences exist between S3 and S4 indicating that either the 3- or 4-hour point can be collected. One limitation of the S1 scheme is that the statistical analysis was based on 65% of the study population; this occurred as AUCinf was not generated unless participants (n=31) had at least three sampling points during the elimination phase for Lambda Z generation. For future encorafenib clinical studies, collecting samples out to six hours, and having at least three time points after the Tmax (2 hours) should be sufficient to reliably calculate AUCinf for the majority of participants and meet criteria for all three statistical endpoints.