PI-086 - POPULATION PHARMACOKINETIC OF BMS-986365, A DUAL ANDROGEN RECEPTOR (AR) LIGAND-DIRECTED DEGRADER AND ANTAGONIST, IN PATIENTS WITH PROGRESSIVE METASTATIC CASTRATION-RESISTANT PROSTATE CANCER

J. Zhou¹, Y. Guo², S. Cheng¹, C. Liu¹, Y. Liu³, Z. Shah¹, M. Ye¹, B. Wang¹, L. Zhu¹, S. Suryawanshi¹; ¹Bristol Myers Squibb, Lawrenceville, NJ, USA, ²Bristol Myers Squibb, Shanghai, China, ³Bristol Myers Squibb, Tempa, FL, USA.

Background: BMS-986365, an orally administered ligand-directed degrader targeting the AR via a first-in-class dual mechanism of AR degradation and antagonism. BMS-986365 is a mixture of diastereomers BMS-986410 (R,S) and BMS-986409 (R,R). Study CC 94676-PCA-001 is a multi center, open-label, FIH, Phase 1, dose-finding study to determine the safety, tolerability, PK, pharmacodynamics, and preliminary efficacy of orally administered BMS-986365 in subjects with progressive mCRPC. BMS-986365 has been well characterized across a broad range of doses and has demonstrated promising and prolonged clinical activity, good tolerability, and manageable safety in heavily pre-treated subjects with mCRPC. The objectives of this analysis were to characterize BMS-986365 PK (tablet formulation) using a popPK approach, and to provide exposure parameters for further exploration of exposure-response relationships of efficacy and safety.

Methods: The popPK analysis population included subjects (N=94) who have measurable concentrations in BMS-986410 and BMS-986409 from study CC 94676-PCA-001. Model evaluation was performed using Goodness-of-fit plots and Visual Predictive Check plots. Covariates were assessed through graphical exploration and a stepwise procedure for comparing objective function value. The popPK analysis was performed using the NONMEM (V7.4). All presentations of data were prepared using R (V4.0.3), while the analysis datasets were prepared using SAS.

Results: The PK data was best described using a two-compartment model with dose-dependent decrease in relative bioavailability at > 400mg, two absorption transit compartments, and first-order absorption. A time varying clearance was incorporated in the model. The model includes a proportional residual error model and random effects on MTT (mean transit time), CL/F, VC/F, and correlation of random effect between CL/F and VC/F. No significant covariate was identified in the final model. The simulations of BMS-986365 exposures over time suggested that the PK profiles achieve steady-state before Day 8.

Conclusion: The PK of BMS-986365 in mCRPC patients was well described by the popPK model. This popPK model was used to predict individual/population subjects’ exposures for further exploration of exposure-response relationships.