Model Predictive Control for Personalized Abiraterone Administration in Metastatic Castrate-Resistant Prostate Cancer

Abstract

Metastatic castrate-resistant prostate cancer (mCRPC) presents significant therapeutic difficulties. This study develops and evaluates a Model Predictive Control (MPC) framework for the personalized administration of Abiraterone in mCRPC. The proposed MPC strategy dynamically optimizes drug dosage over a finite prediction horizon, utilizing a well-established mathematical model of prostate cancer cell populations from the literature (androgen-dependent T +, testosterone-producing T P , and androgen-independent T -). This adaptive approach is designed to delay disease progression and manage tumor burden more effectively than static or pre-optimized schedules. Simulation results obtained in this paper demonstrate the effectiveness of the proposed controller. Specifically, the MPC approach shows potential in prolonging progression metrics, such as the time until resistant T - cells prevail, and in reducing cumulative drug exposure through adaptive dosing, thereby supporting the objective of managing cancer as a chronic condition.