Electrical and Computer Engineering, University of Connecticut, Storrs, 06269, USA.
Sci Rep. 2023 Nov 27;13(1):20850. doi: 10.1038/s41598-023-47406-z.
Bladder cancer is a cancerous disease that mainly affects elder men and women. The immunotherapy that uses Bacillus of Calmette and Guerin (BCG) effectively treats bladder cancer by stimulating the immune response of patients. The therapeutic performance of BCG relies on drug dosing, and the design of an optimal BCG regimen is an open question. In this study, we propose the reparameterized multiobjective control (RMC) approach for seeking an optimal drug dosing regimen and apply it to the design of BCG treatment. This approach utilizes constrained optimization based on a nonlinear bladder cancer model with impulsive drug instillation. We compare the performance of RMC with Koopman model predictive control (MPC) and validate the efficacy of optimal BCG dosing regimens through numerical simulations, demonstrating the efficient elimination of cancerous cells. The proposed control framework holds the potential for generalization to other model-based treatment designs.
膀胱癌是一种主要影响老年男女的癌症。卡介苗（BCG）免疫疗法通过刺激患者的免疫反应，有效地治疗膀胱癌。BCG 的治疗效果依赖于药物剂量，设计最佳的 BCG 方案是一个悬而未决的问题。在这项研究中，我们提出了重新参数化多目标控制（RMC）方法，以寻找最佳的药物剂量方案，并将其应用于 BCG 治疗的设计。该方法利用基于带有脉冲药物灌输的非线性膀胱癌模型的约束优化。我们将 RMC 的性能与 Koopman 模型预测控制（MPC）进行了比较，并通过数值模拟验证了最佳 BCG 剂量方案的疗效，证明了有效消除癌细胞的能力。所提出的控制框架具有推广到其他基于模型的治疗设计的潜力。
