Optimization assisted Kalman filter for cancer chemotherapy dosage estimation

Area of Science:

• Biomedical Engineering
• Control Systems
• Computational Biology

Background:

• Chemotherapy is a vital cancer treatment but carries risks due to potential side effects from improper drug dosage.
• Optimizing drug delivery rates is crucial for maximizing treatment efficacy and minimizing adverse effects in cancer therapy.
• Existing methods for drug dosage control and tumor cell estimation face challenges like overfitting and suboptimal performance.

Purpose of the Study:

• To develop a robust controller for precise chemotherapy drug dosage management.
• To enhance tumor cell estimation accuracy using an advanced filtering technique.
• To improve the overall performance and reliability of drug dosage control strategies.

Main Methods:

• A robust controller was developed for drug dosage control, incorporating parameter estimation.
• A Modified Regularized Error Function-based Extended Kalman Filter (MREF-EKF) was introduced for accurate tumor cell estimation.
• Mean fitness-based Particle Swarm Update (MF-PSU), an enhanced Particle Swarm Optimization (PSO), was used to fine-tune the Extended Kalman Filter's initial state.

Main Results:

• The developed approach demonstrated superior convergence analysis compared to existing methods (GWO, AGWO, PSO) at a noise level of 0.41.
• Error analysis showed that tumor cell estimation achieved a minimum error value of zero across various noise levels (0.41, 0.43, 0.55).
• The MREF-EKF successfully addressed the overfitting issue in drug dosage estimation.

Conclusions:

• The proposed robust controller effectively manages chemotherapy drug dosage, offering a significant advancement in cancer treatment control.
• The MREF-EKF provides a reliable method for tumor cell estimation, applicable to diverse conditions.
• The integration of MF-PSU enhances the performance of the Extended Kalman Filter, leading to more accurate and stable drug dosage control.