Optimization assisted Kalman filter for cancer chemotherapy dosage estimation

Cancer is regarded to be the earth's most deadly disease, with one of the highest mortality rates among people. "Surgery, radiotherapy, chemotherapy, hormone therapy, and immunotherapy" were all options for treat cancer. Chemotherapy is a medication that is most often deployed for treating cancer, as cancer cells develop and proliferate faster than other cells in the body. Even though chemotherapy is an effective method to treatment various kinds of cancers, the treatment includes risk as it causes side effects due to improper drug usage. The application of a controller-based strategy for determining the optimal rate of drug injection during treatment has risen dramatically in recent years. Thereby, this work develops a robust controller for controlling the dosage of drugs that is carried out under parameter estimation. In addition, a Modified Regularized Error Function-based Extended Kalman filter (MREF-EKF) is introduced for estimating the tumor cells and it can be exploited for diverse conditions. Moreover, the overfitting issue that occurs during drug dosage estimation is also solved using this approach. Further, to improve the performance of the developed approach, the initial state of EKF is fine-tuned via Mean fitness-based Particle Swarm Update (MF-PSU), which is the enhanced version of Particle Swarm Optimization (PSO). At last, the supremacy of the presented approach is proved with respect to convergence analysis and error analysis. For instance, our method outperforms existing GWO + e_k + m, AGWO + e_k + m, and PSO + e_k + m approaches in convergence analysis at noise level 0.41 by 0.009%, 0.002%, and 4.9% respectively. In error analysis, the error values for tumor cells have reached a minimum error value of zero for all noise levels (0.41, 0.43, and 0.55). The findings of this study can help for a better understanding of our presented robust controller's effectiveness in controlling the dosage of drugs.