A memetic algorithm for optimization of combination chemotherapy with dose adjustment

Chemotherapy is recognized as one of the most prominent systemic modalities for cancer treatment. The oncologists combine different chemotherapeutic drugs for patient treatment according to clinical practice guidelines, and adjust the drug dose to reduce its toxicity or to improve its efficacy according to the patient's response. This paper introduces a two-drug cancer chemotherapy model to describe the response of the tumor cells and host cells under drug administration. The objective is to cure the patient as soon as possible while the toxicity is maintained below an allowable limit throughout the entire treatment period. A problem-specific memetic algorithm (MA) is designed in this paper to optimize the combination chemotherapy problem with dose adjustment. The proposed algorithm combines the exploration breadth of evolutionary algorithms and the exploitation depth of local search methods. Computational experiments show that the proposed MA outperforms the existing algorithm in terms of solution quality. The results demonstrate the efficacy of treatment with dose adjustment.