Model based chemotherapeutic drug scheduling: A multi-objective particle swarm optimization approach

Cancer chemotherapy is a complicated treatment that needs balancing the cell killing with the undesirable toxic sideeffects caused by the drugs. Several computational methods have been used in finding the right balance. This paper presents a model-based method to design optimum chemotherapy treatment schedule by using multiple objective particle swarm optimization (MOPSO) algorithm. Three dose plans are designed based on a widely used cancer chemotherapy model. MOPSO algorithm is employed for determining drug schedule with suitable parameters that give the dose levels and their occurrences. Besides tumor cell population, the model also shows other important parameters, such as natural killer cells (NK), tumorspecific-cytotoxic-T (CD8+ T) cell population and circulating lymphocyte population. Results show that the designed method can generate efficient drug scheduling that can eliminate the tumor cells with acceptable toxicity. The designed drug schedules also satisfy important treatment constraints and other objectives.