INTRODUCTION AND OBJECTIVES: The hyperthermia treatment of cancer involves the heating of tumor cells in order to make them more susceptible to other kinds of treatment (like chemotherapy or radiotherapy) or to directly kill them with high temperatures and long exposures that cause permanent cell damage. Hyperthermia as a therapeutic technique is not only applied in oncology, but also in physiotherapy, urology, cardiology and ophthalmology. Mathematical modeling is of major importance for the planning and control of any cancer treatment, including hyperthermia. Uncertainties must be appropriately accounted for in numerical simulations and in the optimal design of the cancer treatment. The objective of this work is thus to optimize the hyperthermia treatment of a skin cancer with laser heating, by considering intrinsic uncertainties in model parameters like the physical properties of the tissues. MATERIAL AND METHODS: A thermal damage model was coupled with both bioheat and light propagation models. The bioheat transfer and light propagation problems were solved with the finite volume method, for a one-dimensional region consisting of five tissue layers that represent the skin. The Markov Chain Monte Carlo (MCMC) Method was utilized to obtain an optimal set of design variables, including the volumetric fraction of nanoparticles loaded into the tumor, number of laser applications, power of the incident laser and the duration of each treatment session. The objective function aimed at a safe procedure that causes cellular death in the tumor region, with minimal damage to healthy cells. Values available in the literature were used to specify the prior distributions for thermophysical and optical properties. RESULTS AND CONCLUSION: The optimization was performed with the Metropolis-Hastings algorithm within the Bayesian framework of statistics. Equilibrium Markov chains provided samples of the posterior distribution of the design variables and other model parameters. These samples were then used to simulate the statistical distribution of the thermal damage in the region. The results obtained in this work with the optimal design variables reveal that the thermal damage can be concentrated in the tumor region, as desired.
Keywords: Hyperthermia, Optimization under uncertainties, Skin cancer.
Supported by: UFRJ and CAPES