Analysis of different approaches for the phase function in plasmonic photothermal therapy

Abstract

The Photothermal therapy is a treatment for cancer in which a laser light is used to irradiate a region of a human tissue affected by the disease. The employed laser is such that the irradiated tissue is semitransparent to its wavelength. Therefore, radiation propagates while is partially absorbed, thus heating a volume of tissue. The ideal situation is when the diseased region is heated to a certain temperature and kept at this temperature during a sufficient time period to kill the ill cells; however, avoiding the overheating of surrounding healthy regions. In order to optimize the process, usually they are employed nanoparticles which are highly absorbers in the wavelength of the laser radiation. Such particles are generally either made entirely or recovered by gold, due to the low toxicity of this material. A computer code able to accurately predict the temperature field during a Photothermal therapy is highly desirable, since it could be used to guide medical and experimental procedures and even replace experiments in researches. However, an accurate code requires the implementation of several complex physical phenomena. One of this phenomena is radiation scattering, which is tackled in this research. Usual approximations assume media with nanoparticles as homogeneous, that is, properties of tissue and particles are indistinguishable. In this case, the phase function of particles are assumed as the same as that of the tissue. In the present work, a new algorithm was devised to allow distinct phase functions for particles and host tissue. In addition, different scattering phase functions were implemented: isotropic, Rayleigh and Mie, and analysed to specific cases of skin cancer. Spherical particles made by gold of two different sizes were considered, and also a gold nanoshell with silica core. Usual approximations for the phase function, reduced scattering and Henyey-Greenstein, were also analyzed. Results showed that for the solved cases the errors were not considerable, indicating that the approximations can be used.