Estimation of growth features and thermophysical properties of melanoma within 3-D human skin using genetic algorithm and simulated annealing

Abstract

A study has been performed on human skin model with the motivation to device an effective non-invasive modality to characterize the subsurface skin cancer features such as tumor diameter, penetration depth, blood perfusion and metabolic heat generation based on the thermal response of the skin surface obtained from the thermal images. The work presents the role of data mining algorithms to find the tumor features underneath the skin based on the surface temperature variations obtained from a 3-D model of human skin. The human skin is assumed to be subjected to combined radiative, convective, and evaporative heat flux boundary conditions. The study revealed that, the major variation in the thermal response of tumor is attributed to increase in the volume, blood perfusion and thermogenic capacity. The variations due to inter- and intra-patient variability of tumor properties and size are obvious, which could be explained by the retrieved multiple combinations of variables. Furthermore, the reconstructed surface thermal distributions associated with estimated variables are found to be in a good match with the actual maps. The error <10% in the measured thermal distribution tends to give accurate reconstruction. Present strategy or algorithm along with a thermal camera may prove to be a useful diagnostic tool for the characterization of subsurface skin cancer and reduce the unnecessary biopsy trials.