Numerical uncertainty and perfusion induced instability in bioheat equation: its importance in thermographic interpretation

Abstract

The use of engineering in biomedical sciences has opened new facets in research. The present paper deals with problems arising from numerical simulation, in order to develop an expert system for the diagnosis of breast cancer using thermography. A female breast is modelled in three dimensions and the surface temperature pattern is obtained by solving the bioheat equation. This solution will be used in conjunction with a database of thermograms to develop an intelligent diagnostic tool. The focus of the present paper is to build and check the confidence level of the numerical scheme before proceeding to model the actual problem. The parametric study is done along with a check for mesh insensitivity and wiggle free isotherm contours. This process yielded a benchmark nodal distance, with which the 3D model is generated and isotherm pattern analysed. It can be seen that this enhances the accuracy of the surface temperature distribution. The use of this is tested in a close-to-actual numerical breast model and the results compared with the thermographic results. The outcome is very encouraging. Finally, a typical clinical protocol in conjunction with the use of numerical prediction for breast thermographic interpretation is outlined.