Parametric optimization for tumour identification: Bioheat equation using ANOVA and the Taguchi method

Abstract

Breast cancer is the number one killer disease among women. It is known that early detection of a tumour ensures better prognosis and higher survival rate. In this paper an intelligent, inexpensive and non-invasive diagnostic tool is developed for aiding breast cancer detection objectively. This tool is based on thermographic scanning of the breast surface in conjunction with numerical simulation of the breast using the bioheat equation. The medical applications of thermographic scanning make use of the skin temperature as an indication of an underlying pathological process. The thermal pattern over a breast tumour reflects the vascular reaction to the abnormality. Hence an abnormal temperature pattern may be an indicator of an underlying tumour. Seven important parameters are identified and analysis of variance (ANOVA) is performed using a 2n design (n = number of parameters, 7). The effect and importance of the various parameters are analysed. Based on the above 27 design, the Taguchi method is used to optimize the parameters in order to ensure the signal from the tumour maximized compared with the noise from the other factors. The model predicts that the ideal setting for capturing the signal from the tumour is when the patient is at basal metabolic activity with a correspondingly lower subcutaneous perfusion in a low temperature environment.