Noninvasive estimation of tissue impedance based on an equivalent circuit model

Clinical researches show that the bioelectrical impedance of tumors are different from normal tissues. Noninvasive estimation of the impedance distribution in a tissue is called as electrical impedance tomography (EIT). EIT will be useful for clinical applications such as breast cancer diagnosis. This study aims at developing impedance estimation method for EIT. A tissue equivalent circuit model was used for representing the impedance distribution in a tissue. In our previous study, the impedance estimation performed on a fixed structure model using the Gauss-Newton method. In the present study, we tried to the impedance estimation on a layer structure model using the modified Marquardt method. The estimation of the layer structure is required for the first step for EIT. However, it was difficult to estimate by the Gauss-Newton method. We tested the impedance estimation in computer simulations. A target model was used to simulate the measurements on a real tissue. The simulations were performed with varying the standard deviation of the impedance in the target model. As a result, the Gauss-Newton method failed to estimate the impedance and the modified Marquardt method succeeded to estimate the impedance on the layer structure model. However, large estimation errors were found in the results when the standard deviation of the impedance increased. The modified Marquardt method has an advantage in the impedance estimation. Further improvement is needed for practical use of the method.