Advancements towards Microwave Tomography for Breast Cancer Screening

Breast cancer is nowadays one of the main causes of death in women and its early detection is commonly understood as the most e®ective way to cure this disease. The state-of- the-art screening technique is X-ray mammography which, while sensitive to tumors, has indeed limitations in discriminating amongst malignant tumors and benign ones. Recently, it has been demonstrated that the electromagnetic properties of the malignant tissues at microwave frequencies are very di®erent from those of the normal ones. Accordingly, microwave tomography has emerged as a complementary diagnostic tool, which can be potentially exploited to perform morphological and functional screening. On the other hand, the ambitious aim of achieving quantitative information on the imaged tumors involves the solution of a non-linear and ill-posed inverse scattering problem that, besides not-trivial computational issues, claims for reliable approaches capable of avoiding false alarms or lack of detection. In this communication, in order to address the above problems and proceed towards an e®ective microwave screening of breast cancer, we will show how, by properly acting on the degrees of freedom of the problem (i.e., type of immersion liquid, working frequency and so on), it is possible to achieve accurate tomographic reconstructions and to actually discriminate between di®erent types of tumors when they are contemporary present. In particular, by reasoning on the electromagnetic characteristics of the coupling medium, the desired spatial resolution and the factors a®ecting non-linearity of the underlying inverse scat- tering problem, we ¯rst derive some guidelines for an \optimal" choice of the working frequency and the electromagnetic characteristics of immersion liquid. Then, considering the so determined work- ing conditions, we test the reconstruction capabilities of a Contrast Source inversion ap- proach wherein a regularization by projection is properly exploited in order to increase robustness against false solutions. Numerical examples related to 2D and 3D imaging experiments will be shown at the conference.