We evaluate the effect that fibroglandular tissue has on tumor detection in radar-based microwave imaging (MWI). MWI relies on the contrast between the dielectric properties of healthy and malignant tissues, which may be as low as 10%, thus hindering tumor identification. Here, we performed measurements on a dry MWI setup, in the 2-5 GHz frequency range, to scan anthropomorphic breasts with two fibroglandular tissues, and a tumor placed at different distances from the fibroglandular. Imaging results showed that tumor identification in MWI is highly influenced by the volume and shape of the fibroglandular tissue. Indeed, we find that in most cases the tumor cannot be distinguished with common image metrics because it is indistinguishable from the fibroglandular. Yet, we verify that there are changes in the magnitude of the confocal algorithm, which may open the opportunity to machine learning techniques to detect it.

--

Breast cancer diagnosis,Microwave imaging,Radar imaging,Dry microwave setup,Ultrawideband antennas

• Engenharia Eletrotécnica, Eletrónica e Informática - Engenharia e Tecnologia