Tumour Classification

Recent breast tissue dielectric spectroscopy measurements in [35] suggest that the malignant-to-benign dielectric contrast may not be sufficiently high to allow for tumour classification based on backscatter intensity. Alternatively, it is well known that the architectural distortion in breast parenchyma can aid in distinguishing malignant tumours from benign masses [56, 65]. Mammographic image analysis shows that malignant tumours usually have an irregular shape and are surrounded by a radiating pattern of linear spicules, also their margins are obscured and indistinct. Conversely, benign tumours are roughly elliptical and usually have well-circumscribed margins [56, 65]. Accordingly, microwave backscatter signature-the signal which is reflected when a target is illuminated by microwaves-could be potentially useful for discrimination between benign and malignant tumours, and for inferring their size. Tomographic image reconstruction methods, which solve the inverse scattering problem to obtain a coarse estimation of the breast dielectric profile, have been used in this context [19].