Enhanced Physics Modelling in Radar-Based Microwave Imaging for Breast Cancer Detection

Abstract

Microwave-based breast imaging (MBI) is an emerging modality that may serve as a screening tool due to the relatively large dielectric contrast between malignant and healthy tissues, the relatively low cost and small size of microwave hardware, and the favourable safety profile of non-ionizing microwave imaging. After more than two decades of research into MBI and several published clinical trials, challenges remain before the modality can be used clinically. Existing estimates of the diagnostic specificity are relatively low, between 20–65%. As a result of the limited specificity of the technique, existing radar-based image reconstruction algorithms have not demonstrated sufficient accuracy for breast cancer diagnosis. This article proposes using enhanced physics modelling (EPM) to improve the accuracy of the physics models used in image reconstruction to address the limited diagnostic accuracy. The results obtained in this study indicated that using EPM significantly improved the area under the curve (AUC) of the receiver operating characteristic curve. The AUC was improved from (84 $\pm$ 1)% to (92 $\pm$ 1)% through the use of EPM, demonstrating the potential of physics-informed radar-based image reconstruction in MBI.