Experimental Evaluation of an Axillary Microwave Imaging System to Aid Breast Cancer Staging

Abstract

The number of metastasised Axillary Lymph Nodes (ALNs) is a key indicator for breast cancer staging. Its correct assessment affects subsequent therapeutic decisions. Common ALN screening modalities lack high enough sensitivity and specificity. Level I ALNs produce detectable backscattering of microwaves, opening the way for Microwave Imaging (MWI) as a complementary screening modality. Radar-based MWI is a low-cost, noninvasive technique, widely studied for breast cancer and brain stroke detection. However, new specific challenges arise for ALN detection, which deter a simple extension of existing MWI methods. The geometry of the axillary region is more complex, limiting the antenna travel range required for maximum resolution. Additionally, unlike breast MWI setups, it is impractical to use liquid immersion to enhance energy coupling to the body; therefore, higher skin reflection masks ALNs response. We present a complete study that proposes dedicated imaging algorithms to detect ALNs dealing with the above constraints, and evaluate their effectiveness experimentally. We describe the developed setup based on a 3D-printed anthropomorphic phantom, and the antenna-positioning configuration. To the authors’ knowledge, this is the first ALN-MWI study involving a fully functional anatomically compliant setup. A Vivaldi antenna, operating in a monostatic radar mode at 2-5 GHz, scans the axillary region. Pre-clinical assessment in different representative scenarios shows Signal-to-ClutterRatio higher than 2.8 dB and Location Error lower than 15mm, which is smaller than considered ALN dimensions. Our study shows promising level I ALN detection results despite the new challenges, confirming MWI potential to aid breast cancer staging.