Abstract
Multistatic adaptive microwave imaging (MAMI) methods are presented and compared for early breast cancer detection. Due to the significant contrast between the dielectric properties of normal and malignant breast tissues, developing microwave imaging techniques for early breast cancer detection has attracted much interest lately. MAMI is one of the microwave imaging modalities and employs multiple antennas that take turns to transmit ultra-wideband (UWB) pulses while all antennas are used to receive the reflected signals. MAMI can be considered as a special case of the multi-input multi-output (MIMO) radar with the multiple transmitted waveforms being either UWB pulses or zeros. Since the UWB pulses transmitted by different antennas are displaced in time, the multiple transmitted waveforms are orthogonal to each other. The challenge to microwave imaging is to improve resolution and suppress strong interferences caused by the breast skin, nipple, and so forth. The MAMI methods we investigate herein utilize the data-adaptive robust Capon beamformer (RCB) to achieve high resolution and interference suppression. We will demonstrate the effectiveness of our proposed methods for breast cancer detection via numerical examples with data simulated using the finite-difference time-domain method based on a 3D realistic breast model.
Introduction
Early detection is the primary factor in surviving breast cancer. While X-ray mammography is the current gold standard, it has significant drawbacks: it uses ionizing radiation and often struggles to resolve lesions in "radiologically dense" breast tissue, which is common in younger women.
Microwave imaging is a promising alternative because it is non-ionizing and exploits the significant contrast in dielectric properties (permittivity and conductivity) between healthy and malignant tissues. However, the technical challenge is immense. The reflections from the skin, nipple, and chest wall are orders of magnitude stronger than the signal from a tiny tumor. To find the "needle in the haystack," researchers need algorithms that can suppress these massive interferences while maintaining high spatial resolution.
Methodology
Multistatic Antenna Configuration
The proposed approach uses a real aperture array of 72 antennas arranged in a hemisphere around the breast. Each antenna takes a turn transmitting a UWB pulse while all others receive the backscattered signals. Because antennas transmit sequentially, the UWB pulses are naturally orthogonal in time, simplifying the multistatic processing compared to simultaneous MIMO radar systems.
Robust Capon Beamforming (RCB)
The core of the signal processing is the Robust Capon Beamformer (RCB). Unlike standard "delay-and-sum" methods, RCB is data-adaptive; it adjusts its weights based on the received data to minimize interference while remaining robust against uncertainties in the "steering vector" (the expected signal path). The paper introduces three variations based on how the 3D data cube (Transmitter x Receiver x Time) is sliced:
- MAMI-1: Slices the data cube by time index. It is highly robust against errors in the signal path but uses fewer "snapshots" for its calculations.
- MAMI-2: Slices the data cube by transmitter index. This provides more snapshots, leading to a more precise covariance matrix, ideal for high signal-to-noise environments.
- MAMI-C (Combined): A hybrid approach that feeds the outputs of both MAMI-1 and MAMI-2 into a second stage of RCB. This effectively doubles the "fictitious" array size, providing both precision and robustness.
Simulation Environment
The study relies on high-fidelity numerical simulations using the Finite-Difference Time-Domain (FDTD) method. The 3D realistic breast model includes randomly distributed fatty and glandular tissues, a 2 mm skin layer, a nipple, and a chest wall. To mimic human biology, the researchers introduced 10% Gaussian variance in the dielectric properties and accounted for frequency-dependent dispersion using a Debye model.
Results and Performance
The MAMI-C method successfully localized 4 mm and 6 mm tumors in the realistic 3D model, even when buried under significant clutter. It consistently outperformed MAMI-1 and MAMI-2 across different noise levels. In high-noise scenarios (SNR = -24.5 dB), MAMI-C produced the clearest images with the least amount of background clutter.
The adaptive nature of the RCB allowed for much tighter "focusing" on the tumor location compared to non-adaptive methods, which often produce blurry "blobs" in the presence of skin reflections. By treating the outputs of MAMI-1 and MAMI-2 as snapshots from a "fictitious" array, MAMI-C achieves a level of interference suppression that neither could reach alone.
Conclusion
This research demonstrates that multistatic adaptive microwave imaging, particularly the MAMI-C hybrid method, is a powerful tool for early breast cancer detection. While computational costs are higher due to two stages of RCB processing, the gain in resolution and interference suppression is significant. Future work will involve moving from FDTD simulations to experimental phantom or clinical trials to address real-world coupling issues.
Summary updated