Microwave Imaging in Medical Diagnostics: A Systematic Review of Methods, Applications, and Future Directions

Microwave imaging (MWI) has emerged as a promising non-invasive diagnostic technology in medical applications, offering significant advantages over conventional imaging modalities including reduced radiation exposure, cost-effectiveness, and enhanced patient comfort. This comprehensive survey presents a systematic review of MWI through a novel bifurcated taxonomy addressing both methodological aspects (imaging techniques, deep learning approaches, and antenna design) and application domains (breast imaging, brain imaging, and stroke diagnostics). The methodological analysis reveals substantial progress in reconstruction algorithms, with physics-assisted deep learning frameworks demonstrating superior performance in automated thresholding and bias reduction compared to traditional qualitative imaging methods. Advanced techniques like MDLI-Net and DL-MITAT have shown particular promise in handling sparse data scenarios. In antenna technologies, ultra-wideband antennas and miniaturized millimeter-wave sensors represent significant breakthroughs, with some systems capable of detecting malignancies as small as 1mm. Application-wise, brain imaging applications have achieved remarkable accuracy rates exceeding 94% using YOLOv5 models, while breast imaging systems demonstrate effective tumor detection in dense tissues where conventional mammography faces limitations. Despite these advancements in computational methods, particularly deep learning and sophisticated reconstruction algorithms, several challenges remain in translating MWI technology from experimental settings to widespread clinical adoption. Our contributions include structured analysis of recent advances in MWI techniques, critical insights into challenges across medical applications, and identification of research gaps and future directions. Key findings demonstrate that combining qualitative imaging with deep neural networks achieves reliable real-time reconstruction capabilities. While MWI shows considerable potential as a safe, cost-effective alternative to traditional imaging techniques, particularly for breast and brain imaging, further research in computational efficiency, clinical validation, and standardization protocols is necessary for successful clinical implementation across diverse healthcare settings.