Ischemic strokes of carotid origin represent a significant proportion of cerebrovascular accidents, often leading to severe mortality and disability, making them a major public-health concern. Despite their clinical importance, the absence of accurate diagnostic methods to assess stroke risk related to carotid lesions remains a major challenge. Conventional imaging techniques, such as Doppler ultrasound and CT scans, fail to provide sufficient information on plaque composition to enable precise surgical assessment of the associated risk.
Portable medical devices are now benefiting from advances in mixed analog-digital electronic systems, enabling the development of faster, more reliable, and non-invasive diagnostic solutions. In this context, microwave (MW) techniques stand out for their ability to characterize biological tissues according to their dielectric composition, thus opening new perspectives in the biomedical field.
This thesis focuses on the design of microwave sensors dedicated to the diagnosis of atheromatous plaques in the carotid artery, as well as their integration into a portable system for real-time evaluation of the dielectric properties of biological tissues.
Designed to be positioned on the patient’s neck, at the level of the carotid bifurcation, the sensors use planar technology inspired by metamaterials to reduce size without performance loss. Based on Complementary Split-Ring Resonators (CSRRs), their topology was optimized to enhance the penetration of electromagnetic (EM) waves through tissues and improve both frequency and amplitude sensitivity. EM simulations guided the optimization of resonator geometry, coupling configuration, and substrate properties, leading to the fabrication of several MW sensors.
The optimized CSRRs were then analyzed using an equivalent circuit model, developed by adapting conventional parameter-extraction methodologies. This modeling approach proved particularly effective, allowing not only the prediction of the resonators’ frequency response but also the guidance of their design process.
Experimental evaluations of the MW sensors near the carotid bifurcation were then conducted. Measurements on porcine tissues and single- or multi-layer phantoms enabled the characterization of their electromagnetic properties. In parallel, a 3D model of the carotid artery was developed to simulate the response of optimized CSRRs under realistic conditions and identify the most efficient sensor for atheromatous plaque detection.
Finally, a portable acquisition and processing system, integrating software tools for real-time dielectric property extraction, was implemented for ex vivo tissue measurement using the developed MW sensors. Its performances were validated on dielectric phantoms, confirming the precision, reproducibility, and reliability of the system for dielectric characterization.