The use of microwave (MW) sensors for non-invasive probing of biological tissues presents significant potential in diagnostic medicine. These sensors are applied in various medical applications, such breast cancer detection, brain stroke identification, blood glucose measurement, and more [1], [2]. Among these applications, one with significant clinical impact involves the non-invasive in-situ analysis of atheromatous plaque composition, particularly in the carotid artery.

Atherosclerosis of the carotid artery is characterised by plaque accumulation in the arteries that supply blood to the brain. It is a major cause of ischemic strokes, resulting from the obstruction of blood flow to certain regions in the brain [3]. Diagnosis of atherosclerosis typically relies on imaging techniques, such as ultrasound, computed tomography, and magnetic resonance imaging but these methods do not reliably assess stroke risk linked to a carotid lesion [4]. Surgical decisions are thus based largely on arterial narrowing seen in imaging, without insight into plaque composition [5]. While histology is the gold standard for assessing plaque risk, it requires invasive surgical sampling. Microwave sensors, which differentiate biological tissues by their dielectric properties, offer a non-invasive alternative to provide crucial preoperative data on plaque composition, potentially improving patient care [6], [7].

Frequency variation is likely the most widespread operating principle in planar microwave sensors, particularly in applications dedicated to the dielectric characterization of solids and liquids, material composition determination, biological sample analysis, and the measurement of spatial variables [8]. Frequency-variation sensors can be implemented through planar resonators, whose resonant frequency and quality factor are determined by the dielectric properties of the medium surrounding the resonator. These sensors are therefore well-suited for determining the complex permittivity (dielectric constant and loss tangent) of materials, or for measuring variables related to it, such as material composition. The choice of the planar resonant element is crucial, as it defines the sensor's size, and consequently the size of the sensing area, its sensitivity, and other key parameters. There are various types of planar resonators suitable for sensing applications. Among these, semi-lumped resonators are generally preferred for microwave sensing [9]. They can be designed using microstrip [10], coplanar waveguide or substrate integrated waveguide (SIW) technologies [11], along with innovative metamaterial approaches [12], [13], and may feature interdigital, meander, ladder, or T-shaped structures [14], [15]. Among metamaterial-inspired resonators, split-ring resonators (SRRs) and complementary split-ring resonators (CSRRs) are frequently used for dielectric characterization due to their compact design as sub-wavelength resonators, their simple fabrication process, and their high sensitivity to environmental changes [16], [17]. They can also be combined with SIW [18], [19], interdigital capacitors [20], meander slots [21], [22], [23] to enhance the quality factor and field intensity and provide better sensitivity to the effects of dielectric material. Inter-resonator or inductive coupling mechanism used to excite CSRR can also enhance the sensitivity of dielectric sensors [24], [25]. Meanwhile, CSRR-based microwave sensors offer a more promising option for the intended application. Their sensitivity is higher than that of SRRs because their effective sensing area is larger, as the electric field is primarily localized in the narrow split of the SRR, resulting in limited interaction with the sample [8], [24]. Additionally, the ground plane provides backside isolation, preventing any influence from elements located on the back side of the substrate on the electrical characteristics of the feed line.

This article presents an improved CSRR-based microwave sensor designed for future in-vivo dielectric characterization of atherosclerotic plaques. Optimized to increase EM wave penetration into multilayered tissues, the sensor aims to distinguish between soft, high-risk carotid plaques, rich in necrosis, and lipids highly calcified carotid plaques, which are considered stable and thus less risky. Dielectric measurements are conducted on pig tissue samples (skin, fat, and muscle), both individually and in multilayer configurations, to mimic the EM path through the neck. Experimental results are validated with electromagnetic simulations to estimate wave penetration depth. Finally, a three-dimensional (3D) carotid artery model is developed to assess the sensor's sensitivity to detect atheroma.