Modelling of differential probing analysis for biomedical dielectric spectroscopy

Transcutaneous microwave dielectric spectroscopy holds promise as a modality for non-invasive subcutaneous biomarker monitoring, such as blood glucose levels. Experimental studies have demonstrated a strong correlation between blood glucose concentration and blood permittivity using microwave signals. However, the sample volume of interest, blood vessels, is located underneath skin layers where the microwave signal interaction with other surrounding biomaterials is accumulated. Here, a computational study of the proposed method to mitigate the contribution of unwanted tissue and thereby increase the specificity of subcutaneous spectroscopy, using differential probing analysis (DPA), is presented. By exploiting the relationship between the diameter of an open-ended coaxial probe (OCP) and the size of its protruding electric field, two OCPs with different sizes can sense different volumes beneath the skin. The proposed method computationally subtracts the effect of unwanted tissue, estimated with the smaller probe, from the effective contribution of both blood and unwanted tissue with the larger one. The method is validated by using simulated models and data from the full wave software CST Studio Suite across frequencies ranging from 1 to 10 GHz, achieving errors comparable to state-of-the-art techniques.