Development of finite element simulation and machine learning models relating to Electrical Impedance Spectroscopy

Development of finite element simulation and machine learning models relating to Electrical Impedance Spectroscopy in order to discriminate between different tissue types during thyroid surgery.

Electrical Impedance Spectroscopy (EIS) is a method for identifying different tissue types by its passive electrical characteristics measured over a range of AC frequencies. It has been successfully used to identify early cancers in the cervix (Zedscan, University of Sheffield, Zilico), and early data suggests that it may be applicable as a tool to guide surgical intervention in thyroidectomy by discriminating between visually similar tissue types.

This project aims to use both machine learning on an existing data set, collected during surgery and finite element simulation techniques to answer the following questions:

i) Which characteristics of the impedance spectrum give best discrimination between the tissue types?
ii) What are the characteristics of the tissue that give rise to these features? This will ultimately support the design of a commercially guided instrument for EIS-guided surgery.