Master thesis : Electrical impedance tomography for external wound monitoring : measurements, signal processing and image reconstruction

Abstract

Assessing the state of chronic open wounds and monitoring their healing in the long term is an important but delicate task in the medical world. Manipulating and perturbing traumatized tissues too often hinders the healing process and prolongs the burden for the patient and the medical system. Conversely, ignoring complications such as bacterial infection or necrotic tissue for too long can lead to dramatic consequences, from the need for amputation to septic shock and death. Nowadays, chronic wound monitoring is still a matter of frequent visual inspections, which requires medical expertise and intrusions into the daily lives of patients.

The Vitapatch research project, for which this thesis has been carried out, aims at solving this issue by creating easily deployable smart sensor patches capable of long-term continuous monitoring of a chronic wound in a non-invasive manner.

This thesis explores the use of bioimpedance spectroscopy to assess the state of human skin, and electrical impedance tomography as an image reconstruction tool to provide a non-intrusive visual assessment of wound healing. After briefly reminding the concepts of electrical impedance and conduction of electricity, the physiological and electrical properties of human skin and tissues are presented, and the impacts of a wound on these properties are discussed. Then, the methodology for bioimpedance measurements is explained, with specific care toward long-term medical applications. Following, the prototype bioimpedance spectroscopy circuit created by Microsys is analyzed and simulated. After that, the required signal processing steps to make this circuit work are presented, and the journey towards experimental validation of the processing routine is described. Proceeding, electrical impedance tomography is introduced, different algorithms are assessed through simulations and a preliminary application is presented. Adaptations of the EIT problem to wound imaging are performed, and image reconstruction on a finite-element model of wounded skin is simulated. Finally, an experimental setup for wound assessment on phantom skin is presented. Repeated hardware delays and unfinished or faulty components have prevented the completion of real-life experiments, but simulations show promising prospects for a nontraditional approach to impedance spectroscopy and signal processing, as well as for skin modeling and applications of electrical impedance tomography to non-invasive wound imaging.