par Miskovic, Vanja
Président du jury Delplancke, Marie-Paule
Promoteur Dubois, Frank
Publication Non publié, 2021-09-15
Thèse de doctorat
Résumé : Wound repair is one of the most complex biological processes that occur during a human lifetime. Apart from negatively impacting patient well-being and jeopardizing human life, wounds represent a great burden to the healthcare costs both in developed and developing countries and even more in remote areas and rural communities. Interestingly, those latter regions share commonalities – remoteness, communications’ lags - with the conditions experienced in Space that could be then used as a testing environment for developing effective solutions on Earth. Reliable and wearable monitoring systems could help in reducing these problems. However, despite the significant progress on this topic, critical limitations exist, and the domain is still lacking a firm and complete understanding of the complex network of the involved biophysical phenomena. So far, most of the studies are focusing on the development of wearable sensing device, with limited information on real values and without a clear long-term vision. With the emerging field of Health 4.0, the paradigm is shifted, and the effective use of the wound data should become the primary focus. This study represents a contribution in tackling the conceptual gaps for developing advanced and effective wound monitoring systems. To do so, this thesis consists of two parts dedicated to an innovative configuration to host a wound temperature sensor and a new concept for using actual wound data to foster diagnostics prediction, respectively. In the first part of this thesis a multilayer temperature sensing concept is proposed, consisting of a hydrogel layer acting as a transfer layer for the biophysical signals coming from the wounds and a liquid crystal temperature sensing layer, the temperature being one of the most important biomarkers to determine the wound status. For the hydrogel layer, synthetic and double network hydrogels incorporated with graphene oxide were explored. Hydrogels have been tested on the ground and in hypergravity and microgravity conditions during a parabolic flight. Cholesteric liquid crystals represent the backbone of the first sensing concept that has been developed to create temperature maps of the wound area. This concept targets an easy-to-read, wearable, colorimetric system, having resolution (<0.2 °C) and rapid response time (<1 s) compatible with wound monitoring requirements. In the second part of this thesis, the thermal map of the wound site is generated using a thermal camera. In this case, the 4.0 approaches were used to expand the knowledge on real-life wound data and explore the connection between burn wound temperature and burn wound status. A prospective study is conducted on patients at the Burn Unit in Queen Astrid military hospital. The comparison of the collected and the existing data led to the formulation of a temperature threshold value, based on the difference between temperature at the burn site and temperature of the healthy skin (ΔT), that can be used for the burn healing time prediction, if ΔT > −1.2°C burn will heal in less than 21 days. Results from this study suggest the use of temperature as a reliable wound monitoring marker. Moreover, a real-hospital data was used to train a deep learning algorithm that successfully performed wound image segmentation, which presents a first step in developing a fully automated diagnostic system.