Abstract
Hard field applications are techniques (like X-ray and magnetic resonance imaging (MRI)) which utilize source signals that pass the medium in a predictable manner, independent of the material properties of the medium. Soft field applications such as bioimpedance analysis (BIA), optical tomography (OT), and electrical impedance tomography (EIT), on the other hand, are methods where the path of the source signals is determined by the electromagnetic properties of the medium.
Although soft field applications (SFAs) have the potential to be a useful non-invasive tool in clinical practice, they are not routinely used in hospitals yet. The two main problems of SFA are their low signal-to-noise ratio and their higher computational effort as compared to hard field applications.
This thesis presents two specific applications to illustrate approaches to overcome some of the specific problems in design optimization for SFA. Two articles in peer-reviewed journals provide the foundation for this work.
The first application deals with design optimization for BIA. It demonstrates how the estimation of body fats (BFs) and fat free mass (FFM) using impedance measurements can be improved. Using a realistic three- dimensional finite element (FE) model of the human thorax combined with a hierarchical model which directly incorporates all parameters of interest, a sensitivity analysis was carried out. The analysis revealed a high sensitivity to the subcutaneous fat layer thickness (SFL) and the fat content of the mesentery. The presented structural model is a prerequisite for a comprehensive investigation of body composition measurements and helps to gain further insight into BIA methods. Although this is a promising result, the intrinsic problem with BIA, namely the dependency of the results on the hydration state, still remains. Furthermore, the influence of changes in the model geometry and the effect of models representing different values for the SFL has not been examined yet.
The second application shows the applicability of a novel mathematical method for the optimization of both the location as well as the strength of optical fibers for photodynamic therapy (PDT). This therapy combines light of a specific wavelength with a photosensitizer to chemically destroy tissue cells. Most commonly, it is employed for various oncological and dermatological treatments. PDT requires extremely homogeneous irradia- tion over the whole applicator to avoid ineffective treatment or even lethal overdoses. Simple two-dimensional as well as a realistic three-dimensional model of the human intrathoracic cavity were developed in this work. The coefficient of variation of the photon density on the surface of the applicator was used as an indicator for the homogeneity of illumination. The proposed method produces reasonable optode configurations resulting in homoge- neous irradiation, although the number of required optodes is relatively high. Nevertheless, the algorithm is very efficient and does not need initial configuration. Using the same algorithm for other objectives is straightforward, and thus it could also be used for other design optimization problems without much additional effort.
Although soft field applications (SFAs) have the potential to be a useful non-invasive tool in clinical practice, they are not routinely used in hospitals yet. The two main problems of SFA are their low signal-to-noise ratio and their higher computational effort as compared to hard field applications.
This thesis presents two specific applications to illustrate approaches to overcome some of the specific problems in design optimization for SFA. Two articles in peer-reviewed journals provide the foundation for this work.
The first application deals with design optimization for BIA. It demonstrates how the estimation of body fats (BFs) and fat free mass (FFM) using impedance measurements can be improved. Using a realistic three- dimensional finite element (FE) model of the human thorax combined with a hierarchical model which directly incorporates all parameters of interest, a sensitivity analysis was carried out. The analysis revealed a high sensitivity to the subcutaneous fat layer thickness (SFL) and the fat content of the mesentery. The presented structural model is a prerequisite for a comprehensive investigation of body composition measurements and helps to gain further insight into BIA methods. Although this is a promising result, the intrinsic problem with BIA, namely the dependency of the results on the hydration state, still remains. Furthermore, the influence of changes in the model geometry and the effect of models representing different values for the SFL has not been examined yet.
The second application shows the applicability of a novel mathematical method for the optimization of both the location as well as the strength of optical fibers for photodynamic therapy (PDT). This therapy combines light of a specific wavelength with a photosensitizer to chemically destroy tissue cells. Most commonly, it is employed for various oncological and dermatological treatments. PDT requires extremely homogeneous irradia- tion over the whole applicator to avoid ineffective treatment or even lethal overdoses. Simple two-dimensional as well as a realistic three-dimensional model of the human intrathoracic cavity were developed in this work. The coefficient of variation of the photon density on the surface of the applicator was used as an indicator for the homogeneity of illumination. The proposed method produces reasonable optode configurations resulting in homoge- neous irradiation, although the number of required optodes is relatively high. Nevertheless, the algorithm is very efficient and does not need initial configuration. Using the same algorithm for other objectives is straightforward, and thus it could also be used for other design optimization problems without much additional effort.
| Originalsprache | englisch |
|---|---|
| Qualifikation | Doktor der Technik |
| Gradverleihende Hochschule |
|
| Betreuer/-in / Berater/-in |
|
| Publikationsstatus | Veröffentlicht - 2 Feb. 2016 |