M. Osypka, E. Gersing, and K. Meyer-Waarden. Komplexe elektrische Impedanztomografie im Frequenzbereich von 10 Hz bis 50 kHz. In Zeitschrift für Medizinische Physik, vol. 3, 1993
J. M. Osypka. Application of Deep Neural Networks for Clinical ECG-based Identification of Atrial Fibrosis. Institute of Biomedical Engineering, Karlsruhe Institute of Technology (KIT). Masterarbeit. 2021
Abstract:
Atrial fibrillation (AF) is the globally most prevalent cardiac arrhythmia and is characterized by the chaotic electrical excitation of atrial tissue which results in abnormal atrial contrac- tions. AF causes atrial contractions to lose synchronization with ventricular contractions, effectively reducing cardiac output and thus perpetuating an increased risk of dizziness, fatigue, and even stroke. The onset, development, and maintenance of AF is linked to the presence of fibrosis in the atria, as atrial fibrotic substrate forms conduction barriers and regions of slow conduction which enhance the development of re-entry circuits. These re-entry circuits embody the characteristic atrial electrical activity associated with AF. Since the presence of fibrotic substrate in the atria is linked to AF, the identification of this fibrotic atrial substrate would therefore identify prospective patients who are at risk of developing AF in the future. Preventative treatment of these patients could therefore begin sooner. The state of the art demonstrates the successful application of deep learning (DL) and neural networks for the classification of healthy and pathological electrocardiograms (ECGs). Ad- ditionally, P wave features extracted from ECG signals have been shown to correlate with the presence of atrial fibrosis. This work aimed to compare a P wave feature based neural network with two DL neural networks for the binary classification of healthy and fibrotic P wave signals which stemmed from simulated and clinical ECG datasets. The feature based method served as a benchmark for the two DL methods and required the initial extraction of 15 P wave features: the P wave duration, P wave dispersion, P wave terminal force in lead V1, and the maximum P wave amplitudes of all 12 ECG leads. The first DL method was a long short-term memory neural network (LSTM) and was implemented to investigate its classification performance and its automatic temporal feature extraction capabilities from P wave time series. The second DL method was a convolutional neural network (CNN) in which the pre-trained AlexNet was applied to explore its classification performance as well as its automatic spectral feature extraction capabilites from P wave scalograms. These P wave scalograms were obtained by applying the continuous wavelet transform (CWT) to the input P waves. ...
J. Osypka. EIT sensitivity analysis of local pulmonary blood flow in front of realistic background tissue distributions in a porcine model. Institute of Biomedical Engineering, Karlsruhe Institute of Technology (KIT). Bachelorarbeit. 2019
Abstract:
Electrical Impedance Tomography (EIT) is a radiation-free and non-invasive imaging method suitable for monitoring lung function at the bedside. The interest for EIT in the medical community is grounded in the potential ability to monitor not only pulmonary ventilation, but also pulmonary perfusion. EIT is already a well-established method for monitoring pulmonary ventilation and is currently in a research state for monitoring pulmonary perfusion. The successful simultaneous monitoring of both of these physiological processes would assist in the diagnosis of pulmonary diseases as well as aid in the optimization of mechanical ventilator settings during recruitment maneuvers. This is especially important for patients who’s lungs demonstrate an imbalance of pulmonary perfusion and ventilation, such as those diagnosed with Acute Respiratory Distress Syndrome (ARDS). A significant problem with monitoring pulmonary perfusion with EIT is the influence that the amount of electrical contrast between blood and the background tissue has on the sensitivity of EIT measurements. Therefore, the goal of this thesis was to analyse and quantify the effect that various realistic background tissues have on sensitivity, in response to a local blood volume change. To achieve this goal, EIT forward simulations were performed on three porcine FEM models representing various lung health states: healthy lungs, single-side ventilated lungs, and collapsed lungs. Additionally, each model was separately simulated using pulmonary meshes with a homogenous uniform conductivity or with a realistic heterogenous conductivity distribution. The FEM models were generated from the porcine CT data. Highly conductive spheres (110% of background tissue’s conductivity) were integrated into the pulmonary meshes of each state to simulate a local change in blood volume and therefore a local relative increase in conductivity. The results of the forward simulation were used to calculate the sensitivity to the blood volume changes for each model. Three expectations were formulated and investigated for the sensitivity results. The first stated that pulmonary regions in the proximity of the electrode belt should be more sensitive than regions close to the base or apex of the lung. The results of the simulations were in agreement with this expectation. The second expectation stated that, in comparing the realistic and homogenous pulmonary meshes within individual states, a general difference in spatial sensitivity distributions between the two types of pulmonary tissue mesh should be evident. Additionally the realistic models should be cumulatively less sensitive than the homogenous models. This expectation was also supported by the results. The final expectation was that regions of high conductivity (collapsed lung areas) should be less sensitive than regions of high conductivity. However, the results’ accordance with the third expectation varied from state to state. To model a clinical scenario, an additional comparison between varying background tissues was made to highlight the effect that background tissue might have on sensitivity, and therefore influence the reconstructed images viewed by a physician, evaluating lung function before and after recruitment maneuvers. It was found that by simply changing the background tissue from a well ventilated lung to a collapsed lung in a dorsal region at the electrode belt level, that there was a difference in sensitivity of about 0.0135 mV /S. In a different dorsal lung region at the electrode belt level, a change from collapsed pulmonary tissue to well ventilated pulmonary tissue caused a difference in sensitivity of about 0.0200 mV /S.