BACKGROUND AND OBJECTIVE: Progress in biomedical engineering has improved the hardware available for diagnosis and treatment of cardiac arrhythmias. But although huge amounts of intracardiac electrograms (EGMs) can be acquired during electrophysiological examinations, there is still a lack of software aiding diagnosis. The development of novel algorithms for the automated analysis of EGMs has proven difficult, due to the highly interdisciplinary nature of this task and hampered data access in clinical systems. Thus we developed a software platform, which allows rapid implementation of new algorithms, verification of their functionality and suitable visualization for discussion in the clinical environment. METHODS: A software for visualization was developed in Qt5 and C++ utilizing the class library of VTK. The algorithms for signal analysis were implemented in MATLAB. Clinical data for analysis was exported from electroanatomical mapping systems. RESULTS: The visualization software KaPAVIE (Karlsruhe Platform for Analysis and Visualization of Intracardiac Electrograms) was implemented and tested on several clinical datasets. Both common and novel algorithms were implemented which address important clinical questions in diagnosis of different arrhythmias. It proved useful in discussions with clinicians due to its interactive and user-friendly design. Time after export from the clinical mapping system to visualization is below 5min. CONCLUSION: KaPAVIE(2) is a powerful platform for the development of novel algorithms in the clinical environment. Simultaneous and interactive visualization of measured EGM data and the results of analysis will aid diagnosis and help understanding the underlying mechanisms of complex arrhythmias like atrial fibrillation.
Atrial fibrillation (AF) is the most common arrhythmia of the heart in industrialized countries. Its generation and the transitory behavior of paroxysmal AF are still not well understood. In this work we examine the interaction of two activation sources via an isthmus as possible cause for the initiation of fibrillation episodes. For this study, the electrophysiological model of Bueno-Orovio, Cherry and Fenton is adapted to atrial electrophysiology, both for physiological and electrophysiologically remodeled conditions due to AF. We show that the interaction of the pacemakers, combined with the geometrical constraints of the isthmus, can produce fibrillatory-type irregularities, which we quantify by the loss of spatial phase coherence in the transmembrane voltage. Transitions to irregular behavior occur when the frequencies of the pacemakers exceed certain thresholds, suggesting that AF episodes are initiated by frequency changes of the activating sources (sinus node, ectopic focus).
F. M. Weber, D. U. J. Keller, S. Bauer, O. Dössel, G. Seemann, and C. Lorenz. Predicting tissue conductivity influences on body surface potentials-an efficient approach based on principal component analysis. In IEEE Transactions on Biomedical Engineering, vol. 58(2) , pp. 265-273, 2011
Abstract:
In this paper, we present an efficient method to estimate changes in forward-calculated body surface potential maps (BSPMs) caused by variations in tissue conductivities. For blood, skeletal muscle, lungs, and fat, the influence of conductivity variations was analyzed using the principal component analysis (PCA). For each single tissue, we obtained the first PCA eigenvector from seven sample simulations with conductivities between ±75% of the default value. We showed that this eigenvector was sufficient to estimate the signal over the whole conductivity range of ±75%. By aligning the origins of the different PCA coordinate systems and superimposing the single tissue effects, it was possible to estimate the BSPM for combined conductivity variations in all four tissues. Furthermore, the method can be used to easily calculate confidence intervals for the signal, i.e., the minimal and maximal possible amplitudes for given conductivity uncertainties. In addition to that, it was possible to determine the most probable conductivity values for a given BSPM signal. This was achieved by probing hundreds of different conductivity combinations with a numerical optimization scheme. In conclusion, our method allows to efficiently predict forward-calculated BSPMs over a wide range of conductivity values from few sample simulations.
The anticholinergic antiparkinson drug orphenadrine is an antagonist at central and peripheral muscarinic receptors. Orphenadrine intake has recently been linked to QT prolongation and Torsade-de-Pointes tachycardia. So far, inhibitory effects on I Kr or cloned HERG channels have not been examined. HERG channels were heterologously expressed in a HEK 293 cell line and in Xenopus oocytes and HERG current was measured using the whole cell patch clamp and the double electrode voltage clamp technique. Orphenadrine inhibits cloned HERG channels in a concentration dependent manner, yielding an IC50 of 0.85 μM in HEK cells. Onset of block is fast and reversible upon washout. Orphenadrine does not alter the half-maximal activation voltage of HERG channels. There is no shift of the half-maximal steady-state-inactivation voltage. Time constants of direct channel inactivation are not altered significantly and there is no use-dependence of block. HERG blockade is attenuated significantly in mutant channels lacking either of the aromatic pore residues Y652 and F656. In conclusion, we show that the anticholinergic agent orphenadrine is an antagonist at HERG channels. These results provide a novel molecular basis for the reported proarrhythmic side effects of orphenadrine
In this work an optimization-based method of modeling the cardiac activity is presented. The method employs a personalized anatomical 3D model of the patients thorax provided by the segmentation of MRI data as well as an electrophysiological model of the heart.Cellular automaton is used to model the propagation of depolarization and repolarization fronts through the myocardium. The form of action potential (AP) curves was previously derived from the coupled myocardium cell models developed by Noble, Priebe-Beuckelmann and ten Tusscher. The results provided by these three cell models are compared.A series of body surface potential maps (BSPMs) is calculated, the signals on the nodes representing the electrodes are recorded, providing thus a simulated multichannel ECG. A root-mean-square of the difference between simulated and measured ECGs is taken as a criterion for optimization of heart model parameters.The method provides a time-dependent distribution of transmembrane voltages within the heart muscle of a patient.
Conference Contributions (13)
S. Bauer, T. Oesterlein, J. Schmid, and O. Dössel. Interactive visualization of cardiac anatomy and atrial excitation for medical diagnosis and research. In Current Directions in Biomedical Engineering, vol. 1(1) , pp. 400-404, 2015
Abstract:
State of the art biomedical engineering allows for acquiring enormous amounts of intracardiac data to aid diagnosis and treatment of cardiac arrhythmias. Modern catheters, which are used to acquire electrical information from within the heart, are capable of recording up to 64 channels simultaneously. The software available for data analysis, however, does not provide adequate performance to neither analyze nor visualize the acquired information in an appropriate manner. We present a software package that fascilitates interdisciplinary collaborations between engineers and physicians to adress open questions about pathophysiological mechanisms using data from everyday electrophysiogical studies. Therefore, a package has been compiled that enables algorithm development using MATLAB and subsequent visualization using the VTK C++ class libraries. The resulting application KaPAVIE, which is presented in this paper, is designed to meet the requirements from the clinical side and has been successfully applied in the clinical environment.
S. Bauer, J.-C. Edelmann, G. Seemann, F. B. Sachse, and O. Dossel. Estimating intracellular conductivity tensors from confocal microscopy of rabbit ventricular tissue. In Biomedizinische Technik. Biomedical Engineering, vol. 58(s1) , 2013
Abstract:
Bidomain simulations of the heart need validated parameters to produce realistic data. Therefore, it is nec- essary to develop methods to estimate reliable values for these parameters. We developed an approach to deliver such values by designing an in-silico model of intracellular electrical conduction based on confocal microscopic data of rabbit ventricular tissue. High resolution image data were used to determine the anisotropy of electrical conduc- tivity in the myocardium, which is highly dependent on the specific tissue geometry. Gap junction protein connexin43 and extracellular space were labeled with fluorescent dyes of different spectra. The myocytes were segmented and the gap junction density in-between myocytes was extracted. Assuming conductivities for intracellular liquid and gap junction resistance, a numerical field calculation was per- formed for three principal directions in order to extract in- tracellular conductivity tensors. We calculated 9 tensors by varying the assumed conductivities by ±50%. We esti- mated the intracellular conductivities for the three princi- pal directions σi,x = 0.0653 S/m, σi,y = 0.0042 S/m and σi,z = 0.0033 S/m, respectively. The estimated conductiv- ity values were realistic regarding the electrical anisotropy but need to be improved to fit other experimental data.
S. Bauer, D. U. J. Keller, F. M. Weber, P. Tri Dung, O. Dössel, and G. Seemann. How do tissue conductivities impact on forward-calculated ECGs? An efficient prediction based on principal component analysis. In IFMBE Proceedings World Congress on Medical Physics and Biomedical Engineering, vol. 25/4, pp. 641-644, 2009
C. Lenk, F. M. Weber, M. Bauer, M. Einax, G. Seemann, and P. Maass. Paroxysmal atrial fibrillation caused by interaction of pacemakerwaves and reduced excitability: Insights from the Bueno-Orovio model adapted to atria. In Computing in Cardiology Conference (CinC), pp. 1079-1082, 2013
Abstract:
As possible cause for atrial fibrillation (AF) we study the influence of a reduced excitability on the interaction of pacemaker waves in the Bueno-Orovio model with parameters adapted to atrial electrophysiology (aBO). One of the two pacemakers represents the sinus node and the other one a self-excitatory source in the left atrium. The pacemakers are spatially separated and their waves get in contact via a small bridge. In previous studies based on the FitzHugh-Nagumo (FHN) model it was shown that three different types of irregular activation patterns can occur in this problem. In the aBO model adapted to physiological conditions only one type is observed because, different from the FHN model, a reduction of excitability due to high-frequency pacing does not occur. If the excitability is reduced in the aBO model, all types of irregularities are recovered and, in addition, a further type is found. Because transitions from regular to irregular behavior depend on the pacing frequency, our findings provide a possible explanation for the phenomenon of paroxysmal AF.
D. Farina, Y. Jiang, O. Dössel, C. Kaltwasser, and W. R. Bauer. Model-based method of non-invasive reconstruction of ectopic focus locations in the left ventricle. In Proc. the 4th European Congress for Medical and Biomedical Engineering, pp. 2560-2563, 2008
Y. Jiang, D. Farina, C. Kaltwasser, O. Dössel, and W. R. Bauer. Modeling and reconstruction of myocardial infarction. In Gemeinsame Jahrestagung der Deutschen, der Österreichischen und der Schweizerischen Gesellschaft für Biomedizinische Technik, 2006
O. Dössel, W. Bauer, D. Farina, C. Kaltwasser, and O. Skipa. Imaging of bioelectric sources in the heart using a cellular automaton model. In Conference Proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference, vol. 2, pp. 1067-1070, 2005
Abstract:
The approach to solve the inverse problem of electrocardiography presented here is using a computer model of the individual heart of a patient. It is based on a 3D-MRI dataset. Electrophysiologically important tissue classes are incorporated using rules. Source distributions inside the heart are simulated using a cellular automaton. Finite Element Method is used to calculate the corresponding body surface potential map. Characteristic parameters like duration and amplitude of transmembrane potential or velocity of propagation are optimized for selected tissue classes or regions in the heart so that simulated data fit to the measured data. This way the source distribution and its time course of an individual patient can be reconstructed.
After myocardial infarction, ischemic lesions within the myocardium can be the origin of malignant arrhythmias by the mechanism of re-entry. Surface-ECG and MR-imaging data can be used to detect and classify such re- gions in a non-invasive way. For this purpose a model of the electric conductivity of the tissues within the pa- tients chest and a model of cardiac sources must be constructed out of MR-imaging data. Employing finite- element algorithms the inverse problem of electrocardiology can then be solved, leading to the reconstructionof electrical sources within the myocardium during the process of depolarisation and repolarisation.
O. Skipa, D. Farina, C. Kaltwasser, O. Dössel, and W. R. Bauer. Fast interactive ECG simulation and optimization-based reconstruction of depolarization in the heart. In Biomedizinische Technik, vol. 49-2, pp. 362-363, 2004
N. Bauer. Verifikation und Optimierung eines Gelatine-Zucker-basierten Kopfphantoms für die mikrowellenbasierte Bildgebung. Institut für Biomedizinische Technik, Karlsruher Institut für Technologie (KIT). Bachelorarbeit. 2015
Abstract:
The disease pattern of stroke has a huge impact on today's society. It is the fourth most frequent cause of death in Germany and causes long-lasting effects for the affected person. The degree of irreversible damage majorly depends on the idle time until a differential diagnosis of the disease is performed. Currently such a diagnosis can only be executed by the imaging techniques MRT (Magnetic Resonance Tomography) and CT (Computer Tomography). Both techniques are cost-intensive and can only be used in certain locations, i.e. in hospitals or special stroke units. The transport of the patient to these locations takes a significant amount of valuable time. Therefore research for a technology which is faster, cheaper and usable in a mobile system is brought forward. Microwave imaging shows promise in these aspects. Systems using microwave imaging are under developement and the mode of action needs to be verified through tests on realistic human head phantoms. In a previous bachelor thesis, a phantom of the human head has been developed which has dielectric properties similar to those of the different tissues of a human head. The materials used were clay for the scull and compositions of water, sugar and gelatine for soft tissues.Different permittivity values for different tissues are essential for high-contrast measurements with microwave imaging technology. On this account an important factor for the phantom and the dominant topic of this thesis is diffusion between the materials for soft tissues. In addition skull as well as the border between skin and the measuring antennas were analysed regarding its permittivity.
S. Bauer. Estimating intracellular conductivity tensors from fluorescent labeling and three-dimensional scanning confocal microscopy data of rabbit tissue. Institute of Biomedical Engineering, Karlsruhe Institute of Technology (KIT). Bachelorarbeit. 2012
Abstract:
In this work, an in-silico model of conduction, based on confocal microscopic data of rabbit ventricular tissue, was developed. Confocal microscopy combined with fluorescent tissue labeling are able to image the tissue geometry with high resolution. This information was used to gain further insights into the anisotropy of electrical conduction in the myocardium, which is highly dependent on the specific tissue geometry. The cell nucleus, gap junction connexin43, fibroblasts and collagen were labeled with fluorescent dyes of different spectra. In a previous work, these data were used to quantify the volume fractions of myocytes, fibroblasts and the extracellular space. Additionally, the extracellular conductivity tensor was estimated and the amount of coupling gap junctions in the vicinity of fibroblasts was quantified. In this work, the intracellular conductivity tensors was estimated from the confocal microscopic data. Therefore, the myocytes were segmented and the gap junction density in-between myocytes was extracted. With assuming conductivities for intracellular liquid and gap junction resistance, a numerical field calculation was performed for three principal directions in order to extract intracellular conductivity tensors.
M. Bauer. Simulation von chaotischen Erregungsmustern bei Vorhofflimmern mit einem minimalen Zellmodell. Karlsruhe Institute of Technology (KIT), Institut für Biomedizinische Technik. Bachelorarbeit. 2010
S. Bauer. Efficient reconstruction of BSPMs and optimization of multiple tissue conductivities. Institut für Biomedizinische Technik, Karlsruher Institut für Technologie (KIT). Diplomarbeit. 2009
Abstract:
It is well-known that tissue conductivities impact decisively on the body surface potentials obtained from solving the forward problem of electrocardiography. However, published conductivity values show considerable variations. It is a computationally intensive task to determine the effect of an organs conductivity on the body surface potential map by repetitive forward calculations.In this diploma project an efficient method to reconstruct body surface potentials resulting from forward-calculations with different tissue conductivities was developed and implemented. The method was extended to predict confidence intervals for the BSPM depending on the uncer- tainty of tissue conductivities. Furthermore, a new approach for non-invasively estimating tissue conductivities by using an optimization strategy was evaluated.The presented methods were implemented as computer programs in a mixed Perl and C++ envi- ronment. The aim was to provide a simple handling and to minimize user interactions. Principal Component Analysis was used to identify the effect of an organs conductivity on the body surface potentials. The signal variation caused by a change in tissue conductivities was accounted for by a PCA score and the first principal component. The variation pattern was determined by the first principal component and the variation amplitude by a PCA score. Investigations showed that it was possible to calculate a reliable interpolation function for the scores from a small number of PCA input simulations. Body surface potentials could be reconstructed from an average signal and the first eigenvector scaled by a particular interpolated score. The method achieved a significant reduction in computation time compared to a brute-force method when all possible tissue conductivity combinations would have to be evaluated. The reconstruction method was studied if up to four tissue conductivities were varied simultaneously. In all cases, the reconstruction method showed good results, accurately reconstructing body surface potentials with a minimal root mean squared error compared to the correctly forward-calculated signals. This approach allowed for efficiently predicting the impact of tissue conductivities on the body surface potentials for a wide range of conductivity values from few sample simulations.During this project, a method was developed to specify a confidence interval within which the BSPM signal would lie if a given uncertainty range was assumed for multiple tissue conductivities. The minimum and maximum signals bounding the confidence interval were made up of synthetic signals to account for the unknown positive or negative effect on the signal amplitude if the tissue conductivity was increased. This confidence interval allows for the quantification of a potential model tolerance. With an optimization strategy it was possible to calculate tissue conductivities of several organs simultaneously by comparing a reference signal to a reconstructed signal. A selectable error met- ric was minimized by an optimization algorithm. Tissue conductivities could be calculated with minor errors if the exact cardiac electrophysiology was known and if the considered tissues had a high impact on the body surface potentials. This was true for simulated signals with variant conductivities of blood, lungs and skeletal muscle. The optimization results were less trustful if the cardiac electrophysiology was not exactly known and it failed in the case of in-vivo measured signals. However, this was not a limitation of the optimization method, but the problem was that the modeling of patient electrophysiology was not precise enough. It was also demonstrated that a variation in tissue conductivities could not compensate for errors in the modeling of the pa- tients electrophysiology. This implied that cardiac electrophysiology and tissue conductivities had a different impact on the body surface potentials.