Verbesserung der Ablationsstrategie zur Behandlung von ventrikulären Arrhythmien mit Hilfe der Vielkanal EKG und Lösungen des inversen Problems
Final Report Abstract
ECG imaging is an emerging technique that may be used in the future to visualize the electrical activity of the heart without the need of interventional measurements. Cardiologists may use it for therapy planning of cardiac arrhythmia, a pathological condition of the heart where rhythm disorders result from focal excitation origins of electrical activity or electrical pathways through the substrate of the cardiac muscle. To guide the catheters, which cardiologists use to heat and hence silence the electrically active substrate locally (radio-frequency ablation), these excitation origins are visualized on a 3-D anatomical model of the heart. So far, therapy planning has been guided by the standard 12-lead ECG and by the locally and hence invasively measured catheter signals of endocardial ECGs (electrograms). To go from ECG to ECG imaging 80 ECG channels are used, with electrodes on the entire body surface. The cardiac electrical activity is made visible in 3-D. This requires models of the entire thorax and the heart which are made from MRI scans and which are assigned tissue-specific conductivities. Clinical Application and Validation: To make this project possible, data from MRI and 80-channel ECGs had therefore to be measured and processed. This includes signal filtering, the marking of tissues in the MRI scans and the marking of events in the ECGs. To validate the ECG imaging approach, single events in the ECG were then related to the electrical activity that was actually measured in the heart at the same time. To this end, catheter measurements were recorded, selected by their quality and then integrated into the 3-D heart model using vessels and the shapes of cavities. Results of ECG imaging could then be compared with the origins of the electrical activity measured in the heart. Imaging and measurements were performed for the main chambers of the heart (ventricles), and excitation origins could be localized with 12 to 24 mm accuracy, depending on the patient. The challenging part in this study was to get all contributions to the models from different imaging and measurement techniques to the level of quality and to the accuracy of mutual co-registration that is required for ECG imaging. In particular for the electrodes we achieved advances here, as we made a camera system to identify them automatically and to assign them the respective locations on the thorax model. Challenges, New Algorithms and Methods: To compute electrical activities in the heart from the bodysurface ECG, mathematical solvers have to deal with the underdeterminedness of the solution and the strong attenuation of cardiac signals by the body. This is especially true for cardiac signals of high spatial frequency, and hence the ECG imaging results can usually only image the cardiac activity up to a relatively low resolution, i.e. results are smoothed out. In addition to existing techniques, we have developed different strategies to still find accurate solutions: solvers that introduce constraints from physiological knowledge, such as the typical range of transmembrane voltages, and solvers that make use of models of the cardiac excitation spread, which had to be weighted against the ECG data, which was analyzed for the time interval of the entire cardiac excitation spread. Analyzing all these time steps allows for an imaging of the excitation wavefronts — or the activation times at which each point in the heart is reached by the excitation wavefront. For solvers of activation times, we have compared results with the activation times detected in electrogram measurements with the catheters in the heart. We have also demonstrated how ECG imaging can be combined with such measurements, and we have shown that excitation origins can be detected more accurately with body surface and intracardiac information available. With the demonstrated methods, ECG imaging can be used for completely non-invasive treatment planning of radio-frequency ablation therapies before the intervention, and it can produce improved maps of electrical activity in the entire ventricle from only a small number of catheter measurements when the patient is already undergoing a treatment session.
Publications
- “A new method for choosing the regularization parameter in the transmembrane potential based inverse problem of ECG", In Proc. CinC, pp. 29- 32, 2012
D. Potyagaylo, W. H. W. Schulze, and O. Dössel
- “Activation time imaging in the presence of myocardial ischemia: Choice of initial estimates for iterative solvers", In Proc. CinC, vol. 39, pp. 961–964, 2012
W. H. W. Schulze, D. Potyagaylo, and O. Dössel
- “Electrode arrangements for ECG imaging under practical constraints of a catheter lab setting", In Proc. Biomed Tech, vol. 58(s1), 2013
W. H. W. Schulze, R. Schimpf, T. Papavassiliu, D. Potyagaylo, E. Tulumen, B. Rudic, V. Liebe, C. Doesch, T. Konrad, C. Veltmann, M. Borggrefe, and O. Dössel
(See online at https://doi.org/10.1515/bmt-2013-4156) - “Kalman Filter with Augmented Measurement Model: an ECG Imaging Simulation Study", In Proc. FIMH, LNCS 7945, pp. 200–207, 2013
W. H. W. Schulze, F. Elies Henar, D. Potyagaylo, A. Loewe, M. Stenroos, and O. Dössel
- “Noninvasive Localization of Ectopic Foci: a New Optimization Approach for Simultaneous Reconstruction of Transmembrane Voltages and Epicardial Potentials", In Proc. FIMH, LNCS 7945, pp. 166–173, 2013
D. Potyagaylo, M. Segel, W. H. W. Schulze, and O. Dössel
- “Effect of mesh resolution on forward calculations of the electrocardiogram in a simplified thorax model", In Proc. Biomed Tech, 2014
W. H. W. Schulze, T. Fritz, D. Potyagaylo, J. Trächtler, R. Schimpf, T. Papavassiliu, E. Tülümen, B. Rudic, V. Liebe, C. Doesch, M. Borggrefe, and O. Dössel
(See online at https://doi.org/10.1515/bmt-2014-4392)
