Atrial fibrillation (AFib) is a cardiovascular disease and therefore constitutes one of the most frequent causes of death in the western world. The mechanisms behind AFib are not yet sufficiently understood .
To investigate and treat these arrythmias, electrophysiological (EP) studies are conducted. Catheters are inserted into the heart in a minimally invasive fashion to acquire EP data. Up to 64 poles record intracardiac electro-grams (EGMs), which accumulates to thousands of measurements in the course of a few minutes . The acquired EP data can potentially be used for further analysis to improve the understanding of the underlying pathophysiological mechanisms .
In recent work, EP data was analyzed and evaluated with novel data to approach common clinical questions   . Other research groups presented visualization software for similar problem sets, but with a very specific and narrow field of application .
Our proposed software framework delivers a novel and more universally applicable method to facilitate interdisciplinary communication of clinical expectations and technical inputs provided by engineers. By visualizing the EP data and results of analytical algorithms of the data in an interactive platform, both engineers and clinicians can express their perspectives more easily. The engineer’s desire to produce diagnostically relevant information in the form of descriptive visualizations can be better met by instantly applying the clinician’s judgement of the helpfulness of the data.
In many research facilities, the industry standard for data analysis is MATLAB. However, a common method to provide highly adjustable visualizations is to utilize the powerful C++ class libraries from the Visualization Toolkit (VTK) by Kitware . With our current implementation of an interactive visualization platform, KaPAVIE (Karlsruhe Platform for Analysis and Visualization of Intracardiac Electrograms), we aimed for combining the best of both worlds.
2 Clinical requirements
2.1 Clinical data acquisition
The acquired data from EP studies comprises the atrial anatomy, given as a triangulated surface mesh, the position of all catheters and their electrodes continuously recorded in segments and the electrograms (EGMs) acquired from each electrode. Recordings from different sites along the atrial wall can be acquired during sequential mapping, in which synchronization can be performed using a reference channel with stable location. The analytical data is collected as surface maps. The system used in the clinical institution Staedtisches Klinikum Karlsruhe is the EnSite Velocity mapping system (St. Jude, MN, USA), which acquires electrical information at 2034.5 Hz and locations of electrodes at a sampling rate of about 100 Hz. The data format for later analysis and visualization is given as CSV and XML files.
2.2 Clinical visualization standards
First requirement of clinical usage is the ability to comfortably visualize the raw clinical data. Thus physicians need to be provided with this type of familiar signals if the automatic analysis recommends a certain region for closer surveillance. Atrial anatomy has to be visualized in 3D, together with the position of measurement nodes. These can be either all points of a sequential map or the current electrode positions of multiple catheters. Intracardiac EGMs should be visualized if the user chooses an electrode via mouse click and displayed synchronously to both the surface ECG and the intracardiac reference to aid diagnosis. It should be possible to scale the EGM since the signals may cover two orders of magnitude. Furthermore, it is necessary to visually link both EGMs and measurement positions to aid geometrical understanding of a given situation. With respect to decision support and algorithm development, the results of signal analysis need to be visualized. This comprises both static and dynamic features of the signals and the atrial surface regions, respectively. Colour maps and value ranges need to be adjustable depending on the type of data visualized. Well-known parameters in clinical use already have commonly used colour maps which are to be utilized here as well. For time dynamic data, speed of visualization should be adjustable. An indicator should link the currently displayed time stamp of the three-dimensional animation and the two-dimensional signal plot.
3 Software implementation
3.1 Biomedical signal processing and analysis
Analysis of clinical EP data affected both geometrical and EGM specific tasks. Geometrical analysis focused on the position of measurement electrodes, their spatial stability over time and the amount of atrial surface covered by the multiple measurements. Goal of signal analysis was to extract diagnostically relevant features from the intracardiac EGMs. Signal processing and analysis was done via MATLAB (MATLAB version R2014b) . MATLAB source code was optimized so that the data for visualization could be generated reliably in about 2 minutes. Once data analysis and processing was finished, the resulting 3D geometry along with the processed EGMs and surface maps were saved and exported to a file format that was compatible with the visualization environment. Geometrical data was saved as a polygonal mesh. Data was saved in binary format to optimize hard drive usage for the large amount of clinical data. The work flow established to acquire and visualize the EP data is briefly illustrated in fig. 1.
3.2 Graphical user interface
The graphical user interface (GUI) application is the core feature of our visualization and data analysis platform. It facilitates the clear communication between clinicians and engineers. Analytic methods can thus be efficiently adjusted to diagnostic purposes. Fig. 2 shows a screen snapshot of the GUI application for an example data set. The GUI was implemented with the Qt C++ UI framework (Qt version 5.3.0). It allows for easy integration of fast and responsive GUIs for cross-platform applications. Furthermore, Qt is open-source and contributes from its large developer community. The most significant advantage for our purposes was the possibility to utilize readily available source code to integrate the visualization of the 3D data and plotting of the EGMs. Several plugins were used to develop the GUI application to meet the diversity of features expected by the clinicians. A plugin to include an existing VTK Qt render window was used to visualize the 3D geometry of the atria and their respective surface data maps. EGMs were plotted using the open source plugin QCustom-plot , licensed under the GPL.
3.3 Visualization pipeline
Using the algorithms of the VTK C++ software library (VTK version 6.1.0), the visualization of the 3D geometry was implemented by using a VTK visualization pipeline and displayed in a VTK render widget. The data provided comprised three VTP files (atrial geometry, surface maps and EGM information) and one XML file for data specific settings such as color maps and opacity. Once the user selected an appropriate set of files, the geometrical polygonal data that was defined by point and cell coordinates was loaded via the visualization pipeline. The well-established VTK procedure to input data was used, which includes reading the data and converting it into a data object, which then could be associated with a mapper. The mapper converts the geometrical data object to a “tangible” object, which is then assigned to an actor. The actor serves the purpose of adjusting visible properties of the geometry such as coloring, size and rendering method. The actor is added to a renderer, which is associated with certain interaction styles such as selecting actors and translating and rotating of the geometry. The spatial locations of the catheter measurement points were represented by spherical polygons which could be selected by the user to display the corresponding EGM plots. A customized interaction mode was implemented so that the measurement nodes could be selected via mouse click. Selection is visualized by coloring and enlarging the radii of the spherical polygons.
In order to promote usability for engineers and physicians, one of the main aspects considered during development was to support as many platforms as possible. Open source frameworks like Qt and VTK were chosen to develop the software so that it can be compiled and executed in different operating systems, namely Microsoft Windows, Macintosh OS X and most Linux distributions.
The software does not need any manual configurations or changes on system settings in order to function properly. It can easily be executed in a common way directly in the installation directory. On Macintosh OS X the software was bundled into a single application (*.app) package and can thereby be easily transferred between different systems by moving the package like native Macintosh software. For goal number one of the project, namely visualization of analysis results, no additional software is required on the used computer. If the complete analysis workflow is to be performed using KaPAVIE, a preinstalled MATLAB version is required on the system due to license limitations. Most scientific institutions who will benefit from an improved interdisciplinary communication in a research environment, however, are expected to already own a license for this application.
4 Clinical application
The GUI application was designed so that the user can interactively choose the preferred data representation. The user can choose an arbitrary number of measurement points and display their respective EGMs comparatively in the plotting widget. This allows the clinician to associate EGM data and 3D surface signals. The plotting widget can be adjusted to fit certain resolutions of the time axis (measured in [mm/sec]) common in clinical practice. Clinicians are familiar with these specific resolutions and can therefore more intuitively relate to the plotted EGMs. EGM amplitudes can be normalized for qualitative analysis, displayed with their original magnitude or arbitrarily scaled by the user for a close-up look. Normalization is especially useful when displaying data that varies over the order of two magnitudes. The plotting widget can also display a scale bar that shows the quantitative amplitudes of a selected graph. This ensures quantitative context even when scaled to arbitrary units.
Coloring of the EGM graphs is either matched to the color of the 3D node representation or can be chosen to uniformly represent all EGMs with one color. Fitted coloring is best applied when visualizing measuring points from different regions of the intracardiac surface, whereas uniform EGM graph coloring is useful when displaying measuring nodes close to each other. The user can choose a desired surface data representation, e.g. LAT map, activity map, coverage map and others, depending on which maps are available for the data set at hand. This allows for comparison to evaluate diagnostic value of the different data representations.
Animations like an energy movie or electrical activity distributions can be played in synchrony with the respective EGM plots. Displaying EGM plots during visualization is important because EGMs are the trusted form of diagnostic information for clinicians. Animation speed can be adjusted as the user desires. Animation loops can be played between arbitrary points in time. This gives the user a variety of options to analyze specific time spans of the EP data visualization as well as the aforementioned association of EGMs and 3D visualization not only in space but also in time. Furthermore, the user is given the possibility to adjust the visualization to his or her desire such as the line width of the EGM plots. For demonstrative purposes, video sequences of the visualization can be exported.
The software platform KaPAVIE was successfully applied in the clinical environment of two institutions. Data from ablation procedures of atrial flutter and atrial fibrillation was analyzed retrospectively, with all patients providing informed consent. EGM information of up to 128 channels was successfully analyzed, acquired by multiple catheter types like standard 4 pole ablation catheters, 20 pole double loop spiral catheters and two 64 pole basket catheters during simultaneous biatrial mapping. Several algorithms were implemented in MATLAB and applied to extract diagnostically relevant features of the recorded EGMs.
In this contribution, we presented a new software platform for the analysis and visualization of intracardiac mapping data. A strong collaboration between engineers and physicians is necessary in the field of biomedical engineering to understand the mechanisms behind atrial arrhythmias. Our platform should enable engineers to develop novel algorithms in the common and widely spread MATLAB environment. In the communication between engineers and physicians, extensive visualization plays and important role: It allows engineers to grasp the ideas emerging from clinical practice, discuss the outcomes of data analysis with physicians and ultimately enables physicians to use the results for diagnosis and improved treatment. This platform should be the basis for future development of algorithms which hopefully will aid in diagnosis and treatment of cardiac arrhythmias.
Conflict of interest: Authors state no conflict of interest. Material and Methods: Informed consent: Informed consent has been obtained from all individuals included in this study. Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance the tenets of the Helsinki Declaration, and has been approved by the authors’ institutional review board or equivalent committee.
January, C.T. et al.: AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: Executive Summary: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. Circulation (2014)
Ng, J.; Gordon, D.; Passman, R.S.; Knight, B.P.; Arora, R.; Goldberger, J.J..: Electrogram morphology recurrence patterns during atrial fibrillation. Heart Rhythm. 11 (11), pp. 2027-2034 (2014)
Narayan, S.M. et al.: Clinical mapping approach to diagnose electrical rotors and focal impulse sources for human atrial fibrillation. Journal of Cardiovascular Electrophysiology, 23(5):447–454 (2012
Nakagawa, H. et al.: Rapid high resolution electroanatomical mapping: evaluation of a new system in a canine atrial linear lesion model. Circulation: Arrhythmia and Electrophysiology, 5:417–424 (2012)
Oesterlein, T. et al.: Neue Möglichkeiten zur Diagnose von Arrhythmien durch Visualisierung der zeitlichen Dynamik von Elektrogrammen. Proceedings Jahrestagung der Deutschen Gesellschaft für Kardiologie, 103:V167 (2014)
Ravelli, F. et al.: Computational mapping in atrial fibrillation: how the integration of signal-derived maps may guide the localization of critical sources. Europace, 16(5):714–723 (2012)
Cantwell, C.D. et al.: A software platform for the comparative analysis of electroanatomic and imaging data including conduction velocity mapping. Engineering in Medicine and Biology Society (EMBC), 36th An. Intl. IEE Conf., 1591–1594 (2014)
Schilling C., Keller M., Oesterlein T., Scherr, D., Haissaguerre M., Schmitt C., Luik A.: Fuzzy Decision Tree to Classify Complex Fractionated Atrial Electrograms. Biomedical Engineering/ Biomedizinische Technik, accepted Feb 2015
Schroeder, W. et al.: The Visualization Toolkit, 3rd Edition. Kitware, Inc. (2003)
MATLAB and Signal Processing Toolbox Release 2014b. The MathWorks, Inc., Natick, Massachusetts, United States.