Interact CardioVasc Thorac Surg 2007;6:274-279. doi:10.1510/icvts.2006.142869
© 2007 European Association of Cardio-Thoracic Surgery
Work in progress report - Arrhythmia |
Detection of atrial arrhythmia in superconducting quantum interference device magnetocardiography; preliminary result of a totally-noninvasive localization method for atrial current mapping
Doosang Kima,*,
Kiwoong Kimc,
Yong-Ho Leec and
Hyuk Ahnb
a Department of Thoracic and Cardio-vascular Surgery, Seoul Veterans Hospital, 6-2 Dunchon-dong Kangdong-gu, Seoul, 134-060, Korea
b Department of Thoracic and Cardio-vascular Surgery, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Korea
c Bio-signal Research Center, Korea Research Institute of Standards and Science, Daejeon, Korea
Received 28 August 2006;
received in revised form 28 January 2007;
accepted 29 January 2007
 Presented at the joint 20th Annual Meeting of the European Association for Cardio-thoracic Surgery and the 14th Annual Meeting of the European Society of Thoracic Surgeons, Stockholm, Sweden, September 10–13, 2006.
*Corresponding author. Tel.: +82-2-2225-1346; fax: +82-2-477-5605.
E-mail address: mdksr{at}paran.com; mdksr{at}lycos.co.kr (D. Kim).
 |
Abstract
|
|---|
Map-guided surgery is the goal for treatment of atrial fibrillation (AF), because it minimizes unnecessary incisions or procedures. We propose a totally-noninvasive and even non-contact method to detect atrial arrhythmia with a superconducting quantum interference device magnetocardiography (MCG) system, and report the first clinical application case of MCG map-guided AF surgery. To detect weak atrial excitation, we utilized a high sensitive 64-channel MCG system measuring tangential magnetic field components, which is known to be more sensitive to a deeper current source. We measured the MCG signals from eight patients with chronic AF. Then, we separated the f-wave from the other components by using independent component analysis. The extracted f-wave caused by reentrant myocardial excitation was three-dimensionally localized on the mesh model of a human heart by a novel beamformer technique having a surface action potential activity as its filter output. We localized the abnormal stimulation source of an atrial arrhythmia non-invasively and visualized the current source distribution corresponding to the atrial excitation successfully on the three-dimensional atrial surface, which was separated from the ventricular excitation. Using this atrial mapping, we underwent minimal AF surgery in three patients and converted their AF to sinus rhythm successfully.
Key Words: Atrial fibrillation; Magnetocardiography; Map-guided surgery
|
1. Introduction
|
|---|
Over seven million patients die annually from cardiac arrhythmias worldwide, and many more are disabled [1]. Atrial fibrillation (AF) causes significant morbidities, such as stroke, other thromboembolic events, hemodynamic compromise, ventricular tachycardia, and so on. Approximately 30–40% of patients with mitral valve disease have AF and 5% of patients having cardiac surgery have AF. To decrease this significant morbidity associated with AF, medical therapy, catheter ablation, and surgical intervention such as Cox maze, ablation with radiofrequency or microwave are usually accepted. Drug intolerance, arrhythmia intolerance, and recurrent embolic events are surgical indication for this disease [2]. The patients with symptomatic lone AF and those with AF associated with other organic cardiac disease are generally recommended for surgical intervention. Although the Maze procedure is highly effective, the complexity, long incision, prolonged CPB time, and bleeding risk of the operation prevented widespread application [3]. Current surgical interventions are basically not guided by electro-physiologic findings in an individual patient and thus might include unnecessary incisions. Recently, a 256-channel 3-dimensional dynamic mapping system with intra-operative epicardial patch electrodes was used by Nitta et al. [4], but it is useful only during the operation, so it is still invasive, difficult for follow-up, and does not show normal activity under complete physiological conditions, such as closed chest, no anesthesia, presence of neural inputs, mechanical loading, and normal perfusion in intact human subjects. Electrocardiographic imaging with 224 multi-electrode vest records is studying for map-guide by Ramanathan et al. [5], but it has also limitation from the electrode contact problems, such as missing or noise signals. We propose a totally-noninvasive, non-contact method to detect atrial arrhythmia with a superconducting quantum interference device magnetocardiography (MCG) system, and report the first clinical application case of MCG for map-guided AF surgery.
The high sensitivity of the superconducting quantum interference device has enabled us to measure very weak myocardial magnetic fields generated by the ionic currents flowing inside the myocardium. By solving an inverse problem for the fields measured with a multi-channel superconducting quantum interference device array, we can estimate the electrical excitation of the human myocardium non-invasively, we call the technical device a magnetocardiograph (MCG) [6]. The MCG has many merits over the electrocardiograph (ECG); the MCG is more sensitive to tangential currents, curl currents, and current flow between the endocardium and the epicardium. Particularly, the MCG can measure a closed-looped current on the myocardium while the ECG records it as a zero-potential. In addition, it is less dependent on the variation of the conductance outside of the heart and is a fully non-contact technique, so it does not have an electrode contact problem. Such merits make the MCG a useful diagnosing tool for heart diseases such as coronary artery disease [7, 8].
Examining atrial arrhythmias is another challenging topic with MCG. As the atrial arrhythmias such as AF and atrial flutter generally have irregular rhythm and atrio-ventricular conduction, the MCG signal cannot be improved by QRS-detection averaging; therefore, a superconducting quantum interference device MCG system having a high signal-to-noise ratio is required to measure informative atrial excitation with a single scan. In the case of AF, diminished f-waves are much smaller than normal p-waves because the sources are usually located on the posterior wall of the heart and only small parts of the wall contribute to the signal. Therefore, in this study, we utilize an MCG system measuring tangential field components, which is known to be more sensitive to a deeper current source [9].
In this article, we propose a new localization method for the atrial arrhythmia, which utilizes the undistorted magnetic fields from human heart. We call the method a separative surface potential activation beamformer, which visualizes the activity map of the action potential on the myocardium corresponding to a specific waveform of measured magnetic fields. We apply the separative surface potential activation beamformer to the localization of the rhythmic excitation in the AF. The three-dimensional activity map of action potential is rendered with a color map on a standard heart model.
|
2. Materials and methods
|
|---|
From June 2005 to January 2007, eight patients having chronic AF were checked by the MCG method, and we localized successfully in atrial mapping. The institutional ethical review board approved the study protocols, and all patients gave their informed consent. All procedures were in accordance with institutional guidelines. Among them, three patients underwent the MCG map-guided minimal AF surgery and follow-up examination.
2.1. MCG measurement
The MCG data of all subjects were analyzed by the application software KRISSMCG64 (Biomagnetism Research Center, Korea Research Institute of Standards and Science, Daejeon, Korea). The MCG system utilizes double relaxation oscillation superconducting quantum interference device sensors [10, 11]. The average noise spectral density of the whole system in a magnetic shielded room is 10 fT/ Hz @ 1 Hz and 5 fT/Hz @ 100 Hz. The MCG has 64 planar first-order gradiometers, which measure the tangential components of the cardiomagnetic fields. Measuring tangential field components is effective to obtain the overall heart information with a relatively small area of sensor distributions, that is, feasible to detect deeply-located sources like an electrical activity caused by the atrial excitation [9]. In this system, we change the tri-polar magnetic field map patterns into ordinary dipolar field maps by using the minimum norm estimate [12]. The typical recording time is 30 s and the subjects are in their supine position under the MCG dewar. The signal processing software provides automatic digital filtering, averaging, synthetic gradiometer formation, and baseline correction for the acquired records.
In order to make an exact conductor model and to combine anatomical information with MCG results, we recorded computed tomogram (CT) images for the individual patients. By the co-registration process between MCG and CT, we established the volumetric conductor models. The conductor model explains the volume current effect in calculation of forward problems with the boundary element model, which results in a more accurate localization than in the case of using a simple conductor model such as a horizontally layered model.
2.2. Separative surface potential activation beamformer
Minimum variance beamformer is a kind of spatial filtering technique explaining the covariance matrix of measured fields. The beamformer technique is used for dealing with rhythmic oscillation field like f-wave in atrial arrhythmia because the covariance matrix contains spatiotemporal information. However, minimum variance beamformer cannot separate one source from the other when there are time-correlated multi-sources, especially for successively exciting sources. As another problem, the minimum variance beamformer having been developed so far is based on current dipole sources and is not robust to the dipole orientation. Therefore, current dipole sources are not suitable for explaining the reentrant excitation feature appearing on the myocardium as well.
The basic idea of the proposed separative surface potential activation beamformer is to compensate for the intrinsic drawback of minimum variance beamformer. By using the independent component analysis, we can separate the f-wave from the other activations such as ventricular excitation, and we are also able to separate the features in a time sequence. The surface source model is based on the equivalent double layer model that states all the current sources in a heart can be described by epicardial action potentials on a closed surface [13]. With the boundary element model, we can confine the dependent variables to the action potentials on the boundary element model heart surface and the estimated separative surface potential activation beamformer power is corresponding to the source activities on the heart tissue. Hence, there is no need to estimate the orientation of current dipole at the vertexes.
In the AF study, the path tracing of the excitation as well as localization of the excitation is important to plan an adequate incision for conduction barrier. To do that, we need to separate the waveform feature time by time. This process is done by the independent component analysis procedure in separative surface potential activation beamformer method (Figs. 1 and 2).
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 1. Localization result for a simulated cyclic reentrant myocardial excitation. (a) Counterclockwise cyclic excitation of myocardium. (b) The generated magnetic fields by the reentry plus 10 fTrms random noise. (c) Separative surface potential activation beamformer localization result: The red color implies high activity in the action potential at the corresponding vertex.
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 2. Localization results for time-separated waveforms of successive excitation of current dipoles. (a) Magnetic field traces generated by successive excitation of current dipoles. (b) Independent component analysis – separated source patterns. (c) Localization results corresponding to each trace of (b). The successive excitation of the current dipoles is clearly shown.
|
|
|
3. Results
|
|---|
By using separative surface potential activation beamformer, the activity maps of the action potential for AF demonstrated in patients who have chronic AF. The measurement and analysis are shown in Fig. 3. The oscillatory behavior of f-wave between the T-wave and R-waves can be observed in the multi-channel MCG record (Fig. 3a). Using the independent component analysis process in the separative surface potential activation beamformer, we can separate the independent components of the f-waves and each waveform can be marked visually, time by time (Fig. 3b). In Fig. 3c, the upper row is corresponding to the #14 waveform (early excitation) in Fig. 3b, and the bottom row is corresponding to the #6 waveform (later excitation), respectively. The red color on the myocardium implies the active change in the action potential for each corresponding waveform. From the activity maps of the action potential at the two sequential time instants, we can infer the propagation trace of the atrial excitation. The expected propagation route is drawn in Fig. 3d.
View larger version (84K):
[in this window]
[in a new window]
|
Fig. 3. Preoperative MCG results. (a) The magnetic field trace. Atrial fibrillation time traces recorded by using an MCG. The R- and T-waves were chopped for their relatively large amplitudes. (b) The separated independent components analysis of the f-waves. The time window is corresponding to the red box in Fig a. (c) The activity maps of the action potential found by using separative surface potential activation beamformer; the upper row is corresponding to the #14 waveform (early excitation) in Fig. b, and the bottom row is corresponding to the #6 waveform (later excitation), respectively. The red color implies the active change in the action potential for each corresponding waveform. (d) The inferred propagation trace of the atrial excitation. LAA, left atrial appendage; RAA, right atrial appendage; LPV, left pulmonary vein; RPV, right pulmonary vein.
|
|
According to the MCG map-guide of the patient, who is a 70-year-old male with severe mitral stenosis (0.67 cm2), large Lt atrium (60 mm), ejection fraction 53%, and chronic AF, the atrial impulse activation starts at Rt atrial appendage and impulse propagates to Rt and Lt atrium via around the Rt inferior pulmonary vein and finished at Lt atrial appendage. So, we planned to block the propagation around the Rt inferior pulmonary vein. We conducted mitral valve replacement with a St Jude 29 mm prosthesis, dissection deeply between Rt and Lt atrium, especially at around Rt superior and inferior pulmonary veins, and only Rt-side pulmonary vein isolation using the cut-and-saw method with a 4-0 prolene continuous double layer suture. We do not conduct any more unnecessary procedures; Lt-side pulmonary vein isolation, Lt atrial additional linear connecting incisions to mitral valve annulus, coronary sinus, between pulmonary veins and Lt atrial appendage, excision of Lt and Rt atrial appendages, and any Rt atrial incisions. After the MCG map-guided minimal AF surgery, the patient's AF rhythm converted sinus successfully (Fig. 4). With six more months follow-up, his sinus rhythm is well preserved.
View larger version (54K):
[in this window]
[in a new window]
|
Fig. 4. Postoperative MCG results from the same patient of Fig. 3. The patient is a 70-year-old male with severe mitral stenosis, large Lt atrium (60 mm size) and chronic AF. He underwent mitral valvular replacement with St Jude 29 mm prosthesis and Lt side pulmonary vein isolation only. After minimal AF surgery, his rhythm converted sinus successfully and maintained at six months follow-up. MCG results show regular p-wave on the magnetic field trace instead of preoperative coarse and uneven f-wave (Fig. 3) on magnetic field trace, and show normal propagation on the activity maps of the action potential.
|
|
|
4. Discussion
|
|---|
Nakai et al. reported to develop a 3D recovery time dispersion map using 64-channel MCG and MRI [14]. However, their target is QRS component, relatively big signals on MCG for myocardial injury patients. Koskinen et al. reported to detect a weak atrial signal on MCG in AF patients successfully and showed the good reproducibility of MCG signals using signal time duration and signal size [15] not using a source localization information at 3D heart.
The proposed separative surface potential activation beamformer successfully visualized the activity maps of the action potential corresponding to the f-wave in the MCG records of a patient with AF. In addition, a prompt visualization of the electrical excitation of the myocardium is crucial for a clinical application of the MCG. By confining the source power calculation of the separative surface potential activation beamformer to the coordinates on the endocardial and epicardial surfaces, we could obtain the activity maps of the action potential on the heart boundary in a couple of minutes.
The obtained f-waves show periodic oscillatory behaviors. By using the separative surface potential activation beamformer, we separate the f-wave from the other activations and localize the position of a reentry circuit corresponding to the f-wave. By separating the f-wave time by time and visualizing the activity maps of the action potential for each time-separated waveform, we were able to infer the propagation trace of the AF.
Map-guided AF surgery facilitates less-extensive procedures and has an ultimate benefit to the patient. Smaller incision, simple minimal procedure, reduced postoperative bleeding, shortened cardiopulmonary bypass time, etc., could be the advantages of map-guided surgery. MCG is a totally non-invasive and even non-contact method for analyzing AF. The visualization of the myocardial current distribution corresponding to the reentrant excitation would be a great help for planning the AF surgery and for the follow-up examination. However, more technical advances in sensitivity of MCG system and image processing solutions are required for more accurate source localization and propagation.
|
Conference discussion
|
|---|
Dr. J. Melo (Lisbon, Portugal): I have two questions. One is regarding the concept and the other is the feasibility. Regarding the concept, what you have shown with your electrophysiologic assessment in permanent atrial fibrillation. My difficulty with this is that in paroxysmal atrial fibrillation or in permanent that you cardiovert, it is easy to know where A-Fib starts from. When we have permanent AF because of different refractory periods, because of conduction velocities, and because of those reentry mechanisms, where are you finding your starting point. This is a continuous circle movement. So the question concerns the concept of this method for permanent AF without having a reference point. My second question is feasibility. I could not understand your first human case. Does this magnetocardiographic machine go to the operating room or does the patient go to the room where the machine is. How do you do this
Dr. Kim: For the second question first, when the patient visits my clinic, we take the past medical history and surface EKG from the patient. After confirming that the patient has A-Fib, I explain to the patient about this study protocol and send the patient to the MCG center to be checked preoperatively at resting state. We plan the operation with the patient, including the A-Fib surgery, using the results from the MCG map-guided information, and then we confirm the sinus rhythm conversion during the surgery by surface EKG. After the surgery, we check surface EKG daily and send the patient to the MCG center again to recheck MCG and follow up the patient.
Your first question is about the concept. There are so many causes of A-Fib; macro-reentry and autonomic ganglia and so on. As we know, in the patient with normal atria, focal triggers are important, and in the enlarged atria with organic mitral valvular disease like my patients, triggers and macro-reentry as a substrate are important. And the cardioversion is important step to differentiate the persistent and permanent A-Fib and to know where A-Fib starts from. However, in my institute, I do not use cardioversion electrically or pharmacologically to the A-Fib patient. Instead of it, I use the patient's history and repeated surface EKG at resting state. So I use the term ‘chronic’ instead of ‘permanent’ or ‘persistent’. (If I confuse the term at my article, I want to correct it chronic.) As we know the atrial wave is very weak and I think standard tool is still insufficient to detect triggers and reentry at resting state. My focus is the abnormal atrial excitation and propagation at normal resting state without any manipulation like cardioversion, pacing, or isoproterenol infusion, especially very high current wave from the normal level. The normally excited atrial wave propagates from the right atrial SA node to the ventricular side, and the other atrial waves propagate to whole atria and disappear. However, in the A-Fib, abnormally high current atrial wave developed, propagated along abnormal route and did not disappear. So, with this MCG system, we check the atrial surface current, the initiation, propagation and disappearance, we detect the abnormal signal and abnormal route, and we plan the surgical method to remove or block the abnormals.
|
Acknowledgements
|
|---|
The authors thank Prof. Eun-Bo Shim in Kangwon National University and Prof. Seung-Pyung Lim in Chungnam National University Hospital.
|
References
|
|---|
- Furberg CD, Psaty BM, Manilo TA, Gardin JM, Smith VE, Rautaharju PM. Prevalence of atrial fibrillation in elderly subjects. Am J Cardiol 1994; 74:236–241.[CrossRef][Medline]
- Prasad SM, Maniar HS, Camillo CJ, Schuessler RB, Boineau JP, Sundt III TM, Cox JL, Damiano RJ Jr. The Cox maze III procedure for atrial fibrillation: long-term efficiency in patients undergoing lone versus concomitant procedures. J Thorac Cardiovasc Surg 2003; 126:1822–1828.[Abstract/Free Full Text]
- Gillinov AM, McCarthy PM. Advances in the surgical treatment of atrial fibrillation. Cardiol Clin 2004; 22:147–157.[CrossRef][Medline]
- Nitta T, Ohmori H, Sakamoto S, Miyagi Y, Kanno S, Shimizu K. Map-guided surgery for atrial fibrillation. J Thorac Cardiovasc Surg 2005; 129:291–299.[Abstract/Free Full Text]
- Ramanathan C, Jia P, Ghanem R, Ryu K, Rudy Y. Activation and repolarization of the normal human heart under complete physiological conditions. Proc Natl Acad Sci USA 2006; 103:6309–6314.[Abstract/Free Full Text]
- Vrba J. Weinstock H. (ed) Multichannel SQUID biomagnetic systems in applications of superconductivity2000;Dordrecht: Kluwer Academic Publishers 61–138. In:.
- Park JW, Jung F. Qualitative and quantitative description of myocardial ischemia by means of magnetocardiography. Biomed Technik 2004; 49:267–273.[Medline]
- Takala P, Hänninen H, Montonen J, Korhonen P, Mäkijärvi M, Nenonen J, Oikarinen L, Toivonen L, Katila T. Heart rate adjustment of magnetic field map rotation in detection of myocardial ischemia in exercise magnetocardiography. Basic Res Cardiol 2002; 97:88–96.[CrossRef][Medline]
- Kim K, Lee YH, Kwon H, Kim JM, Kim IS, Park YK. Optimal sensor distribution for measuring the tangential field components in MCG. Neurol Clin Neurophysiol 2004; 1:60.
- Lee YH, Kwon H, Kim JM, Park YK, Park JC. Double relaxation oscillation SQUID with high flux-to-voltage transfer and its application to a biomagnetic multichannel system. J Korean Phys Soc 1998; 32:600.
- Kim K, Lee YH, Kwon H, Kim JM, Park YK, Kim IS. Correction in the principal component elimination method for neuromagnetic-evoked field measurements. J Korean Phys Soc 2004; 44:980.
- Numminen J, Ahlfors S, Ilmoniemi R, Montonen J, Neonen J. Transformation of multichannel magnetocardiographic signals to standard grid form. IEEE Trans Biomed Eng 1995; 42:72–78.[CrossRef][Medline]
- Geselowitz DB. Description of cardiac sources in anisotropic cardiac muscle. Application of bidomain model. J Electrocardiol 1992; 25:Suppl65–67.[Medline]
- Nakai K, Izumoto H, Kawazoe K, Tsuboi J, Fukuhiro Y, Oka T, Yoshioka K, Shozushima M, Itoh M, Suwabe A, Yoshizawa M. Three-dimensional recovery time dispersion map by 64-channel magnetocardiography may demonstrate the location of a myocardial injury and heterogeneity of repolarization. Int J Cardiovasc Imaging 2005; 21:505–565.[CrossRef][Medline]
- Koskinen R, Lehto M, Vaananen H, Rantonen J, Voipio-Pulkki L, Makijarvi M, Lehtonen L, Montonen J, Toivonen L. Measurement and reproducibility of magnetocardiographic filtered atrial signal in patients with paroxysmal lone atrial fibrillation and in healthy subjects. J Electrocardiol 2005; 38:330–336.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
R. Jurkko, V. Mantynen, J. M. Tapanainen, J. Montonen, H. Vaananen, H. Parikka, and L. Toivonen
Non-invasive detection of conduction pathways to left atrium using magnetocardiography: validation by intra-cardiac electroanatomic mapping
Europace,
February 1, 2009;
11(2):
169 - 177.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
Top
Abstract
1. Introduction
