Electroanatomic Mapping of the Left Ventricle in a Porcine Model of Chronic Myocardial Infarction With Magnetic Resonance–Based Catheter Tracking
Background— X-ray fluoroscopy constitutes the fundamental imaging modality for catheter visualization during interventional electrophysiology procedures. The minimal tissue discriminative capability of fluoroscopy is mitigated in part by the use of electroanatomic mapping systems and enhanced by the integration of preacquired 3-dimensional imaging of the heart with computed tomographic or magnetic resonance (MR) imaging. A more ideal paradigm might be to use intraprocedural MR imaging to directly image and guide catheter mapping procedures.
Methods and Results— An MR imaging–based electroanatomic mapping system was designed to assess the feasibility of navigating catheters to the left ventricle in vivo using MR tracking of microcoils incorporated into the catheters, measuring intracardiac ventricular electrograms, and integrating this information with 3-dimensional MR angiography and myocardial delayed enhancement images to allow ventricular substrate mapping. In all animals (4 normal, and 10 chronically infarcted swine), after transseptal puncture under fluoroscopic guidance, catheters were successfully navigated to the left ventricle with MR tracking (13 to 15 frames per second) by both transseptal and retrograde aortic approaches. Electrogram artifacts related to the MR imaging gradient pulses were successfully removed with analog and digital signal processing. In all animals, it was possible to map the entire left ventricle and to project electrogram voltage amplitude maps to identify the scarred myocardium.
Conclusions— It is possible to use MR tracking to navigate catheters to the left ventricle, to measure electrogram activity, and to render accurate 3-dimensional voltage maps in a porcine model of chronic myocardial infarction, completely in the MR imaging environment. Myocardial delayed enhancement guidance provided dense sampling of the proximity of the infarct and accurate localization of complex infarcts.
Received September 1, 2007; accepted June 24, 2008.
Recent advances in catheter ablation of cardiac arrhythmias have involved the integration of computed tomography and magnetic resonance (MR) imaging (MRI) with conventional fluoroscopic catheter guidance to enhance the acquisition of electroanatomic images and to guide radiofrequency ablation.1–5 In these procedures, computed tomography images or MRIs are typically acquired days to weeks before the ablation procedure, and the segmented 3-dimensional (3D) images are subsequently registered to x-ray fluoroscopy and the electroanatomic mapping (EAM) systems using internal or surface fiducial landmarks. Although this approach provides better visualization of cardiac structures and less fluoroscopy exposure, it has some disadvantages. The preacquired images may not accurately reflect the anatomy at the time of the interventional procedure; ionizing x-ray radiation exposure, although less, is not eliminated; and the direct assessment of ablation lesions is not possible.
Clinical Perspective p 862
The use of intraprocedural MRI, without x-ray fluoroscopy, could theoretically overcome these deficiencies. With MRI, vascular roadmaps for catheter navigation can be created with a variety of MR angiography (MRA) techniques. Abnormal myocardium such as infarcted tissue can be readily visualized with myocardial delayed enhancement (MDE) imaging.6 Radiofrequency ablation lesions may be visualized after ablation to assess the extent of myocardial injury.7,8 Furthermore, MRIs acquired at the time of the procedure provide accurate representations of the true anatomy and are, by definition, registered without the use of any fiducial structures because the images and navigational data are acquired in the same coordinate frame. High-speed MRI has already been successfully used to navigate catheters through the body and cardiac chambers and has been shown to be accurate enough to perform transseptal puncture and stenting of aortic coarctation.9,10
Several strategies for navigating catheters, sheaths, and needles with MRI have been proposed. Catheter position and orientation can be tracked either passively, on the basis of the catheter magnetic material properties, or actively, through the incorporation of radiofrequency coils in the catheters. In the latter approach, the coils are capable of receiving MR signals, either at selected points on the shaft (MR tracking) or through the entire length of the catheter.11–13 The incorporation of MR-compatible electrodes into the catheter to measure intracardiac electrograms and to perform ablations requires care in selecting materials to minimize undesirable interactions with the MR system.13
MRI-based electrophysiology procedures have not yet been performed in a clinically relevant manner. One important aspect of electrophysiology procedures is the ability to navigate a catheter throughout the relevant cardiac chamber and to annotate the contact electrogram information to each location, ie, perform EAM. To this end, an interventional electrophysiology MRI-based EAM system was designed to acquire MRIs, to use these images to navigate catheters to and throughout the left ventricle (LV) with MR tracking of microcoils embedded in the shaft of the catheters, to measure the contact electrogram activity, and to integrate these electric and anatomic data to allow 3D ventricular substrate mapping in a porcine model of healed myocardial infarction.
This study was approved by the Massachusetts General Hospital Subcommittee of Research Animal Care. Fourteen pigs (4 normal, 10 with healed myocardial infarction) were used in these experiments.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
The principles of MR tracking have been described previously.12 The hardware implementation (GE Global Research, Niskayuna, NY) uses 8 MR system receiver channels, allowing simultaneous tracking of up to 8 MR-tracking coils. Radiofrequency signals received from the tracking coils during an MR-tracking pulse sequence are converted into x, y, and z positional information using dedicated software running in the MR-tracking processor. The MR-tracking rate was 13 to 15 frames per second (fps), with a tracking spatial resolution of 1.4×1.4×1.4 mm. The MR-tracking pulse sequence receiver bandwidth was ±16 or ±32 kHz.
MR-Tracked Catheter Design
Deflectable MRI-compatible catheters were specifically designed for these experiments (St Jude Medical, Minnetonka, Minn). Mechanical and material aspects of the catheter design evolved during the course of this study to improve performance. The final version (Figure 1A) is a 7F deflectable catheter made with a variable-stiffness design consisting of a metallic braided proximal section and a 7- to 12-cm–long polymer distal section housing five 4-mm–long solenoid MR-tracking coils. The MR-tracking coils are spaced at 1.5- to 2.0-cm increments starting 1 cm from the distal tip and are each connected to the proximal catheter with a 50-gauge coaxial cable. The catheter tip section consists of a gold-coated copper 4-mm-tip electrode and a low-magnetic-susceptibility copper/platinum ring electrode to allow measurement of bipolar electrograms. The catheter tip also incorporates a nitinol pull wire for deflection and a thermocouple for temperature measurement during catheter ablation.
MRI-Based Interventional Electrophysiology System
An MRI-based electrophysiology system (see Figure 1B for a schematic representation) was constructed using a combination of commercial products and specifically constructed prototype parts. System components within the MRI scan room include a commercial cardiac MRI coil, the prototype MR-tracked catheter, a prototype isolation module to minimize leakage currents into the catheter, a commercial multichannel electrogram amplifier (1 component of the GE Healthcare CardioLab 7000, Waukesha, Wis), 6 MR-compatible displays, and a set of prototype system-activating foot pedals. Components outside the MRI scan room include the commercial MRI computer and system electronics (GE Healthcare) with integrated prototype MR tracking, the electrogram data logging system (second part of Cardiolab 7000), and a data handler computer that is used to integrate the detected electrogram data with the instantaneous catheter position and the MRIs. The isolation module limits electric leakage currents into the catheter to <10 mA, meeting the industry current protection standards (UL 60601-1) for invasive devices.
The MR-tracked catheters propagate 2 types of signals to the MR-guided electrophysiology system. The first signals are low-frequency (<2 kHz) contact electrogram activity from the electrodes at the catheter tip in contact with the myocardial wall. These signals are sent from the catheter to a multichannel electrogram amplifier, digitized, and sent to the electrogram data logging system via a fiberoptic link. The signals are then relayed to the data handler computer for processing and display.
The second set of signals are the MR-tracking signals. These low-amplitude signals occur at the MR scanner Larmor frequency (≈63.8 MHz). MR-tracking signals flow from the catheter to the MR-tracking receiver unit and are amplified and digitized in the MRI system electronics. The digitized data are then sent to the MR-tracking processor running on the MRI system computer, where it is processed to give the x, y, and z coordinates of each MR-tracking coil in the catheter. The real-time coordinates of the catheter location are sent to the data handler computer for integration with the electrogram and MRI data and subsequent display.
Electrophysiology Recording System
The electrogram signals obtained from the electrophysiology catheter bipolar tip, as well as signals from the surface ECG electrodes, are sent to the multichannel electrogram amplifier located in the MRI scan room. The signals first pass through an analog radiofrequency low-pass filter to minimize interference from the MRI radiofrequency pulses. The shielded design and fiberoptic connection of the amplifier to the electrogram data logging computer provided an acceptably low level of interference with the MRI scanner. Selected ECG and intracardiac electrogram signals recorded in the electrogram data logger are converted back to an analog signal, which is then sent to an analog-to-digital converter in the data handler computer. Digital filters in the electrogram data logging computer (commercial 60-Hz band reject and 400-Hz low pass) and the data handler computer (30-Hz band reject and adaptive impulse response) are used to minimize the effects of artifacts created by currents induced by the magnetic field gradient subsystem of the MR system (Figure 1C and 1D).
Foot Pedal Data Capture System
The data capture system consists of 3 pneumatic pedals on the floor of the MRI scan room. Pressing the middle pedal causes the data handler computer to freeze its display and capture the instantaneous catheter position and electrogram data (voltage or activation time in relation to a reference surface ECG or internal electrogram signal). A second middle pedal depression causes data acceptance, allowing the data handler to combine the positional and electrogram data with the 3D MRI surface renditions, thus creating an EAM. The left pedal causes data rejection, and the right pedal turns MR tracking on or off.
Data Handler System
The data handler system (GE Global Research) is a dual-xenon-processor computer (Dell Precision 670, Round Rock, Tex) with a dual-monitor display, an analog-to-digital converter card (National Instruments, Austin, Tex), and custom software to integrate the electrogram data, catheter tracking, and preacquired MRI data. The analog-to-digital converter card senses the state of the MRI floor pedals and receives the analog output of the electrogram data logger. The data handler computer also receives images from the MRI system and uses virtual tool kit algorithms for surface extraction and display. The 3D MRA images can be segmented to provide renditions of the endocardial surfaces of the heart, which can be seen via an endoluminal or external view. Cutting and thresholding tools are used to remove unwanted features, providing a high-quality view of selected anatomy such as the endoluminal surface of the LV.
During MR tracking, catheter position information is sent to the data handler computer from the MRI computer via Ethernet at the rate at which it is acquired. The data handler computer renders the catheter in real-time using a cubic-spline interpolation. This rendering is merged with the anatomic surface images to provide real-time visualization of the catheter with respect to the anatomy.
Electrograms and surface ECGs also are transferred to the data handling computer from the electrophysiology recording system. During MR tracking and MRI, the MR system magnetic field gradients generate sharp spikes in the electrogram signal that correlate with the tracking rate (Figure 1D). Digital radiofrequency signal processing removes this artifact and restores electrogram fidelity. Signal processing of the surface ECG signals provides automatic detection of the QRS complex, which in turn is used to define a period of interest in which amplitudes and/or delays are detected from the electrogram signal. These amplitudes and/or delays are converted to colors and projected onto the endoluminal surface reconstructions. The data handler system, along with the MR scanner and the electrogram data logger, constitute an MRI analog of an electromagnetically or electrically tracked EAM system such as CARTO (Biosense Webster Inc, Diamond Bar, Calif) or NavX (St Jude Medical).
MRI was performed in a 1.5-T MRI (GE Healthcare CV/i) equipped with a 4-channel or an 8-channel cardiac array coil. Sagittal-plane 3D contrast-enhanced (0.44 cm3/kg Gd-DTPA) ECG-gated MRA was performed, spanning from the liver to the aortic arch, to include the entire heart and the aorta and inferior vena cava. The 3D MRA scan parameters were as follows: repetition time/echo time/flip angle, 2.9 ms/0.8 ms/37°; matrix, 256×160; field of view, 36 to 32×32 to 28 cm; bandwidth, ±83.3 kHz; slice width, 2.8 to 3.2 mm interpolated to 1.4 to 1.6 mm; timing, middiastolic; and slices per breath hold, 40 to 46. The 3D MRA acquisition was followed by 12 to 14 2-dimensional (2D) fast imaging employing steady-state acquistion with steady-state preparation (FIESTA-SP)14 short-axis wall-motion cine scans covering from the aortic arch to the apex of the heart. The FIESTA-SP scan parameters were as follows: repetition time/echo time/flip angle, 2.9 ms/1.0 ms/50°; field of view, 224×224 or 28×28 cm; bandwidth, ±125 kHz; slice width, 6.0 to 7.0 mm; cardiac frames per cardiac cycle, 20; and slices per breath hold, 1. For infarcted animals, 3D navigator-echo, respiratory-triggered, 3D MDE imaging was used, covering the entire LV. The 3D MDE scan parameters were as follows: repetition time/echo time/inversion time/flip angle, 3.5 ms/0.9 ms/150 ms/25°; field of view, 256×160 or 32 to 30×32 to 30 cm; bandwidth, ±62.5 kHz; slice width, 2.0 to 2.4 mm interpolated to 1.0 to 1.2 mm; timing, middiastolic; slices, 20 to 30; and acquisition time, 4 to 6 minutes.
MR-Tracked Navigational and EAM Techniques
The 3D MRA and 3D MDE data were reformatted into sagittal, coronal, axial, long-axis, and short-axis slice sets of 6-mm slice width covering the anatomy of interest (Reformat, GE Healthcare Advantage Windows 4.2). Real-time catheter manipulation was performed using simultaneous displays of these slice sets and the 3D segmented LV models created by the data handler system.
MR Catheter Tracking on Multicontrast, Multidirectional 2D Slice Data Sets
MR catheter navigation was performed on the just-acquired MRIs. Real-time catheter location was displayed using icons and graphic lines superimposed on the 2D slice data sets. The tip of the catheter was computed using a linear extrapolation of the 2 distal MR-tracking coils. As many as 6 data sets from among the reformatted 3D MRA and 3D MDE slice sets, as well as the wall-motion cine studies, were transferred into the MR-tracking system and simultaneously presented inside the MRI magnet room (Figures 2 and 3⇓). Each data set was overlaid with a cubic-spline interpolation of the distal catheter coil locations and displayed with a line width that was scaled to the actual thickness of the catheter. In addition to the catheter presentation, each of the multislice displays can independently update the displayed image according to the catheter position, thereby automatically displaying the slice from within the given slice set that intersects with the catheter tip position (Figure 2). Data sets from the wall motion studies can be updated in real-time using both the slice and the instantaneous cardiac trigger delay (Figure 3A, upper left). The instantaneous cardiac phase was provided via 1 of the 4 surface ECG leads.
MR-Tracked Navigation on the Data Handler 3D Endoluminal or Surface-Rendered Models
The segmented 3D MRA, with a 3D MDE overlay (infarcted animals only), was overlaid with the instantaneous 3D catheter shape, position, and orientation, as provided by the tracking coils on the catheter, using custom display algorithms on the data handler computer. The data handler renders the catheter as a flexible 3D surface appropriately scaled to the size of the catheter and with the location of each MR-tracking coil displayed as a blue ball. As with the MR-tracking display, the tip of the catheter is computed and displayed using a linear extrapolation algorithm. This information was also displayed on the in-room monitors (Figure 3A, bottom left).
Porcine Infarct Model Preparation
By methods previously described, a closed-chest myocardial infarction was created in 10 swine by ethanol injection within the mid left anterior descending or mid left circumflex arteries.15–17 The infarcts generated by this procedure, as later observed by MDE and histology, were primarily (8 of 10) single large lesions spanning portions of the apical and septal walls. In 1 animal, a patchy infarct was created in the apical/septal region; in another animal, a small infarct was created behind the lateral papillary muscle. The pigs were then housed in an animal facility for a mean of 37±7 days before MR-tracking and EAM experiments.
After premedication and induction of general anesthesia, intravascular introducer sheaths were placed in the femoral arteries and veins bilaterally. With the use of an 8.5F Mullins sheath, transseptal puncture was performed under fluoroscopic guidance (GE Healthcare OEC 9800, Salt Lake City, Utah). Once access to the left atrium was obtained, intravenous heparin (100-U/kg initial bolus, followed by 30 U · kg−1 · h−1) was administered. The 4 normal swine and 4 of 10 infarcted pigs were then transported to the interventional MRI suite for MR-guided navigation and EAM experiments. The remaining 6 of 10 infarcted animals underwent standard electroanatomic LV substrate mapping (with CARTO) before transfer to the MRI suite.
LV mapping was performed via retrograde aortic or transseptal approaches with a 4-mm-tip 7F deflectable catheter (Navistar, Biosense Webster, Inc). A multichannel electrophysiology recording system (Bard, Lowell, Mass) was used with the electrograms filtered from 10 to 400 Hz and the peak-to-peak bipolar voltage amplitude computed by CARTO. The LV was mapped and a 3D voltage map was constructed displaying peak-to-peak bipolar voltage electrogram amplitude with a fill threshold of <15 mm. Myocardium with a voltage amplitude ≤1.5 mV was defined as infarcted tissue, and a voltage >1.5 mV was defined as normal myocardium.15–17
MR-Tracked Navigation to the LV and EAM
After CARTO mapping, the animals were immediately transferred to the interventional MRI suite. At this point, preprocedural MRI was conducted. After the imaging session, the data were segmented for use in navigation. The MR-tracked catheters were then inserted into the transseptal sheath, beginning navigation from the left atrium into the LV. This stage of navigation relied on either the 2D slice set displays or the 3D MRA surface display. After passing through the mitral valve into the LV, the catheter was navigated to various endocardial locations where electrograms were recorded point by point with the pneumatic pedals to capture peak-to-peak voltages. The color overlays on the 3D display of the data handler were the same as with CARTO mapping. In the case of infarcted pigs, the 3D MDE slice data sets, along with the 3D surface-rendered MRA, played a key role in guiding the catheter to the borders of the infarct in the mapping stage of the procedure.
Postexperimental Gross Histology
After the mapping procedure, all the animals were sacrificed, and gross pathology was performed to visualize the infarct regions. The MRI infarct area was obtained by segmenting the 3D MDE images (Volume Viewer, GE Healthcare Advantage Windows 4.2, Waukesha, Wis). The animals and hearts also were inspected to record possible sites of inadvertent heating or puncture resulting from the experiments.
The MR-tracking pulse sequence used a data acquisition bandwidth of ±16 or ±32 kHz, which is less than the ±125-kHz bandwidth commonly used in real-time MRI.10 As a result, MR tracking was ≈40 dB lower in acoustic noise than real-time image-based tracking, allowing standard operation without closing the MRI radiofrequency door or using ear protection. The MR-tracking approach also uses a projection technique, which has a much higher signal-to-noise ratio (≈20 times) than MRI.
Tracking of the MR-compatible catheter at 13 to 15 fps at a 1.4×1.4×1.4-mm spatial resolution was possible in all animals. This may prove to be faster and more spatially accurate than the 1.5×1.5×5.0-mm resolution, 4- to 5-fps tracking rate reported with real-time MRI.7,10 In all animals, after transseptal puncture under fluoroscopic guidance, it was possible to navigate the tracked catheter to the LV by both retrograde aortic and transseptal approaches using preprocedural MRA roadmaps (Figure 2) without any appreciable time differences in the approaches (≈3 minutes). The operators who performed the studies effectively used the combination of reformatted slice data sets and 3D rendered LV maps for navigation (Figure 3). Visualization of the tracking coil, combined with the overlaid catheter display, permitted real-time visualization of the orientation and 3D deflection of the distal portion of the catheter in a manner similar to standard fluoroscopy-based procedures.
MR-tracking–guided EAM of the LV was successful in all animals. With design changes, the catheter deflection and torque-ability properties improved during the course of this study, allowing access to all parts of the LV in 10 of 14 experiments. In the 4 initial experiments, access to the posterior LV wall was difficult but resolved with catheter improvements. In most infarcted animals, MR-tracking–guided EAMs were comparable to CARTO maps in terms of the range of peak-to-peak voltages observed, the spatial position, and the areas (±20%) of the infarct (Figure 4). The MR-tracking–based infarct maps agreed well with the infarct shapes and locations, as seen in the pathological specimens (Figure 5).
The advantages of MR-tracking–guided procedures were evident in 2 animals. In 1 animal, CARTO mapping revealed no infarct (Figure 6). However, on MDE imaging, a lateral wall infarct was seen behind the papillary muscle and later confirmed at necropsy. Using the MDE image, MR tracking was used to navigate the catheter to this difficult-to-reach location to obtain a more accurate map of the infarct that may have been missed completely with standard mapping technologies.
In the second animal, MDE imaging identified a patchy nontransmural infarct. The MR-tracking–guided voltage map provided greater detail in the multiregion infrastructure than did the CARTO map. This suggests that the use of MRIs as references may be advantageous during mapping of complex infarcts. The endocardial infarct areas obtained by MR-tracking–guided mapping agreed well with the segmented infarct areas obtained from 3D MDE images (Figure 7).
The initial MRI sessions took 2.5±1 hours, with the later experiments completed in ≈1.5 hours. Shorter imaging times were achieved by performing contrast-enhanced 3D MRA as the first scan; this allowed preintervention image preparation to be performed in parallel with subsequent imaging. The complete MR-guided LV mapping procedure required 4±1 hours, with procedure duration progressively decreasing as catheter performance and operator experience improved.
Three animals died in the midst of the procedure as a result of ventricular tachycardia. This complication was alleviated in later experiments with the use of intravenous lidocaine. No significant endocardial trauma or LV perforations were noted at autopsy, suggesting that the mechanical function of the MRI-compatible catheters was similar to conventional catheters. No internal lesions were found during pathological examination, suggesting that inadvertent radiofrequency-induced heating did not occur, although no systematic study of in vivo catheter heating was conducted. For the animals that died, no evidence was found of any surface insulation failures of the catheters. In the event of an insulation failure, the electric isolation of the electrogram amplifier and the prototype isolation module were designed to limit the leakage currents to <10 mA (the level of leakage current protection found in invasive monitoring systems approved for clinical use).
In this study, we demonstrated that it is possible to perform 3D EAM in an MRI environment using MR tracking of catheters. First, we demonstrated that a strategy using multiple tracking coils on the catheter allowed real-time visualization of a 3D rendering of the catheter during in vivo manipulation. Second, this deflectable catheter could be manipulated into the porcine LV using both the retrograde aortic and the transseptal approaches (although fluoroscopy was required for transseptal puncture and sheath placement in the left atrium). Third, the catheter could be manipulated to all of the endocardial locations to project electrophysiological electrogram information onto the ventricular shell. Finally, this combination of electrophysiological and anatomic information correctly located chronically infarcted myocardium.
In this study, MR-tracking–based vascular and intracardiac navigation was similar in quality to that commonly observed with conventional x-ray–guided EAM and at temporal 3D catheter tracking rates approaching x-ray fluoroscopy. Using 4 to 5 tracking coils (with a relatively simple spline interpolation of the intervening catheter shaft) provided enough of an estimate of the 3D curvature, shape, and orientation of the distal catheter to successfully negotiate the various complex anatomic obstacles during navigation to and within the LV. However, extending the tracked catheter region from 12 to 20 cm would have provided even better resolution of complex mapping catheter configuration such as tight catheter tip encirclements (eg, >360° tip wraps).
Once within the LV, a number of 2D and 3D images were available to the operator to help guide ventricular mapping. The 3D surface reconstruction was very useful for providing a volumetric understanding of the chamber and the relative location of the catheter. This visualization paradigm was similar to that used with conventional mapping systems. However, the soft tissue discriminative property of MRI provides much more information than simply the endocardial chamber volume, including wall thickness, the presence of myocardial scar by MDE, and surrounding structures. This information was much better appreciated with the orthogonal 2D slice series. Additional MRI information such as T2-based postablation edema maps or strain-based wall-motion maps could enhance this advantage. This could certainly provide a possible superiority of MR-tracking–guided mapping over the use of conventional x-ray–based EAM of complex infarcts.
In this study, animal motion did not occur because all of the animals were under general anesthesia and intubated. In nonanesthetized subjects, gross motion could occur, requiring nonrigid body reregistration, as performed in CARTO, with its known associated error. Relative to techniques relying on continuous real-time MRI, device positional tracking by MR allows higher spatial resolution, faster update rates, and lower acoustic noise. Construction of MR-tracked catheters that do not produce sizable susceptibility artifacts while providing mechanical properties similar to conventional electrophysiology catheters was shown to be possible. However, it should be noted that additional MRI-compatible devices such as deflectable and trackable sheaths are needed. Use of these devices will allow greater procedure speed and flexibility and will be essential for performing the entire electrophysiology procedure in the MRI environment.
Although most of these procedures were performed entirely within the MRI environment without x-ray fluoroscopy, the transseptal puncture procedure did use fluoroscopy. However, other investigators have demonstrated that MRI can indeed guide this procedure.
MRI-based procedures may provide a number of potential advantages over fluoroscopy-based procedures, including the absence of exposure to ionizing radiation and enhanced localization of the catheter tip in relation to the tissue. However, the actual relative advantages of each of these approaches can actually be fully assessed only in a comparative fashion.
Despite the favorable findings of this study, it remains clear that a number of additional challenges, including patient motion, monitoring, safety, and catheter ablation, must be overcome before MRI-based procedures can be used in standard clinical practice.
Sources of Funding
This study was supported in part by research grants from General Electric Co and St Jude Medical, Inc, as well as a National Institutes of Health K23 award (HL 68064) to Dr Reddy.
Dr Mallozzi, Dr Schmidt, R. Guhde, R.D. Darrow, Dr Slavin, M. Fung, Dr Foo, and Dr Dumoulin are employees of General Electric Co. G. Kampa and J.D. Dando are employees of St Jude Medical, Inc. Dr Reddy has served as a consultant to St Jude Medical, Inc and Biosense-Webster, Inc. Dr d'Avila has served as a consultant to St Jude Medical, Inc. The remaining authors report no conflicts.
Reddy VY, Malchano ZJ, Holmvang G, Schmidt EJ, D'Avila A, Houghtaling C, Chan RC, Ruskin JN. Integration of cardiac magnetic resonance imaging with three-dimensional electroanatomic mapping to guide left ventricular catheter manipulation: feasibility in a porcine model of healed myocardial infarction. J Am Coll Cardiol. 2004; 44: 2202–2213.
Dong J, Calkins H, Solomon SB, Lai S, Dalal D, Lardo A, Brem E, Preiss A, Berger RD, Halperin H, Dickfeld T. Integrated electroanatomic mapping with three-dimensional computed tomographic images for real-time guided ablations. Circulation. 2006; 113: 186–194.
Ector J, De Buck S, Adams J, Dymarkowski S, Bogaert J, Maes F, Heidbuchel H. Cardiac three-dimensional magnetic resonance imaging and fluoroscopy merging: a new approach for electroanatomic mapping to assist catheter ablation. Circulation. 2005; 112: 3769–3776.
Sra J, Krum D, Malloy A, Vass M, Belanger B, Soubelet E, Vaillant R, Akhtar M. Registration of three-dimensional left atrial computed tomographic images with projection images obtained using fluoroscopy. Circulation. 2005; 112: 3763–3768.
Malchano ZJ, Neuzil P, Cury R, Holmvang G, Weichet J, Schmidt E, Ruskin JN, Reddy VY. Integration of cardiac CT/MR imaging with 3-dimensional electroanatomical mapping to guide catheter manipulation in the left atrium: implications for catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol. 2006; 17: 251–255.
Mahrholdt H, Wagner A, Holly TA, Elliott MD, Bonow RO, Kim RJ, Judd RM. Reproducibility of chronic infarct size measurement by contrast-enhanced magnetic resonance imaging. Circulation. 2002; 106: 2322–2327.
Lardo AC, McVeigh ER, Jumrussirikul P, Berger RD, Calkins H, Lima J, Halperin HR. Visualization and temporal/spatial characterization of cardiac radiofrequency ablation lesions using magnetic resonance imaging. Circulation. 2000; 102: 698–705.
Raval AN, Telep JD, Guttman MA, Ozturk C, Jones M, Thompson RB, Wright VJ, Schenke WH, DeSilva R, Aviles RJ, Raman VK, Slack MC, Lederman RJ. Real-time magnetic resonance imaging-guided stenting of aortic coarctation with commercially available catheter devices in swine. Circulation. 2005; 112: 699–706.
Callans DJ, Ren JF, Michele J, Marchlinski FE, Dillon SM. Electroanatomic left ventricular mapping in the porcine model of healed anterior myocardial infarction: correlation with intracardiac echocardiography and pathological analysis. Circulation. 1999; 100: 1744–1750.
During catheter ablation of cardiac arrhythmias, the importance of delineating the cardiac anatomy is undisputed. Currently, image integration with either computed tomographic or magnetic resonance (MR) imaging is used both to enhance the acquisition of 3-dimensional electroanatomic mapping and to guide radiofrequency ablation. Typically, the imaging is performed before the procedure and registered to the patient’s anatomy at the time of the procedure using fiducial landmarks. However, a more ideal paradigm would be to perform these procedures completely in an MR imaging environment. Both errors in registration and exposure to ionizing radiation could theoretically be eliminated. Superior anatomic visualization could be possible with the generation of vascular roadmaps by MR angiography and visualization of abnormal infracted or ablated myocardium with myocardial delayed enhancement imaging. In this porcine study, MR tracking of microcoils embedded in electrophysiology catheters was used to navigate the catheters to the left ventricle at rates approaching that of x-ray fluoroscopy (13 to 15 frames per second) using MR angiography vascular roadmap guidance. With myocardial delayed enhancement images, it was possible to accurately maneuver the catheter within the chamber, to measure intracardiac electrograms, and to project this spatial and electrogram information onto the MR angiography–generated 3-dimensional left ventricle models. In infracted animals, MR imaging–based voltage maps were compared with standard x-ray–based electroanatomic voltage maps to establish the equivalency of this MR-tracking approach to ventricular mapping. This article demonstrates that it may be possible to use these imaging techniques with active MR tracking to perform electrophysiology procedures in a clinically relevant manner completely in an MR imaging environment.