View-Synchronized Robotic Image-Guided Therapy for Atrial Fibrillation Ablation
Experimental Validation and Clinical Feasibility
Background— A robotic catheter navigation system has been developed that provides a significant degree of freedom of catheter movement. This study examines the feasibility of synchronizing this robotic navigation system with electroanatomic mapping and 3-dimensional computed tomography imaging to perform view-synchronized left atrial (LA) ablation.
Methods and Results— This study consisted of a porcine experimental validation phase (9 animals) and a clinical feasibility phase (9 atrial fibrillation patients). Preprocedural computed tomography images were reconstructed to provide 3-dimensional surface models of the LA pulmonary veins and aorta. Aortic electroanatomic mapping was performed manually, followed by registration with the corresponding computed tomography aorta image using custom software. The mapping catheter was remotely manipulated with the robotic navigation system within the registered computed tomography image of the LA pulmonary veins. The point-to-surface error between the LA electroanatomic mapping data and the computed tomography image was 2.1±0.7 and 1.6±0.1 mm in the preclinical and clinical studies, respectively. The catheter was remotely navigated into all pulmonary veins, the LA appendage, and circumferentially along the mitral valve annulus. In 7 of 9 animals, circumferential radiofrequency ablation lesions were applied periostially to ablate 11 pulmonary veins. In patients, all of the pulmonary veins were remotely electrically isolated in an extraostial fashion. Adjunctive ablation included superior vena cava isolation in 6 patients, cavotricuspid isthmus ablation in 5 patients, and ablation of sites of complex fractionated activity and atypical LA flutters in 3 patients.
Conclusions— This study demonstrates the safety and feasibility of an emerging paradigm for atrial fibrillation ablation involving the confluence of 3 technologies: 3-dimensional imaging, electroanatomic mapping, and remote robotic navigation.
Received November 19, 2006; accepted February 16, 2007.
A number of approaches to catheter ablation of atrial fibrillation (AF) exist, including ostial pulmonary vein (PV) isolation, “extraostial” PV isolation, circumferential PV ablation without isolation, ablation of complex fractionated atrial electrograms, placement of linear ablation lesions, and ablation of autonomic targets.1–6 Although the ideal ablation strategy is still in evolution, it is clear that facile left atrial (LA)-PV catheter manipulation is important for safety and effectiveness.
Clinical Perspective p 2714
To facilitate intracardiac mapping and ablation, a series of technological advances have been introduced to the field over the past decade. The first and most significant was the introduction of real-time electroanatomic mapping (EAM) to precisely localize the catheter within the heart.7 More recently, the integration of high-resolution 3-dimensional (3D) computed tomography (CT)/magnetic resonance imaging with EAM permits real-time catheter manipulation within a more highly featured patient-specific anatomy.8,9 Despite the aid of this image-guided therapy paradigm, considerable technical skill is required for catheter manipulation. To this end, a remote robotic navigation system (RNS) has recently been developed.10,11 After a mapping/ablation catheter is placed within this 2-sheath system, a software interface allows the operator to remotely manipulate this catheter tip with precision.
The present study examines the feasibility of combining this remote navigation system with both EAM and 3D CT imaging. The confluence of these 3 technologies was used in an experimental porcine model to assess the feasibility of safely navigating within the LA-PVs and performing circumferential PV ostial ablation. These data provided the experimental groundwork for subsequent view-synchronized remote LA-PV ablation in a series of AF patients.
This protocol was divided into 2 phases: the preclinical phase, including mapping/ablation in 9 pigs, which was approved by the Massachusetts General Hospital Subcommittee of Research Animal Care, and the clinical phase, which involved catheter ablation in 9 patients and was approved by the Massachusetts General Hospital Human Research Committee and the Homolka Hospital Ethical Committee.
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.
Contrast-enhanced cardiac scanning was performed under general anesthesia on 9 normal pigs (25 to 50 kg) with a 64-slice CT scanner (Sensation 64, Siemens, Inc, Germany). Up to 7 days before the procedure, CT images were acquired with cardiac gating during a single breathhold at held expiration. Patient imaging was performed within 2 weeks of the procedure with a 16-slice CT scanner (Sensation 16, Siemens, Inc) with cardiac gating at end expiration.9
Segmentation and 3D Reconstruction of the CT Imaging Data Sets
The LA-PVs and aorta were manually segmented separately with the research software package Cardiac++ (GE Medical Systems, Inc, Milwaukee, Wis). These segmented data sets were converted into 3D surface reconstructions (Magics, Materialise Inc, Ann Arbor, Mich). Then, the 3D models of the aorta and LA-PVs were imported into the image-guided intervention software (see Registration below).8,9
The Robotic Navigation System
The RNS (Hansen Medical, Inc, Mountain View, Calif) consists of a “master-slave” electromechanical system that controls an internal steerable guide sheath and an external steerable sheath (Figure 1). The internal sheath contains 4 pull wires located at each quadrant; the range of motion includes deflection in 360° and the ability to insert/withdraw. The external sheath contains a single pull wire to permit deflection and to provide the ability to rotate and insert/withdraw. These movements allow a broad range of motion in virtually any direction. This steerable sheath system (SSS) can be used in the same way as conventional sheaths with different mapping/ablation catheters inserted through the guide lumen. Thus, with the mapping/ablation catheter fixed so that it just protrudes beyond the tip of the inner system, remotely driving the SSS translates to remote navigation of the catheter tip.
The SSS is attached to the remote robotic arm unit that can be fixed to the foot of a standard x-ray procedure table. A “3D” joystick allows the operator to remotely drive the catheter tip using a software interface. This interface translates movements of the joystick into the complex series of manipulation by the pull wires governing sheath motion. The operator can decide whether to individually manipulate the internal or external sheath.
EAM procedures were performed using a magnetic localization system (CARTO 7x, Biosense-Webster, Inc, Diamond Bar, Calif) with an EAM point location export function to permit real-time export of both any acquired EAM point and the real-time catheter tip location. With this EAM system,7 mapping was performed with a 4-mm-tip, 8-mm-tip, or saline-irrigated 3.5-mm-tip catheter (Navistar, DS Navistar, or Thermocool, respectively; Biosense-Webster, Inc).
A separate workstation with custom imaging software was used to combine the real-time EAM information with the preprocedural imaging data.8,9 This software can visualize, manipulate, and register the 3D CT images and EAM data. In addition to receiving real-time EAM catheter position information, the software also interfaces with the RNS. The RNS view governing the perspective used by the 3D joystick is sent to the custom software so that the registered 3D CT image also is displayed in an identical view. This view synchronization provides the operator an intuitive means to manipulate the catheter (see Figure 2).
As shown in Figure 3, the registration strategy first involves manual mapping of the endoluminal surface of the aorta using a retrograde aortic approach, which takes ≈5 minutes.8,9 These EAM points were registered to the aorta 3D CT model using an “iterative closest points” distance minimization algorithm. With this initial registration pose, the LA was robotically mapped with careful attention to minimize chamber distortion from excessive catheter-tissue pressure. PV mapping was performed using the “vessel tagging” function to the level of the LA-PV junction. These additional LA-PV EAM points were incorporated into the registration algorithm to further refine the quality of the CT registration. The degree of registration error was calculated as the mean point-to-surface error of the EAM points to the nearest surface location on the registered 3D CT image.9
Preclinical Electrophysiology Study
Under general anesthesia, transseptal punctures were performed with fluoroscopic guidance either manually in standard fashion with a Brockenbrough needle or remotely with the RNS using a specialized transseptal needle under intracardiac echocardiography (ICE) guidance.11 When the puncture was performed manually, an exchange-length guidewire was used to pass the SSS into the LA. Remote LA-PV and right atrial (RA) mapping/ablation was performed with and without 3D CT image integration, respectively, but always with EAM guidance. At the conclusion of the procedure, the animals were killed, and the hearts immediately explanted and examined.
Clinical Electrophysiology Study
The clinical procedures were performed with conscious sedation. First, manual aorta mapping was performed as described above. Then, dual transseptal access was achieved using a standard 8F long transseptal sheath and a puncture performed remotely with the RNS under ICE guidance.11 Intravenous heparin was given before the transseptal punctures to achieve an activated clotting time >300 seconds. LA-PV mapping and registration were performed as described above with the 3.5-mm-tip irrigated catheter. Then, remote ablation lesions were placed to electrically isolate the PVs as ipsilateral pairs in an extraostial fashion. To minimize the possibility of esophageal damage, power was limited along the posterior LA to 30 W; elsewhere, power was titrated (20 to 70 W) to achieve a 5% to 10% drop in the impedance. The irrigation flow rate was 2 mL/min during mapping and 30 mL/min during ablation. Radiofrequency energy was terminated once the electrogram was no longer visible or after 60 seconds. A circular multielectrode mapping catheter (Lasso, Biosense-Webster Inc) was used to verify both entrance and exit block. Electrical PV isolation also was confirmed during the infusion of 20 μg/mL isoproterenol.
Remote Porcine Atrial Mapping and Ablation With View-Synchronized 3D CT Imaging
After manual aorta mapping and registration with the CT scan, EAM of the LA was performed with 3D CT view synchronization. A total of 65±29 LA EAM points (range, 25 to 119) were acquired (see Figure 4). Registration using the aorta EAM data alone resulted in an LA EAM point-to-surface error of 2.1±0.7 mm; incorporation of the LA EAM data further refined the registration accuracy to an error of 2.0±0.4 mm (P=NS). Thus, LA-PV image integration was highly accurate using the aorta EAM registration alone and guided robotic LA-PV mapping; inclusion of LA-PV EAM data had minimal impact on the registration accuracy. Despite the relatively small size of the porcine atria and the typical short distance between the transseptal puncture and the posterior LA wall, it was possible to enter and navigate into all 4 PVs in all animals; an example of the tubular renderings of the PVs is shown in Figure 4. Although difficult to quantify, the stability of the RNS allowed facile manipulation along the circumference of the mitral valve (note the periannular EAM points acquired in Figure 4) despite the characteristic vigorous motion of this annular region.
Through the use of either an 8-mm-tip or irrigated 3.5-mm-tip ablation catheter, periostial radiofrequency lesions were robotically placed in 7 of 9 animals. A total of 11 PVs were targeted for ablation: 7 right superior PVs and 4 left superior PVs. With the robotic system, a total of 13.6±3.6 and 11.8±1.3 lesions per PV were placed to periostially ablate the right and left superior veins, respectively. On postmortem pathological examination, no cardiac perforation was noted, and the radiofrequency lesions were uniformly situated at the PV ostia.
Although 3D CT images of the RA were not available for use during these procedures, the RNS was used in concert with the EAM system to construct full RA maps in 3 animals. As shown in Figure 4, detailed RA mapping was possible without complications. Similar to the perivalvular stability observed on the left side of the heart, the catheter was easily manipulable circumferentially along the tricuspid annulus. The His bundle electrogram was identified in all animals.
As shown in the Table, A total of 9 drug-resistant AF patients underwent AF catheter ablation: paroxysmal/persistent AF, 3 of 6; age, 56.7±15.7 years; male/female, 3/6; weight, 82±16 kg (range, 65 to 110 kg); mean duration of AF, 5.9±6.7 years (range, 0.5 to 20 years); and left ventricular ejection fraction, 65±14% (range, 36% to 86%).
Remote-Guided Transseptal Puncture in Humans
Using a combination of fluoroscopic and ultrasonographic guidance, the SSS could be remotely manipulated to probe the RA aspect of the interatrial septum to identify the appropriate septal puncture site. Transseptal needle puncture was possible in all 9 patients (Figure 5). In 7 of the 9 patients, the dilator and sheath were advanced over the needle into the left atrium. In the remaining 2 patients, technical issues unassociated with the puncture procedure—a software system fault requiring computer reset of the system and manual placement of a long guidewire and exchange in standard fashion—prevented remote advancement of the SSS into the LA. No complications associated with transseptal access, such as pericardial effusion/tamponade, occurred.
Remote LA-PV Mapping and Ablation in AF Patients With View-Synchronized 3D CT Imaging
After manual aorta mapping and registration, the view-synchronized 3D CT scan guided detailed LA mapping; a total of 116±49 EAM points were acquired (range, 60 to 193). When these additional LA points were incorporated into the registration scheme, the mean LA EAM point-to-surface error was 2.1±0.3 mm. As with the experimental study, the view-synchronized CT scan allowed facile manipulation into the PVs, the LA appendage, and the mitral valve annulus in all patients. ICE imaging was performed manually to verify the proper localization of the catheter tip at each of these locations.
With the view-synchronized CT scan, an irrigated radio frequency ablation catheter was robotically manipulated to perform extraostial PV isolation as shown in Figures 6 and 7⇓. This was performed during sinus rhythm in 6 of 9 patients and during AF in 3 of the 9 patients. Along the posterior LA, the lesion set was extended to include a significant portion of the PV antrum; during ablation in this region, the maximum amount of energy was limited to 30 W. During isolation of the left-sided PVs, lesions were placed on the PV side of the LA appendage–PV ridge. In all patients, the integrity of the isolating lesion set was confirmed by manually manipulating a circular mapping catheter into each of the PVs; entrance and exit conduction block was confirmed in all 36 PVs.
Adjunctive ablation lesions were placed in both the paroxysmal and persistent AF patients. In the paroxysmal patients, electrical isolation of the superior vena cava was attempted and successfully performed in 4 patients. Damage to the pericardiac right phrenic nerve was avoided by careful pacing before radiofrequency energy delivery. Cavotricuspid isthmus flutter was inducible in 1 paroxysmal AF patient; the mechanism was confirmed by entrainment from the isthmus. The robotically driven irrigated catheter was used to achieve bidirectional block across the cavotricuspid isthmus. A number of linear atrial lesions also were deployed in all 3 persistent AF patients: a cavotricuspid isthmus line, an LA “roof” line connecting the right and left PVs, and a mitral isthmus line (see Figure 8). Conduction block across each of these lines was achieved in all patients; for the mitral isthmus line, ablation within the coronary sinus also was required to achieve block in 2 of 3 patients. Finally, robotically driven ablation lesions also were placed at sites of complex fractionated activity, mostly along the left-sided interatrial septum, along the inferior wall, and at the base of the LA appendage. The rhythm converted to either sinus (1 patient) or an organized atrial flutter (2 patients) during catheter ablation in all 3 patients with persistent AF.
The overall “skin-to-skin” procedure time was 338±89 minutes (range, 260 to 480 minutes), and the total duration of x-ray fluoroscopy exposure to the patient was 29±14 minutes (range, 14 to 55; see the Table). A total of 63±17 ablation lesions (range, 45 to 104) were placed, some of which were “drag” lesions (ie, the ablation catheter was moved during ablation). The duration of time from the first to the last of these lesions was 180±75 minutes (range, 104 to 360 minutes). No acute or chronic complications related to the robotic system or procedure occurred, including thromboembolic complications such as stroke, pericardial effusion with tamponade (assessed by transthoracic echocardiography the day after the procedure), atrioesophageal fistula formation, and PV stenosis.
The accuracy of mapping and ablation during catheter ablation of AF is limited in part by the patient-to-patient variability and complexity of the LA-PV anatomy and in part by the electrophysiologist’s ability to manipulate the ablation catheter manually to the desired intracardiac locations. To overcome these limitations, the present study examined the feasibility of combining preacquired 3D CT imaging with real-time EAM and robotic navigation. The fundamental utility of this combination is the ability to robotically navigate on a view-synchronized 3D CT rendering of the patient-specific LA-PV anatomy. In this study, the feasibility of this approach was demonstrated in a porcine experimental protocol, followed by a clinical series demonstrating safety and clinical utility.
One of the potential advantages of the RNS is increased and more precise instinctive control of the catheter tip. Existing catheters (that have only a single- or dual-deflection mechanism) require a significant degree of operator skill and experience to efficiently navigate the LA-PV anatomy. The SSS attempts to address this challenge by providing an instinctive catheter control mechanism; movement of a simple 3D joystick is electronically transduced to a 1:1 movement of the catheter tip. Alternatively, the movement can be decimated to provide a greater degree of control (ie, physical movement of the joystick by 1 cm can be decimated 4:1 to result in only a 0.25-cm movement of the catheter tip). This provides a degree of control not easily achieved with a standard manually driven catheter.
Robotic navigation is controlled by the clinician to direct the catheter tip to a desired location on the basis of feedback from the standard imaging modalities currently used to guide electrophysiology procedures: fluoroscopy and EAM technologies. Of importance, although existing mapping/ablation catheters typically have only single- or dual-deflection capability, the range of motion by the SSS is considerably greater. Because the inner guiding sheath has 4-quadrant deflection capability, it can navigate in virtually any direction. One limitation is a single fulcrum for these deflections. To overcome this limitation, the deflection and rotation capability of the external sheath provides a moveable fulcrum from which the inner sheath is manipulated.
View-Synchronized LA-PV Mapping and Ablation
The workflow for CT-EAM registration was similar to that used previously by our group. A manually derived EAM of the aorta was registered to the corresponding 3D CT reconstruction. At this point, the registration is largely complete, so LA-PV mapping can be performed atop a view-synchronized 3D CT image. Although difficult to quantify, the initial experience with porcine LA-PV mapping emphasized the intuitiveness of this view synchronization (ie, the capability for the operator to move the catheter tip to the left by simply moving the 3D joystick in the same direction regardless of the orientation of the view being used). Given the small size of the porcine LA, the ability to robotically create accurate and realistic renderings of the LA-PVs is itself an important result. The ability to place periostial PV ablation lesions also was established in this initial experimental phase. Porcine RA mapping was performed to generate realistic renderings of the chamber. Because a 3D CT rendering of the RA was not available, view synchronization was not used. Instead, the RNS and EAM views were manually oriented to the same cardinal view. For example, both could be placed in a standard anterior view so that a rightward movement of the 3D joystick results in a similar rightward movement of the catheter tip on the EAM screen. One qualitative observation from these mapping studies was the stability of the catheter tip when positioned at various intracardiac locations such as the His electrogram and atrioventricular valve annuli.
The clinical experience with view-synchronized LA-PV mapping and ablation was similarly favorable. Although only a limited number of subjects were included, the variety of LA ablation lesion sets normally deployed clinically was evaluated. For example, an important procedural end point to catheter ablation of paroxysmal AF is electrical isolation of the PVs, an end point achieved in all patients in the present study regardless of whether the ablation was performed during sinus rhythm (n=6) or AF (n=3). Although a circular mapping catheter was used to confirm electrical PV isolation, it was not necessary to manipulate this mapping catheter constantly during lesion delivery. In the typical procedure workflow, after the circular catheter was manually placed in a PV, this catheter was not manipulated until both ipsilateral PVs were electrically isolated robotically with the ablation catheter. Then, the circular catheter was moved manually to the other ipsilateral PV to verify isolation, followed by movement to the contralateral side to monitor robotically driven isolation of those veins. Thus, ablation could be completely performed remotely and by a single operator. Electrical superior vena cava isolation was similarly performed.
In addition to PV and superior vena cava isolation, ablation of chronic AF typically involves the ablation of other targets such as sites of complex fractionated activity and placement of linear lesions like the LA roof and mitral isthmus lines, all of which were successfully performed in the present study. The electric continuity of the lines was verified by differential pacing. However, the small number of patients studied precludes an assessment of the relative effectiveness of robotic linear lesion creation with manual linear lesion formation. A larger experience is required to fully appreciate the effectiveness of robotic chronic AF ablation.
Both the overall procedure and ablation times during these studies were considerable. However, it is important to recognize that completion of the ablation lesion sets was not the only goal of these procedures. A considerable amount of time was expended for detailed robotic mapping of the LA-PVs, collecting data to assess the quality of image integration, and verifying tissue contact during selected ablation lesions using ICE. The time spent for these various maneuvers was not separately recorded, but the overall procedure times would certainly have been shorter if the only goal was ablation to isolate the PVs or to place linear lesions.
The Safety of View-Synchronized Robotic Ablation
No complications were observed either during or after the ablation procedure. However, a number of potential complications are worth discussing. First, the size of the SSS mandates that careful attention be paid to achieving hemostasis after removal. Second, one of the important qualitative advantages of robotic ablation appears to be the relative stability of the catheter tip. However, this could prove to be a disadvantage if excessive catheter pressure/power is applied to the tissue during ablation; in this situation, one may anticipate a greater chance for critical subsurface heating, leading to impedance “pops” and cardiac tamponade. This suggests the importance of future work in 2 areas: developing strategies for sensing tissue contact pressure and assessing the amount of power that should be delivered on the basis of the degree of contact. Third, no clinical evidence of esophageal damage could be observed in the present study, but neither intraprocedural esophageal temperature monitoring nor postprocedural esophageal endoscopy was performed. Fourth, embolism-related stroke related to either thrombus or air did not occur, probably because, although various radiofrequency ablation catheter designs were used in the experimental phase, only the irrigated catheter was used clinically—in large part because of its improved safety profile for thromboembolism.
Finally, catheter-related perforation of the heart did not occur. Because this was the first clinical experience with AF ablation, view synchronization with 3D CT images was used in all cases, a fact that may have contributed to the favorable safety record in the present study. However, it would be inappropriate to interpret this to mean that view synchronization is necessary to perform robotic LA mapping and ablation. Note that porcine RA mapping was performed without view-synchronized CT images, and in other preliminary clinical work, we have performed AF ablation without CT synchronization (unpublished observation, V.R. and P.N.). Similarly, other CT image registration strategies may prove clinically useful. Ultimately, although the present study establishes that robotic LA-PV mapping and ablation are feasible, the clinical superiority of this paradigm remains to be proved. Ultimately, randomized comparisons of robotic image-guided therapy with standard approaches are necessary to definitively establish its utility.
Prior Experience With Remote Navigation
On the use of the robotic system, 1 ex vivo and 1 in vivo canine study exist.10,11 However, both of these simply report movement to various general regions in the heart (eg, right ventricular outflow tract, left ventricular septum), not precise locations. Thus, creating full EAMs of the atrial chambers (in both animals and patients) represents a qualitative demonstrative leap. Second, it is important to recognize that electric PV isolation has never been described using this particular RNS. Furthermore, although a single clinical report exists on the use of an alternative remote magnetic navigation system for AF ablation, it is important to note that the procedural end point in that report was not electrical PV isolation in the standard electrophysiological sense.12 Instead, its end point was “>90% reduction in the bipolar electrogram amplitude, and/or peak-to-peak bipolar electrogram amplitude <0.1 mV inside the line”.12 Thus, this is not only the first report of electrical PV isolation using this particular RNS but also the only report of electrical PV isolation using any remote navigation system, robotic or magnetic. Third, the concept of CT view-synchronized remote image-guided therapy is, to the best of our knowledge, unique and has never been reported, including the aforementioned AF ablation study with the magnetic navigation system. On the other hand, a number of reports have been published on using the magnetic navigation system to map and ablate the slow pathway or accessory pathways in patients with supraventricular tachycardias.13,14
It is important to recognize that a leftward movement of the 3D joystick does not always translate to an exact leftward movement in the in vivo situation. As with manual catheter movement, anatomic obstacles must be properly negotiated. In addition, the resistance exerted by contact with the cardiac tissue alters whether the catheter tip will arrive at the exact location desired by the movement. However, this limitation was not restrictive enough to preclude achieving all of the mapping and ablation objectives of the present study.
Although certain technical limitations of manual mapping are overcome with robotic mapping, one must still master other skills required to perform remote mapping, including how to use the software for the RNS. Robotic mapping also requires the purchase of an additional expensive piece of equipment, which may not be feasible for any given institution.
Although only 2D ICE is currently available clinically, strategies are being developed for 3D ICE imaging. When available, these may be incorporated into the procedure to further minimize radiation exposure to the patient. However, to fully capitalize on these advancements, future iterations of the RNS will require a means for separately manipulating the 2D/3D ICE catheter in tandem with the ablation catheter.
Sources of Funding
This work was supported in part by Hansen Medical, Inc and a National Institutes of Health K23 award (HL68064) to Dr Reddy.
Dr Reddy has served as a consultant to Biosense-Webster, Inc. Drs Reddy and Neuzil have received research support from Hansen Medical, Inc. Dr Abbara serves as a consultant to E-Z-EM, Inc, and is on the speakers’ bureau for Siemens Medical Systems, Inc. The other authors report no conflicts.
Chen SA, Hsieh MH, Tai CT, Tsai CF, Prakash VS, Yu WC, Hsu TL, Ding YA, Chang MS. Initiation of atrial fibrillation by ectopic beats originating from the pulmonary veins: electrophysiological characteristics, pharmacological responses, and effects of radiofrequency ablation. Circulation. 1999; 100: 1879–1886.
Marrouche NF, Dresing T, Cole C, Bash D, Saad E, Balaban K, Pavia SV, Schweikert R, Saliba W, Abdul-Karim A, Pisano E, Fanelli R, Tchou P, Natale A. Circular mapping and ablation of the pulmonary vein for treatment of atrial fibrillation: impact of different catheter technologies. J Am Coll Cardiol. 2002; 40: 464–474.
Pappone C, Oreto G, Rosanio S, Vicedomini G, Tocchi M, Gugliotta F, Salvati A, Dicandia C, Calabro MP, Mazzone P, Ficarra E, DiGioia C, Gulletta S, Nardo S, Santinelli V, Benussi S, Alfieri O. Atrial electroanatomical remodeling after circumferential radiofrequency pulmonary vein ablation: efficacy of an anatomic approach in a large cohort of patients with atrial fibrillation. Circulation. 2001; 104: 2539–2544.
Gepstein L, Hayam G, Ben-Haim SA. A novel method for nonfluoroscopic catheter-based electroanatomical mapping of the heart: in vitro and in vivo accuracy results. Circulation. 1997; 95: 1611–1622.
Mikaelian BJ, Malchano ZJ, Neuzil P, Weichet J, Doshi SK, Ruskin JN, Reddy VY. Integration of 3-dimensional cardiac CT images with real-time electroanatomical mapping to guide catheter ablation of atrial fibrillation. Circulation. 2005; 112: e35–e36.
Ernst S, Ouyang F, Linder C, Hertting K, Stahl F, Chun J, Hachiya H, Bansch D, Antz M, Kuck KH. Initial experience with remote catheter ablation using a novel magnetic navigation system: magnetic remote catheter ablation. Circulation. 2004; 109: 1472–1475.
The past decade has seen a tremendous evolution in our understanding of how to treat atrial fibrillation with catheter ablation. However, these scientific advances have not been widely applied in clinical practice, in part because of the technical expertise required to safely and effectively perform the procedure. Significant patient-to-patient variability and complexity to the left atrial and pulmonary venous anatomy exist, and manual manipulation of the ablation catheter to these locations can be challenging. To overcome these limitations, this study examined the feasibility of performing atrial fibrillation ablation by remote robotic navigation of the catheter on a view-synchronized 3-dimensional computed tomographic rendering of the subject-specific left atrial pulmonary vein anatomy. The initial porcine experimental phase (9 animals) validated the ability of view-synchronized robotic navigation to guide atrial mapping and ablation. The consequent clinical phase (9 atrial fibrillation patients) established that this paradigm could be successfully applied to perform all of the major aspects of catheter ablation of paroxysmal or chronic atrial fibrillation: electrical PV isolation in an extraostial fashion, isolation of the superior vena cava, and linear atrial ablation of typical and atypical atrial flutters. Although the ideal ablation strategy to use in any given patient is still in evolution, this study demonstrates the safety and feasibility of an emerging paradigm for atrial fibrillation ablation involving the confluence of 3 technologies: volumetric 3-dimensional imaging, real-time electroanatomic mapping, and remote robotic navigation. Technological advances such as these will likely pave the way for a more generalized use of catheter ablation to treat this troublesome arrhythmia.
Presented in part at the 2006 Annual Scientific Sessions of The Heart Rhythm Society, Boston, Mass, May 19, 2006, and published in abstract form (Heart Rhythm. 2006;3:S91–S92).
The online-only Data Supplement, consisting of 2 movies and a figure, is available with this article at http://circ.ahajournals.org/cgi/content/ full/CIRCULATIONAHA.106.677369/DC1.