Electrophysiological Effects of Long, Linear Atrial Lesions Placed Under Intracardiac Ultrasound Guidance
Background A curative atrial fibrillation procedure will most likely rely on creating transmural linear ablative lesions. However, it is currently unknown whether endocardial radiofrequency lesions can create lines of conduction block.
Methods and Results In six pigs, intracardiac echocardiography was used to guide the positioning of multiple coil array catheters to bridge endocardial structures in three right atrial locations: (1) from the crista terminalis to the tricuspid annulus; (2) from the fossa ovalis to the crista terminalis; and (3) from the inferior vena cava to the tricuspid annulus. Once the catheter was positioned, linear lesions were made by radiofrequency energy applied sequentially to each of the four coils. After 15 days, the chest was opened and a 112-electrode epicardial plaque was positioned over the atrial free wall lesion to determine activation patterns. Three lesions were placed in each animal, with a mean procedure time of 47±11 minutes. Once adequate contact was determined by intracardiac echocardiography, a single series of radiofrequency application was required to achieve tissue heating (65±4°C) with a power of 21±10 W. Epicardial mapping demonstrated complete conduction block across the lesions in all animals, with split potentials and disparate activation times (64±16 ms) across the lesion. At autopsy, all lesions were discrete, continuous, and without evidence of charring. The lesions were within 0.3±0.5 mm of their targeted anatomic locations and measured 21±4 mm long and 2.8±0.6 mm wide. Histology revealed transmural fibrosis throughout the length of each lesion.
Conclusions Linear lesions that are electrophysiologically transmural and continuous can be placed in the right atrium of normal pigs. With intracardiac echocardiography, adequate tissue contact over several coil electrodes can be ensured, resulting in short procedure times, efficient energy application, and accurate anatomically linked lesion placement.
Atrial fibrillation has been treated with surgical procedures such as the maze, corridor, or compartment operation.1 2 3 4 Although the maze procedure has a high success rate, morbidity and mortality are associated with the open-heart surgical procedure. Recently, the principle behind the maze procedure—segmenting the atrium into areas too small to sustain reentry—has been adapted for transvenous ablation techniques in humans and animals.5 6 7 8 These early attempts have resulted in very long procedure times and recurrent atrial arrhythmias.
The surgical cures of atrial fibrillation rely on creating surgically placed lines of conduction block. To adapt this approach to a catheter-based ablation procedure, technology must be developed to readily create discrete, long, linear lesions at anatomically defined sites. In addition, these lesions must result in continuous lines of conduction block, which presumably need to be transmural. To date, it has not been demonstrated whether continuous lesions resulting in conduction block can be made with an endocardial approach using radiofrequency energy, particularly in the trabeculated right atrium. In addition, the electrophysiological effects of creating long, linear atrial lesions have not yet been described. In particular, on the basis of well-known animal models of atrial flutter, linear lesions placed in the right atrial free wall might be proarrhythmic by creating a substrate for macroreentry.9 10 11
For any lesion to be effectively created by radiofrequency energy application, contact between the electrode and the myocardium must be firmly established to minimize convective heat loss and maximize energy delivery to the tissue. Although minor deficiencies in electrode-tissue contact can be overcome by increasing the delivered power, poor (or absent) contact will render lesion formation impossible. Such problems with efficient energy delivery to the endocardium are likely to be particularly acute when contact must be established over the length of a multisegmented array of electrodes. Furthermore, fluoroscopy provides only very indirect estimates of catheter-endocardial contact. Alternatively, intracardiac echocardiography has been shown to reliably predict catheter-endocardial contact and has been used to guide radiofrequency ablations.12 13 14 15 16 17 18
We hypothesized that long, continuous lesions made with intracardiac echo guidance would be accurately guided to anatomic sites and would result in continuous, transmural lines of conduction block. We tested a novel multielectrode array catheter to determine whether histological and electrophysiological transmural, continuous lesions could be made and to determine whether these lesions are proarrhythmic.
The study was approved by the Committee on Animal Research of the University of California San Francisco. Six pigs weighing between 25 and 30 kg were intubated and anesthetized with isoflurane. After hemostatic sheaths were placed in the femoral veins, a 7F standard quadripolar electrophysiology catheter (Mansfield) with 2-mm electrodes was advanced into the high right atrium for pacing and recording. The animals did not receive anticoagulating agents at any time during the study.
A baseline electrophysiology study was performed, consisting of determination of SNRTs and programmed atrial stimulation. All pacing was bipolar, with a pulse width of 2 ms and at twice diastolic threshold. SNRTs were determined at paced cycle lengths of 400 and 600 ms after pacing for 30 seconds. Corrected SNRTs were calculated by subtracting the sinus cycle length from the return cycle length of the first sinus beat after pacing ceased. Programmed atrial stimulation consisted of single, double, and triple extrastimuli at basic drive cycle lengths of 400 and 300 ms at twice diastolic threshold. Extrastimuli were introduced in decrements of 10 ms until atrial refractoriness. After extrastimulation, burst pacing at twice diastolic threshold was performed at cycle lengths of 280 ms down to 20 ms in 10-ms decrements for 10 seconds in each drive. The programmed stimulation and burst pacing were repeated during isoproterenol infusion (to achieve a 20% increase in baseline sinus rate).
A custom-made, multiple coil (four 5-mm coils separated by 2 mm) array catheter (Medtronic-Cardiorhythm) was used to make three anatomically positioned continuous lesions, guided by intracardiac echo. The catheters have two planes of steering: a distal curve and a more proximal orthogonally directed curve. Unmodulated radiofrequency energy at 500 MHz was applied through a radiofrequency generator with temperature feedback (Medtronic-Cardiorhythm). Temperature feedback was via a thermocouple mounted in the center of each electrode. Ablations were performed with a set temperature of 65°C and with automatic adjustment of the power (up to a limit of 50 W) for 60 seconds through each coil. Radiofrequency energy was applied to each coil individually in sequence without the catheter being moved until the entire lesion was completed. Once contact and positioning were confirmed by intracardiac echo (see below), the catheter was not repositioned until energy application was completed through each coil. Each lesion was attempted only once in each animal.
The intracardiac echo system used has been described in detail elsewhere.15 16 17 Briefly, a 10F, 10-MHz intracardiac echo catheter (CVIS) was used for imaging of the right atrium. The catheter was advanced through the femoral vein sheath into the right atrium. Images were recorded on SVHS tape for later review. Intracardiac echo was used to guide the anatomic placement of the ablation catheter and to assess catheter-tissue contact. Lesions were targeted at the following anatomic sites: (1) in the trabeculated right atrium, from the CT coursing anteriorly to the TA; (2) on the smooth right atrium, from the FO coursing posteriorly to the CT; and (3) from the IVC to the TA. These sites were chosen because they represent potential anatomic lines of conduction block that may be appropriate targets for ablation of atrial fibrillation and because they are readily identified with intracardiac echo.
The use of intracardiac echo to assess catheter-tissue contact for ablation through single standard-tip electrodes (4 mm) has been reported previously.12 Similar assessment was used but applied to all four long coil electrodes. Good tissue contact was characterized by visualizing the entire length of the ablation segment (all four coil electrodes) adjacent to the atrial endocardial surface without a space between catheter and endocardium and without evidence of the catheter bouncing off the surface. In addition, good contact was characterized by the lack of catheter movement (sliding) on the endocardial surface. If intracardiac echo revealed any portion of the catheter not to be in good contact, the catheter steering and torque were used to improve the intracardiac echo–determined contact. Radiofrequency energy was applied only when good contact was confirmed for all electrodes. Once good contact was confirmed, the catheter was not repositioned until radiofrequency energy was applied to all coils. After ablation, the lesion was imaged with intracardiac echo to confirm anatomic placement.
The animals were allowed to recover for 14 days, at which time they were returned to the laboratory and underwent repeat electrophysiology testing as described above (same protocol as baseline). After the electrophysiology testing, the chest was opened via a median sternotomy, and an atrial plaque was sewn onto the anterior surface of the right atrium, with care taken to place the plaque generally over the visible lesion on the right atrial free wall. The plaque consisted of 112 electrodes (7×16) with a 3.5-mm interelectrode spacing covering an area of 56×21 mm (Prucka Engineering). Simultaneous bipolar recordings from all electrodes were made and stored digitally on a CardioMap system (Prucka Engineering). Pacing was performed from a bipole pair of the plaque through a custom-made switch box. Activation maps were constructed from plaque recordings obtained during pacing with the CardioMap system. Activation times were marked at the maximum positive or negative deflection. When split electrograms were recorded, activation times to the larger of the two components were marked. A difference in activation time of ≥50 ms in adjacent electrodes was considered to indicate conduction block.19
After the above studies, the animals were euthanized and the heart removed. The corners of the plaque were marked on the epicardial surface, and the plaque was removed. The anterior lesion (CT-TA) was identified on the epicardial surface of the heart. A clear plastic reproduction of the plaque was placed in the precise orientation and position of the actual plaque during the study. The lesion was traced on the plastic reproduction for correlation to the activation maps obtained. The atria were then opened and the lesions identified on the endocardial surface of the atrium. The lesion dimensions (length and width) and the distances from the targeted anatomic sites were measured.
After the excised hearts were fixed for 24 hours in formalin, each of the three lesions was blocked and sectioned parallel to the long axis of the lesion in a plane perpendicular to the endocardial surface (cut from endocardium to epicardium along the lesion length). Sections were stained with hematoxylin-eosin and Masson’s trichrome for microscopic examination of the lesions.
Values are expressed as mean±SD. Paired t tests or ANOVA was used to determine statistical significance.
All radiofrequency applications were made with good catheter-tissue contact, as assessed by intracardiac echo. The CT, FO, TA, and IVC were identified in all pigs by intracardiac echo. Fig 1⇓ demonstrates the intracardiac echo image of the region of the trabeculated right atrium between the CT and TA. In Fig 1B⇓, the catheter is positioned between these structures before ablation. Fig 1C⇓ demonstrates the ablation lesion after radiofrequency application in the region between the CT and TA. Fig 2⇓ shows the corresponding fluoroscopy images of this catheter position.
A total of 18 long, linear lesions were made in six pigs. All three lesions in each animal were made with a mean total fluoroscopy time of 11.3±4 minutes and a total procedure time of 47±11 minutes. There was a learning curve associated with the procedure, with the first three studies requiring 53 to 60 minutes to create the three lesions and the latter three studies requiring 36 to 39 minutes (P=.0009) (Fig 3⇓). Every radiofrequency application resulted in adequate tissue temperature (>60°C) for all lesions in every animal. No “test applications” or catheter repositioning after attempted radiofrequency was necessary, once the catheter position and tissue contact were confirmed on intracardiac ultrasound.
The mean temperature achieved during all radiofrequency applications was 65±2.2°C (range, 60°C to 67°C). The mean power required to achieve these temperatures was 20.3±6.7 W (range, 14 to 35 W). The average efficiency-of-heating index (calculated as temperature/power) was 4.36 (range, 2 to 6) (Fig 4⇓). There was no statistical difference in the temperature achieved (65±3.1°C for CT-TA, 65±2°C for FO-CT, 65±1.7°C for TA-IVC), in power required (20.5±8.6 W for CT-TA, 19.7±7 W for FO-CT, 20.8±5.9 W for TA-IVC), or in efficiency-of-heating index (4.6±1.6 for CT-TA, 4.6±1.3 for FO-CT, 3.9±1.6 for IVC-TA) among the three lesion locations (Fig 4⇓).
There were no long- or short-term complications of the procedure. There was no difference in the baseline heart rate or corrected SNRTs at baseline compared with 2 weeks after ablation (Table 1⇓). One pig had an inducible atrial tachycardia at baseline. This was inducible with burst pacing (cycle length, 50 ms) and persisted for 3 minutes before spontaneously converting to sinus rhythm. The cycle length of this tachycardia was 110 ms. This animal had an inducible atrial tachycardia at follow-up testing with a cycle length of 165 ms, induced with burst pacing (cycle length, 40 ms) during isoproterenol infusion only. This was nonsustained, however, lasting only 25 seconds. Mapping with the epicardial plaque demonstrated that this was not due to reentry in the free wall (or around the CT-TA lesion). In another animal, nonsustained atrial fibrillation (<1 minute in duration) was induced with rapid burst pacing (cycle length, 20 ms) before ablation. At follow-up testing, no arrhythmia was induced in this animal. No other animal had any inducible arrhythmias. No animal that was not inducible before ablation had an inducible arrhythmia after ablation.
The linear lesion on the anterior aspect of the right atrium (CT-TA) was mapped. A line of conduction block was found in all animals during pacing from a bipole pair on the plaque (Fig 5⇓). A significant difference in activation times (64±16 ms) between adjacent bipoles on either side of the lesion was found (P<.00001). In addition, split electrograms or fragmented electrograms were recorded from electrodes in the region of this conduction block in every animal. The fragmented electrograms were recorded at the edges of the lesions (one instance) or adjacent to electrodes where split electrograms were recorded (four instances). The mean difference between the components of the split electrograms was 60±14 ms. In addition, activation on the left side of the apparent line of block opposite the pacing site (on the right of the conduction block) occurred left to right instead of the expected right to left. This is further evidence of a line of conduction block, as opposed to conduction slowing. However, because more complete mapping was not possible, the source of this retrograde wave front was not determined.
The lines of conduction block on the activation maps were at the location of the ablation lesion on the anterior right atrium in all animals.
Three lesions were visually apparent in all animals. They appeared as discrete pale discolorations in the atrium (Fig 6⇓). There was no evidence of charring or thrombus formation on any lesion. The mean lesion length was 21±4 mm, and lesion width was 2.8±0.6 mm. The mean lesion lengths for each of the three locations are shown in Table 2⇓. All lesion borders were well demarcated and continuous on gross examination. The lesions were within 0.3±0.5 mm of their targeted anatomic sites. In addition, there was 100% correlation of the location of the CT-TA lesion in relation to the epicardial plaque placement and the sites of electrophysiological conduction block on the activation map.
Histological examination of the lesions revealed loss of myocytes with collagen and fat deposition in all lesions. All lesions were histologically continuous and transmural throughout their length (Fig 7⇓). An organized endocardial thrombus <0.5 mm thick was seen on one lesion.
We have shown that long, linear lesions can be placed in the right atrium with multielectrode coil catheters. Using high-density epicardial plaque mapping, we have demonstrated that these lesions are continuous and transmural electrophysiological lines of conduction block as well as being histologically transmural and continuous. These lesions were accurately targeted to specific anatomic structures by intracardiac echo. In addition, when intracardiac echo was used to prospectively assess catheter-tissue contact, adequate tissue heating of >60°C was achieved at low power (<50 W). This resulted in no coagulum formation or charring.
Because lesions were made in pigs without atrial fibrillation, no statement can be made as to whether the lesions created would be palliative or curative for atrial fibrillation; however, we have demonstrated the principle of surgical atrial segmentation to create transmural linear scar with catheter ablation and shown that it blocks propagation, such that wandering wave fronts are abolished. We used a high-density multielectrode epicardial plaque to confirm that linear lesions placed endocardially in the trabeculated right atrium are indeed continuous and transmural lines of conduction block. Kadish and Spear19 have demonstrated that the electrogram morphology and isochronal activation patterns are useful in distinguishing conduction block created in a tissue bath with cut lesions with multielectrode plaques. In particular, the recording of split electrograms, a large gap in activation times between adjacent sites, and disparate activation sequences are all evidence of conduction block.19 In our study, all of these criteria were met. Marked differences in activation times (>60 ms) occurred between adjacent electrodes over the lesions. Split potentials that spanned the gap in activation times were recorded from electrodes overlying the lesion. In addition, disparate activation sequences were observed on either side of the lesion. Although fragmented electrograms were rarely recorded, they were recorded at the edges of the lesion or at sites adjacent to where split electrograms where recorded. This suggests that slowed conduction may occur before the lesion or at the edges but does not indicate slow conduction across the lesion.
The lesions that were placed were not proarrhythmic. No new tachycardia was induced 2 weeks after ablation despite a very aggressive programmed stimulation protocol. Animal models of atrial flutter and patients with incisional reentry involving surgical repair of congenital heart disease have demonstrated that artificial obstacles placed in the right atrium may lead to macroreentrant arrhythmias.9 10 11 20 21 22 23 24 When these obstacles are extended to bridge two barriers, macroreentry does not occur.9 10 20 We placed each of our lesions in such a way as to bridge two anatomic barriers, and as such, macroreentry did not result from the ablation lesions. Intracardiac echo was used to confirm that the lesions bridged anatomic obstacles.
In our study, no sinus node dysfunction was observed after ablation. Previous studies have demonstrated that extensive ablation along the CT results in suppressed sinus node activity.13 18 However, because our lesions were shown by intracardiac echo to intersect the CT perpendicular to its long axis, only a small portion of our lesions were on the CT and, hence, sinus node dysfunction was not observed. It should be noted that the only way to identify the CT is with intracardiac echo, because fluoroscopy does not delineate such endocardial structures.
The creation of linear lesions, at least in the right atrium, appears to be safe. No complications, such as perforation or valvular damage, were observed during or after the procedure.
Intracardiac echo was used both to guide the anatomic placement of the lesions and to ensure adequate catheter-tissue contact of all coil electrodes. Intracardiac echo was accurate (within 0.3 mm) at guiding the anatomic targeting of linear ablation lesions. Intracardiac echo has been shown to be highly accurate at guiding anatomic placement of standard ablation lesions with a single 4-mm-tip ablation electrode and in some cases, visualizing the actual ablation lesion.15 16 However, this is the first study to assess the accuracy and use of intracardiac echo at guiding the placement of long, linear lesion by ablation through multiple long electrodes along the shaft of a catheter. We have shown that intracardiac echo can confirm location of long, linear lesions after energy application. Such “real-time” confirmation of lesion continuity might have a particular advantage over fluoroscopic guidance when lesions cannot be seen.
By direct visualization of the catheter-tissue interface, intracardiac echo was used to ensure adequate tissue contact of each coil electrode on the multielectrode catheter before energy application. Because catheter-tissue contact was assessed before ablation, there were no electrophysiological or histological gaps in the lesions. That a temperature of 65°C was achieved at low power (<30 W) for each energy application with no charring or coagulum formation is further evidence of the utility of intracardiac echo in prospectively predicting firm tissue contact.
Comparison With Other Studies
Few studies have been published regarding the creation of long, linear lesions in the atrium. Two techniques for creating such lesions have been suggested. One is a drag technique, whereby a standard-tip catheter is dragged over several centimeters during radiofrequency application.6 The other technique involves sequential ablation through multiple small ring electrodes along a catheter.5 The efficacy of either of these techniques at producing lines of conduction block has not been demonstrated. The dragging technique requires very long procedure and fluoroscopy times.6 Ablation through small ring electrodes is also time-consuming and has been shown to be ineffective at creating continuous lesions in the trabeculated right atrium.25 It is unclear whether the high failure rate is due to ineffective lesion formation or whether the lesion set is ineffective.8
Under guidance by intracardiac echo, discrete lesions can be made with a short procedure time (40 minutes) and fluoroscopy time (11 minutes). In addition, each lesion was made with a single series of energy applications without repositioning, once location and contact were determined by intracardiac echo. Therefore, the amount of atrial tissue that was destroyed to create three lines of conduction block was minimal. This may have implications in the clinical setting for preservation and restoration of atrial mechanical function.
Although we did not place linear lesions in an animal model of atrial fibrillation, we did demonstrate that discrete lines of conduction block can be made with radiofrequency ablative techniques. Further study is needed to demonstrate the utility of this technique in large atria and in the left atrium. In addition, we did not demonstrate that the set of lesions made would cure atrial fibrillation. We did, however, demonstrate that anatomically based lesions can be accurately placed under intracardiac echo guidance. It is clear from the work of Cox et al3 and Swartz et al6 that an ablative cure of atrial fibrillation will be at least in part anatomically based. However, further study will need to focus on appropriate lesion sets to cure atrial fibrillation.
Only the anterior lesion (CT-TA) was mapped with the epicardial plaque; therefore, the other two lesions were not shown to be electrophysiological lines of block. The FO-CT and IVC-TA lesions are not accessible with an epicardial plaque. However, the histological appearance was continuous and transmural for all lesions. In addition, because the entire atrium was not mapped, it was not determined how the tissue distal to the lesion was activated.
We have demonstrated that with proper tools, discrete linear lesions that create lines of conduction block can be readily placed. With intracardiac echo guidance, these lesions can be anatomically targeted and created with low power when contact is prospectively assessed. In addition, with these tools, the procedure in the right atrium appears to be safe and not to be proarrhythmic, which has been a major complication in previous studies.6 7 8 Because the procedure time was relatively short and because no charring or coagulum formation was seen, similar techniques in the left atrium would probably be safe if intracardiac echo were used to guide ablation. However, current technology is limited in its ability to image the left atrium. As the technology improves to allow imaging of the left atrium from the right atrium with better depth of penetration and steerability, further studies will need to be performed to assess the safety and efficacy of such a procedure. However, for right-side lesions, intracardiac echo is useful and results in discrete linear lesions being accurately placed with brief procedure times.
Selected Abbreviations and Acronyms
|IVC||=||inferior vena cava|
|SNRT||=||sinus node recovery time|
This study was supported in part by a grant from Medtronics-Cardiorhythm, Inc. Dr Olgin is supported by the California Heart Association and is the recipient of a California Heart Association postdoctoral fellowship grant.
- Received December 9, 1996.
- Revision received June 3, 1997.
- Accepted June 6, 1997.
- Copyright © 1997 by American Heart Association
van Hemel N, Defauw JJ, Kingma JH, Jaarsma W, Vermeulen FE, de Bakker J, Guiraudon GM. Long-term results of the corridor operation for atrial fibrillation. Br Heart J. 1994;71:170-176.
Feinberg MS, Waggoner AD, Kater KM, Cox JL, Lindsay BD, Perez JE. Restoration of atrial function after the maze procedure for patients with atrial fibrillation: assessment by Doppler echocardiography. Circulation. 1994;90(pt 2):II-285-II-292.
Swartz J, Pellersels G, Silvers J, Patten L, Cervantez D. A catheter-based curative approach to atrial fibrillation in humans. Circulation. 1994;90(suppl I):I-335. Abstract.
Haines D, McRury I. Primary atrial fibrillation ablation in a chronic atrial fibrillation model. Circulation. 1995;92(suppl I):I-265. Abstract.
Jais P, Haissaguerre M, Shah D, Gencel L, Garrigue S, Lavergne T, Clementy J. Staged approach for paroxysmal atrial fibrillation ablation. Pacing Clin Electrophysiol. 1996;19:581. Abstract.
Frame L, Page R, Hoffman B. Atrial reentry around an anatomic barrier with a partially refractory excitable gap: a canine model of atrial flutter. Circ Res. 1986;58:495-511.
Kalman JM, Lee RJ, Fisher WG, Chin MC, Ursell P, Stillson CA, Lesh MD, Scheinman MM. Radiofrequency catheter modification of sinus pacemaker function guided by intracardiac echocardiography. Circulation. 1995;92:3070-3081.
Chu E, Fitzpatrick AP, Chin MC, Sudhir K, Yock PG, Lesh MD. Radiofrequency catheter ablation guided by intracardiac echocardiography. Circulation. 1994;89:1301-1305.
Olgin J, Kalman J, Fitzpatrick A, Lesh M. The role of right atrial endocardial structures as barriers to conduction during human type I atrial flutter: activation and entrainment mapping guided by intracardiac echocardiography. Circulation. 1995;92:1843-1848.
Lee RJ, Kalman JM, Fitzpatrick AP, Epstein LM, Fisher WG, Olgin JE, Lesh MD, Scheinman MM. Radiofrequency catheter modification of the sinus node for ‘inappropriate’ sinus tachycardia. Circulation. 1995;92:2919-2928.
Kadish A, Spear J. Identification of conduction block in cardiac muscle: in vitro observations in canine epicardium. Cardiovasc Res. 1994;28:259-269.
Frame LH, Page RL, Boyden PA, Fenoglio JJ, Hoffman B. Circus movement in the canine atrium around the tricuspid ring during experimental atrial flutter and during reentry in vitro. Circulation. 1987;76:1155-1175.
Kalman J, Van Hare G, Olgin J, Stark S, Lesh M. Ablation of ‘incisional’ reentrant atrial tachycardia complicating surgery for congenital heart disease: use of entrainment to define a critical isthmus of conduction. Circulation. 1996;93:502-512.
Lesh MD, Van Hare GF, Fitzpatrick AP, Griffin JC, Chu E. Curing reentrant atrial arrhythmias: targeting protected zones of slow conduction by catheter ablation. J Electrocardiol. 1993;20:194-203.
Cronin CS, Nitta T, Mitsuno M, Isobe F, Schuessler RB, Boineau JP, Cox JL. Characterization and surgical ablation of acute atrial flutter following the Mustard procedure: a canine model. Circulation. 1993;88(suppl II):II-461-II-471.
He D, Mackey S, Marcus F, Graham A, Haissaguerre M, Lampe L. Radiofrequency energy for cardiac ablation using a multipolar catheter in sheep. Circulation. 1995;92(suppl I):I-265. Abstract.