The role of gangliar plexi in the treatment of atrial fibrillation

Introduction: Gangliar plexi (GP) are bundles of interconnected neurons found embedded in FPs (FPs) surrounding the atria of the heart. The function of these neurons is controversial but they may form a localized control system for the heart and play a role in the induction of atrial fibrillation (AF), the most common form of heart rhythm disturbance. During the treatment of AF, catheters inserted into the heart via the venous system apply radio frequency (RF) energy to sites on the inner surface of the heart in order to disrupt the triggering of AF. In many cases, the sites of ablation overlap the locations of the GP and we hypothesize that the resulting damage to the myocardial tissue adjacent to the GP may contribute to success of the ablation procedure. Methods: To identify the sites of ablation and locations of the FPs containing GP, we acquired late gadolinium enhanced (LGE) MRI scans and dark blood (DB) MRI scans and analyzed them using segmenting software from the SCI Institute at the University of Utah called Seg3D. Regions of enhanced intensity in LGEMRI scans identified the lesion locations, which we segmented to create a mask or template. The DBMRI scans provided locations of the FPs, which we also segmented into six separate locations in each heart. To determine overlap between ablation sites and GP, the LGEMRI and DBMRI images were registered (aligned) and the extent of overlap quantified. Conclusion: The findings do not support the hypothesis in that ablating the myocardial tissue adjacent to the GP in the FPs surrounding the left atrium results in improved success of the ablation procedure for curing AF. The mechanism of initiation and maintenance of AF is still a mystery, but through this study we were able to narrow the search for truth regarding the origin of AF.

THE ROLE OF GANGLIAR PLEXI IN THE TREATMENT OF ATRIAL FIBRILLATION by Tyler Marler Major contributions made by Shawn Tate, Josh Blauer, Eugene Komolvski, Nassir Marrouche, and Rob MacLeod A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In Biomedical Engineering Approved: ____________________ Rob MacLeod Supervisor ____________________ Patrick Tresco Chair, Department of Bioengineering ____________________ Bradley Greger Department Honors Advisor ____________________ Sylvia Torti Dean, Honors College May 2012 ABSTRACT Introduction: Gangliar plexi (GP) are bundles of interconnected neurons found embedded in FPs (FPs) surrounding the atria of the heart. The function of these neurons is controversial but they may form a localized control system for the heart and play a role in the induction of atrial fibrillation (AF), the most common form of heart rhythm disturbance. During the treatment of AF, catheters inserted into the heart via the venous system apply radio frequency (RF) energy to sites on the inner surface of the heart in order to disrupt the triggering of AF. In many cases, the sites of ablation overlap the locations of the GP and we hypothesize that the resulting damage to the myocardial tissue adjacent to the GP may contribute to success of the ablation procedure. Methods: To identify the sites of ablation and locations of the FPs containing GP, we acquired late gadolinium enhanced (LGE) MRI scans and dark blood (DB) MRI scans and analyzed them using segmenting software from the SCI Institute at the University of Utah called Seg3D. Regions of enhanced intensity in LGEMRI scans identified the lesion locations, which we segmented to create a mask or template. The DBMRI scans provided locations of the FPs, which we also segmented into six separate locations in each heart. To determine overlap between ablation sites and GP, the LGEMRI and DBMRI images were registered (aligned) and the extent of overlap quantified. Conclusion: The findings do not support the hypothesis in that ablating the myocardial tissue adjacent to the GP in the FPs surrounding the left atrium results in improved success of the ablation procedure for curing AF. The mechanism of initiation and maintenance of AF is still a mystery, but through this study we were able to narrow the search for truth regarding the origin of AF. ii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 METHODS 4 RESULTS 8 DISCUSSION 11 REFERENECES 13 iii 1 I. Introduction In a normal heart, the propagation of electrical signals is uniform and efficient. The sinoatrial (SA) node is responsible for pacing the beat of the heart, and is often known as the pacemaker. The cells contained within the sinoatrial (SA) node are automatic meaning that they are spontaneously excited and generate an electrical response that is propagated throughout the heart eliciting one synchronized beat of the heart. Nonautomatic myocardial cells demonstrate a rapid depolarization, followed by repolarization and then a stable resting potential persisting until the next action potential arrives. “All cells which demonstrate… slow diastolic depolarization are said to be automatic [8].” The initiation and propagation of abnormal electrical signals from multiple locations can cause irregular beating of the heart; when the irregularity is widespread and completely uncoordinated, it is termed fibrillation. Fibrillation can occur in all parts of heart, but the most common is in the atria, or upper chambers, where it is known as atrial fibrillation (AF). AF occurs in over 2.3 million Americans and increases dramatically the risk of stroke, as it also reduces quality of life [1]. The source of these abnormal electrical signals is not known. There are three types of classifications for AF. The mildest form of AF is called paroxysmal AF and occurs in patients who have recurrent episodes of AF lasting longer than 30 seconds and self-terminate in less than 7 days. The next form of AF is more difficult to cure and is called persistent AF. This is where the patient has recurrent episodes of AF that last more than 7 days. Long-standing persistent AF is an ongoing long-term episode. Those with persistent or long-standing persistent AF need to be ablated more extensively. This is because the longer a person has AF, without it being cured, the worse the disease gets; “AF begets AF” [12].Those with paroxysmal AF have a much higher chance of being cured after the first ablation. The MAZE procedure was developed in the 1990’s and is still used today to cure longstanding persistent AF. The MAZE procedure consists of “cutting and sewing the atria in complex pieces” [4]. There are many limitations to this approach. The procedure is very complicated and requires open heart surgery. The patient then has a large incision in their chest, has to be placed on a cardiopulmonary bypass machine during the surgery, finally the patient has a long recovery time in and out of the hospital. The MAZE procedure was the “role model” for a more non-invasive procedure called ablation [4]. Radio frequency (RF) ablation for cardiac arrhythmias is the standard treatment for patients with AF where AF cannot be controlled by medication. An ablation procedure requires the use of multiple modalities. The first modality is fluoroscopy used at the beginning of the procedure when placing catheters and throughout the procedure. The second modality is echocardiogram (ECHO) ultrasound used to rule out thrombus and establish cardiac function. Sheaths are inserted into the femoral vein on both legs of the 2 patient for multiple catheter insertions. The third modality consists of multiple diagnostic catheters used to induce and characterize arrhythmias. Arrhythmias are electrically induced in order to identify locations of abnormal electrical activity within the left atrium during the ablation procedure. This technique enables the electrophysiologist (EP) to ablate areas within the left atrium that would otherwise be undetectable. A needle is then inserted through the catheter to enter into the right atrium and puncture the septum to gain access into the left atrium. The tip of the RF catheter is cooled to reduce the risk of clot formation. A CARTO Lasso Variable Circular Mapping catheter is then inserted. The electrodes on the CARTO catheter are then used to create a 3D geometry or model of the left atrium used during the ablation procedure. This is a better technique than fluoroscopy because fluoroscopy uses radiation in order to create a real time image inside the body (real-time x-ray). After this 3D image is created the EP can then start burning heart tissue in the left atrium. This can be performed by manual manipulation of the ablation catheter or using magnets to maneuver the ablation catheter throughout the left atrium. The advantage to the magnetic maneuvering is that it is virtually impossible to poke a hole through the left atrium with the tip of the ablation catheter. These systems have only been available in the past 6 years. The pulmonary vein sleeves are a region of the left atrium where abnormal electrical signals are known to originate [3]. The mechanism by which these abnormal electrical signals initiate is not known. Ablation procedures have emerged in order to isolate and prevent these abnormal signals from entering the left atrium. Ablating the pulmonary veins alone, termed pulmonary vein isolation (PVI), results in a 50% success rate in patients with AF [4]. This led to research in other areas where these abnormal signals can originate. “Posterior Wall Debulking” (PWD) is a secondary procedure ablating the posterior wall of the left atrium and was conceived by Dr. Nassir Marrouche [6] to achieve a higher success rate. Dr. Marrouche believes that the middle region of myocardial tissue between the pulmonary veins located on the posterior wall of the left atrium is a center for abnormal electrical signals to originate using re-entry [8]. PWD increases the success rate from 50% to 70% [6]. The areas of the left atrium that are ablated using the PWD technique are variable from patient to patient. There is also additional risk to the patient when PWD is performed. The posterior wall of the left atrium is often located next to the esophagus and if ablated too aggressively could burn through the tissue pouring blood into the patient’s esophagus. There is a temperature probe that is inserted into the patient’s esophagus to monitor the esophageal temperature and attempt to prevent the formation of an esophageal fistula. The esophageal temperature probes are not perfect, in fact, the HFS radio frequency waves emitted from the ablation catheter tip cause inflammation in the myocardial tissue which could result in an ulcer post operation. This is the main reason why many EPs do not use the PWD ablation technique. It also requires a great deal of EP experience in order to know where on the posterior wall of the left atrium to ablate and not cause holes to form. 3 Some cardiologists believe that there are electrical intersections, or “convergence points” in the atria [2], places where abnormal signals can originate. Some of these convergence points are proposed to exist embedded in the pads of fat that lie on the outside of the heart and form what are known as ganglionated plexi (GP), networks of interconnected nerves. These cardiologists use ablation procedures that focus on ablating the GP surrounding the heart [5], but have not focused directly on the FPs surrounding the heart. The intrinsic cardiac autonomic nervous system (ICANS) controls the electrical activity in the heart. It forms a neural network and has been shown to be a crucial part of the initiation and maintenance of atrial fibrillation. There are four critical GP localized into four distinct areas namely, superior left (SL) GP, inferior left (IL) GP, inferior right (IR) GP, anterior right (AR) GP which are identified and characterized using high-frequency stimulation (HFS) [2]. The parasympathetic subdivision of the autonomic nervous system slows the heart rate. When atrial fibrillation occurs there is something wrong with the signals that are generated by the parasympathetic nervous system, because signals may be firing faster or slower than the SA node is pacing. A research group in Oklahoma presents convincing evidence of the GP being the source, or cause of AF: In the first 83 patients with either paroxysmal or persistent GPablation alone decreased the incidence of spontaneous PV firing (without isoproterenol) from 54 of 83 (65.1%) patients to 12 of 83 (14.5%) patients (P < 0.01), suggesting that the activity of the intrinsic cardiac ANS plays a significant role in spontaneous PV firing. These 83 patients underwent both GP ablation and PV antrum isolation and were followed for a mean of 22 months. The percent of patients free of symptomatic atrial fibrillation, or atrial tachycardia after a single ablation procedure was 80% at 12 months and 86% at 22 months. [2] In a study done at the Tokyo Medical and Dental University, Tokyo, Japan canine hearts were used to understand the mechanism of autonomic innervations. “Vagal stimulation converted the PV tachycardia into sustained AF consistently in the two different models only when the tachycardia rates reached a critically high frequency … This study suggested that vagal stimulation operates as a modulator on the substrate for perpetuating AF as well as on the AF- trigger in normal canine hearts” (Horikawa-Tanami, et al. 540). By stimulating the vagus nerve with high frequency stimulation (HFS) these researchers were able to induce AF in these dogs. The signal that was generated by the researchers propagated down the parasympathetic pathway and triggered AF at the site of the PV Ostia. This supports the hypothesis that GP play a large role in the initiation of AF and are the primary cause of AF, because they are an integral part of the parasympathetic pathway within the ICANS. Dr. Dana Peters has identified fat areas surrounding the left atrium of AF patients. These areas of fat have been termed FPs. The FPs surrounding the left atrium have been 4 histologically shown by Dr. Jackman to contain autonomic ganglia [2]. The FPs characterized by Dr. Peters will be identified in my patient cohort in order to determine their relevance in the initiation of AF. The purpose of this correlation study is to analyze MRI scans and quantify the overlap of scar tissue and FPs to see if any patterns emerge. If there is a high percentage of overlap between scar and fat in a specific region (for example FP 2) then there is evidence for abnormal electrical signals originating in the GP. It is hypothesized that there will be a higher overlap percentage of FP and scar in cured (non-recurrent) patients than uncured (recurrent) patients and that by ablating those regions of greater overlap percentage in the uncured patients during the second ablation this will lead to the curing of AF in those patients. II. Methods A. Cohort Approval for this study was obtained by the CARMA Center from the IRB (approval number – 00020347). A cohort of 26 AF patient’s were considered for this study. The patient cohort consisted of all three AF types: paroxysmal, persistent, and long-standing persistent. Sixteen of them had a left atrium ablation once with no recurrences 11 months on average, with a standard deviation of 6.17 months. The other ten patients had recurrences after the first ablation. These patients were ablated a second time with no recurrences after 10 months on average, with a standard deviation of 6.74 months. The cohort was based solely on MRI scan quality for the late gadolinium enhanced (LGE) and dark blood (DB) MRI scans. All of the patients received PVI and PWD ablations. GP ablation was not performed on any of the patients in this study. The majority of MRI images were recorded using a 1.5 Tesla MRI machine. The rest were recorded using a Tesla MRI machine. The 1.5 and 3 Tesla MRI Machines were made by Seimens Healthcare in Erlangen, Germany. B. Image Types Two types of MRI scans were used to correlate fat and myocardial scar tissue in the left atrium. Dark Blood (DB), also known as Turbo-spin Echo (TSE) MRI scans revealed the fat surrounding the left atrium and Late Gadolinium Enhanced (LGE) MRI scans postablation revealed the scar tissue. The LGE-MRI and DB-MRI scans were acquired for all patients months post-ablation (scar is completely formed). The fat surrounding the left atrium has been categorized into regions known as FPs as shown in figure 1 [7]. 5 Figure 1. DB-MRI scan depicting FPs 1-6 surrounding the left atrium in two patients. Four axial slices were used to show all six FPs. The FPs are represented by higher intensity or brighter signal. C. Image Segmentation Segmentation is a process in which certain anatomical features are outlined using software to highlight the desired regions. Seg3D, invented by the SCI Institute at the University of Utah, was used to segment the left atrium (LGE-MRI) and the FPs (DBMRI). The scar is identified in the LGE-MRI scans as bright or intensified resolution above a certain threshold specific to each patient scan (fig. 2). The FP segmentations, as seen in the bottom left quadrant of figure 2, were identified using a pixel intensity threshold specific to each patient DB-MRI scan. In order to capture the scar tissue in the scan the endocardium and epicardium of the left atrium need to be segmented to determine the middle region of scar. 6 A B FP #1 FP #3 FP #5 FP #2 FP #6 D C Raw MRI Scans Segmented MRI Scans Figure 2. Seg3D segmentations of the left atrium of a single patient. A) and B) depict DB MRI scans. C) and D) depict LGE MRI scans. A) Raw DB MRI scan. B) Segmented DB MRI scan depicting the FPs (FPs) segmented (only 5 out of 6 analyzed are shown). Different colors depict FP regions (i.e. blue is FP #2, etc.). C) Raw LGE MRI scan. D) Segmented epicardium surface of the left atrium. D. Image Processing Registration and Pipeline First, segmented patient scans were registered (aligned) using transformation matrix software that took two like images as an input and outputted a transformation matrix which mapped the images on top of one another. Second, these transformation matrices were imported into SCIRun, a program created by the SCI Institute as a general purpose problem solving environment. Third, a pipeline was created in SCIRun that required four inputs: (1) FPs from the DB-MRI images, (2) scar tissue from the LGE-MRI images, (3) 7 Epicardium segmentations from the LGE-MRI scans, and (4) transformation matrices. These four inputs were used to create a 3D rendering of the FPs overlaid onto the epicardial surface of the left atrium with the scar identified on the surface of the left atrium (fig. 3). Figure 3. Depiction of FP and scar from a posterior and anterior view of the left atrium of a patient in the study. Both images were rendered in SCIRun. The FP are colored according to the color scale in the bottom left of the two images (blue – FP 1, teal – FP 2, etc.). The scar is the lighter gray. Normal heart tissue is the darker gray color. This shows the overlap of FP and scar in a patients’ heart that has underwent an ablation procedure. Quantification of FP and Scar Overlap The final stage of image processing was to quantify the amount of correlation that occurred between the scar and fat. Using a novel algorithm created in SCIRun, orthogonal vectors (5 mm in length) pointing inward and outward from the epicardial surface were constructed to search for scar from the LGE-MRI scans and fat from the DB-MRI scans [13]. The epicardial surface was separated into voxels, which are 3dimensional cubes that define a specific location. There were 31,716 voxels created from the surface of the epicardium. The scar profile data collected by the SCIRun algorithm consisted of 0’s and 1’s representing each voxel where a “1” represented scar. The FP profile consisted of numbers 0-6 where “0” represented a voxel with no FP and 1-6 represented FPs 1-6 respectively. The 1’s from the scar profile were matched up with the 8 1-6 numbered voxels from the FP profile using MATLAB in order to find a percentage of FP within scar correlation. E. Statistics A two sample t-test for independent samples with unequal variances was performed to compare the non-recurrent patients (one time ablation) to the recurrent patients (2 ablations) at two different time points (recurrent patient set only) to determine if there is a significant difference between the two data sets. III. Results There were two main patient sets that were compared and contrasted against one another. They consisted of non-recurrent (one ablation) and recurrent patients (two ablations). The percentage of overlap between the scar and FP was quantified using SCIRun and MATLAB for the non-recurrent and recurrent patient sets post 1st ablation. The average percentage of scar within the FPs surrounding the left atrium after the 1st ablation of both non-recurrent and recurrent patient sets is shown below in figure 4. 9 Figure 4. Comparing average percentage of scar within FP between non-recurrent and recurrent patients. The two patient sets shown above consisted of patients who were ablated once for atrial fibrillation (AF) and were cured (non-recurrent, n=16), and patients who had an AF episode lasting longer than 30 seconds leading to a second ablation (recurrent, n=10). Figure 5 is depicting the recurrent patients that had to undergo two ablation procedures before they were cured of AF. The FPs to scar overlap percentages were compared for each FP to determine if there was more a higher FP to scar overlap percentage after the second ablation compared to the first. This would suggest that GP found in the FP, after the first ablation, were still able to generate abnormal electrical signals and cause the left atrium to fibrillate causing there to be a need for a second ablation. Therefore if there was a higher overlap percentage after the 2nd ablation this would show that during the 2nd ablation more myocardial tissue was ablated adjacent to the FPs causing the formation of scar tissue and blocking abnormal electrical signals from entering into the left atrium resulting in abatement of AF. 10 Figure 5. The red bars represent the recurrent patient’s FP’s post 1st ablation with the same demographics as in figure 4. The green bars represent the average percentage of FP and scar overlap for the recurrent patients post 2nd ablation. The p-values are shown below in Figure 6 for the comparison of the non-recurrent patient overlap averages to recurrent patient overlap averages (post 1st and 2nd ablation). The size of FPs surrounding the left atrium of each patient varied and therefore a normalization of FP size did not take place in this study. These p-values indicate that there is not a statistically significant difference (p-value less than 0.05) between the non-recurrent and recurrent patient (post 1st ablation) sets as well as between both recurrent patient sets. 11 Figure 6. Relates the p-value for FP 1-6 of a student T test found when comparing the overlap percentages of non-recurrent patients after the first ablation to the recurrent patients overlap percentages after the first ablation in the blue bars. The red bars indicate the comparison between recurrent patients post 1st and 2nd ablation. There was variability in the amount of fat surrounding the left atrium for each patient. This variability is independent of all risk factors including age, sex, hypertension, valvular heart disease, diabetes mellitus, and body mass index (BMI) [9]. All the patients in this cohort were ablated using the PVI and PWD techniques for their first and second ablation. The regions of fat characterized as FPs were shown not to play a role in the successful ablation of the recurrent patients. IV. Discussion Ablation of the myocardial tissue adjacent to the FPs of the recurrent patients was not found to be greater after the second ablation procedure when compared to the first. This is made evident when comparing the overlap percentages of the recurrent patients post 1st ablation and post 2nd ablation as shown in figure 5. The first criteria that must be met in order to validate the hypothesis is to observe a lower scar to FP overlap percentage in the recurrent patient set post 1st ablation compared with a higher overlap percentage in the 12 non-recurrent patient set. This would indicate that myocardial tissue adjacent to FPs would have been ablated in non-recurrent patients preventing abnormal electrical signals from entering the left atrium and curing AF in those patients. Meeting this first criterion would also suggest that recurrent patients needed a second ablation because the myocardial tissue adjacent to the FPs containing GP were not ablated. This first criterion seems to be satisfied in figure 4 where the recurrent patient’s FP1 and FP3 have less overlap than the non-recurrent patients FP1 and FP3, but the p-values for FP1 and FP3 in the non-recurrent vs. recurrent (post 1st ablation) are 0.44 and 0.62 respectively. If the first criterion was met, the second criterion would require a higher scar to FP overlap percentage in the recurrent patient set post 2nd ablation. The second criterion is also shown to be flawed in this study because the p-values when comparing the recurrent patient sets are all greater than 0.05, as shown in table I. If the first two criteria were met for any one of the FPs then it could be stated that the hypothesis is correct. There would be evidence to ablate whatever region of fat (i.e. FP1, FP2, etc.) that passed these two criteria in the first ablation procedure thus achieving a higher success rate. This turned out to not be the case in this study. The FP with the lowest p-value during this comparison was FP2 (p=0.12). FP2 was ablated more in the recurrent patient set (post 1st ablation) than in the non-recurrent patient set. If this was reversed and FP2 was ablated more in the non-recurrent patient set then the first criterion would have been met and the second criterion could be evaluated for FP2. The limitations of this study include: 1) Standardization of segmentation of the FPs including: a) FP size normalization, b) voxel intensity threshold determined for what is considered fat and what is considered normal tissue in the DBMRI scans. This will create less noise when comparing the overlap of scar and fat since the fat will be better defined for each patient. 2) As MRI technologies improve over time scan quality will increase resulting in better resolution, less noise, and more accurate data. 3) Improvements in the algorithms used to align the DBMRI scans and the LGEMRI scans. A closer look at GP and how they work will reveal some important concepts to take into consideration. “Intrinsic cardiac neurons have been reported to be located in small ganglia scattered primarily over the posterior surfaces of the atria… [11].” There are a number of different types of neurons that are found within the intrinsic cardiac nervous system. The locations and connectivity of the neurons in the ICANS are areas of focus for possible abnormal signals originating to cause AF. “Intrinsic cardiac ganglia were typically associated with interconnecting nerves that formed ganglionated plexuses. Ganglionated plexuses were consistently identified in five atrial and five ventricular locations [11].” Ganglia in the heart contain anywhere from one neuron to 200 neurons. There was an average of 458 ± 43 ganglia of varying dimensions identified in atrial tissues. The exact anatomical configuration of the GP varied greatly from specimen to 13 specimen, as did the size of the ganglia [11]. “There were extensive regions of atrial or ventricular fat where no ganglia were identified” (Armour, et al. 293). This is evidence that it is not a good assumption to say that all fat surrounding the heart is innervated with GP. Even though FPs themselves do not seem to be correlated with the success of AF, GP could still be a major player in the initiation and propagation of abnormal electrical signals causing AF. It is also good to note that the success rate went up over time in the study done by Dr. Jackman’s group [2] suggesting that there could continue to be spontaneous firing of the PV for a period of time post ablation, but because the GP were eliminated the PV stopped receiving abnormal electrical input from the GP and stopped firing spontaneously. This evidence would promote identification of these four main GP regions in patients who have already been ablated and perform a similar study as this one instead of looking at the FPs alone. There is significant evidence that the GP are the main source for these abnormal electrical signals that are causing AF. This is supported by two sources, the Berenfeld and Sergerson papers. The authors of these papers argue that the source of the abnormal signals is located at high dominant frequency (HDF) locations which can be detected with an electrode placed in the proximity of the left atrium detecting voltage potentials. These HDF locations are localized in the posterior wall of the left atrium. Posterior wall debulking is effective because it is blocking the abnormal signal which is being generated by the GP. But this does not get at the heart of the problem which lies within the GP themselves. By ablating the GP directly there would be less of a chance that the abnormal signal could find another way into the circuitry of the heart and cause AF. This is supported by the fact that the more you ablate the myocardial tissue of the heart the more scar tissue is formed, which results in the blocking of abnormal signals due to the abnormal structure of the scar tissue. At the forefront of research for identifying the locations of electrical signals causing AF is analyzing complex fractionated atrial electrograms (CFAEs) recorded during AF. Areas in the posterior left atrium showing “the highest incidence of wavebreaks and beatto-beat variability in the direction and velocity of propagation are associated with highly fractionated signals…” [10]. These areas where there is dominant, fast frequency in the CFAEs have been termed rotors. These rotors were identified in sheep hearts induced with persistent AF and ablated to achieve better results than previous methods in AF ablation procedures [10]. In summary, this study has suggests that there is not a correlation between ablation of myocardial tissue adjacent to the FPs in AF patients and success of the ablation procedure. When analyzing the recurrent patients’ post 1st and post 2nd ablation there was not a statistically significant difference and therefore could not draw any conclusions. 14 Because of these findings future work can be done studying rotor ablation in conjunction with PWD and PVI techniques to ultimately improve the success rate of the ablation procedure in the treatment of AF. References [1] O. Berenfeld. “Toward discerning the mechanisms of atrial fibrillation from surface electrocardiogram and spectral analysis.” Journal of Electrocardiology. 43.6 (2010) 509–514 [2] S. Po, Hiroshi Nakagawa, and Warren M. Jackman."Localization of left atrial ganglionated plexi in patients with atrial fibrillation." Journal of Cardiovascular Electrophysiology 20.10 (2009): 1186-9. Web. 5 Apr 2010. [3] T. Horikawa-Tamani M.D., Hirao, Kenzo M.D., Furukawa, Toshiyuki M.D., Mistsuaki Isobe, M.D. “Mechanism of the Conversion of a Pulmonary Vein Tachycardia to Atrial Fibrillation in Normal Canine Hearts: Role of Autonomic Nerve Stimulation.” Journal of Cardiovascular Electrophysiology. 18.5 (2007): 534-541. Web. Mar 2011. [4] J. Koebe, Paulus Kirchhof. “Novel non-pharmacological approaches for antiarrhythmic therapy of atrial fibrillation” Europace. 10.4 (2008): 433-437. Web Mar 2011. [5] Z. Lu, Scherlag, Benjamin J., Lin, Jiaxiong, Yu, Lilei, Guo, Ji-Hong, Niu, Guodong, Jackman, Warren M., Lazzara, Ralph, Jiang, Hong, and Sunny S. Po. "Autonomic mechanism for initiation of rapid firing from atria and pulmonary veins: evidence by ablation of ganglionated plexi."European Society of Cardiology 84.2 (2009): 245-52. Web. 5 April 2009. [6] N. Segerson, Daccarett, Marcos, Badger, Troy J., Shabaan, Akram, Ch., M.B.B., Akoum, Nazem, Fish, Eric N., Rao, Swati, Burgon, Nathan S., Adjei-Poku, Yaw, Kholmovski, Eugene, Vijayakumar, Sathya, Dibella, Edward V.R., Macleod, Rob S., and Nassir F. Marrouche. "Magnetic Resonance Imaging-Confirmed Ablative Debulking of the Left Atrial Posterior Wall and Septum for Treatment of Persistent Atrial Fibrillation: Rationale and Initial Experience." Journal of Cardiovascular Electrophysiology. 21.2 (2010): 126-32. Print. [7] D. Peters, James W. Goldfarb, Reza Nezafat, Nisha I. Parikh, Michael L. Chuang, and Warren J. Manning. "Left Atrial Fat with Cardiac MR." Journal of Cardiovascular Magnetic Resonance 11.Suppl 1 (2009): P128. Print. [8] P. Cranefield, Wit, Andrew L., and Brian F. Hoffman. "Genesis of Cardiac Arrhythmias." Circulation. 47.1 (1973): 190-204. Web. 15 March 2011. [9] M. Chekakie, Christine C. Wells, Raymond Metoyer, Ahmed Ibrahim, Adam R. Shapira, Joseph Cytron, Peter Santucci, David J. Wilber, and Joseph G. Akar. “Pericardial Fat Is Independently Associated With Human Atrial Fibrillation.” Journal of the American College of Cardiology. 56.10 (2010): 784-788. Print [10] J. Kalifa, Kazuhiko Tanaka, Alexey V. Zaitsev, Mark Warren, Ravi, Vaidyanathan, David Auerbach, Sandeep Pandit, Karen L. Vikstrom, Robert, Ploutz-Snyder, Arkadzi Talkachou, Felipe Atienza, Gérard Guiraudon, José Jalife and Omer Berenfeld. “Mechanisms of Wave Fractionation at 15 Boundaries of High-Frequency Excitation in the Posterior Left Atrium of the Isolated Sheep Heart During Atrial Fibrillation.” Circulation 113 (2006): 626-633. Web. 22 March 2012 [11] J. Armour, David A. Murphy, Bing-Xiang Yuan, Sara Macdonald, and David A. Hopkins. “Gross and Microscopic Anatomy of the Human Intrinsic Cardiac Nervous System."Anatomical Record. 247.2 (1997): 289-298. Web. 5 April 2010. [12] M. C. Wijffels and C. J. Kirchhof and R. Dorland and M. A. Allessie. "Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats." Circulation 92.7 (1995) :1954-1968. [13] C. J. McGann, MD, Eugene G. Kholmovski, PhD*, Robert S. Oakes, BS, Joshua J.E. Blauer, BS , Marcos Daccarett, MD, Nathan Segerson, MD, Kelly J. Airey, MD, Nazem Akoum, MD, Eric Fish, Troy J. Badger, MD, Edward V.R. DiBella, PhD, Dennis Parker, PhD, Rob S. MacLeod, PhD, and Nassir F. Marrouche, MD. “New Magnetic Resonance Imaging-Based Method for Defining the Extent of Left Atrial Wall Injury After the Ablation of Atrial Fibrillation.” Journal of the American College of Cardiology 52 (2008): 1263-1271. Web. 4 April 2012.