Extent and Spatial Distribution of Left Atrial Arrhythmogenic Sites, Late

Aus dem Universitäts-Herzzentrum Freiburg-Bad Krozingen

Klinik für Kardiologie und Angiologie II

Abteilung für Rhythmologie

Gadolinium Enhancement at MRI and Low Voltage Areas in Patients with

Persistent Atrial Fibrillation

– Comparison of Imaging vs. Electrical Parameters of Fibrosis and Arrhythmogenesis

I N A U G U R A L – D I S S E R T A T I O N

zur Erlangung des Medizinischen Doktorgrades

der Medizinischen Fakultät

der Albert-Ludwigs-Universität Freiburg im Breisgau

Vorgelegt 2018 von Juan Chen geboren in Hubei, V.R. China

Dekan: Professor Dr. Norbert Südkamp

1. Gutachter: Professor Dr. med. Thomas Arentz

2. Gutachter: Professor Dr. med. Peter Kohl

Jahr der Promotion: 2019

Widmung

An dieser Stelle möchte ich mich bei allen Beteiligten herzlich bedanken, ohne deren Beitrag diese Arbeit unmöglich gewesen wäre:

Prof. Dr. Thomas Arentz, Chefarzt der Abteilung Rhythmologie am Universitäts-Herzzentrum Freiburg-Bad Krozingen für die Möglichkeit der Anfertigung dieser Doktorarbeit in seiner Abteilung, der Überlassung des Themas und der Betreuung der Arbeit sowie für das entgegengebrachte Vertrauen.

Dr. Amir Jadidi, für seine Begeisterung und kompetente Betreuung und Unterstützung, für seine reichlichen Ideen und die Einführung in das wissenschaftliche Arbeiten, für die sorgfältigen Korrekturen und vor allem das freundliche und angenehme Arbeitsklima.

Dr. Zoraida Moreno Weidmann, für ihre großzügige Unterstützung meiner Forschungsarbeit sowie die statistische Auswertung und die konstruktiven Vorschläge und die Hilfe durch mein Leben.

Dr. Bjoern Mueller-Edenborn, für die Korrekturen und statistischen Auswertung und freundliche Hilfe.

Meine Familie danke ich für die Motivation, sowohl in der Durchführung als auch in jedem Aspekt des Lebens.

Table of Contents

1. INTRODUCTION ...... 1

1.1 History of Atrial Fibrillation ...... 1

1.2 Definition of Atrial Fibrillation ...... 3

1.3 Incidence and Prevalence of Atrial Fibrillation ...... 4

1.4 The Electrophysiological Mechanism of Atrial Fibrillation – Three Distinct Hypothesis: Multiple Wavelets, Localized Rotational Sources (Rotors) and Focal Triggers ...... 4 1.4.1 Focal or Localized Reentrant Circuits Develop within Pulmonary Veins and Their Antral Areas and Constitute the Arrhythmogenic Sources of Human Paroxysmal Atrial Fibrillation ...... 6 1.4.2 Non-Pulmonary Vein Arrhythmogenic Sources with Rotational or Focal Activity Originate from the Atrial Body and Maintain Persistent Atrial Fibrillation ...... 6 1.4.3 Role of Arrhythmogenic Atrial Substrate for Development of Persistent Atrial Fibrillation ...... 8

1.5 Reversibility of Arrhythmogenic Left atrial Substrate by Treatment of Cardiovascular Risk Factors and Reduction of Left Atrial Hypertension / Stretch ...... 10

1.6 Management of Atrial Fibrillation ...... 12 1.6.1 Antiarrhythmic Drug Therapy ...... 12 1.6.2 Catheter Ablation Strategy ...... 13

1.7 Atrial Substrate Detection Techniques...... 17 1.7.1 Delayed Gadolinium Enhancement Magnetic Resonance Imaging for Detection of Atrial Substrate ...... 17 1.7.2 Electroanatomic Voltage Mapping ...... 18

1.8 Correlation between CFAE, Atrial Delayed Enhanced Areas at MRI and Low Voltage Substrate ...... 20

1.9 Summary ...... 21

2. MATERIALS AND METHODS ...... 22

2.1 Patients ...... 22

2. 2 Magnetic Resonance Imaging ...... 22

2.3 Electrophysiological mapping ...... 24

2.4 Statistical Analyses ...... 29

3. RESULTS ...... 29

3.1 Patient Characteristics ...... 29

3.2 Left Atrial High-density Mapping ...... 29

3.3 LA Size, Extent of Delayed Enhanced Areas, Low Voltage Areas and Arrhythmogenic Sites ...... 31

3.4 Regional Distribution of Delayed Enhanced Areas and Low Voltage Areas ...... 32

3.5 Correlation of Delayed Enhanced Areas and Low Voltage Areas ...... 34

3.6 Relationship of Arrhythmogenic Sites to Delayed Enhanced Areas and Low Voltage Areas ...... 36

3.7 Electrogram Characteristics of Arrhythmogenic Sites ...... 37

4. DISCUSSION ...... 39

4.1 The Main Findings of Our Study ...... 39

4.2 Atrial Voltage Mapping for Identification of Arrhythmogenic Substrate ...... 40

4.3 The Spatial Distribution of Dense/Patchy DE and of Atrial Low Voltage Areas in AF ..... 42 4.3.1 Differences in the Extent of Substrate Detection in DE-MRI and Voltage Mapping .. 42 4.3.2 Potential Causes for Mismatch between MRI and Voltage Mapping for Identification of Fibro-fatty Arrhythmogenic Atrial Substrate ...... 43

4.4 Electrophysiological Criteria to Localize Sources of AF: Electrogram Voltage and Duration at Atrial Areas with Continuous or Rotational Activity ...... 44

4.5 Implication for Catheter Ablation ...... 46

4.6 Limitations ...... 47

4.7 Summary ...... 47

5. ZUSAMMENFASSUNG / ABSTRACT...... 48

6. REFERENCES ...... 50

7. PUBLICATION ...... 59 III

8. ANHANG ...... 60

EIDESSTATTLICHE VERSICHERUNG ...... I

ERKLÄRUNG ZUM EIGENANTEIL ...... II

ACKNOWLEDGEMENTS ...... III

Abbreviations

3D three-dimensional

AAD antiarrhythmic drug

AF atrial fibrillation

CA continuous activity

CFAE complex fractionated atrial electrogram

CL cycle length

CPAP continuous positive airway pressure

CS coronary sinus

CT computer tomography

CV conduction velocity

DE delayed enhancement

EAS electroanatomic system

EAVM electroanatomical voltage mapping

ECG electrocardiogram/electrocardiography

EGM electrogram

FIRM focal impulse and rotor modulation

IIR MRI image intensity ratio

IVSDD interventricular septum diastolic diameter

LA left atrium

LAA left atrial appendage

LAD left atrial diameter

LGE late gadolinium enhancement

LIPV left inferior pulmonary vein

LSPV left superior pulmonary vein

LVA low voltage area

LVEDD left ventricular end-diastolic diameter

LVEF left ventricular ejection fraction

LVESD left ventricular end-systolic diameter

MRI magnetic resonance imaging

PA prolonged activity

PV pulmonary vein

PVI pulmonary vein isolation

RIPV right inferior pulmonary vein

RotA rotational activity

RSPV right superior pulmonary vein

SR sinus rhythm

1. Introduction

1.1 History of Atrial Fibrillation

The earliest description of atrial fibrillation (AF) was allegedly found in the Yellow Emperor’s Classic of Internal Medicine (Huang Ti Nei Ching) 4000 years ago in China (Lip and Beevers 1995). However, from history record, William Harvey (1578 - 1657) who first described the circulatory system appropriately, was first to describe fibrillation of the auricles in animals in 1628 (Fazekas 2007).

In 1783, Jean Baptiste Sénac (1693 - 1770) firstly connected ‘rebellious palpitation’ and mitral stenosis (Fazekas, et al. 2003). William Stokes (1804 - 1878) and Karel Wenckebach (1864 - 1940) described an irregular pulse corresponding to AF (Prystowsky 2008). James Mackenzie (1853 - 1925) demonstrated that a presystolic wave cannot be seen during "pulsus irregularis perpetuus", from an analysis of simultaneously recorded arterial and venous pressure curves (Fazekas 2007).

The great diagnostic breakthrough was the invention of the electrocardiograph showing atrial fibrillation by Willem Einthoven (1860 - 1927) in 1906 (Einthoven 1906). The proof of a direct connection between absolute arrhythmia and auricular fibrillation was established by two Viennese physicians, Carl Julius Rothberger and Heinrich Winterberg in 1909 (Rothberger and Winterberg 1909).

The major discoveries relating to the pathomechanism and the clinical features of AF were in the 20th century:

• George Ralph Mines (1886 - 1914; “Description of wavefronts with reentrant activity and the vulnerable period during which cardiac fibrillation may occur”) (Prystowsky 2008), • Walter Eugene Garrey (1874 - 1951; “Discovered that a critical mass is needed to sustain AF”) (Prystowsky 2008), • Thomas Lewis (1881 - 1945; “ AF is from auricular origine; AF affects the entire auricular surface with varying path of exications”) (Prystowsky 2008),

• Gordon Moe (1915 - 1989; “Development of a computer model of AF”) (Moe and Abildskov 1959), • Maurits Allessie (1945 - Physiologist from Maastrich; “AF begets AF”) (Wijffels, et al. 1995), • Michel Haissaguerre (1955 - Cardiologist from Bordeaux, France; “Discovery of the pulmonary veins as the arrhythmia sources/triggers of paroxysmal human atruial fibrillation and development of pulmonary vein isolation as a novel therapeutic concept”) have made great contributions to build our current knowledge about pathophysiological and therapeutical aspects of human AF (Haissaguerre, et al. 1998).

A current search in the PubMed database on publications about “atrial fibrillation” reveals (Figure 1) an impressive increase of research and publications in this field over the last three decades. The amount of publications within the last eight years has reached nearly one-half of all recorded publications.

Figure 1. PubMed search in September 2018 for publications on the topic of “atrial fibrillation”. 2

1.2 Definition of Atrial Fibrillation

Atrial fibrillation (AF) is an abnormal heart rhythm, characterized by rapid irregular electrical activity within the atria. The atrial activation rate lies between 300 and 500 per minute (120ms – 200ms cycle length); resulting in an irregular ventricular (heart) rate. AF is the most common cardiac supraventricular arrhythmia and becomes more prevalent with age (Go, et al. 2001). AF is associated with thrombo-embolism, ischemic stroke, cognitive impairment, heart failure, decline in quality of life and increased mortality (Benjamin, et al. 1998). The diagnosis of AF is based on the electrocardiogram (ECG) showing (1) absolutely irregular RR intervals (in the absence of complete atria-ventricular block), (2) no distinct P waves on the surface ECG or P- waves with changing morphology in the same ECG lead, and (3) an atrial cycle length (when visible) that is usually variable and less than 200 milliseconds. In 2016, ESC guidelines (Kirchhof, et al. 2016) for management of AF distinguish five types of AF based on the presentation, duration, and form of termination of AF episodes in: first diagnosed, paroxysmal, persistent, long-standing persistent and permanent AF (Table 1).

Table 1. Clinical Classification of AF Patterns According to ESC Guidelines 2016* AF pattern Definition

First diagnosed AF that has not been diagnosed before, irrespective of the duration of the arrhythmia or the presence and severity of AF-related AF symptoms Self-terminating, in most cases within 48 hours. Some AF paroxysms Paroxysmal AF may continue for up to 7 days. AF episodes that are cardioverted within 7 days should be considered paroxysmal AF that lasts longer than 7 days, including episodes that are Persistent AF terminated by cardioversion, either with drugs or by direct current cardioversion, after 7 days or more. Long-standing Continuous AF lasting for ≥ 1years when it is decided to adopt a persistent AF rhythm control strategy AF that is accepted by the patient (and physician). Hence, rhythm Permanent AF control interventions are, by definition, not pursued in patients with permanent AF. Should a rhythm control strategy be adopted, the arrhythmia would be re-classified as long-standing persistent AF.

AF = atrial fibrillation

*From Eur Heart J 2016,37:2893-2962. 3

1.3 Incidence and Prevalence of Atrial Fibrillation

In the Framingham heart study 8725 patients were followed up during 20 years, 936 patients developed AF. The lifetime risk (at age 80 years) for AF development was 22% both in men and women (Lloyd-Jones, et al. 2004).

In Europe, because of the demographic shift in age, the number of AF prevalence in 2010 is estimated at 8.8 million among adults older than 55 years and is expected to double by 2060 if age- and sex-specific prevalence remains stable (Krijthe, et al. 2013). Especially, the number of adults with AF older than 75 years will increase from 5.6 million in 2010 to 13.8 million in 2060. AF is independently associated with a 1.5- to 1.9-fold increased risk of all-cause mortality in men and women, respectively (Andersson, et al. 2013). The increased morbidity of AF is also associated with heart failure, stroke (Stewart, et al. 2002; Wolf, et al. 1991), aortic and mitral valve disease, left atrial enlargement, hypertension, and advanced age, obesity and obstructive sleep apnea (Verrier and Josephson 2009). As a reason of ischemic stroke (Henriksson, et al. 2012), cognitive impairment (Ball, et al. 2013), decreased quality of life (Marzona, et al. 2012) and depressed mood (von Eisenhart Rothe, et al. 2015) in AF patients, 10% to 40% of AF patients are hospitalized each year (Kotecha, et al. 2014). The costs on AF management have already taken between 6.0 - 26.0 billion US dollars (AF-related costs alone and AF-related costs plus other cardiovascular costs and non-cardiovascular costs) in the US for 2008 (Kim, et al. 2011).

1.4 The Electrophysiological Mechanism of Atrial Fibrillation – Three Distinct Hypothesis: Multiple Wavelets, Localized Rotational Sources (Rotors) and Focal Triggers

In the 1960s, Moe et al. reported (by using their computer model of AF) that chaotic fibrillatory activity might be self-perpetuating even in the absence of focal sources/triggers (Moe and Abildskov 1959; Moe, et al. 1964). Hereby, they provided the hypothesis of “multiple meandering and changing wavelets” sustaining AF.

The first experimental evidence for small localized stable rotational sources (so called “rotors”) which sustain AF was provided by Skanes et al. in Langendorf perfused sheep hearts in 1998 (Skanes, et al. 1998). In the same year, Haissaguerre et al. discovered rapid focal triggers

4 originating from the pulmonary vein (PV), responsible for initiation and maintenance of human paroxysmal AF (Haissaguerre, et al. 1998). In 2006, Haissaguerre and colleagues described spatially stable localized rotational or focal sources, which occurred within the left atrial body and maintained persistent AF (Haissaguerre, et al. 2006).

The electrophysiological mechanisms underlying AF are complex: They involve interaction among multiple factors, including- PV and non-PV rotational or focal triggers, which are responsible for AF initiation and maintenance. Moreover, the arrhythmogenic atrial substrate is important for AF maintenance and progression of the arrhythmia from paroxysmal to the persistent AF (Calkins, et al. 2017) (Figure 2).

Figure 2. Schematic drawing showing various hypotheses and proposals concerning the mechanisms of human atrial fibrillation. A. Multiple wavelets hypothesis. B. Rapidly discharging automatic foci. C. single re-entrant circuit with fibrillary conduction. D. Functional re-entry resulting from rotors or spiral waves. (From Calkins H et al. J Arrhythm 2017:369-409.).

1.4.1 Focal or Localized Reentrant Circuits Develop within Pulmonary Veins and Their Antral Areas and Constitute the Arrhythmogenic Sources of Human Paroxysmal Atrial Fibrillation

In a landmark study, Haissaguerre and his team could demonstrate that rapid focal activity originating from the PV initiates and sustains paroxysmal AF in humans (Haissaguerre, et al. 1998). Since then, multiple clinical studies demonstrated the efficacy of pulmonary vein isolation (PVI) in maintaining sinus rhythm in patients with paroxysmal AF (Arentz, et al. 2003). Multiple studies demonstrated that PV and left PV-atrium junction area constitutes the major affected structures, that are involved in triggering and maintenance of human AF (Arentz, et al. 2007).

The complex microscopic anatomy at the PV antra – with sub-endocardial strands crossing in longitudinal and oblique directions into the PV - facilitate re-entrant arrhythmias in this area (Ho, et al. 2012). Compared to atrial myocytes, pulmonary venous cells have a shorter action potential duration, a slower upstroke velocity and a more depolarized resting membrane potential (Mahida, et al. 2015). These electrophysiological characteristics predispose the PV to focal and reentrant arrhythmia sources (Ehrlich, et al. 2003).

Following the discovery of the PV as the main arrhythmogenic sources of paroxysmal AF (Haissaguerre, et al. 1998), Arentz et al. performed a high-density mapping study of the PV and their ostial regions during ongoing AF using a 64-electrode basket-catheter. The study revealed both rotational activation patterns (“rotors”) and focal sources, originating within the PV and their surroundings (Arentz, et al. 2007; Arentz, et al. 2003; Arentz, et al. 2003).

1.4.2 Non-Pulmonary Vein Arrhythmogenic Sources with Rotational or Focal Activity Originate from the Atrial Body and Maintain Persistent Atrial Fibrillation

Further experimental animal (Hocini, et al. 2002; Ryu, et al. 2006) studies conducted in 2002 - 2006, suggested focal and reentrant arrhythmia sources within the LA may be responsible for development of AF. The first clinical description of focal and reentrant rotational sources within the left atrial body - without anatomical relationship to the PV – came from Haissaguerre and colleagues in 2006: After isolation of the PV, the authors demonstrated additional focal and 6 rotational sources from the left atrial body that maintained persistent and long-standing persistent human AF (Haissaguerre, et al. 2006). Focal radiofrequency application at these areas acutely converted AF into organized atrial tachycardia or directly into sinus rhythm (Haissaguerre, et al. 2006).

Recently, Haissaguerre (Haissaguerre, et al. 2014) and Lim (Lim, et al. 2017) using a 252- electrode vest for noninvasive body surface mapping found increasing number of arrhythmogenic areas with reentrant or focal sources in their study cohort of 105 patients with persistent or long-standing persistent AF. Importantly, 85% of patients presented reentrant drivers at the left PV antra and further 87% at right PV antra. Focal drivers mostly located in left PV and left atrial appendage area (LAA). The PV constituted the predominant arrhythmogenic regions. Total number and distribution regions of reentrant rotations and focal discharges were positive related to duration of AF.

The number of areas with arrhythmogenic sources increased from four to six and seven for each subgroup of patients with persistent AF presenting with sinus rhythm vs. persistent AF since <12 months and long-standing persistent AF since ≥12 months. Importantly, 31% vs. 55% vs. 75% of each AF severity group had non-PV sources from the right- or left atrium. Ablation targeting arrhythmogenic sources resulted in acute AF termination in 70% of patients in persistent AF (Lim, et al. 2017).

In summary, Lim et al. demonstrated rapid discharges from the PV and their antral regions are the most common triggers of paroxysmal AF. Even in patients with persistent AF, the PV plays a pivotal role. However 50% vs. 75% of patients with persistent vs. long-standing persistent AF present additional non-PV atrial arrhythmogenic sources (Sanders, et al. 2006). This finding explains why PVI-only approach is effective in up to 70% vs. 50% of patients with paroxysmal vs. persistent AF (Kuck, et al. 2016; Verma, et al. 2015).

Another recent clinical mapping approach – the so called: `FIRM (for focal impulse and rotor modulation)´ - was introduced by Narayan et al. in 2012 (Narayan, et al. 2012; Narayan, et al. 2012). The authors brought the evidence for spatially stable localized reentry circuits (measuring <2 cm in diameter), which were responsible for maintenance of persistent human AF (Narayan, et al. 2012; Narayan, et al. 2012). Focal ablation at these areas was associated

7 with conversion of AF into sinus rhythm (Narayan, et al. 2012; Narayan, et al. 2012). Narayan and colleagues used simultaneous whole atrial mapping using a 64-electrode basket catheter in combination with automatic ECG signal analysis using phase-mapping to search for rotational activities during ongoing AF (Narayan, et al. 2012). Although, the initially reported results were very encouraging with high AF termination rates and low (18%) arrhythmia recurrences after two years (Narayan, et al. 2012; Narayan, et al. 2012), recent multi-center trials using the same technology could not reproduce the high efficacy rates (Buch, et al. 2016).

1.4.3 Role of Arrhythmogenic Atrial Substrate for Development of Persistent Atrial Fibrillation In paroxysmal AF, a PVI-only ablation strategy has a 70% success rate (defined as freedom of arrhythmias one year after PVI) (Calkins, et al. 2017). In contrast, in patients with persistent AF, rate of arrhythmia freedom one year after PVI-only is about 50% only (Calkins, et al. 2017; Verma, et al. 2015). In a histological study, Platonov, Ho and colleagues analyzed the microscopic structure of atrial myocardial tissue in patients without history of AF vs. patients with paroxysmal and long-standing persistent AF (Figure 3). Presence of AF and increasing AF duration were associated with increasing interstitial collage deposition and fibro-fatty replacement of atrial cardiomyocytes (Platonov, et al. 2011).

Moreover, recent clinical studies revealed that arrhythmia recurrences occur especially in patients with arrhythmogenic LA substrate. Presence or absence of the arrhythmogenic left atrial substrate can be assessed either by voltage mapping during the PVI procedure or by use of gadolinium enhanced atrial MRI (Marrouche, et al. 2014; Oakes, et al. 2009). Further studies demonstrated that targeted catheter ablation of atrial low voltage areas – either identified in sinus rhythm (Rolf, et al. 2014; Yang, et al. 2016) or during AF (Jadidi, et al. 2016; Yagishita, et al. 2016) – was associated with improved arrhythmia freedom rates, when compared to a PVI-only strategy in patients with persistent AF.

Figure 3. Light microscopy of crista terminalis specimens shows extent of fibrosis in atrial wall. Important fibro-fatty substrate (Fibrosis extent 51%, fat 15%) is found in patient with long-persistent AF (A), which is less abundant (Fibrosis extent 14%, fat 24%) in paroxysmal AF (B) and healthy atrial myocardium (C) with little interstitial fibrotic tissue (Fibrosis extent 5%, fat 1% as blue and white areas) between the cardiomyocytes (red cells) (A). (Masson’s trichrome stain; original magnification ×200). (From Platonov P et al. JACC 2011,58:2225-2232).

In addition, Jadidi, Arentz et al. revealed that arrhythmogenic AF sources co-localize to low voltage areas <0.5mV in AF (using high-density voltage mapping with 20-polar circumferential mapping catheter (Lasso and AFocus II)). Within these low voltage areas, sites with repetitive rotational activity or continuous activity were targeted by radiofrequency energy, leading to acute AF termination in 73% of patients and an improved arrhythmia freedom rate one year after the procedure (success rate 69% with PVI and selective low voltage ablation vs. 47% in PVI-only group (Jadidi, et al. 2016). Another study by Cochet, Haissaguerre et al. demonstrated co-localization of arrhythmogenic AF sources (as identified by the 252-electrode bodysurface ECG-Imaging system CardioInsight) to DE areas at atrial MRI (Cochet, et al. 2018; Haissaguerre, et al. 2016).

From various experimental studies and clinical data, it can be assumed that the histological correlate of myocardial replacement by fibro-fatty tissue corresponds to atrial areas with slow conduction, anisotropy and low voltage substrate, predisposing to focal and reentrant arrhythmia sources of human AF. The arrhythmogenic substrate can be identified using voltage mapping during the electrophysiological study or DE areas at high-resolution atrial MRI. 9

1.5 Reversibility of Arrhythmogenic Left atrial Substrate by Treatment of Cardiovascular Risk Factors and Reduction of Left Atrial Hypertension / Stretch

The arrhythmogenic atrial AF substrate contains two aspects: electrical remodeling and structural remodeling of the atria (Nattel, et al. 2008; Nattel and Harada 2014). A number of studies demonstrated that the cardiovascular risk factors age, female gender, heart failure, hypertension, diabetes mellitus, obesity, sleep apnea are associated with development of arrhythmogenic atrial low voltage substrate in AF patients (`arrhythmogenic remodeling´) (Huo, et al. 2018; Kosiuk, et al. 2015; Prabhu, et al. 2018).

On the other hand, reverse electrical and structural remodeling may occur if sinus rhythm can be restored and maintained (Chalfoun, et al. 2007; Rivard, et al. 2012). These results are also supported by experimental observations from Wijffels and Allessie et al. who reported that “AF begets AF”(Wijffels, et al. 1995), meaning the longer AF persists, the more the electrical and structural atrial remodeling favoring AF occurs.

Moreover, in a recent clinical study Lavie and Sanders et al. demonstrated reverse cardiac remodeling with reduction of AF burden (episodes and duration) in obese patients who achieved an increase of 2 METS in their cardiovascular fitness and weight loss by participation to a tailored exercise program (Lavie, et al. 2017; Pathak, et al. 2015) (Figure 4, 5).

In another study, Müller et al. (Muller, et al. 2016) showed reverse atrial remodeling in patients with obstructive sleep apnea syndrome in whom atrial hypertension was reduced after 30 days use of continuous positive airway pressure (CPAP)-therapy: The electro-mechanic atrial contractility duration reduced from 131.4 ± 16 ms at baseline to 124.6 ± 16 ms after CPAP-therapy. In addition, the authors reported concomitant reduction of brain natriuretic peptide in the CPAP-treated patients (Muller, et al. 2016).

Notably, reverse ventricular structural remodeling has been reported recently after restoration of sinus rhythm by catheter ablation in heart failure patients with persistent AF (Prabhu, et al. 2018).

The recent multicenter Catheter Ablation vs. Standard Conventional Therapy in Patients with Left Ventricular Dysfunction and Atrial Fibrillation (CASTLE-AF) study (Marrouche, et al. 2018) showed catheter ablation for AF in patients with heart failure correlated with a significantly lower rate of a composite endpoint of death from any cause and hospitalization for heart failure than medical therapy. In addition, catheter ablation reduced the burden of AF and improved the left ventricle ejection faction (LVEF).

Figure 4. Explaining the increased risk of atrial fibrillation in obesity.

CHD = coronary heart disease; DM = diabetes mellitus; EAT = epicardial adipose tissue; HF = heart failure; HTN = hypertension; LVH = left ventricular hypertrophy; OSA = obstructive sleep apnea. (From Lavie C et al. J Am Coll Cardiol. 2017,70:2022-35.)

However, persistence of cardiovascular comorbidities responsible for LA hypertension and LA fibrotic substrate development may continue to cause atrial structural remodeling and may lead to future arrhythmia recurrences, even after an initially successful AF ablation procedure. Therefore, the combined therapeutic approach with reduction/elimination of cardiovascular risk 11 factors responsible for LA hypertension followed by catheter ablation of AF may be the ideal way to reduce atrial substrate development and AF recurrences.

Figure 5. Illustration of relationship of obesity, weight loss and exercise in atrial fibrillation. CV = cardiovascular (From Lavie C et al. J Am Coll Cardiol. 2017,70:2022-35.)

1.6 Management of Atrial Fibrillation

1.6.1 Antiarrhythmic Drug Therapy The management of AF patients includes two strategies: rate control (in which the arrhythmia (AF) is accepted and only an optimal heart rate target is preferred) and rhythm control (which aims to restore and maintain normal sinus rhythm) (Kirchhof, et al. 2016). Rate control is considered as first-line treatment by physicians for patients with acute AF (Kirchhof, et al. 2016). Beta-blockers, calcium channel blockers (Verapamil or diltiazem), cardiac glycosides (Digoxin) and specific indications (amiodarone) are common and effective anti-arrhythmia drugs (AADs) for rate control therapy. Amiodarone is a useful drug for rate control. These drugs can reduce AF-related symptoms like palpitations, dizziness und shortness of breath.

Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) (Wyse, et al. 2002) and Rate Control Efficacy in Permanent Atrial Fibrillation (RACE) (Groenveld, et al. 2009) trials demonstrated that there was no significant difference in mortality when rhythm control strategy was compared to heart rate control strategy (23% vs. 21%, p = 0.008). However, the rate of re-hospitalizations and adverse drug effects were higher in rhythm control group (Wyse, et al. 2002). It is assumed that the pro-arrhythmogenic effects of antiarrhythmic drug therapy (with increased risk for Torsade-de-pointes, polymorphic VT and VF) may be one the causes of the lacking benefit of a rhythm control strategy in the AFFIRM trial (Corley, et al. 2004).

Recent clinical studies, using catheter ablation for maintenance of sinus rhythm, have revealed a reduction of mortality in persistent AF patients with successful sinus rhythm maintenance after the procedure, compared to the control group (rate control strategy) (Di Biase, et al. 2016; Marrouche, et al. 2018).

Some animal studies show that (Bhatia, et al. 2018) upstream therapy like angiotensin- converting enzyme inhibitors, angiotensin receptor blockers, statins, or omega-3 polyunsaturated fatty acids, modify the atrial substrate or target specific mechanisms to prevent the occurrence or recurrence of AF (Liu, et al. 2016; Savelieva, et al. 2011). However, translation of these findings to human demands more clinical prospective randomized trials to prove the role in the prevention of atrial substrate remodeling and fibrosis in human.

1.6.2 Catheter Ablation Strategy

Rhythm control can be achieved by means of pharmacological treatment or catheter ablation. According to the current guidelines, in symptomatic patients who are refractory or intolerance to antiarrhythmic drug therapy or experience a decline on left ventricular ejection fraction (ie. arrhythmia-related tachymyopathy), catheter ablation is recommended over medical treatment due to a more effectiveness in restoring and maintaining sinus rhythm (Figure 6).

Figure 6. Schematic of Common Lesion Sets Employed in AF Ablation. 6A. Proximal circumferential pulmonary veins isolation (PVI). 6B shows most common sites of linear ablation: LA roof line; lateral mitral isthmus line (between the lateral mitral valve and the left PVs; anterior mitral line between the roof line or the left or right PVs to the mitral annulus anteriorly. The RA cavo-tricuspid isthmus lien is shown. 6C shows additional ablation within the carina of the PVs. Circumferential SVC isolation is also shown. 6D shows some of the most common sites of ablation lesions when complex fractionated electrograms (CFAE) are targeted. Notably, during the “stepwise ablation approach” the combination of PVI (A), linear ablation sets (B) and ablation of CFAE (D) is applied during the same procedure to treat persistent AF. (From Calkins H et al. Heart Rhythm 2012,9:632-696.)

1.6.2.1 Pulmonary Vein Isolation (PVI)

Haissagurre et al. (Haissaguerre, et al. 1998) found that ectopic beats from the pulmonary veins may trigger AF, which could be suppressed by radiofrequency ablation. PVI can be accomplished by either ostial ablation or proximal antral PV isolation (1 to 2 cm away from the PV ostia. Arentz et al. demonstrated a higher rate of arrhythmia freedom with proximal antral PVI vs. segmental PVI (Arentz, et al. 2007). One-year success rate of PVI in paroxysmal vs. persistent AF lies about 70% vs. 50%. Most patients with paroxysmal AF experiencing arrhythmia recurrences after a first PVI procedure have reconnection of previously isolated PV rather than a new arrhythmia substrate outside of the PV or from non-targeted PV (Baber, et al. 2011; Steinberg, et al. 2015). 14

Electrical gaps in the lesions maybe one of the reasons of recurrence of arrhythmia. In addition, some non-PV arrhythmogenic foci (Arentz, et al. 2003), and extend/localization of non-uniform anisotropic low voltage atrial substrate determine future recurrence of AF (Kottkamp, et al. 2015; Po, et al. 2006; Scherlag, et al. 2005). Arrhythmogenic low voltage substrate plays an important role especially in patients with persistent AF (Jadidi, et al. 2016).

1.6.2.2 Stepwise Ablation Therapy for Persistent Atrial Fibrillation: Combined PVI, Ablation of Fractionated Electrograms and Linear Ablation

Previous studies suggested that a stepwise ablation approach including PVI, additional ablation of fractionated EGMs and linear lesions (during the same procedure) might increase freedom of AF recurrences in persistent AF (O'Neill, et al. 2009; Verma 2011).

However, in the last years, data from multiple experienced centers on stepwise ablation approach revealed necessity for multiple long re-do procedures following the index extensive ablation procedure, because of multiple recurrent atrial flutter and atrial tachycardia, some of which arising from the extensive atrial ablation lesions (Scherr, et al. 2015).

Scherr et al. reported after a long radiofrequency application time of 264 ± 74 minutes a single- procedural success rate of 35%, 28% and 17% after 1, 2 and 5 years in patients with persistent AF. They reported a multi-procedural success rate (after 2 ± 1 procedures) was 90%, 80% and 63% after 1, 2 and 5 years of follow up after the last procedure (Scherr, et al. 2015).

1.6.2.3 Ablation of Complex Fractionated Atrial Electrograms

In 2004, Nademanee et al. were the first to describe `complex fractionated atrial electrograms (CFAE)´ recorded during ongoing AF as the electrophysiological substrate and ablation target of human AF (Nademanee, et al. 2004). In a seminal work, he tagged atrial EGMs with prolonged fractionated activation complexes using the electro-anatomical mapping system CARTO. He defined fractionated activation as EGMs with two or more deflections and/or perturbation of baseline with continues deflection and short cycle length (≤120 ms) during a minimal recording interval of 10 seconds (Nademanee, et al. 2004). They found CFAEs mostly

15 located in atrial septum. Ablation of those areas with CFAEs resulted in conversion of AF to sinus rhythm in most (95%) patients, in order to remove the substrate of AF, instead of removing the triggering foci (Nademanee, et al. 2004). Following CFAE-ablation, Nademanee et al. reported arrhythmia freedom in 91% of patients with paroxysmal and long-standing persistent AF after one-year follow up.

In order to establish CFAE mapping for ablation therapy of persistent AF, automatic algorithms were developed, integrated and used in the clinical AF mapping systems from 2004 to 2015. The “CFEmean algorithm” was integrated to the NavX electro-anatomic system (EAS), which is one of the most spread EAS mapping systems in the world and frequently used in the clinical setting of ablative treatment of AF patients. The Randomized Ablation Strategies for the Treatment of Persistent Atrial Fibrillation (RASTA) study was the first single-center study targeting CFAE areas (as identified by the CFEmean algorithm) in addition to PVI. The study revealed higher arrhythmia recurrences in patients with PVI and `CFAE´ (CFEmean <120 ms) ablation after one year, when compared to standard ablation (PVI) in patients with persistent AF (Dixit, et al. 2012).

It is important to note that complex fractionated atrial electrograms during AF consist of very different types of electrical activities: CFAE may display (1) high or (2) low voltage EGMs amplitudes (> or <0.5 mV). Moreover, they may be of short or prolonged duration (< or > 70% of local AF cycle length). Lastly, CFAE may occur intermittently at a given atrial site or be continuously present at that site (Jadidi, et al. 2016).

Targeting low voltage CFAEs, Seitz et al. reported high arrhythmia freedom rates one year after catheter ablation in patients with persistent AF (Seitz, et al. 2016; Seitz, et al. 2017).

Again, using the automatic algorithm within the NavX electro-anatomic system, Verma et al. demonstrated in a large prospective randomized multi-center study, the lack of clinical benefit of additional CFEmean-guided CFAE ablation, when compared to PVI-only strategy in patients with persistent AF (Verma, et al. 2015). Moreover, the authors evaluated the effect of PVI plus linear ablation to PVI-only or PVI plus CFAE ablation. Arrhythmia recurrence did not differ among three approaches. During 18 months of follow up, arrhythmia freedom rate was 50% in

16 patients with PVI-only, compared with 60% of patients with additional ablation therapy (Verma, et al. 2015).

1.7 Atrial Substrate Detection Techniques

1.7.1 Delayed Gadolinium Enhancement Magnetic Resonance Imaging for Detection of Atrial Substrate

Several clinical studies (Mahnkopf, et al. 2010; McGann, et al. 2014; Oakes, et al. 2009) have been able to demonstrate presence of atrial DE areas using gadolinium-enhanced magnetic resonance imaging in patients with AF. In generally, it is assumed that atrial DE areas represent regions with increased atrial fibrosis (increase in the interstitial tissue and collagen fibers), which is the reason for a delayed clearance of the magnetic contrast dye gadolinium, compare to its clearance from healthy myocardium.

The Utah group was the first to use a free breathing MRI sequence taking 20 - 30 minutes acquisition time, in order to image the atria at high resolution with a voxel size of 1.25 × 1.25 × 2.5 mm (interpolated to 0.625 × 0.625 × 1.25 mm). The LA wall volume is detected and manually segmented as the number of pixels within epicardial and endocardial atrial borders. The delayed enhanced areas are identified by thresholding method to detect pixels with higher image intensity than the mean LA wall intensity (Oakes, et al. 2009). Oakes and Marrouche et al. demonstrated in a single-center study and later on in the Delayed-Enhancement MRI (DE- MRI) Determinant of Successful Radiofrequency Catheter Ablation of Atrial Fibrillation (DECAAF) multi-center study (Marrouche, et al. 2014) that patients with extensive DE areas within the LA had high arrhythmia recurrence rates one year after PVI-only ablation approach. Patients with little DE extent had good outcomes with low arrhythmia recurrence rates after PVI.

However, respiration and heart movements and the movement of the adjacent organs (aorta, pulmonary trunk) may affect the accuracy of detection and quantification of atrial MRI signal intensity and lead to potentially false categorization of adjacent structures or noise signals as `atrial fibrosis´. Moreover, the current MRI resolution is limited to a voxel size of 1.25 × 1.25

× 2.5 mm, which is not enough to accurately detect atrial remodeling in detail. MRI in LA wall imaging should be improved with enhanced spatial resolution and signal-to-noise ratio.

In an anatomical study on fixed human cardiac tissue, Yen-Ho et al. (Ho, et al. 2012) found the thickness of left atrium walls is not homogeneous. The thickness of left atrial muscle measured from 1.5 to 6.5 mm: the lateral LA wall ranges 2.5 - 4.9 mm, anterior LA wall ranges 1.5 - 4.8 mm, and the area between superior pulmonary vein to be the thinnest area measured 2.3 ± 0.69 mm. The thickness of wall in the middle and between the inferior venous orifices was thinner in patients with AF than without AF (Platonov, et al. 2008).

Clinical studies revealed even thinner left atrial wall thicknesses (ranging from 1.2 mm to 2.4 mm) (Takahashi, et al. 2015; Whitaker, et al. 2016), when using cardiac CT in beating hearts and living humans. The CT measurements are in disagreement with study results obtained from fixed anatomical preparations with artificially thicker walls due to fixation process, protein denaturation and tissue shrinkage.

Therefore, further advanced high resolution in MRI based fibrosis imaging would potentially enable higher accuracy and evaluation of the transmurality of atrial fibrosis/remodeling. The reproducibility of MRI detected atrial DE areas is still under investigation in other centers.

1.7.2 Electroanatomic Voltage Mapping

Atrial structural remodeling is associated with myocardial cell-to-cell uncoupling and reduction of the regional conduction velocity. Therefore, less than normal myocardial cells are electrically depolarized, leading to a reduction of voltage amplitudes recorded at the atrial endocardium during the electrophysiological study. Using 3D elelctro-anatomical voltage mapping (EAVM) systems as CARTO (Biosense-Webster, US) or NavX (Abbott, US) the 3D atrial geometry and the electrical voltage amplitudes are simultaneously recorded during the electrophysiological study.

Atrial voltage mapping can be conducted during either sinus rhythm (SR) (Teh, et al. 2012) or AF. Completely fibrotic and electrically silent atrial tissue (electrical scar) displays bipolar voltage amplitudes <0.05 mV and low voltage areas are defined as bipolar voltages <0.5 mV

18 during SR. Verma et al. (Verma, et al. 2005) first highlight the correlation between left atrial scar area (areas displaying <0.05 mV) and outcome in patients with PVI only. The extent and localization of the fibrotic atrial substrate is variable from patient to patient. Compared with age, duration of AF, ejection fraction, LA size and heart disease, low voltage area is a more powerful predictor of arrhythmia recurrences after catheter ablation for AF (Verma, et al. 2005).

A recent study (Rolf, et al. 2014) analyzed 178 patients with paroxysmal or persistent AF. The proportion of patients with low voltage area was higher in persistent AF than paroxysmal AF. The authors performed radiofrequency catheter ablation of atrial low voltage sites in all patients with low voltage substrate <0.5 mV in SR with the aim to homogenize the diseased LA tissue. Linear lesions were applied when extensive regional ablation might create critical isthmus for potential macro-reentrant tachycardia. After one year follow-up, the success rates in patients with PVI plus additional low voltage ablation was higher (70%) than in patients with PVI-only (27%). Yang et al. (Yang, et al. 2016) showed that PVI plus ablation targeting atrial regions with low voltage and abnormal EGMs during sinus rhythm improved AF ablation outcomes, compared to a stepwise AF ablation approach (PVI + CFAE + lines).

Moreover, we recently demonstrated that PVI plus selective ablation of rotational or continuous activities within low voltage areas (defined as areas with <0.5 mV in AF) was more effective than PVI only for persistent AF (Jadidi, et al. 2016). Seventy-three percent of patients with low voltage area exhibited prolonged electrical activity (>70% of AF cycle length) on a single electrode or multiple electrodes. Importantly, 80% and 20% of AF termination sites co-localized to low voltage areas and 20% to low voltage borders, emphasizing a mechanistic proarrhythmic role of low voltage substrate in maintenance of AF. Moreover, Yagishita et al. (Yagishita, et al. 2017; Yagishita, et al. 2016) showed that there is not significant difference in the long-term outcome between the patients undergoing low voltage-guided substrate modification in addition to PVI and the patients without atrial low voltage substrate undergoing PVI only.

Voltage mapping for identification of arrhythmogenic atrial substrate necessitates consideration of a number of physical and technical aspects, in order to be reproducible and specific: Electrogram voltage amplitudes crucially depend on electrode size, orientation of the bipole with regard to the wavefront and therefore depend on both the equipment and the underlying

19 rhythm (sinus rhythm, atrial fibrillation, pacing rhythm). Besides transmurality of atrial fibrosis, voltage amplitudes depend on catheter stability, electrode-tissue contact and atrial tissue thickness. Currently no clear voltage thresholds have been defined for specific identification of atrial arrhythmogenic substrate during different rhythms (AF vs. SR). Recent studies suggest comparable extent of low voltage areas in the same patients between AF and SR mapping, when the low voltage threshold is set to <0.5 mV in AF and to 1.0 - 1.5 mV during SR (Yagishita, et al. 2016).

1.8 Correlation between CFAE, Atrial Delayed Enhanced Areas at MRI and Low Voltage Substrate

Oakes et al. (Oakes, et al. 2009) showed delayed enhancement-MRI correlated with regions of low voltage during sinus rhythm in a small series of patients. Our previous study (Jadidi, et al. 2013) found that a majority (48%) of CFAE sites (as identified by the NavX electro-anatomic mapping system using the threshold of “CFEmean <80 ms”) are not related to DE areas, 41% of CFAE at regions with patchy DE. Only 19% of CFAE present within or around dense DE regions.

Currently, it is not clear which substrate should be targeted by ablation: DE areas or low voltage areas or both. A current multi-center studies evaluates the impact of PVI plus ablation of DE areas at MRI for persistent AF (DECAAF 2 trial). Another multi-center study assessed the benefit of low voltage-guided substrate ablation beyond PVI in persistent AF patients (Selective Ablation of Low Voltage Areas for Persistent AF (SOLVE-AF)).

Moreover, our recent study showed that the arrhythmogenic sites are related to low voltage areas with slow conduction properties. PVI and additional selective ablation of prolonged electrical activities (covering >70% of the local AF cycle length) which localize to atrial low voltage areas <0.5 mV, is associated with high AF termination rate (70%) and improved arrhythmia freedom rates, compared to PVI-only strategy (Jadidi, et al. 2016).

The current work investigates the extent and spatial correlation between (1) DE areas at high- resolution MRI, (2) low voltage areas during AF and localization of typical arrhythmogenic AF sources with repetitive rotational activities or continuous activities in AF.

1.9 Summary

In addition to the PV, non-PV focal and rotational sources within the atrial body play an important role in the perpetuation of persistent AF (Haissaguerre, et al. 2014; Jadidi, et al. 2016; Nademanee, et al. 2004; Narayan, et al. 2012; Yagishita, et al. 2016; Yang, et al. 2016). 1-year outcome after PVI for persistent AF is unsatisfactory with arrhythmia recurrences in about 50% of patients, even after multiple PVI procedures (Tilz, et al. 2012; Verma, et al. 2015). Recent studies suggest a spatial relationship of atrial arrhythmogenic sources to atrial low voltage areas, which correspond to fibro-fatty structural substrate (Cochet, et al. 2018; Haissaguerre, et al. 2016; Jadidi, et al. 2016). The arrhythmogenic substrate can be identified by atrial mapping as low voltage areas or by detection of Gadolinium enhanced atrial areas on high-resolution MRI (Oakes, et al. 2009).

Recent studies have provided initial evidence of the superiority of a combined PVI and additional ablation of low voltage areas to conventional PVI in patients with persistent AF (Jadidi, et al. 2016; Rolf, et al. 2014; Yagishita, et al. 2017; Yang, et al. 2016). Moreover, atrial DE-MRI has been used by several groups as a non-invasive tool to estimate the degree of atrial substrate. The DECAAF-trial revealed that increasing amounts of left atrial DE-MRI identified patients at risk for arrhythmia recurrence after a PVI-only approach (Marrouche, et al. 2014). Large prospective multi-center studies are currently evaluating the effect of PVI plus additional ablation of DE areas (DECAAF II trial) or PVI and additional ablation of low voltage areas (SOLVE-AF trial) to a PVI-only approach for persistent AF. However, correlations and differences in the spatial distribution between the atrial substrate – as identified by DE-MRI vs. voltage mapping – vs. arrhythmogenic atrial areas (with rotational or continuous activity) is not known.

Therefore, in the current study, we sought to correlate the distribution of LA ablation targets demonstrating repetitive rotational activity or continuous activity (prolonged electrical activity >70% of local AF cycle length) to (1) the delayed enhanced areas on high-resolution MRI and (2) low voltage areas <0.5 mV during AF on 3D-electroanatomical maps in patients with persistent AF. In addition, the distribution of low voltage areas is compared to atrial DE areas. We further characterized the electrogram characteristics of these target sites with regard to bipolar voltage and electrogram duration.

2. Materials and Methods

2.1 Patients

Sixteen consecutive patients referred for first ablation of persistent AF were enrolled after informed consent. Inclusion criteria were the presence of persistent (lasting >7 days) or long- standing persistent (lasting ≥12 months) AF prior to ablation. Exclusion criteria were contraindications to MRI and the presence of atrial thrombi on pre-procedural transesophageal echocardiography or previous atrial ablation or surgery. Antiarrhythmic medications were discontinued for 5 half-lives before the ablation procedure. All patients underwent cardiac gadolinium enhanced MRI 1 to 2 days before the procedure. High-density LA mapping was performed during AF before ablation.

2. 2 Magnetic Resonance Imaging

Image Acquisition. MRI studies were conducted as previously reported (Jadidi, et al. 2013) on a 1.5-T clinical scanner (Avanto, Siemens Medical Solutions, Erlangen, Germany) equipped with a 32-channel cardiac coil. DE-MRI was performed 15 min after the administration of 0.2 mmol/kg gadoterate dimeglumine (Dotarem, Guerbet, France). Imaging was acquired with the use of 3-D, inversion recovery-prepared, respiration-navigated, ECG-gated, gradient-echo pulse sequence with fat-saturation. ECG gating was set to 50% of the mean RR interval to acquire signal in mid-diastole. Acquisition parameters were as follows: voxel size, 1.25 × 1.25 × 2.5 mm (reconstructed to 0.625 × 0.625 × 1.25 mm with in-plane interpolation); flip angle, 22°; repetition time/echo time, 5.4/2.3 ms; inversion time, 260 to 320 ms (depending on the results of a previously acquired TI scouting sequence); parallel imaging with GRAPPA technique R = 2; and number of reference lines, 44. Scan time ranged 5 to 10 min depending on the patient’s heart and respiratory rates.

Image Post-processing. Image post-processing was performed by two persons independently. One experienced cardiac radiologist (H.C., University Hospital of Bordeaux) performed the LA wall segmentation and the quantitative analysis of voxel intensities from LA wall, in order to identify the DE areas. MRI-image processing was performed before the electrophysiology (EP) 22 procedure. Segmentation was performed with OsiriX 3.6.1 (OsiriX Foundation, Geneva, Switzerland). For modeling Cardio Viz3D (INRIA, Sophia Antipolis, France) was used by H.C.. LA was segmented manually by contouring the endocardial and epicardial borders of the atrium, including the PV ostia and the first 2 cm of each PV. Areas of DE were detected and segmented by performing a slice-by-slice histogram analysis with a semiautomatic signal threshold, as described previously (Jadidi, et al. 2013; Oakes, et al. 2009) (Figure 7).

From the segmented images, the LA blood pool volume was quantified (corresponding to LA endocardial volume). Two series of binary images were produced for modeling purposes: the first one corresponding to the areas of DE and the second one corresponding to the blood pool volume, including PVs and coronary sinus to provide landmarks for subsequent registration with Ensite NavX (Abbott) LA geometry. Three-dimensional meshes of LA blood pool volume and DE volumes were computed using CardioViz3D (INRIA, Sophia Antipolis, France) and exported as XML for registration with mapping geometry of the LA in the NavX system (Abbott, St. Paul, Minnesota). S.K., an experienced specialist for the cardiac mapping system Ensite, (Abbott), performed the 3D registration of the imported 3D-reconstructed meshes of the LA and DE areas with the mapping geometry of the LA, as described previously (Jadidi, et al. 2013). The LA geometry was carefully merged/registered with the LA blood pool volume derived from MRI. This process was achieved by performing a point-by-point registration of anatomic landmarks (PV ostia, LAA ostium). (Haissaguerre, et al. 2006; Jadidi, et al. 2016).

For further analysis of the spatial relationship of low voltage areas to DE areas, we differentiated DE areas into: 1) dense/continuous DE regions; and 2) patchy/intermittent DE regions if the DE was non-continuous and non-enhanced tissue was found in between the DE sites. We considered atrial tissue as dense/continuous enhanced if > 90% of that regional atrial surface area contained DE. Patchy/intermittent DE was defined as inhomogeneous infiltration of atrial tissue by DE (DE content 20% to 70% of regional surface area). If the atrial tissue did not contain any DE on MRI, it was qualified as non-DE areas (Figure 7). Dense DE areas and patchy DE areas were annotated on the LA geometry in the Ensite system, with the operator being blinded to mapping data regarding localization of LA low voltage areas.

Figure 7. Segmentation of atrial delayed enhancement (DE). On the trans-axial late gadolinium- enhanced imaging volume (A), the atrial wall was manually defined (blue surface in B) from endocardial and epicardial segmentations. The histogram of pixels contained within the LA wall was analyzed (C), as described previously (Jadidi, et al. 2013; Oakes, et al. 2009). This resulted in a segmentation of DE within the LA wall (yellow areas in D), and in a 3-dimensional representation of DE distribution over the LA surface that was subsequently imported in the NavX electro-anatomical mapping system (E).

2.3 Electrophysiological mapping

Mapping and analysis of surface areas. LA high-density mapping (>1000 points per chamber) were acquired in AF using a 20-pole catheter (AFocus II HD, Abbott) and the Ensite Velocity Electro-anatomical mapping system (Abbott). Surface ECG and intracardiac ECGs were recorded using a surface digital amplifier/recorder system (Labsystem Pro, Bard Electrophysiology). ECGs were recorded with 0.05 to 100 Hz (high- and low- pass filter). Intracardiac bipolar electrograms were filtered at 30 to 300 Hz and bipolar voltage during a 4- s recording period at each site. To ensure highest accuracy only mapping sites that were within a distance of 5 mm from the acquired atrial geometry contributed to the voltage map, high adjustment setting: 12 on Ensite Velocity (Abbott) and respiratory gating were performed.

The following catheters were introduced via the right femoral vein: (1) a steerable decapolar catheter was located in the CS; (2) a 20-pole mapping catheter able to cover an area of 2-cm 24 diameter (AFocus II HD: 1-mm electrodes with 4-mm spacing) was used to map LA in addition to providing atrial geometry; (3) an irrigated-tip quadripolar catheter with a distal 3.5-mm tip and three 1-mm proximal electrodes with interelectrode distance of 2, 5 and 2 mm (TactiCath, Abbott). During atrial mapping, considering temporal EGM amplitude variability, the circumferential catheter was kept in the same position for at least 6 to 8 seconds.

Low voltage area in AF was considered as sites displaying <0.5 mV peak-to-peak bipolar voltage (determined as the maximum bipolar voltage within a 4 seconds window of interest during AF). Low voltage areas in AF <0.5 mV were confirmed by contact force enabled catheters (>5 g), in order to exclude ‘false low voltage areas’ due to low electrode-tissue contact.

Spatial relationships between low voltage areas <0.5 mV in AF and the regions of DE were quantified by manually contouring both areas, as well as the region of overlap (between low voltage areas and dense DE and between low voltage areas and patchy DE), on the registered volume (Jadidi, et al. 2013). Each of these surface areas was measured separately and expressed as a percentage of the total LA area, total low voltage area <0.5mV in AF, total dense DE area, and total patchy DE area (Jadidi, et al. 2016).

AF Ablation. The procedure was performed under heparin anticoagulation with a targeted ACT value of 300 to 350 s after transseptal access. The procedural end point of ablation was completion of PVI. PVI was achieved by circumferential ablation around PV ostia. Irrigated- tip catheter ablation was performed using 25-30 W power. We applied 25 W at a maximum duration of 25 seconds of radiofrequency application at each ablation site at the posterior LA and at 28-30 W for 45 seconds at other LA areas.

Electrogram Characteristics of Arrhythmogenic Sites. Electrograms were analyzed according to their voltage, duration and sequence of activation compared to other adjacent electrodes. EGM analysis was performed in a blinded manner, i.e. without knowledge about the distribution of low voltage areas or DE area.

Potential arrhythmogenic sources were defined as electric activity with following characteristics:

1) Intermittently repeated rotational activity (RotA) was considered as repeated sequential activation around the circumferential mapping catheter that exceeded 8 adjacent electrodes (corresponding >250° pivoting), displayed >70% of the local AF cycle length (AFCL). Moreover, the consistency of the observed rotational activity had to exceeded 2/10 consecutive AF beats (Figure 8) in order to qualify as a rotational site.

2) Continuous activity (CA) was defined as the coverage of electric activity ≥70% of local AFCL on at least 1 bipolar (and <8 bipolar) on at least 7/10 consecutive AF beats (Figure 9).

3) Within the 20 bipolar recordings with repetitive rotational patterns, the bipolar electrogram that displayed the longest duration (longest local conduction time with assumable slow conduction) was considered as the site with prolonged activity (PA), if its duration exceeded 50% of the local AFCL (Figure 8).

These sites were annotated on the map for comparison to the spatial distribution of DE areas and low voltage area.

For analysis of regional AF patterns, AF electrogram recordings with the double-spiral catheter were reassessed from all mapped LA areas, where LA geometry was acquired. Each LA area corresponds to a position of the double-spiral 20-polar mapping catheter that was positioned to cover the respective LA area, in order to allow best electrode contact. A mean of 48+/-4 LA areas were analyzed in each patient. At each of the 48 LA areas with catheter-tissue contact and high quality electrograms, rotational activity or continuous activity was searched within 10 consecutive AF beats on the 19 bipolar electrogram recordings. This resulted in analysis of a total of 480 AF wavelet patterns in each patient, corresponding to 7680 AF patterns (and to145920 AF EGMs) in the whole 16 patients.

Analysis of Electrograms at Arrhythmogenic Sites. For the analysis of electrogram characteristics of arrhythmogenic sites, we exported electrogram from each of these arrhythmogenic regions from the maps.

The exported electrogram were analyzed according to their mean voltage in AF (4-s recording period). Electrogram criteria defining target areas were established during AF. High-mapping density (2 - 3 acquisitions per site) allowed integration of voltage variations that occur in AF into the map. The AFCL was measured by averaging the CL of 10 consecutive AF beats on a given bipolar electrode within the left atrial appendage or CS or the current location of mapping catheter within the atria.

Figure 8. Spatial distribution of atrial low voltage areas (A), delayed enhancement areas (B), and the relationship to an atrial site with rotational activity (RotA; white arrows in C). The double-spiral catheter is positioned at the left posterior LA within low voltage and DE area, representing one of the 48 mapped LA areas in this patient. Rotational activity is identified by electrical activity exceeding 70% of the local cycle length and occurring around ≥8 electrodes of the circumferential catheter (≥270 degree rotation; white arrow in A, B, E and gray arrows in C). The example illustrates co-localization of a RotA to LVA and DE sites at the posterior LA. The electrogram with prolonged activity is highlighted by the white box in C. (Figure 8A shows the voltage map of the LA with purple area representing voltages >0.5mV in AF; non-purple areas represent low voltage <0.5 mV in AF.)

RotA was detected, as mentioned above, by sequential activation on the double-spiral catheter. Therefore, the surface area of a RotA-source was assumed to correspond to the area covered at least by 250° of the spiral catheter (3.1 cm2). Similarly, the surface area of CA was assumed to take the electrodes involved in the bipolar recordings that displayed CA. If a single bipole (4 mm spaced adjacent electrodes with a sensing field measuring 5 mm radius) displayed CA, the surface area of CA was calculated as follows: A = π * (0.5 cm)2 = 0.8 cm2.

Figure 9. Relationship of atrial DE and atrial low voltage areas (LVA) to sites with continuous activity. Illustration of the relationship between low voltage area (A), delayed enhancement (DE) area (B), and an area displaying continuous activity (highlighted by the white boxes in C) in a patient with persistent AF. The double-spiral catheter is positioned at the roof/posterior LA, representing one of the 48 mapped LA areas in this patient. The heterogeneous area at the high posterior LA displays low voltage (A) and patchy DE (B) and contains a region with continuous electrical activity. (Figure 9A shows the voltage map of the LA with purple area representing voltages >0.5 mV in AF; non-purple areas represent low voltage <0.5 mV in AF.)

2.4 Statistical Analyses

Statistical analyses were performed using SPSS version 22.0 (IBM Corporation, Armonk, NY, USA). Continuous variables are presented as mean ± SD or as and 25th and 75th percentiles, according to their distribution. Categorical data was expressed as a fraction or percentage. Because EGMs characteristics (areas and voltage) showed a positively skewed distribution, data were analyzed using nonparametric tests. 2-sided Mann-Whitney U test or Kruskal-Wallis 1-way analysis were performed for comparison of continuous variables, as appropriate. Categorical data were compared between different groups using the chi-square test. A p value < 0.05 was considered statistically significant.

3. Results

3.1 Patient Characteristics

The clinical characteristics of the study group (n = 16; age 62 ± 12 years; 13 men) are outlined in Table 2 (see next page). Nine of sixteen (56%) patients presented with long-standing persistent AF (≥12 months duration), the remaining six patients had persistent AF (>7 days, but <12 months). The total history of AF was 81 ± 65 months, with an uninterrupted AF duration of 25 ± 11 months. The left ventricular ejection fraction was 53 ± 13%, and 9 of 16 (56%) patients presented left ventricular dysfunction (LVEF <50%). Structural heart disease, coronary artery disease and hypertension were present in 44%, 6% and 56% of patients, respectively. The average number of prior electrical cardioversions was 1.8 ± 0.8 per patient. The number of failed anti-arrhythmic drug therapies prior to AF ablation were 2.4 ± 1.0, and 10 of 16 (63%) patients were on oral amiodarone treatment.

3.2 Left Atrial High-density Mapping

A total of 16550 points were acquired from AF maps for bipolar voltage and electrogram duration (>1000 points per LA map). For analysis of regional AF patterns, 10 consecutive AF beats were analyzed at each mapped LA area. AF wavelet patterns were mapped and analyzed at a median of 48 LA areas in each patient using the double-spiral 20-pole catheter. We analyzed 29 a total of 480 AF patterns in each patient, corresponding to 9120 AF electrograms and 145920 recorded electrograms in the 16 patients. Each of the 48 LA regions was scrutinized for repetitive propagation patterns during AF displaying rotational activity (RotA; Figure 8) or continuous activity (CA; Figure 9), as defined under the methods. Analysis of 10 consecutive AF beats at each mapped LA area revealed 1 rotational (IQR: 0.25 - 4) and 3 continuous activities (IQR: 1 - 5) per patient. In the total study population, we analyzed 7680 AF wavelet patterns (recorded by the 20-pole double-spiral catheter).

Table 2. Clinical characteristics of study population (n = 16) Age (years) 62 ± 12 Sex (male; n / %) 13 / 81 History of AF (months) 81 ± 65 Duration of continuous AF (months) 25 ± 11 Long-persistent AF (≥ 12 months) (n / %) 9 / 56 LVEF (%) 53 ± 13 Left ventricular dysfunction (LVEF<50%; n / %) 9 / 56 LVEDD (mm) 54 ± 7 LVESD (mm) 36 ± 9 IVSDD (mm) 12 ± 2 LAD (mm) 46 ± 7 Stuctural heart disease (n / %) 7 / 44 Coronary artery disease (n / %) 1 / 6 Hypertension (n / %) 9 / 56 No. of electrical cardioversions 1.8 ± 0.8 No. of failed AAD 2.4 ± 1.0 Administration of amiodarone (n / %) 10 / 63

Values are number/percentage or mean ± SD.

AAD = antiarrhythmic drug; AF = atrial fibrillation; LVEDD = left ventricular end-diastolic diameter; LVEF = left ventricular ejection fraction; LVESD = left ventricular end-systolic diameter; IVSDD = interventricular septum diastolic diameter; LAD = left atrial diameter (anteroposterior)

3.3 LA Size, Extent of Delayed Enhanced Areas, Low Voltage Areas and Arrhythmogenic Sites

The median LA surface area was 112 cm2 (IQR: 97 – 136 cm2). On average, 61 cm2 (IQR: 54 – 74 cm2) of the LA demonstrate delayed enhancement, accounting for 55% of the total LA surface (Figure 10). Of all DE-areas, 28% demonstrated dense and 72% patchy pattern of DE (p <0.001) (Figure 10). In contrast, low voltage areas <0.5 mV in AF covered a smaller LA area: 30 cm2 (IQR: 14 -53 cm2), corresponding to the 24% of the LA surface (Figure 10). Rotational and continuous activities took 2 cm2 (IQR: 1.5 – 4 cm2) and 2.5 cm2 (IQR: 1 – 5 cm2), corresponding each to 2% of the total LA surface, respectively (Figure 10).

Figure 10. Extent of atrial delayed enhanced (DE) areas (left), low voltage areas (LVA; middle) and arrhythmogenic areas with rotational (RotA) or continuous (CA) activity (right).

There was a weak correlation between LA size (total LA surface area in cm2) and area of pathological substrate, independent of the method used (correlation to total DE areas: r = 0.41, p=0.11; correlation to dense DE areas: r = 0.42, p=0.11; correlation to patchy DE areas: r = 0.21, p=0.43 and correlation to low voltage areas <0.5 mV: r = 0.52, p=0.04). (Figure 11)

Figure 11. Correlation between left atrium (LA) size (LA surface area in cm2) and extent of pathological substrate area as expressed by: (A) total delayed enhancement (DE), (B) dense DE area; (C) patchy DE area; (D) low voltage area (<0.5 mV).

3.4 Regional Distribution of Delayed Enhanced Areas and Low Voltage Areas

Differences in the regional distribution of DE areas vs low voltage areas <0.5 mV were noted in all patients.

Dense DE predominantly located to the left posterior LA, with only few patients (13%) demonstrating dense DE at the antero-septal LA (Table 3). Patchy DE was found in 94% of cases at the right posterior LA and, to small extents, at the antero-septal LA in 69% of patients (Table 3).

In contrast, low voltage areas <0.5 mV were mainly found at the anterior and septal LA (88% and 56% of patients, respectively), followed by posterior LA (50% of patients) and LA roof (44% of patients). Patients displayed low voltage areas at the lateral wall of the LA only in 13% of cases (Table 3).

The greatest mismatch between DE-MRI and low voltage mapping was found at the posterior and lateral LA, with significantly more DE as compared to low voltage: at the left posterior LA, 100% of patients presented DE, whereas low voltage was found in 50% of patients. Again, at the LA lateral wall DE was present in 88% of patients vs. low voltage in 13% of patients. Furthermore, at the antero-septal LA wall we frequently found important low voltage substrate as compared to the smaller extent of DE areas. Figure 12 gives an example of a patient with typical DE and low voltage substrate distribution, leading to a mismatch between DE areas and low voltage areas <0.5 mV.

Table 3. Percentage of patients presenting delayed enhancement and low voltage area in the indicated left atrial regions.

Posterior Anterior Roof Septal Lateral Left Right Dense DE 100% 31% 38% 6% 13% 25% Patchy DE 31% 94% 88% 75% 69% 69% DE 100% 88% 75% 69% 88% LVA 50% 88% 44% 56% 13%

DE = delayed enhancement; LVA = low voltage area

Figure 12. Representative example of a patient with persistent AF demonstrating spatial mismatch between atrial low voltage substrate (mainly localizing to the anteroseptal LA wall in A) and DE areas (mainly present at the posterior LA wall in B). Notably, dense DE is present in abundance at the left posterior wall (region with high voltage electrograms in B). In contrast, high anterior LA with low voltage electrograms does not show DE at MRI. A&B left panel show the voltage map of the LA with purple area representing voltages >0.5 mV in AF; non-purple areas represent low voltage <0.5 mV in AF.

3.5 Correlation of Delayed Enhanced Areas and Low Voltage Areas

Discrepancies in the spatial distribution of DE areas and low voltage area were observed: DE was present at 61% of low voltage area, whereas low voltage area was present at 28% of DE areas (Figure 13A, 14). Seventy-two percent of DE areas displayed bipolar voltages >0.5 mV in AF (Figure 13A, 14). Thirty-one percent of dense DE overlapped with low voltage area, whereas 23% of patchy DE overlapped with low voltage area (Figure 13A, C).

Sixty-one percent of low voltage areas overlapped with DE areas (Figure 13B, C). Twenty-two percent of low voltage area overlapped with dense DE and 39% overlapped with patchy DE (Figure 13B).

Figure 13. Overlap between low voltage area (LVA) and DE area. (A) LVA as percentage of DE region: LVAs were displayed in 42% of dense DE (in dark green), 28% of patchy DE (in dark orange). (B) DE as percentage of LVA: Sixty-one percent of LVA overlapped with DE areas. 39% of LVA overlapped with patchy DE and 22% with dense DE areas. Notably, non-DE areas contained 39% of low voltage areas. (C) DE and LVA vs total LA surface: DE was present at 55% of left atrium surface whereas LVA was present at 24% of left atrium surface.

Figure 14. Spatial relationship between (1) total LA surface (gray rectangle), (2) arrhythmogenic AF sources (rotational activity (RotA), continuous activity (CA) and prolonged activity (PA) as the small orange, green and red ovals), (3) DE areas (large pink oval area) and low voltage areas (large yellow circle): DE (pink area) covers 55% of total LA surface and low voltage (yellow area) covers 24% of total LA surface. Rotational and continuous activities cover a small percentage of total LA area and mostly overlap with low voltage areas and partially with DE areas.

3.6 Relationship of Arrhythmogenic Sites to Delayed Enhanced Areas and Low Voltage Areas

The majority of the rotational (82%) and continuous (88%) activities located on the left atrial body, and the remaining were found within the PV antra. While 75% of rotational activities were detected within low voltage area, only 64% of were localized within DE areas. In a comparable fashion, the vast majority of continuous activities were found within low voltage area (80%), while localization within DE areas was found in 61% (Figure 15A). Healthy atrium did not contain any rotational activities (Figure 15A).

Rotational activities co-localized preferentially with patchy DE (46%) than with dense DE (18%). Similarly, continuous activities were found predominantly in patchy DE (45% compared to 16% in dense DE). Healthy atrium contained 2% of continuous activities (Figure 15A).

Thirty-nine and 43% of rotational and continuous activities were located within areas displaying both DE and low voltage (Figure 15A).

Forty percent of sites with prolonged electrical activities located to both DE and low voltage areas (Figure 15A). Notably, both DE areas and low voltage areas had a low specificity for identification of the identified arrhythmogenic sites: only 6% of DE areas vs. 8% of low voltage areas harbored the arrhythmogenic sites displaying rotational or continuous activities (Figure 15B, C).

3.7 Electrogram Characteristics of Arrhythmogenic Sites

The consistency of RotA was 60% (6/10 consecutive beats displayed RotA). The consistency of CA was 90% (9/10 consecutive AF beats displayed CA). Electrogram voltage of the rotational activities (on the spiral catheter) was 0.64 ± 0.47 mV, while sites with continuous activities displayed 0.58 ± 0.51 mV.

Local prolonged activity was found at all 28 rotational activities and 56 continuous activities sites, which displayed low voltage (<0.5 mV) electrograms on ≥ 2 of 19 bipolar recordings (Figure 16). Prolonged electrical activity (prolonged duration with fractionation) was observed in all patients on ≥ 1 bipolar recordings at sites with rotational activity. These sites displayed prolonged electrical activity exceeding 50% of AF cycle length on a single bipole and >70% of AFCL within the mapping areas of the spiral catheter (all electrodes of the 20-pole catheter). The longest electrogram duration (on a single bipole) measured 115 ± 14 ms and displayed an electrogram voltage of 0.34 ± 0.11 mV.

Figure 15. Relationship of DE area and low voltage area with three arrhythmic activities. A illustrates percentage of LA sites with rotational activity (RotA), continuous activity (CA) or prolonged activity (at sites with RotA), which co-localize with low voltage area (yellow bars) vs. with delayed enhanced areas (pink bars) vs. both low voltage areas (LVA) and DE-areas (orange bars) vs. neither with LVA nor DE-MRI (gray bars). It shows that most RotA (75%) and CA (80%) areas co-localize with low voltage areas. In contrast, 64% of RotA and 61% of CA sites co-localize with DE areas at MRI. All RotA and prolonged activities co-localize either with LVA or DE-MRI. 2% CA colocalize neither with low voltage. B&C illustrate that the sites displaying CA (green bars), RotA (orange bars), prolonged activity at RotA(yellow bars) account for DE area 4%, 2%, and 1% respectively, whereas account for low voltage area 7%, 1%, 0.5%, respectively. (** p<0.01)

Figure 16. Characteristics of three arrhythmogenic activities. Maximum bipolar voltage at LA areas with rotational activity (RotA: orange box), continuous activity (CA: green box) and prolonged activity of RotA (yellow box). Data are box plot with min/max whiskeis. (*** p <0.001 vs. prolonged activity of RotA).

4. Discussion

4.1 The Main Findings of Our Study

The current study adds important new insights on atrial DE-MRI and voltage mapping in persistent AF:

First, the left atrial surface area that is considered pathologic/fibrotic using DE-MRI is more than twice as large as when voltage mapping with a conservative cutoff of <0.5 mV is used (55% vs. 24%).

Second, the regional distribution of affected areas is inconsistent between the two methods, with the majority of DE-areas being located at the posterior and lateral LA wall (88%), while low-voltage areas primarily co-localize to the anterior and septal region (88% and 56%).

Third, both methods share a relatively low specificity to detect AF sources, with repetitive continuous or rotational electrical activity occurring in only 6% and 8% of pathologic areas in DE-MRI and voltage mapping, respectively.

And fourth, despite a larger overall surface area identified by DE-MRI, arrhythmogenic sites are mostly located within low-voltage areas (78% vs. 63%). 2% CA located within healthy atrium, whereas no RotA located within healthy atrium.

4.2 Atrial Voltage Mapping for Identification of Arrhythmogenic Substrate

High-density Voltage Mapping in Sinus Rhythm and Atrial Fibrillation Identification of pathological substrate by voltage mapping depends on a predefined voltage threshold to distinguish diseased from healthy atrial myocardium. In sinus rhythm a peak-to- peak bipolar amplitude <0.5 mV (on 3.5 mm ablation catheter) has been extensively used as diseased and <0.1 - 0.2 mV as scar tissue with “loss of capture” upon high-output pacing (Squara, et al. 2014) and atrial areas with voltages 0.2 -0.5 mV have been attributed to scar borderzones.

Histologically validated transmural radiofrequency ablation lesions revealed ablated tissue displayed voltages <0.3 mV, when a 3.5 mm tip catheter was used (Harrison, et al. 2014). Notably, in the same atrium, using a 20-pole (1 mm electrode) mapping catheter, peak-to-peak bipolar electrogram voltages are 1.5 to 1.7 fold higher than voltage mapping using a 3.5 mm tip catheter (Anter, et al. 2015; Liang, et al. 2017). Moreover, Anter and Josephson et al. found smaller extent of the LA being affected by low voltage areas in the same patient, when comparing voltage maps by the small-sized electrodes using the 20-pole PentaRay mapping catheter, compared to the voltage maps created by use of the larger-sized electrodes (3.5mm tip) of the ablation catheter (Anter, et al. 2015).

Further reports from Yang and Lin et al. have reported abnormal low voltage tissue during sinus rhythm as left atrial areas displaying ≤1.3 mV, when using 20-polar circumferential mapping catheter (Lin, et al. 2014; Yang, et al. 2016). Further studies have demonstrated that voltage amplitudes also depend on the rate and the direction of the activation fronts: with increasing activation rate (e.g. during AF), the maximum voltage values decrease (Williams, et al. 2017).

In addition voltage values may vary depending on the direction of the main activation front (Williams, et al. 2017). Therefore, evaluation of atrial substrate by high-density voltage mapping must take into consideration both the underlying rhythm/rate (AF vs. sinus rhythm or paced rhythm) and the electrode size of the mapping catheter, which determines the mapping resolution and the extent of low voltage substrate.

Recent clinical studies from our laboratory and from Yagishita and Arruda et al. used a voltage cutoff of <0.5 mV in AF to define low voltage tissue. Although electrogram voltages during AF may underlie variations (depending on rate and direction of the activation front during AF), both independent studies demonstrated that selective ablation of low voltage areas <0.5 mV in AF (displaying repetitive prolonged or rotational or rapid electrical activity) in addition to PVI is associated with improved arrhythmia freedom-rates, when compared to a PVI-only approach in patients with persistent AF (Jadidi, et al. 2016; Yagishita, et al. 2017).

The above-mentioned studies therefore suggest that left-atrial areas that display a bipolar voltage <0.5mV in AF (by 20-pole 1mm electrode mapping catheters) are regions involved in arrhythmogenesis and maintenance of AF. In addition, Yagishita et al. showed a good correlation between low voltage areas <0.5 mV in AF and low voltage areas <1.5 mV in sinus rhythm, using the 3.5 mm tip electrode of ablation catheter (Yagishita, et al. 2016). In the current study, low voltage areas in AF <0.5 mV were confirmed by contact force enabled catheters, in order to exclude ‘false low voltage’ due to low electrode-tissue contact.

In the current study, atrial voltage mapping was performed during clinical persistent AF, which enabled us to assess the distribution of LA areas with arrhythmogenic rotational or continuous activities. In our experience, atrial areas displaying low-voltage (<0.5 mV) in sinus rhythm also display low-voltage (<0.5 mV) in AF (due to local fibrosis/conduction disturbance in affected areas which reduces electrogram amplitudes in both rhythms). Thus, the identified low voltage areas in AF in the current study include the low voltage substrate that is identifiable during sinus rhythm. Besides atrial fibrosis, low voltage in AF may reflect dynamic phenomena as wavelet collisions or functional refractory tissue. However, high voltage areas during AF (>0.5 mV) are consistent with high voltage areas during sinus rhythm. In the current analysis, we found large areas of LA high voltage (>0.5 mV) tissue overlapping with DE areas at MRI in 15

41 of 16 (94%) study patients. Therefore, the current mismatch between DE areas occurring at high voltage LA areas demonstrates difficulties with current MRI resolution and technology to identify atrial fibrosis.

4.3 The Spatial Distribution of Dense/Patchy DE and of Atrial Low Voltage Areas in AF

The greatest mismatch between low voltage area and DE areas was found in the posterior and anterior LA regions. We observed significantly more DE at MRI in the left posterior LA as compared to low voltage (100% patients with DE vs. 50% with low voltage areas), and a higher frequency of DE as compared to low voltage substrate in the lateral LA (DE in 88% of patients vs. low voltage areas in 13% of patients).

4.3.1 Differences in the Extent of Substrate Detection in DE-MRI and Voltage Mapping

We demonstrate important differences in the overall substrate area when comparing voltage mapping vs. DE-MRI. Our study used the high-resolution MRI sequence from Marrouche et al. (Utah group) as it is used in clinical practice and the DECAAF-trails, with previously published segmentation methods to identify DE areas (Jadidi, et al. 2013; Oakes, et al. 2009): Specifically, in the current study, DE area were identified as such if the voxel intensity was ≥4SD of the mean LA wall intensity. The histological substrate of these areas is thought to be replacement of cardiomyocytes by fibro-fatty tissue (Platonov, et al. 2011).

In a recent study, Higushi and Marrouche et al. analyzed the Utah results on the spatial distribution of DE area of the LA from 160 patients with AF, which are in complete agreement with the DE distribution patterns we report in our manuscript (Higuchi, et al. 2018). However, alternative methods have been introduced for identification of DE area at MRI: A study from Khurram and colleagues showed a high correlation between DE areas with low voltage areas during sinus rhythm, when the “image intensity ratio IIR >0.97” (IIR = MRI image intensity ratio; comparing mean intensity of regional atrial wall to mean intensity of total blood pool) was used to define pathological atrial enhancement (Khurram, et al. 2014; Khurram, et al. 2016).

The authors showed that IIR >0.97 and >1.61 corresponded to bipolar voltage values during sinus rhythm <0.5 mV and <0.1 mV, respectively (Khurram, et al. 2014).

Another study by Zghaib et al. found good correlation between DE-MRI and low voltage areas using IIR-thresholding with a cutoff value of IIR >0.74 corresponding to low voltage in sinus rhythm (Zghaib, et al. 2018). The good correlations reported in both studies however must be seen with the important limitation that 57% and 65% of patients, respectively, had undergone prior PVI. As a result, low voltage areas mainly restricted to pre-ablated atrial areas or the PV and their antral areas. The good correlation in both studies accounts for the important gadolinium enhancement at these vascular structures or pre-ablated atrial areas, which is more reliably identifiable than pathological patchy fibrosis within the atrial body/musculature.

In contrast to these studies, we used the high-resolution Utah DE-MRI sequence together with the reported methodology to detect the reconstruct DE areas (Figure 7), as it is performed in the DECAAF trails. Importantly, we focused on study on 16 ablation-naïve patients with history of (long-standing)-persistent AF, in whom significant arrhythmogenic low voltage areas were found within the atrial body.

4.3.2 Potential Causes for Mismatch between MRI and Voltage Mapping for Identification of Fibro-fatty Arrhythmogenic Atrial Substrate

Currently, it is unclear why MRI identifies DE area most frequently at the left posterior LA (all 16 patients in the current study and in the study from Higuchi et al. (Higuchi, et al. 2018) whereas low voltage areas are most frequently found at anterior LA (88% of patients) and septal LA (56% of patients). Again, our current results about distribution of low voltage substrate (<0.5 mV in AF) are in accordance with our previous observations with voltage mapping in AF (Jadidi, et al. 2016) and reports by other groups, using voltage mapping in sinus rhythm (Huo, et al. 2018; Yagishita, et al. 2016). a) Differences exist in the left atrial wall thickness between the posterior LA/pulmonary vein junction and the anterior LA, which may also contribute to voltage variations. While both regions demonstrate anatomical proximity to the descending and ascending aorta, respectively, wall thickness measures only 0.8 to 3 mm at the posterior vs. 3 to 7 mm at the anterior LA wall. 43

The current resolution of DE-MRI might be inaccurate particularly in the thin-walled segments of the LA, which may explain discrepancies with low voltage substrate at the left posterior LA. b) DE area at left posterior LA may be due to artefact/increased signal intensity at DE-MRI due to the presence of adjacent descending aorta with high signal intensity values (typical for vascular wall structures). Further ameliorations and developments of the MR imaging and fibrosis detection at higher resolution may modify the current observations. c) Atrial DE may depict/identify epicardial fibro-fatty tissue that is not detected by high-density (small-electrode) endocardial voltage mapping. d) The used voltage threshold of 0.5 mV during AF is not sensitive enough to detect all fibrotic areas, which might be more accurately detected by gadolinium enhancement at MRI.

Currently, none of the above-mentioned hypothesis can be validated with certainty, as the gold standard to evaluate presence of fibrosis (histological analysis of atrial walls) is not possible in vivo.

4.4 Electrophysiological Criteria to Localize Sources of AF: Electrogram Voltage and Duration at Atrial Areas with Continuous or Rotational Activity

We investigated three criteria of arrhythmogenic sites in AF: repetitive rotational activity (RotA >250 degree on ≥3/10 consecutive AF beats), continuous activity (CA on ≥7/10 consecutive AF beats) and locally persisting prolonged activity (slow conduction) at a site with rotational activity. Detection of localized reentrant AF sources displaying >70% of local AF cycle length with rotational activity was first described by Haissaguerre et al. in 2006 using 20- pole PentaRay catheter in persistent AF patients after AF organization was achieved by ablation (Haissaguerre, et al. 2006). Moreover, Takahashi and Haissaguerre et al. (Takahashi, et al. 2008) revealed AF termination occurring at sites of continuous activity during AF.

We recently showed that PVI plus selective ablation of these areas displaying rotational or continuous activity within/at low voltage borderzones is associated with high AF termination rate and ameliorated arrhythmia freedom rates, compared to PVI only approach (Jadidi, et al. 2016). Use of small (1 mm) multi-electrode mapping catheters (PentaRay or the spiral-like 44 catheters with a simultaneous mapping field of 9 cm2 and 3.1 cm2, respectively) enables detection of low amplitude signals, enhancing the chance for identification of rotational or focal AF sources. However, the detection of these small AF sources remains challenging despite use of multi-electrode mapping catheters, because of the intermittent nature of the AF sources with changing activation patterns during AF and influences of neighboring wavelets on local and regional atrial activation.

Using the currently described methods by visual scrutinizing of regional AF activation patterns recorded by the 20-pole spiral catheter, we observed in each patient a median of one left atrial areas with repetitive RotA and three areas with CA. In the current study population of persistent and long-standing persistent AF patients the majority of the rotational (82%) and continuous (88%) activities located on the left atrial body and the remaining were found within the PV antra. All potential arrhythmogenic sites with rotational or continuous activity displayed reduced bipolar peak-to-peak voltage values (0.64 ± 0.47 mV) with prolonged electrical activity duration (115 ± 14 ms), as a potential underlying slow conduction site.

In AF, the reduced bipolar voltage may be caused by increased fibrosis, functional tissue refractoriness or dissociation of endo-mid-epicardial myocardial tissue layers. Mechanistically, the phenomenon of intermittent repetitive rotational activity may correspond to a localized reentrant source maintaining AF, which depends on a slow conduction isthmus displaying low voltage and prolonged activity. Alternatively, changes in the local tissue architecture and fiber orientations may favor pivoting of AF wavelets at that site, which may lead to intermittent alternating pivoting in one or another direction and intermittent pass-through activations of neighboring wavelets. Both potential mechanisms may contribute to maintenance of persistent AF.

In the current study, all identified rotational and continuous activities presented at least one bipole displaying low voltage electrograms. Most (78%) of the rotational and continuous activities located within low voltage areas <0.5 mV in AF as displayed by the electroanatomical map, suggesting a mechanistical role of atrial low voltage areas and increased atrial fibrosis (62% of RotA and CA located within DE areas) for development of AF perpetuating arrhythmogenic sources. These current findings are supported by recent studies, reporting local fibrosis as the histological substrate of localized reentrant spatially stationary AF sources,

45 which could be detected by panoramic epicardial optical mapping in explanted human atria (Hansen, et al. 2015; Zhao, et al. 2017).

In the current study only 6% of DE areas vs. 8% of low voltage areas harbored the arrhythmogenic sites with rotational or continuous activities. The low specificity of DE areas and low voltage areas for identification of the arrhythmogenic sites may be explained by the fact that not all arrhythmia sources (located within the low voltage areas) were identified by our current methods.

Another observation from regular atrial tachycardias allows to explain the low specificity of low voltage areas for a single rotational source measuring 1 cm in diameter: during ongoing atrial tachycardia (with a small localized rotational reentry as underlying mechanism) only a small portion of the low voltage substrate is involved with the arrhythmia mechanism. After successful ablation of this small area (measuring 0.5-1cm2) and termination of the atrial tachycardia, other atrial tachycardias can frequently be re-induced. These sources locate to the remaining low voltage sites. Thus a small portion of the low voltage substrate is responsible / involved for a given arrhythmia source. However, the remaining low voltage areas may be important for development of further/other arrhythmia sources.

4.5 Implication for Catheter Ablation

The discrepancies in the extent and distribution of atrial delayed enhanced areas vs. low voltage areas are important to consider, when substrate-based ablation strategies are applied to treat patients with persistent AF. Specifically, in the current study - as well as in the recent study by Higuchi and Marrouche et al. (Higuchi, et al. 2018) – DE area were most consistently present at the left posterior LA. Thus, compared to previous ablation strategies, substrate-based catheter ablation for persistent AF that targets DE area detected by MRI will result in significantly more radiofrequency delivery to the posterior LA wall, with increased risk for collateral damage to the esophagus. Therefore, in addition to low voltage or DE as ablation targets, further electrophysiological criteria should be used to guide ablation of arrhythmogenic substrate (late potentials, fractionated potential, slow conduction areas, rapid/continuous or repetitive rotational activity).

4.6 Limitations

There is no gold standard neither for detection of the fibrotic LA substrate nor for detection of sources of AF by available multi-electrode catheters. It may be that adjustment of the MRI technique and voltage mapping is necessary to better characterize the substrate. Thresholding in MRI makes a major difference in results and is varying between centers and even between patients in a given center. Voltage is sensitive to far-field effect, electrode size and spacing, activation front orientation. Other more global mapping techniques such as body surface ECG- imaging (Lim, et al. 2015) or FIRM mapping (Narayan, et al. 2012) may yield different results.

4.7 Summary

This study shows that a significant spatial mismatch exists between delayed Gadolinium enhanced areas and LA areas displaying low voltage in AF. Only a small portion of DE areas and low voltage areas harbor potential arrhythmogenic sources with repetitive rotational and continuous activity. Further studies are necessary to explain why different imaging methods such as DE at MRI and low voltage mapping for detection of the fibro-fatty substrate showed a significant spatial mismatch.

5. Zusammenfassung / Abstract

Hintergrund: Arrhythmie-Rezidive kommen am häufigsten bei den Patienten vor, die ein arrhythmogenes teil-fibrosiertes Vorhofsubstrat aufweisen, das als „Niedrig-Voltage“ Areale beim elektrophysiologischen Voltage-Mapping erfaßt wird oder durch späte Gadoliniumanreicherung („delayed Gadolinium enhancement“ (DE)-Areale) in der MRI auffällt. Die räumliche Beziehung zwischen „low voltage“ Arealen, DE-Arealen und möglichen Vorhofflimmer-Arrhythmieherden ist bisher unbekannt.

Methoden: Sechzehn Patienten mit persistierendem AF (neun lang anhaltendes persistierendes AF) wurden vor dem elektrophysiologischen Mapping und Katheterablation mittels Gadolinium-Kontrast-MRI untersucht. Vor der PVI, wurde in jedem Patienten ein hochauflösendes Voltage-Map (>1000 Stellen) des LA erstellt. Pro-arryhthmische linksatriale Areale mit rotierender (RotA) oder kontinuierlicher (CA) Aktivität wurden nach ihrer Elektrogram-Voltage-Amplitude und -Dauer charakterisiert und ihre anatomische Beziehung zu DE-Arealen und Niedrig-Voltage-Arealen <0.5 mV bestimmt.

Ergebnisse: DE-Areale und Niedrig-Voltage-Areale belegten 55% und 24% (p<0.01) der gesamten linksatrialen (LA) Oberfläche.Einundsechzig Prozent der Niedrig-Voltage-Areale überlappten mit DE-Arealen aus der MRI, während 28% der DE-Areale eine niedrige Voltage <0.5 mV aufwiesen. Rotierender und kontinuierlicher Aktivität kamen häufiger an Niedrig- Voltage-Arealen als an DE-Arealen vor (78% vs. 62%, p = 0.02). Die bipolare Voltage- Amplitude von RotA- vs. CA-Arealen betrug 0.64 ± 0.47 mV vs. 0.58 ± 0.51 mV. Alle 28 RotA- und 56 CA-Areale enthielten Elektrogramme mit verlängerter Dauer (115 ± 14 ms) und niedriger Voltage (0.34 ± 0.11 mV).

Schlussfolgerung: Nur 6% der atrialen DE-Areale in der MRI und 8% der Niedrig-Voltage- Areale beherbergen die arrhythmogenen Quellen mit sich wiederholenden rotierenden oder kontinuierlichen Aktivitäten. Allerdings treten die meisten (78%) der Arrhythmieherde mit kontinuierlicher (CA) oder rotierender Aktivität (RotA) innerhalb von Niedrig-Voltage Arealen (<0.5 mV) auf. Es gibt eine große Diskrepanz zwischen der anatomischen Lokalisation von DE- Arealen aus MRI und Niedrig-Voltage-Arealen des linken Vorhofs, die bei der Substrat- basierten Katheterablation von persistierendem Vorhofflimmern berücksichtigt werden muss.

Abstract

Background: Arrhythmia recurrence following pulmonary vein isolation (PVI) is high in persistent AF due to atrial arrhythmogenic fibro-fatty substrate that can be identified as low voltage areas or atrial delayed Gadolinium enhancement (DE) at MRI. The relationship between low voltage areas, DE areas and AF triggers is unknown.

Methods: Sixteen patients with persistent AF (9 long-standing) underwent DE-MRI (1.25 mm ×1.25 mm ×2.5 mm) prior to PVI. LA voltage-mapping was acquired in AF (>1000 sites/LA) and the regional activation patterns of 7680 AF wavelets were analyzed. Sites with rotational activity (RotA) or continuous activity (CA) were characterized (voltage, duration) and their spatial relationship to DE and low voltage areas <0.5 mV was assessed.

Results: DE and low voltage areas covered 55% and 24% (p <0.01) of total LA surface, respectively. DE was present at 61% of low voltage areas, whereas low voltage was present at 28% of DE areas. RotA and CA more frequently co-localized with low voltage areas than with DE areas (78% vs. 62%, p = 0.02). Regional bipolar voltage of RotA vs. CA was 0.64 ± 0.47 mV vs. 0.58 ± 0.51 mV. All 28 RotA and 56 CA areas contained electrograms with prolonged duration (115 ± 14 ms) displaying low voltage (0.34 ± 0.11 mV).

Conclusion: Only 6% of DE and 8% of low voltage areas harbor the arrhythmogenic areas displaying repetitive rotational or continuous activity. Most (78%) continuous or rotational activities co-localized with low voltage areas, while there was less co-localization with DE areas. There is an important mismatch between DE areas and low voltage areas which needs to be considered when used as target for catheter ablation.

6. References

Andersson T, Magnuson A, Bryngelsson IL, Frobert O, Henriksson KM, Edvardsson N, Poci D (2013), All- cause mortality in 272,186 patients hospitalized with incident atrial fibrillation 1995-2008: A swedish nationwide long-term case-control study, Eur Heart J, 34, 1061-1067. Anter E, Tschabrunn CM, Josephson ME (2015), High-resolution mapping of scar-related atrial arrhythmias using smaller electrodes with closer interelectrode spacing, Circ Arrhythm Electrophysiol, 8, 537-545. Arentz T, Haegeli L, Sanders P, Weber R, Neumann FJ, Kalusche D, Haissaguerre M (2007), High-density mapping of spontaneous pulmonary vein activity initiating atrial fibrillation in humans, J Cardiovasc Electrophysiol, 18, 31-38. Arentz T, Von Rosenthal J, Blum T, Burkle G, Stockinger J, Kalusche D (2003), 3-d mapping of pulmonary veins using a multipolar basket catheter. Implications for catheter ablation of atrial fibrillation, Herz, 28, 566-574. Arentz T, von Rosenthal J, Blum T, Stockinger J, Burkle G, Weber R, Jander N, Neumann FJ, Kalusche D (2003), Feasibility and safety of pulmonary vein isolation using a new mapping and navigation system in patients with refractory atrial fibrillation, Circulation, 108, 2484-2490. Arentz T, Weber R, Burkle G, Herrera C, Blum T, Stockinger J, Minners J, Neumann FJ, Kalusche D (2007), Small or large isolation areas around the pulmonary veins for the treatment of atrial fibrillation? Results from a prospective randomized study, Circulation, 115, 3057-3063. Baber U, Howard VJ, Halperin JL, Soliman EZ, Zhang X, McClellan W, Warnock DG, Muntner P (2011), Association of chronic kidney disease with atrial fibrillation among adults in the united states: Reasons for geographic and racial differences in stroke (regards) study, Circ Arrhythm Electrophysiol, 4, 26-32. Ball J, Carrington MJ, Stewart S (2013), Mild cognitive impairment in high-risk patients with chronic atrial fibrillation: A forgotten component of clinical management?, Heart, 99, 542-547. Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB, Levy D (1998), Impact of atrial fibrillation on the risk of death: The framingham heart study, Circulation, 98, 946-952. Bhatia S, Sugrue A, Asirvatham S (2018), Atrial fibrillation: Beyond rate control, Mayo Clin Proc, 93, 373- 380. Buch E, Share M, Tung R, Benharash P, Sharma P, Koneru J, Mandapati R, Ellenbogen KA, Shivkumar K (2016), Long-term clinical outcomes of focal impulse and rotor modulation for treatment of atrial fibrillation: A multicenter experience, Heart Rhythm, 13, 636-641. Calkins H, Hindricks G, Cappato R, Kim YH, Saad EB, Aguinaga L, Akar JG, Badhwar V, Brugada J, Camm J, Chen PS, Chen SA, Chung MK, Nielsen JC, Curtis AB, Davies DW, Day JD, d'Avila A, de Groot N, Di Biase L, Duytschaever M, Edgerton JR, Ellenbogen KA, Ellinor PT, Ernst S, Fenelon G, Gerstenfeld EP, Haines DE, Haissaguerre M, Helm RH, Hylek E, Jackman WM, Jalife J, Kalman JM, Kautzner J, Kottkamp H, Kuck KH, Kumagai K, Lee R, Lewalter T, Lindsay BD, Macle L, Mansour M, Marchlinski FE, Michaud GF, Nakagawa H, Natale A, Nattel S, Okumura K, Packer D, Pokushalov E, Reynolds MR, Sanders P, Scanavacca M, Schilling R, Tondo C, Tsao HM, Verma A, Wilber DJ, Yamane T (2017), 2017 hrs/ehra/ecas/aphrs/solaece expert consensus statement on catheter and surgical ablation of atrial fibrillation: Executive summary, J Arrhythm, 33, 369- 409. 50

Chalfoun N, Harnick D, Pe E, Undavia M, Mehta D, Gomes JA (2007), Reverse electrical remodeling of the atria post cardioversion in patients who remain in sinus rhythm assessed by signal averaging of the p-wave, Pacing Clin Electrophysiol, 30, 502-509. Cochet H, Dubois R, Yamashita S, Al Jefairi N, Berte B, Sellal JM, Hooks D, Frontera A, Amraoui S, Zemoura A, Denis A, Derval N, Sacher F, Corneloup O, Latrabe V, Clement-Guinaudeau S, Relan J, Zahid S, Boyle PM, Trayanova NA, Bernus O, Montaudon M, Laurent F, Hocini M, Haissaguerre M, Jais P (2018), Relationship between fibrosis detected on late gadolinium-enhanced cardiac magnetic resonance and re-entrant activity assessed with electrocardiographic imaging in human persistent atrial fibrillation, JACC Clin Electrophysiol, 4, 17-29. Corley SD, Epstein AE, DiMarco JP, Domanski MJ, Geller N, Greene HL, Josephson RA, Kellen JC, Klein RC, Krahn AD, Mickel M, Mitchell LB, Nelson JD, Rosenberg Y, Schron E, Shemanski L, Waldo AL, Wyse DG, Investigators A (2004), Relationships between sinus rhythm, treatment, and survival in the atrial fibrillation follow-up investigation of rhythm management (affirm) study, Circulation, 109, 1509-1513. Di Biase L, Mohanty P, Mohanty S, Santangeli P, Trivedi C, Lakkireddy D, Reddy M, Jais P, Themistoclakis S, Dello Russo A, Casella M, Pelargonio G, Narducci ML, Schweikert R, Neuzil P, Sanchez J, Horton R, Beheiry S, Hongo R, Hao S, Rossillo A, Forleo G, Tondo C, Burkhardt JD, Haissaguerre M, Natale A (2016), Ablation versus amiodarone for treatment of persistent atrial fibrillation in patients with congestive heart failure and an implanted device: Results from the aatac multicenter randomized trial, Circulation, 133, 1637-1644. Dixit S, Marchlinski FE, Lin D, Callans DJ, Bala R, Riley MP, Garcia FC, Hutchinson MD, Ratcliffe SJ, Cooper JM, Verdino RJ, Patel VV, Zado ES, Cash NR, Killian T, Tomson TT, Gerstenfeld EP (2012), Randomized ablation strategies for the treatment of persistent atrial fibrillation: Rasta study, Circ Arrhythm Electrophysiol, 5, 287-294. Ehrlich JR, Cha TJ, Zhang L, Chartier D, Melnyk P, Hohnloser SH, Nattel S (2003), Cellular electrophysiology of canine pulmonary vein cardiomyocytes: Action potential and ionic current properties, J Physiol, 551, 801-813. Einthoven W (1906), Le télécardiogramme, Arch Int Physiol, 4, 132–164. Fazekas T (2007), [the concise history of atrial fibrillation], Orvostort Kozl, 53, 37-68. Fazekas T, Liszkai G, Bielik H, Luderitz B (2003), [history of atrial fibrillation], Z Kardiol, 92, 122-127. Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, Singer DE (2001), Prevalence of diagnosed atrial fibrillation in adults: National implications for rhythm management and stroke prevention: The anticoagulation and risk factors in atrial fibrillation (atria) study, JAMA, 285, 2370-2375. Groenveld HF, Crijns HJ, Rienstra M, Van den Berg MP, Van Veldhuisen DJ, Van Gelder IC (2009), Does intensity of rate control influence outcome in persistent atrial fibrillation? Data of the race study, Am Heart J, 158, 785-791. Haissaguerre M, Hocini M, Denis A, Shah AJ, Komatsu Y, Yamashita S, Daly M, Amraoui S, Zellerhoff S, Picat MQ, Quotb A, Jesel L, Lim H, Ploux S, Bordachar P, Attuel G, Meillet V, Ritter P, Derval N, Sacher F, Bernus O, Cochet H, Jais P, Dubois R (2014), Driver domains in persistent atrial fibrillation, Circulation, 130, 530-538. Haissaguerre M, Hocini M, Sanders P, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu LF, Jonsson A, O'Neill MD, Bordachar P, Reuter S, Roudaut R, Clementy J, Jais P (2006), Localized sources maintaining atrial fibrillation organized by prior ablation, Circulation, 113, 616-625.

Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le Mouroux A, Le Metayer P, Clementy J (1998), Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins, N Engl J Med, 339, 659-666. Haissaguerre M, Shah AJ, Cochet H, Hocini M, Dubois R, Efimov I, Vigmond E, Bernus O, Trayanova N (2016), Intermittent drivers anchoring to structural heterogeneities as a major pathophysiological mechanism of human persistent atrial fibrillation, J Physiol, 594, 2387-2398. Hansen BJ, Zhao J, Csepe TA, Moore BT, Li N, Jayne LA, Kalyanasundaram A, Lim P, Bratasz A, Powell KA, Simonetti OP, Higgins RS, Kilic A, Mohler PJ, Janssen PM, Weiss R, Hummel JD, Fedorov VV (2015), Atrial fibrillation driven by micro-anatomic intramural re-entry revealed by simultaneous sub-epicardial and sub-endocardial optical mapping in explanted human hearts, Eur Heart J, 36, 2390-2401. Harrison JL, Jensen HK, Peel SA, Chiribiri A, Grondal AK, Bloch LO, Pedersen SF, Bentzon JF, Kolbitsch C, Karim R, Williams SE, Linton NW, Rhode KS, Gill J, Cooklin M, Rinaldi CA, Wright M, Kim WY, Schaeffter T, Razavi RS, O'Neill MD (2014), Cardiac magnetic resonance and electroanatomical mapping of acute and chronic atrial ablation injury: A histological validation study, Eur Heart J, 35, 1486-1495. Henriksson KM, Farahmand B, Asberg S, Edvardsson N, Terent A (2012), Comparison of cardiovascular risk factors and survival in patients with ischemic or hemorrhagic stroke, Int J Stroke, 7, 276- 281. Higuchi K, Cates J, Gardner G, Morris A, Burgon NS, Akoum N, Marrouche NF (2018), The spatial distribution of late gadolinium enhancement of left atrial magnetic resonance imaging in patients with atrial fibrillation, JACC Clin Electrophysiol, 4, 49-58. Ho SY, Cabrera JA, Sanchez-Quintana D (2012), Left atrial anatomy revisited, Circ Arrhythm Electrophysiol, 5, 220-228. Hocini M, Ho SY, Kawara T, Linnenbank AC, Potse M, Shah D, Jais P, Janse MJ, Haissaguerre M, De Bakker JM (2002), Electrical conduction in canine pulmonary veins: Electrophysiological and anatomic correlation, Circulation, 105, 2442-2448. Huo Y, Gaspar T, Pohl M, Sitzy J, Richter U, Neudeck S, Mayer J, Kronborg MB, Piorkowski C (2018), Prevalence and predictors of low voltage zones in the left atrium in patients with atrial fibrillation, Europace, 20, 956-962. Jadidi AS, Cochet H, Shah AJ, Kim SJ, Duncan E, Miyazaki S, Sermesant M, Lehrmann H, Lederlin M, Linton N, Forclaz A, Nault I, Rivard L, Wright M, Liu X, Scherr D, Wilton SB, Roten L, Pascale P, Derval N, Sacher F, Knecht S, Keyl C, Hocini M, Montaudon M, Laurent F, Haissaguerre M, Jais P (2013), Inverse relationship between fractionated electrograms and atrial fibrosis in persistent atrial fibrillation: Combined magnetic resonance imaging and high-density mapping, J Am Coll Cardiol, 62, 802-812. Jadidi AS, Lehrmann H, Keyl C, Sorrel J, Markstein V, Minners J, Park CI, Denis A, Jais P, Hocini M, Potocnik C, Allgeier J, Hochholzer W, Herrera-Sidloky C, Kim S, Omri YE, Neumann FJ, Weber R, Haissaguerre M, Arentz T (2016), Ablation of persistent atrial fibrillation targeting low-voltage areas with selective activation characteristics, Circ Arrhythm Electrophysiol, 9, e002962. Khurram IM, Beinart R, Zipunnikov V, Dewire J, Yarmohammadi H, Sasaki T, Spragg DD, Marine JE, Berger RD, Halperin HR, Calkins H, Zimmerman SL, Nazarian S (2014), Magnetic resonance image intensity ratio, a normalized measure to enable interpatient comparability of left atrial fibrosis, Heart Rhythm, 11, 85-92.

Khurram IM, Habibi M, Gucuk Ipek E, Chrispin J, Yang E, Fukumoto K, Dewire J, Spragg DD, Marine JE, Berger RD, Ashikaga H, Rickard J, Zhang Y, Zipunnikov V, Zimmerman SL, Calkins H, Nazarian S (2016), Left atrial lge and arrhythmia recurrence following pulmonary vein isolation for paroxysmal and persistent af, JACC Cardiovasc Imaging, 9, 142-148. Kim MH, Johnston SS, Chu BC, Dalal MR, Schulman KL (2011), Estimation of total incremental health care costs in patients with atrial fibrillation in the united states, Circ Cardiovasc Qual Outcomes, 4, 313-320. Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, Castella M, Diener HC, Heidbuchel H, Hendriks J, Hindricks G, Manolis AS, Oldgren J, Popescu BA, Schotten U, Van Putte B, Vardas P, Group ESCSD (2016), 2016 esc guidelines for the management of atrial fibrillation developed in collaboration with eacts, Eur Heart J, 37, 2893-2962. Kosiuk J, Dinov B, Kornej J, Acou WJ, Schonbauer R, Fiedler L, Buchta P, Myrda K, Gasior M, Polonski L, Kircher S, Arya A, Sommer P, Bollmann A, Hindricks G, Rolf S (2015), Prospective, multicenter validation of a clinical risk score for left atrial arrhythmogenic substrate based on voltage analysis: Dr-flash score, Heart Rhythm, 12, 2207-2212. Kotecha D, Holmes J, Krum H, Altman DG, Manzano L, Cleland JG, Lip GY, Coats AJ, Andersson B, Kirchhof P, von Lueder TG, Wedel H, Rosano G, Shibata MC, Rigby A, Flather MD (2014), Efficacy of beta blockers in patients with heart failure plus atrial fibrillation: An individual-patient data meta- analysis, Lancet, 384, 2235-2243. Kottkamp H, Bender R, Berg J (2015), Catheter ablation of atrial fibrillation: How to modify the substrate?, J Am Coll Cardiol, 65, 196-206. Krijthe BP, Kunst A, Benjamin EJ, Lip GY, Franco OH, Hofman A, Witteman JC, Stricker BH, Heeringa J (2013), Projections on the number of individuals with atrial fibrillation in the european union, from 2000 to 2060, Eur Heart J, 34, 2746-2751. Kuck KH, Brugada J, Furnkranz A, Metzner A, Ouyang F, Chun KR, Elvan A, Arentz T, Bestehorn K, Pocock SJ, Albenque JP, Tondo C, Fire, Investigators ICE (2016), Cryoballoon or radiofrequency ablation for paroxysmal atrial fibrillation, N Engl J Med, 374, 2235-2245. Lavie CJ, Pandey A, Lau DH, Alpert MA, Sanders P (2017), Obesity and atrial fibrillation prevalence, pathogenesis, and prognosis: Effects of weight loss and exercise, J Am Coll Cardiol, 70, 2022- 2035. Liang JJ, Elafros MA, Muser D, Pathak RK, Santangeli P, Supple GE, Schaller RD, Frankel DS, Dixit S (2017), Comparison of left atrial bipolar voltage and scar using multielectrode fast automated mapping versus point-by-point contact electroanatomic mapping in patients with atrial fibrillation undergoing repeat ablation, J Cardiovasc Electrophysiol, 28, 280-288. Lim HS, Derval N, Komatsu Y, Zellerhoff S, Denis A, Shah AJ, Sacher F, Hocini M, Jais P, Haissaguerre M (2015), Is ablation to termination the best strategy for ablation of persistent atrial fibrillation? Persistent atrial fibrillation is best ablated by a strategy that terminates the arrhythmia: Procedural termination is associated with improved long-term outcomes, Circ Arrhythm Electrophysiol, 8, 963-971. Lim HS, Hocini M, Dubois R, Denis A, Derval N, Zellerhoff S, Yamashita S, Berte B, Mahida S, Komatsu Y, Daly M, Jesel L, Pomier C, Meillet V, Amraoui S, Shah AJ, Cochet H, Sacher F, Jais P, Haissaguerre M (2017), Complexity and distribution of drivers in relation to duration of persistent atrial fibrillation, J Am Coll Cardiol, 69, 1257-1269.

Lin Y, Yang B, Garcia FC, Ju W, Zhang F, Chen H, Yu J, Li M, Gu K, Cao K, Callans DJ, Marchlinski FE, Chen M (2014), Comparison of left atrial electrophysiologic abnormalities during sinus rhythm in patients with different type of atrial fibrillation, J Interv Card Electrophysiol, 39, 57-67. Lip GY, Beevers DG (1995), Abc of atrial fibrillation. History, epidemiology, and importance of atrial fibrillation, BMJ, 311, 1361-1363. Liu T, Korantzopoulos P, Shao Q, Zhang Z, Letsas KP, Li G (2016), Mineralocorticoid receptor antagonists and atrial fibrillation: A meta-analysis, Europace, 18, 672-678. Lloyd-Jones DM, Wang TJ, Leip EP, Larson MG, Levy D, Vasan RS, D'Agostino RB, Massaro JM, Beiser A, Wolf PA, Benjamin EJ (2004), Lifetime risk for development of atrial fibrillation: The framingham heart study, Circulation, 110, 1042-1046. Mahida S, Sacher F, Derval N, Berte B, Yamashita S, Hooks D, Denis A, Amraoui S, Hocini M, Haissaguerre M, Jais P (2015), Science linking pulmonary veins and atrial fibrillation, Arrhythm Electrophysiol Rev, 4, 40-43. Mahnkopf C, Badger TJ, Burgon NS, Daccarett M, Haslam TS, Badger CT, McGann CJ, Akoum N, Kholmovski E, Macleod RS, Marrouche NF (2010), Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced mri: Implications for disease progression and response to catheter ablation, Heart Rhythm, 7, 1475-1481. Marrouche NF, Brachmann J, Andresen D, Siebels J, Boersma L, Jordaens L, Merkely B, Pokushalov E, Sanders P, Proff J, Schunkert H, Christ H, Vogt J, Bansch D, Investigators C-A (2018), Catheter ablation for atrial fibrillation with heart failure, N Engl J Med, 378, 417-427. Marrouche NF, Wilber D, Hindricks G, Jais P, Akoum N, Marchlinski F, Kholmovski E, Burgon N, Hu N, Mont L, Deneke T, Duytschaever M, Neumann T, Mansour M, Mahnkopf C, Herweg B, Daoud E, Wissner E, Bansmann P, Brachmann J (2014), Association of atrial tissue fibrosis identified by delayed enhancement mri and atrial fibrillation catheter ablation: The decaaf study, JAMA, 311, 498-506. Marzona I, O'Donnell M, Teo K, Gao P, Anderson C, Bosch J, Yusuf S (2012), Increased risk of cognitive and functional decline in patients with atrial fibrillation: Results of the ontarget and transcend studies, CMAJ, 184, E329-336. McGann C, Akoum N, Patel A, Kholmovski E, Revelo P, Damal K, Wilson B, Cates J, Harrison A, Ranjan R, Burgon NS, Greene T, Kim D, Dibella EV, Parker D, Macleod RS, Marrouche NF (2014), Atrial fibrillation ablation outcome is predicted by left atrial remodeling on mri, Circ Arrhythm Electrophysiol, 7, 23-30. Moe GK, Abildskov JA (1959), Atrial fibrillation as a self-sustaining arrhythmia independent of focal discharge, Am Heart J, 58, 59-70. Moe GK, Rheinboldt WC, Abildskov JA (1964), A computer model of atrial fibrillation, Am Heart J, 67, 200-220. Muller P, Grabowski C, Schiedat F, Shin DI, Dietrich JW, Mugge A, Deneke T, Walther JW, Kara K (2016), Reverse remodelling of the atria after treatment of obstructive sleep apnoea with continuous positive airway pressure: Evidence from electro-mechanical and endocrine markers, Heart Lung Circ, 25, 53-60. Nademanee K, McKenzie J, Kosar E, Schwab M, Sunsaneewitayakul B, Vasavakul T, Khunnawat C, Ngarmukos T (2004), A new approach for catheter ablation of atrial fibrillation: Mapping of the electrophysiologic substrate, J Am Coll Cardiol, 43, 2044-2053.

Narayan SM, Krummen DE, Rappel WJ (2012), Clinical mapping approach to diagnose electrical rotors and focal impulse sources for human atrial fibrillation, J Cardiovasc Electrophysiol, 23, 447-454. Narayan SM, Krummen DE, Shivkumar K, Clopton P, Rappel WJ, Miller JM (2012), Treatment of atrial fibrillation by the ablation of localized sources: Confirm (conventional ablation for atrial fibrillation with or without focal impulse and rotor modulation) trial, J Am Coll Cardiol, 60, 628- 636. Narayan SM, Patel J, Mulpuru S, Krummen DE (2012), Focal impulse and rotor modulation ablation of sustaining rotors abruptly terminates persistent atrial fibrillation to sinus rhythm with elimination on follow-up: A video case study, Heart Rhythm, 9, 1436-1439. Nattel S, Burstein B, Dobrev D (2008), Atrial remodeling and atrial fibrillation: Mechanisms and implications, Circ Arrhythm Electrophysiol, 1, 62-73. Nattel S, Harada M (2014), Atrial remodeling and atrial fibrillation: Recent advances and translational perspectives, J Am Coll Cardiol, 63, 2335-2345. O'Neill MD, Wright M, Knecht S, Jais P, Hocini M, Takahashi Y, Jonsson A, Sacher F, Matsuo S, Lim KT, Arantes L, Derval N, Lellouche N, Nault I, Bordachar P, Clementy J, Haissaguerre M (2009), Long- term follow-up of persistent atrial fibrillation ablation using termination as a procedural endpoint, Eur Heart J, 30, 1105-1112. Oakes RS, Badger TJ, Kholmovski EG, Akoum N, Burgon NS, Fish EN, Blauer JJ, Rao SN, DiBella EV, Segerson NM, Daccarett M, Windfelder J, McGann CJ, Parker D, MacLeod RS, Marrouche NF (2009), Detection and quantification of left atrial structural remodeling with delayed- enhancement magnetic resonance imaging in patients with atrial fibrillation, Circulation, 119, 1758-1767. Pathak RK, Elliott A, Middeldorp ME, Meredith M, Mehta AB, Mahajan R, Hendriks JM, Twomey D, Kalman JM, Abhayaratna WP, Lau DH, Sanders P (2015), Impact of cardiorespiratory fitness on arrhythmia recurrence in obese individuals with atrial fibrillation: The cardio-fit study, J Am Coll Cardiol, 66, 985-996. Platonov PG, Ivanov V, Ho SY, Mitrofanova L (2008), Left atrial posterior wall thickness in patients with and without atrial fibrillation: Data from 298 consecutive autopsies, J Cardiovasc Electrophysiol, 19, 689-692. Platonov PG, Mitrofanova LB, Orshanskaya V, Ho SY (2011), Structural abnormalities in atrial walls are associated with presence and persistency of atrial fibrillation but not with age, J Am Coll Cardiol, 58, 2225-2232. Po SS, Scherlag BJ, Yamanashi WS, Edwards J, Zhou J, Wu R, Geng N, Lazzara R, Jackman WM (2006), Experimental model for paroxysmal atrial fibrillation arising at the pulmonary vein-atrial junctions, Heart Rhythm, 3, 201-208. Prabhu S, Voskoboinik A, McLellan AJA, Peck KY, Pathik B, Nalliah CJ, Wong GR, Azzopardi SM, Lee G, Mariani J, Ling LH, Taylor AJ, Kalman JM, Kistler PM (2018), Biatrial electrical and structural atrial changes in heart failure: Electroanatomic mapping in persistent atrial fibrillation in humans, JACC Clin Electrophysiol, 4, 87-96. Prystowsky EN (2008), The history of atrial fibrillation: The last 100 years, J Cardiovasc Electrophysiol, 19, 575-582. Rivard L, Hocini M, Rostock T, Cauchemez B, Forclaz A, Jadidi AS, Linton N, Nault I, Miyazaki S, Liu X, Xhaet O, Shah A, Sacher F, Derval N, Jais P, Khairy P, Macle L, Nattel S, Willems S, Haissaguerre

M (2012), Improved outcome following restoration of sinus rhythm prior to catheter ablation of persistent atrial fibrillation: A comparative multicenter study, Heart Rhythm, 9, 1025-1030. Rolf S, Kircher S, Arya A, Eitel C, Sommer P, Richter S, Gaspar T, Bollmann A, Altmann D, Piedra C, Hindricks G, Piorkowski C (2014), Tailored atrial substrate modification based on low-voltage areas in catheter ablation of atrial fibrillation, Circ Arrhythm Electrophysiol, 7, 825-833. Rothberger C, Winterberg H (1909), Vorhofflimmern und arhythmia perpetua, Wien Klin Wochenschr, 22, 839 - 844. Ryu K, Sahadevan J, Khrestian CM, Stambler BS, Waldo AL (2006), Use of fast fourier transform analysis of atrial electrograms for rapid characterization of atrial activation-implications for delineating possible mechanisms of atrial tachyarrhythmias, J Cardiovasc Electrophysiol, 17, 198-206. Sanders P, Nalliah CJ, Dubois R, Takahashi Y, Hocini M, Rotter M, Rostock T, Sacher F, Hsu LF, Jonsson A, O'Neill MD, Jais P, Haissaguerre M (2006), Frequency mapping of the pulmonary veins in paroxysmal versus permanent atrial fibrillation, J Cardiovasc Electrophysiol, 17, 965-972. Savelieva I, Kakouros N, Kourliouros A, Camm AJ (2011), Upstream therapies for management of atrial fibrillation: Review of clinical evidence and implications for european society of cardiology guidelines. Part i: Primary prevention, Europace, 13, 308-328. Scherlag BJ, Yamanashi W, Patel U, Lazzara R, Jackman WM (2005), Autonomically induced conversion of pulmonary vein focal firing into atrial fibrillation, J Am Coll Cardiol, 45, 1878-1886. Scherr D, Khairy P, Miyazaki S, Aurillac-Lavignolle V, Pascale P, Wilton SB, Ramoul K, Komatsu Y, Roten L, Jadidi A, Linton N, Pedersen M, Daly M, O'Neill M, Knecht S, Weerasooriya R, Rostock T, Manninger M, Cochet H, Shah AJ, Yeim S, Denis A, Derval N, Hocini M, Sacher F, Haissaguerre M, Jais P (2015), Five-year outcome of catheter ablation of persistent atrial fibrillation using termination of atrial fibrillation as a procedural endpoint, Circ Arrhythm Electrophysiol, 8, 18- 24. Seitz J, Bars C, Ferracci A, Maluski A, Penaranda G, Theodore G, Faure J, Bremondy M, Curel L, Beurtheret S, Avula UMR, Kalifa J, Pisapia A (2016), Electrogram fractionation-guided ablation in the left atrium decreases the frequency of activation in the pulmonary veins and leads to atrial fibrillation termination: Pulmonary vein modulation rather than isolation, JACC Clin Electrophysiol, 2, 732-742. Seitz J, Bars C, Theodore G, Beurtheret S, Lellouche N, Bremondy M, Ferracci A, Faure J, Penaranda G, Yamazaki M, Avula UM, Curel L, Siame S, Berenfeld O, Pisapia A, Kalifa J (2017), Af ablation guided by spatiotemporal electrogram dispersion without pulmonary vein isolation: A wholly patient-tailored approach, J Am Coll Cardiol, 69, 303-321. Skanes AC, Gray RA, Zuur CL, Jalife J (1998), Effects of postshock atrial pacing on atrial defibrillation outcome in the isolated sheep heart, Circulation, 98, 64-72. Squara F, Frankel DS, Schaller R, Kapa S, Chik WW, Callans DJ, Marchlinski FE, Dixit S (2014), Voltage mapping for delineating inexcitable dense scar in patients undergoing atrial fibrillation ablation: A new end point for enhancing pulmonary vein isolation, Heart Rhythm, 11, 1904-1911. Steinberg BA, Hellkamp AS, Lokhnygina Y, Patel MR, Breithardt G, Hankey GJ, Becker RC, Singer DE, Halperin JL, Hacke W, Nessel CC, Berkowitz SD, Mahaffey KW, Fox KA, Califf RM, Piccini JP, Committee R-AS, Investigators (2015), Higher risk of death and stroke in patients with persistent vs. Paroxysmal atrial fibrillation: Results from the rocket-af trial, Eur Heart J, 36, 288- 296.

Stewart S, Hart CL, Hole DJ, McMurray JJ (2002), A population-based study of the long-term risks associated with atrial fibrillation: 20-year follow-up of the renfrew/paisley study, Am J Med, 113, 359-364. Takahashi K, Okumura Y, Watanabe I, Nagashima K, Sonoda K, Sasaki N, Kogawa R, Iso K, Ohkubo K, Nakai T, Hirayama A (2015), Relation between left atrial wall thickness in patients with atrial fibrillation and intracardiac electrogram characteristics and atp-provoked dormant pulmonary vein conduction, J Cardiovasc Electrophysiol, 26, 597-605. Takahashi Y, O'Neill MD, Hocini M, Dubois R, Matsuo S, Knecht S, Mahapatra S, Lim KT, Jais P, Jonsson A, Sacher F, Sanders P, Rostock T, Bordachar P, Clementy J, Klein GJ, Haissaguerre M (2008), Characterization of electrograms associated with termination of chronic atrial fibrillation by catheter ablation, J Am Coll Cardiol, 51, 1003-1010. Teh AW, Kistler PM, Lee G, Medi C, Heck PM, Spence SJ, Sparks PB, Morton JB, Kalman JM (2012), Electroanatomic remodeling of the left atrium in paroxysmal and persistent atrial fibrillation patients without structural heart disease, J Cardiovasc Electrophysiol, 23, 232-238. Tilz RR, Rillig A, Thum AM, Arya A, Wohlmuth P, Metzner A, Mathew S, Yoshiga Y, Wissner E, Kuck KH, Ouyang F (2012), Catheter ablation of long-standing persistent atrial fibrillation: 5-year outcomes of the hamburg sequential ablation strategy, J Am Coll Cardiol, 60, 1921-1929. Verma A (2011), The techniques for catheter ablation of paroxysmal and persistent atrial fibrillation: A systematic review, Curr Opin Cardiol, 26, 17-24. Verma A, Jiang CY, Betts TR, Chen J, Deisenhofer I, Mantovan R, Macle L, Morillo CA, Haverkamp W, Weerasooriya R, Albenque JP, Nardi S, Menardi E, Novak P, Sanders P, Investigators SAI (2015), Approaches to catheter ablation for persistent atrial fibrillation, N Engl J Med, 372, 1812-1822. Verma A, Wazni OM, Marrouche NF, Martin DO, Kilicaslan F, Minor S, Schweikert RA, Saliba W, Cummings J, Burkhardt JD, Bhargava M, Belden WA, Abdul-Karim A, Natale A (2005), Pre- existent left atrial scarring in patients undergoing pulmonary vein antrum isolation: An independent predictor of procedural failure, J Am Coll Cardiol, 45, 285-292. Verrier RL, Josephson ME (2009), Impact of sleep on arrhythmogenesis, Circ Arrhythm Electrophysiol, 2, 450-459. von Eisenhart Rothe A, Hutt F, Baumert J, Breithardt G, Goette A, Kirchhof P, Ladwig KH (2015), Depressed mood amplifies heart-related symptoms in persistent and paroxysmal atrial fibrillation patients: A longitudinal analysis--data from the german competence network on atrial fibrillation, Europace, 17, 1354-1362. Whitaker J, Rajani R, Chubb H, Gabrawi M, Varela M, Wright M, Niederer S, O'Neill MD (2016), The role of myocardial wall thickness in atrial arrhythmogenesis, Europace, 18, 1758-1772. Wijffels MC, Kirchhof CJ, Dorland R, Allessie MA (1995), Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats, Circulation, 92, 1954-1968. Williams SE, Linton NWF, Harrison J, Chubb H, Whitaker J, Gill J, Rinaldi CA, Razavi R, Niederer S, Wright M, O'Neill M (2017), Intra-atrial conduction delay revealed by multisite incremental atrial pacing is an independent marker of remodeling in human atrial fibrillation, JACC Clin Electrophysiol, 3, 1006-1017. Wolf PA, Abbott RD, Kannel WB (1991), Atrial fibrillation as an independent risk factor for stroke: The framingham study, Stroke, 22, 983-988.

Wyse DG, Waldo AL, DiMarco JP, Domanski MJ, Rosenberg Y, Schron EB, Kellen JC, Greene HL, Mickel MC, Dalquist JE, Corley SD (2002), A comparison of rate control and rhythm control in patients with atrial fibrillation, N Engl J Med, 347, 1825-1833. Yagishita A, Gimbel JR, S DEO, Manyam H, Sparano D, Cakulev I, Mackall J, Arruda M (2017), Long-term outcome of left atrial voltage-guided substrate ablation during atrial fibrillation: A novel adjunctive ablation strategy, J Cardiovasc Electrophysiol, 28, 147-155. Yagishita A, S DEO, Cakulev I, Gimbel JR, Sparano D, Manyam H, Manrique-Garcia A, Arredondo M, Mackall J, Arruda M (2016), Correlation of left atrial voltage distribution between sinus rhythm and atrial fibrillation: Identifying structural remodeling by 3-d electroanatomic mapping irrespective of the rhythm, J Cardiovasc Electrophysiol, 27, 905-912. Yang G, Yang B, Wei Y, Zhang F, Ju W, Chen H, Li M, Gu K, Lin Y, Wang B, Cao K, Kojodjojo P, Chen M (2016), Catheter ablation of nonparoxysmal atrial fibrillation using electrophysiologically guided substrate modification during sinus rhythm after pulmonary vein isolation, Circ Arrhythm Electrophysiol, 9, e003382. Zghaib T, Keramati A, Chrispin J, Huang D, Balouch MA, Ciuffo L, Berger RD, Marine JE, Ashikaga H, Calkins H, Nazarian S, Spragg DD (2018), Multimodal examination of atrial fibrillation substrate: Correlation of left atrial bipolar voltage using multi-electrode fast automated mapping, point- by-point mapping, and magnetic resonance image intensity ratio, JACC Clin Electrophysiol, 4, 59-68.

Zhao J, Hansen BJ, Wang Y, Csepe TA, Sul LV, Tang A, Yuan Y, Li N, Bratasz A, Powell KA, Kilic A, Mohler PJ, Janssen PML, Weiss R, Simonetti OP, Hummel JD, Fedorov VV (2017), Three-dimensional integrated functional, structural, and computational mapping to define the structural "fingerprints" of heart-specific atrial fibrillation drivers in human heart ex vivo, J Am Heart Assoc, 6.

7. Publication

Publication in Proof Journal:

Juan Chen, Thomas Arentz, Hubert Cochet, Steven Kim, Zoraida Moreno-Weidmann, Bjoern Mueller-Edenborn, Jan Minners, Heiko Lehrmann, Juergen Allgeier, Dietmar Trenk, Meleze Hocini, Pierre Jais, Michel Haissaguerre, Amir Jadidi. Extent and Spatial Distribution of Left Atrial Arrhythmogenic Sites, Late Gadolinium Enhancement at MRI and Low Voltage Areas in Patients with Persistent Atrial Fibrillation – Comparison of Imaging vs Electrical Parameters of Fibrosis and Arrhythmogenesis. Accepted in April 2019 by Europace (Manuscript in proof).

Contributions:

Juan Chen, Amir Jadidi, Thomas Arentz and Michel Haissaguerre: initiator of the clinical study, study coordinator, data analysis (MRI data and Mapping system), drafting article and revision of article.

Hubert Cochet and Steven Kim: data processing of MRI acquisition, segmentation and registration and data collection.

Zoraida Moreno-Weidmann, Bjoern Mueller-Edenborn, Jan Minners, Heiko Lehrmann, Juergen Allgeier, Meleze Hocini, Pierre Jais, Dietmar Trenk: critical revision of article, statistics.

8. Anhang

- Eidesstattliche Versicherung

- Erklärung zum Eigenanteil

- Acknowledgements

Eidesstattliche Versicherung

Hiermit versichere ich, die vorliegende Abschlussarbeit selbstständig und nur unter

Verwendung der von mir angegebenen Quellen und Hilfsmittel verfasst zu haben. Sowohl inhaltlich als auch wörtlich entnommene Inhalte wurden als solche kenntlich gemacht.

Die Arbeit hat in dieser oder vergleichbarer Form noch kein anderes Prüfungsgremium vorgelegen.

Datum: Unterschrift: ______

Erklärung zum Eigenanteil

Die für diese Dissertation relevanten Daten wurden in Räumlichkeiten des Herzzentrums Bad Krozingen im Zeitraum von Juli 2016 bis August 2017 erhoben. Dieses Projekt wurde durch Herrn Prof. Dr. med. Arentz und Herrn Dr. med. Jadidi ins Leben gerufen und betreut. Die intrakardialen EKG-Aufzeichnungen, drei-dimensional rekonstruierte Fibroseareale aus MRI- Untersuchungen und die links-atrialen drei-dimensionalen Voltage-Kartographien waren im Rahmen elektrophysiologischer Untersuchungen erstellt und durch Prof. Dr. T. Arentz und Dr. A. Jadidi zur weiteren Analyse bereitgestellt worden.

Mein Beitrag an diesem Projekt war die detaillierte Analyse und anatomische Kartographie von EKG-Mustern während Vorhofflimmerns, die potentielle Vorhofflimmer-Arrhythmieherde darstellen. Die Lokalisation dieser Arrhythmieherde sollte mit den Lokalisationen von Niedrig- Voltage-Arealen in der elektrophysiologischen Kartographie des linken Vorhofs sowie mit Fibrose-Arealen aus den MRI-Untersuchungen der Patienten verglichen werden. Zudem wurde die Übereistimmung und Diskrepanz zwischen Low Voltage-Arealen und DE-Arealen aus MRI quantifiziert, um diese in der Klinik eingesetzten Methoden zu vergleichen.

Während der Durchsichtung und Kategorisierung der intrakardialen EKGs sowie der weiteren Schritte zur Korrelation der Verteilungsmuster von Arrhythmieherden zu Fibrose-Arealen aus der MRI und zu atrialen Niedrig-Voltage Regionen standen mir meine Betreuer bei Fragen oder Problemen jederzeit zur Verfügung. Bei der statistischen Auswertung bekam ich freundlicherweise Unterstützung von Dr. Zoraida Moreno Weidmann und Dr. Bjoern Müller-Edenborn.

Acknowledgements

The present work was performed in Prof. Arentz´s group in the electrophysiology lab, department arrhythmia, clinic of cardiology and angiology II, University Heart Center Freiburg- Bad Krozingen, Germany.

First of all, I would like to express my deep appreciation to Prof. Dr. Thomas Arentz and Dr. Amir Jadidi for giving me the opportunity to work in their lab, for continuous great supervision and support, as well as for giving me the chance to present data in international and national congresses in this field. They are helpfulness, consideration and kindness. Special thanks to Prof. Dr. Thomas Arentz and Dr. Amir Jadidi for funding and generous support and advice on my research.

I am grateful to Prof. Dr. Dietmar Trenk for his great support and advices! Thanks to his concerns, guidance and patience. I successfully overcame many difficulties in my life and work.

I would further like to express my thankfulness to Dr. Bjoern Mueller-Edenborn and Dr. Zoraida Moreno-Weidmann for their guidance, advices, constructive critic and motivation!

I would also like to thank my colleagues from electrophysiology lab: Dr. Heiko Lehrmann, Dr. Juergen Allgeier, Dr. Reinhold Weber for their help and the friendly and productive working atmosphere!

My profound thanks are to Frau Nicole Geier and Frau Nadira El Ajouz for a fabulous friendship, which brought sunshine to the disappointing days.

I would like to thank my family for their support, devotion and encouragement! Very special thanks to my lovely sister for always being on my side and taking care of our parents.

Finally, I would like to express thankfulness to German Cardiology Society, Heart Rhythm Society, European Society of Cardiology, European Heart Rhythm Association and Chinese Heart Rhythm Society for giving me the chance to present the data in the congresses!

iii