Europace (2014) 16, 743–749 doi:10.1093/europace/euu045
FOCUSED ISSUE: ORIGINAL RESEARCH
Non-invasive imaging of cardiac electrophysiology in a cardiac resynchronization therapy defibrillator patient with a quadripolar left ventricular lead Michael Seger 1†, Friedrich Hanser 1†, Wolfgang Dichtl 2, Markus Stuehlinger 2, Florian Hintringer 2, Thomas Trieb3, Bernhard Pfeifer 1, and Thomas Berger 2,4*
Received 14 July 2013; accepted after revision 7 February 2014
Aims
The present study was aimed to assess epi- and endocardial ventricular electroanatomical activation during cardiac resynchronization therapy (CRT) by means of non-invasive imaging of cardiac electrophysiology (NICE) in a patient with a novel quadripolar LV lead. ..................................................................................................................................................................................... Methods Non-invasive imaging of cardiac electrophysiology is a novel imaging tool which works by fusing data from high-resolution electrocardiogram (ECG) mapping with a model of the patient’s individual cardiothoracic anatomy created from magnetic and results resonance imaging. This was performed in a cardiac resynchronization therapy defribrillator (CRT-D) patient with a quadripolar left ventricular (LV) lead. Beat-to-beat endocardial and epicardial ventricular activation sequences were computed using NICE during intrinsic conduction as well as during different pacing modes with different LV and biventricular (biV) pacing vectors. The spatial resolution of NICE enabled discrimination of the different pacing vectors during LV and biV pacing. Biventricular pacing resulted in a marked shortening of the total activation duration (TAD) of both ventricles when compared with intrinsic conduction and RV and LV pacing. ..................................................................................................................................................................................... Conclusion Non-invasive imaging of cardiac electrophysiology facilitates non-invasive imaging of ventricular activation, which may be useful in CRT patients to locate the area of latest ventricular activation as the target area for LV lead placement. Moreover, especially in non-responders to CRT NICE may be further useful to determine the best electrical repositioning option.
----------------------------------------------------------------------------------------------------------------------------------------------------------Keywords
Non-invasive imaging † Electroanatomical mapping † NICE † Cardiac resynchronization therapy † Ventricular activation
Introduction Cardiac resynchronization therapy (CRT) is a well established therapy modality in patients with severe heart failure and marked ventricular conduction delay. Despite all current inclusion criteria, non-responder rates of up to one-third of patients are still unsatisfactory for such a cost-intensive treatment.1 Therefore, careful selection of patients to ensure maximal responder rates is crucial. This implies a better knowledge of the underlying anatomical as well as electrical substrate. Novel developments, such as quadripolar left ventricular (LV) leads which enable electrical repositioning by individual pacing †
vector options, may help to increase the response to CRT in selected patients. For example, in patients with ischaemic cardiomyopathy insufficient LV pacing due to high thresholds within scar tissue are frequent problems. Due to the given anatomy of the target cardiac vein, implantation of such a multipolar lead allows electrical repositioning and these individual programmable pacing vectors may result in a better LV-pacing response.2 In all CRT patients, one of the major goals is an improvement of the ventricular electrical conduction delay with consecutive resynchronization of the LV and the right ventricle (RV). Nevertheless, little is known about the electroanatomical properties in these patients. Most of the currently available data were obtained during invasive
MS and FH contributed equally to this paper.
* Corresponding author. Tel: +43 6582 790 71200; fax: +43 6582 790 71290, E-mail address: [email protected] Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2014. For permissions please email: [email protected].
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
1 Institute of Electrical, Electronic and Bioengineering, University for Health Sciences, Medical Informatics and Technology (UMIT), 6060 Hall i. Tirol, Austria; 2Department of Internal Medicine III, Division of Cardiology, Medical University Innsbruck, 6020 Innsbruck, Austria; 3Department of Radiology I, Medical University Innsbruck, 6020 Innsbruck, Austria; and 4 Department of Cardiology, Ludwig Boltzmann Institute for Rehabilitation of Internal Diseases, SKA-RZ Saalfelden, Thorerstraße 26, 5760 Saalfelden, Austria
744
What’s new? † Non-invasive imaging of cardiac electrophysiology (NICE) enables assessment of ventricular electrical activation mapping at a sufficient spatio-temporal resolution. The discrimination of different pacing vectors when using a novel quadripolar left ventricular-lead is clinically feasible applying NICE. Moreover, NICE enables targeted LV-lead placement as well as optimization of electrical repositioning in case of CRT.
Methods
Table 1 Patient’s demographic and clinical data Age (years) Sex
56 M
QRS duration (ms)
183
LVEF (%) Co-morbidities
28 CAD, LBBB, HT, DM, DL
Medications
BB, ARB, Diur, ASA
M, male; LVEF, left ventricular ejection fraction; CAD, coronary artery disease; LBBB, left bundle branch block; HT, hypertension; DM, diabetes mellitus; DL, dyslipidemia; BB, betablocker; ARB, angiotensin receptor blocker; Diur, diuretics; ASA, acetylsalicylic acid.
Template Graphics Software, France) was adapted for this purpose. Liquid-filled anatomical markers on the patient’s torso were used to couple the MRI geometric data with the data obtained from ECG mapping (Figure 1). The electrode positions were transformed by a rigid body transformation (rotation matrix and displacement vector) using a magnetic digitizer (Fastrak; Polhemus, Inc.) as previously described.7 – 10
Non-invasive imaging of cardiac electrophysiology
Prior to CRT-D implantation, patient-specific anatomical data were obtained by magnetic resonance imaging (MRI) with a Magnetom Vision Plus 1.5-T scanner (Siemens). Ventricular end-diastolic geometry and torso geometry were assessed in an ECG-gated cine mode during breath hold as previously described.7 – 9
Non-invasive imaging of cardiac electrophysiology was applied using a bidomain model-based boundary element formulation to relate step-like local activation at the endocardial and epicardial source points to the simulated potentials at the electrode locations. A Wilson terminal defined the reference for measured and computed unipolar signals. For inverse computation of the ventricular activation sequence, the target beats were selected using an automated signal processing algorithm. The model-based computation of the activation sequence of a single beat can be described as a continuous approximation approach. For an assumed model activation sequence, the ECG was simulated and compared with the measured ECG. The activation sequence was then systematically changed to minimize the difference between the simulated and the measured data. The initial estimation of the activation sequence was computed applying the critical point theorem. As this imaging problem was ill-posed, an additional regularization term of the Tikhonov second-order type was considered in the optimization process. This regularization imposes the constraint that the neighbouring source points have similar activation times (smooth activation pattern). The coupled regularization strategy was applied to solve the non-linear optimization problem. In a previous study, the spatial resolution of NICE respectively the localization accuracy was 18 + 5 mm.6 Details of the NICE method have been published previously.7 – 10
Electrocardiogram mapping
Data acquisition
A high-resolution 65-lead electrode array was applied 15 h after CRT-D implantation. The ECG recordings were performed using a batterypowered (6 V) high-precision ECG amplifier (Mark-8 system, Biosemi V.O.F.) with a sampling rate of 2048 Hz (0.3 – 400 Hz band pass filter) and an alternating current resolution of 500 nV/bit.
Non-invasive imaging of cardiac electrophysiology was performed during native sinus rhythm (intrinsic conduction) as well as during single RV, left (LV), and biventricular (biV) pacing. For each pacing mode atrial synchronous pacing was performed. The AV interval was optimized echocardiographically and the VV interval was programmed at 0 ms. For LV and biV pacing modes, different stimulation vectors via the different electrodes of the LV lead (e.g. prox1-dist4) were tested (Figure 2). For all pacing modes and the different pacing vectors right and LV total activation duration, earliest septal, endocardial and epicardial breakthrough sites, and the endocardial/epicardial activation sequences were analysed.
Patient Non-invasive imaging of cardiac electrophysiology (NICE) was performed in a 61-year-old male patient with a complete left bundle branch block who underwent CRT-D implantation (Promote Quadra, St. Jude Medical) due to ischaemic heart failure. The RV lead was implanted in an apical position. For LV pacing a quadripolar LV lead (Quartet, St. Jude Medical) was placed in a posterolateral cardiac vein. During implantation adequate pacing and sensing properties were tested. At all pacing sites, a safety margin of .2× pacing threshold was accepted as adequate. The pacing output was initially set to 3.5 V at 0.5 ms impulse duration at the time of implantation due to safety reasons. The patient’s characterisitics are shown in Table 1.
Magnetic resonance imaging
Model building Magnetic resonance imaging data were used to construct a patientspecific volume conductor model of the patient’s torso, including compartments of different conductivity (heart, lung, blood mass, and chest surface). A commercial software package (AMIRA Developer, TGS
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
studies and primarily endocardial ventricular activation was measured, as information of epicardial activation is limited to a small area of the left ventricle which can be obtained only by mapping in the coronary sinus.3 There have been only a few small studies regarding non-invasive imaging of ventricular electroanatomical activation with limited data on the effects of simultaneous RV endocardial and LV epicardial pacing (CRT) on biventricular endo/epicardial electrical activation.4 – 6 This is the first study investigating the impact of a quadripolar LV-lead on epicardial and endocardial ventricular activation using a novel non-invasive mapping technique.
M. Seger et al.
745
Non-invasive imaging of ventricular activation
2 Voltage leads (mV)
1 0 –1 –2 –3 –4 –5 –6 0
100
200
300 Time (ms)
400
500
600
0
100
200
300 Time (ms)
400
500
600
14
10 8 6 4 2 0 –2
Figure 1 Left panel: volume conductor model (VCM) comprising lungs, ventricles, and atria with two MRI-slices (long axis scan). Right upper panel: high-resolution body surface 65-lead ECG during of intrinsic activation. Right lower panel: standard 12-lead ECG during intrinsic rhythm.
The breakthrough sites were identified as the sites on the ventricular epicardium/endocardium where depolarization first appeared. Ventricular total activation duration was defined as the interval from the earliest breakthrough to the latest observed electrical activation of the RV and LV. The root-mean-square QRS duration was measured during native sinus rhythm as well as during different pacing modes as previously described.11
Results Intrinsic endocardial and epicardial ventricular activation The earliest ventricular activation was located in the RV endocardium. The earliest septal breakthrough site was located in the midseptum. The right septum was activated from apical/mid-septal to basal. In accordance with complete left bundle branch block, the LV endocardial and epicardial activation was markedly delayed. The earliest left septal endocardial breakthrough site was located in a mid-septal position. The LV activation wavefront turned around the apex and spread radially across the inferior/anterior wall towards the lateral wall of the left ventricle. The latest ventricular activation was located epicardially in the lateral wall of the left ventricle (Table 2, Figure 3). There was a marked correlation between the delayed electrical activation of the left lateral wall as obtained by NICE and the MRI-derived LV dysynchrony. The LV lead was
Proximal 4 Mid 3
Mid 2 RV Coil
Distal 1
Figure 2 Ten possible pacing vectors are programmable using the three ring electrodes and distal tip of the Quartet lead and the RV coil (image courtesy of St Jude Medical—already published in Europace Online ISSN 1532-2092 - Print ISSN 1099-5129).
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
Voltage RMS (mV)
12
746
M. Seger et al.
Table 2 Effect of different pacing modes on ventricular electrical properties as derived from NICE Intrinsic
RV
LVdist1-prox4
LVmid3-prox4
biVdist1-prox4
biVmid2-prox4
............................................................................................................................................................................... RVAD (ms)
0– 81
0 –130
74– 213
101–208
0 –165
0–150
LVAD (ms) Breakthrough site
49– 183 Left lat. Epi 183
33– 190 Left apical endo 197
0–94 Septal apical epi
0–142 Left lateral epi
224
222
0 –137 Right apical endo; left septal apical epi 171
0–133 Right apical endo; left lateral epi 164
184
188
198
219
150
148
184
191
214
209
166
151
QRS duration 12-lead (ms) QRS duration RMS (ms) TAD (ms)
Intrinsic, intrinsic activation; RV, right ventricular pacing; LVdist1-prox4, single left ventricular pacing from distal 1 to proximal 4; LVmid3-prox4, single left ventricular pacing from mid 3 to proximal 4; biVdist1-prox4, biventricular pacing from distal 1 to proximal 4; biVmid2-prox4, biventricular pacing from mid 2 to proximal 4; RVAD, right ventricular activation duration in relation to earliest activation onset; LVAD, left ventricular activation duration in relation to earliest activation onset; epi, epicardial; endo, endocardial; RMS, root mean square; TAD, total ventricular activation duration.
Right ventricular pacing Right ventricular pacing showed no significant effects on septal and LV activation when compared with native rhythm. Right ventricular pacing resulted in a decrease of the time course of RV endocardial breakthrough, a reversal of RV endocardial and epicardial activation, and a prolongation of the total activation duration of the LV and the RV.
Left ventricular pacing Left ventricular pacing resulted in a reversal of LV endocardial and epicardial activation of the left ventricle when compared with intrinsic activation. Moreover, LV pacing resulted in a marked shortening of epicardial and endocardial LV activation when compared with RV pacing and intrinsic activation. There were no significant effects on septal activation when compared with native rhythm. In addition, there was a marked prolongation of the total biventricular activation duration (TAD; Table 2, Figure 3).
Biventricular pacing Biventricular pacing showed a marked prolongation of RV activation time when compared with intrinsic activation. During biV pacing, the onset of LV endocardial as well as epicardial activation was markedly shortened when compared with native sinus rhythm. Moreover, during biV pacing the ventricular activation sequence markedly changed. In concordance to LV pacing, the earliest activation was located epicardially in the LV lateral free wall. This was followed by an early endocardial right ventricular apical breakthrough (Table 2, Figure 3).
Pacing vector-dependent changes of left ventricular and biventricular stimulation There was no marked effect on right ventricular activation duration (RVAD) and LV activation duration (LVAD) during biV pacing with a LV dist1-prox4 and a LVmid2-prox4 vector (Table 2, Figure 3).
Single-LV pacing with a dist1-prox4 vector resulted in an LVAD of 94 ms, whereas single-LV pacing with a mid3-prox4 vector resulted in an LVAD of 142 ms. Also the LV epicardial activation sequence changed. During dist1-prox4 pacing, the earliest LV lateral wall activation was located close to the apex as well as close to the LV lateral basis with a distinct multisite LV activation pattern. During mid3-prox4 LV pacing, this distinct multisite LV pacing pattern disappeared with a fused area of mid-lateral epicardial LV activation (Table 2, Figure 3).
Discussion This is the first study using an NICE for visualization of (bi)ventricular electroanatomical activation in a CRT patient with a novel quadripolar LV lead. This multipolar LV lead allows programming of several different pacing vectors. The spatial resolution of NICE also facilitated visualizing the impact of the different LV and biV pacing vectors on ventricular electroanatomical activation properties. There are only few studies on non-invasive imaging of cardiac electroanatomical activation from surface ECG mapping and most of these studies are either limited with respect to clinical applicability (conducted in animals) or limited by an incomplete depiction of cardiac electrical activation (e.g. visualization of either epi- or pericardial activation only).4,12 Non-invasive imaging of cardiac electrophysiology enables single-beat visualization of both endocardial and epicardial electroanatomical activation with sufficient spatial resolution as previously described.6,9 The model building for NICE can be performed using MRI or computed tomography scans prior to CRT implantation to guide the LV lead placement and the same anatomical model can be used after implantation to determine the most appropriate pacing vector to guide electronical repositioning and optimize electroanatomical activation. In this patient, RV pacing resulted in an increase of total RV activation duration when compared with intrinsic activation (Table 2). This increase may be due to the unphysiological activation of the RV as the activation wavefront propagates by cell-to-cell coupling until connecting to the intrinsic conduction system and spreading via the Purkinje system. As previously described, left bundle branch block
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
placed within the corresponding area of delayed electroanatomical activation as indicated by NICE.
747
Non-invasive imaging of ventricular activation
Anterior-Posterior view
Cranial view
Posterior oblique view
Intrinsic
91.5 183
0
95 190
0
91.5 183
0
95 190
0
106.5 213
0
104 208
0
91.5
183
RV
0
95
190
LVdist1 -prox4
0 106.5 213
0
106.5 213
LVmid3 -prox4
0 104 208
0
104
208
0
82.5
165
0
76
biVdist1 -prox4
0
82.5 165
0 82.5 165
biVmid2 -prox4
0 75 150
0 75 150
150
Figure 3 Colour-coded electroanatomic activation map of the RV and RV during intrinsic activation (intrinsic), RV pacing, LV pacing with cathode to anode stimulation from distal 1 to proximal 4 (LVdist1-prox4) and from mid 3 to proximal 4 (LVmid3-prox4) as well as during biV stimulation pacing with cathode to anode stimulation from distal 1 to proximal 4 (biVdist1-prox4) and from mid 2 to proximal 4 (biVmid2-prox4). Red colour illustrates earliest activation, blue (purple) illustrates the area of late (latest) ventricular activation. Time scale numbers reflect the activation duration in milliseconds. Head icon indicates the point of view.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
0
748
and often result in suboptimal pacing thresholds with suboptimal pacing response.13 Therefore, the clinical concept of multipolar pacing leads is the availability of different stimulation configurations due to different programmable pacing vectors. Using this quadripolar LV lead, 10 different biV and LV pacing vector options are available.2 If there are some local capture problems, individual electrical repositioning may help to find acceptable pacing thresholds by programming a different pacing vector. Moreover, implementation of information about scar tissue (as obtained by late enhancement MRI) into the electroanatomical model building may lead to a further refinement of non-invasive electroanatomical imaging. Novel approaches such as cellular automaton models for inverse computation of the electroanatomical activation may be suitable developments for this purpose. Also other limitations such as a suboptimal target vein anatomy may be overcome by the option of individual electrical repositioning.
Conclusions Non-invasive imaging of cardiac electrophysiology facilitates the identification of the target area for LV lead placement as it allows visualization of LV endo/epicardial activation. The combination of electrical and anatomical data may help to increase responders to CRT as this allows an individualized ‘patient-tailored’ pacing therapy. Assessment of this information prior to implantation may help to refine the implantation strategy. It may also help to identify if the electrical and anatomical properties are suitable for an endovascular approach or (e.g. if there is no suitable cardiac vein) if an epicardial surgical approach may be more promising. Moreover, in patients with a quadripolar LV lead NICE may even facilitate post-procedural CRT optimization by identification of the best electrical repositioning option. Conflict of interest: none declared.
Funding This study was financially supported by the Austrian Science Fund (FWF, Fonds zur Fo¨rderung der wissenschaftlichen Forschung), Grant START Y144 and by Oesterreichische Nationalbank Jubila¨umsfond (Jubila¨umsfondsprojekt Nr. 11183). TB received an one year EP-fellowship grant from the European Heart Rhythm Association (European Society of Cardiology). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References 1. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino J et al. Results of the Predictors of Response to CRT (PROSPECT) trial. Circulation 2008;117: 2608 –16. 2. Asbach S, Hartmann M, Wengenmayer T, Graf E, Bode C, Biermann J. Vector selection of a quadripolar left ventricular pacing lead affects acute hemodynamic response to cardiac resynchronization therapy: a randomized cross-over trial. PloS ONE 2013; 8:e67235. 3. Lambiase PD, Rinaldi A, Hauck J, Mobb M, Elliott D, Mohammad S et al. Non-contact left ventricular endocardial mapping in cardiac resynchronisation therapy. Heart 2004;90:44– 51. 4. Ramanathan C, Ghanem RN, Jia P, Ryu K, Rudy Y. Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia. Nat Med 2004; 10:422 – 8. 5. Jia P, Ramanathan C, Ghanem RN, Ryu K, Varma N, Rudy Y. Electrocardiographic imaging of cardiac resynchronization therapy in heart failure: observation of variable electrophysiologic responses. Heart Rhythm 2006;3:296 –310.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
results in a typical electroanatomical activation pattern with latest activation of the epicardial lateral free wall.6 Also in this patient, singleleft and biV pacing resulted in a marked change of the LV activation pattern with resynchronization of the delayed left lateral wall. During single-LV pacing the RV activation duration was decreased when compared with biV pacing. However, the TAD was even shorter during biV pacing when compared with single-LV pacing (Table 2). This indicates that the RV activation timing relative to the LV activation is of particular relevance for the TAD. This effect may also explain the varying impact of the different pacing modes on haemodynamic response. In this patient, epicardial pacing close to the area of latest LV activation as determined by NICE and simultaneous endocardial pacing at the right ventricular apex resulted in the shortest TAD. A limitation of this study is that the results are derived from only a single patient undergoing multipolar LV stimulation. Nevertheless, these results are in accordance with data from another study from our working group, showing similar effects of CRT on (bi)ventricular activation in patients with standard LV stimulation (no quadripolar leads).6 A major limitation of CRT is the high rate of non-responders. There have been several studies to find suitable parameters to predict the response to CRT. Despite some interesting results of smaller cohort studies, larger clinical trials failed to confirm these initially promising predictors such as some sophisticated echocardiographic parameters.1 The main goal of CRT is preceding the delayed LV by stimulation at the left lateral epicardium in patients with left bundle branch block. Thus, optimal placement of the LV lead is crucial. Stimulation as close as possible to the site of latest LV activation should be aimed at.13 Two recent trials nicely demonstrated that targeted LV lead placement close to sites of latest LV mechanical activation resulted in best CRT response.14,15 Both the TARGET and the STARTER trial used echocardiographic parameters for the determination of latest LV activation.14,15 Given the data from the PROSPECT trial, echocardiographic parameters of ventricular dyssynchrony show only a modest sensitivity and specificity for identification of responders to CRT.1 This may be due to the relatively high inter/intraobserver variabilities of these highly sophisticated analyses. Hence, identification of the area of latest LV activation by echocardiography may be difficult in clinical practice. Non-invasive imaging of cardiac electrophysiology may help to overcome these difficulties in the identification of the area of latest ventricular activation as the technical and interpretative factors are well defined. Up to now parameters obtained by standard 12-lead ECG mapping, such as QRS duration and left bundle branch block pattern, still seem to be one of the best predictors for identification of potential responders to CRT. Meta-analysis showed that these ECG criteria were still the strongest predictors for response to CRT. Therefore, the European Society of Cardiology revised in 2012 the guidelines for the management of heart failure and put even more focus on standard ECG criteria.16 Non-invasive imaging of cardiac electrophysiology may have further potential for improving responder identification rates as it works by fusing electrical data obtained from high-resolution body surface ECG with an individual model of the patient’s cardiac anatomy. In patients with severe heart failure, differences in local capture thresholds due to fibrotic changes in the myocardium are common
M. Seger et al.
Non-invasive imaging of ventricular activation
6. Berger T, Pfeifer B, Hanser FF, Hintringer F, Fischer G, Netzer M et al. Single-beat noninvasive imaging of ventricular endocardial and epicardial activation in patients undergoing CRT. PloS ONE 2011;6:e16255. 7. Modre R, Tilg B, Fischer G, Hanser F, Messnarz B, Seger M et al. Atrial noninvasive activation mapping of paced rhythm data. J Cardiovasc Electrophysiol 2003;14: 712 – 9. 8. Tilg B, Fischer G, Modre R, Hanser F, Messnarz B, Schocke M et al. Model-based imaging of cardiac electrical excitation in humans. IEEE Trans Med Imaging 2002;21: 1031– 9. 9. Berger T, Fischer G, Pfeifer B, Modre R, Hanser F, Trieb T et al. Single-beat noninvasive imaging of cardiac electrophysiology of ventricular pre-excitation. J Am Coll Cardiol 2006;48:2045 –52. 10. Fischer G, Hanser F, Pfeifer B, Seger M, Hintermuller C, Modre R et al. A signal processing pipeline for noninvasive imaging of ventricular preexcitation. Methods Inf Med 2005;44:508 –15. 11. Berger T, Hanser F, Hintringer F, Poelzl G, Fischer G, Modre R et al. Effects of cardiac resynchronization therapy on ventricular repolarization in patients with congestive heart failure. J Cardiovasc Electrophysiol 2005;16:611 –7.
749 12. Han C, Liu Z, Zhang X, Pogwizd S, He B. Noninvasive three-dimensional cardiac activation imaging from body surface potential maps: a computational and experimental study on a rabbit model. IEEE Trans Med Imaging 2008;27: 1622 –30. 13. Liu J, Adelstein E, Saba S. Targeting left ventricular lead placement to improve cardiac resynchronization therapy outcomes. Curr Cardiol Rep 2013;15:390. 14. Khan FZ, Virdee MS, Palmer CR, Pugh PJ, O’Halloran D, Elsik M et al. Targeted left ventricular lead placement to guide cardiac resynchronization therapy: the TARGET study: a randomized, controlled trial. J Am Coll Cardiol 2012;59: 1509 –18. 15. Saba S, Marek J, Schwartzman D, Jain S, Adelstein E, White P et al. Echocardiographyguided left ventricular lead placement for cardiac resynchronization therapy: results of the Speckle tracking assisted resynchronization therapy for electrode region trial. Circ Heart Fail 2013;6:427 –34. 16. Sipahi I, Carrigan TP, Rowland DY, Stambler BS, Fang JC. Impact of QRS duration on clinical event reduction with cardiac resynchronization therapy: meta-analysis of randomized controlled trials. Arch Intern Med 2011;171:1454 –62.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
FOCUSED ISSUE: ORIGINAL RESEARCH
Non-invasive imaging of cardiac electrophysiology in a cardiac resynchronization therapy defibrillator patient with a quadripolar left ventricular lead Michael Seger 1†, Friedrich Hanser 1†, Wolfgang Dichtl 2, Markus Stuehlinger 2, Florian Hintringer 2, Thomas Trieb3, Bernhard Pfeifer 1, and Thomas Berger 2,4*
Received 14 July 2013; accepted after revision 7 February 2014
Aims
The present study was aimed to assess epi- and endocardial ventricular electroanatomical activation during cardiac resynchronization therapy (CRT) by means of non-invasive imaging of cardiac electrophysiology (NICE) in a patient with a novel quadripolar LV lead. ..................................................................................................................................................................................... Methods Non-invasive imaging of cardiac electrophysiology is a novel imaging tool which works by fusing data from high-resolution electrocardiogram (ECG) mapping with a model of the patient’s individual cardiothoracic anatomy created from magnetic and results resonance imaging. This was performed in a cardiac resynchronization therapy defribrillator (CRT-D) patient with a quadripolar left ventricular (LV) lead. Beat-to-beat endocardial and epicardial ventricular activation sequences were computed using NICE during intrinsic conduction as well as during different pacing modes with different LV and biventricular (biV) pacing vectors. The spatial resolution of NICE enabled discrimination of the different pacing vectors during LV and biV pacing. Biventricular pacing resulted in a marked shortening of the total activation duration (TAD) of both ventricles when compared with intrinsic conduction and RV and LV pacing. ..................................................................................................................................................................................... Conclusion Non-invasive imaging of cardiac electrophysiology facilitates non-invasive imaging of ventricular activation, which may be useful in CRT patients to locate the area of latest ventricular activation as the target area for LV lead placement. Moreover, especially in non-responders to CRT NICE may be further useful to determine the best electrical repositioning option.
----------------------------------------------------------------------------------------------------------------------------------------------------------Keywords
Non-invasive imaging † Electroanatomical mapping † NICE † Cardiac resynchronization therapy † Ventricular activation
Introduction Cardiac resynchronization therapy (CRT) is a well established therapy modality in patients with severe heart failure and marked ventricular conduction delay. Despite all current inclusion criteria, non-responder rates of up to one-third of patients are still unsatisfactory for such a cost-intensive treatment.1 Therefore, careful selection of patients to ensure maximal responder rates is crucial. This implies a better knowledge of the underlying anatomical as well as electrical substrate. Novel developments, such as quadripolar left ventricular (LV) leads which enable electrical repositioning by individual pacing †
vector options, may help to increase the response to CRT in selected patients. For example, in patients with ischaemic cardiomyopathy insufficient LV pacing due to high thresholds within scar tissue are frequent problems. Due to the given anatomy of the target cardiac vein, implantation of such a multipolar lead allows electrical repositioning and these individual programmable pacing vectors may result in a better LV-pacing response.2 In all CRT patients, one of the major goals is an improvement of the ventricular electrical conduction delay with consecutive resynchronization of the LV and the right ventricle (RV). Nevertheless, little is known about the electroanatomical properties in these patients. Most of the currently available data were obtained during invasive
MS and FH contributed equally to this paper.
* Corresponding author. Tel: +43 6582 790 71200; fax: +43 6582 790 71290, E-mail address: [email protected] Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2014. For permissions please email: [email protected].
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
1 Institute of Electrical, Electronic and Bioengineering, University for Health Sciences, Medical Informatics and Technology (UMIT), 6060 Hall i. Tirol, Austria; 2Department of Internal Medicine III, Division of Cardiology, Medical University Innsbruck, 6020 Innsbruck, Austria; 3Department of Radiology I, Medical University Innsbruck, 6020 Innsbruck, Austria; and 4 Department of Cardiology, Ludwig Boltzmann Institute for Rehabilitation of Internal Diseases, SKA-RZ Saalfelden, Thorerstraße 26, 5760 Saalfelden, Austria
744
What’s new? † Non-invasive imaging of cardiac electrophysiology (NICE) enables assessment of ventricular electrical activation mapping at a sufficient spatio-temporal resolution. The discrimination of different pacing vectors when using a novel quadripolar left ventricular-lead is clinically feasible applying NICE. Moreover, NICE enables targeted LV-lead placement as well as optimization of electrical repositioning in case of CRT.
Methods
Table 1 Patient’s demographic and clinical data Age (years) Sex
56 M
QRS duration (ms)
183
LVEF (%) Co-morbidities
28 CAD, LBBB, HT, DM, DL
Medications
BB, ARB, Diur, ASA
M, male; LVEF, left ventricular ejection fraction; CAD, coronary artery disease; LBBB, left bundle branch block; HT, hypertension; DM, diabetes mellitus; DL, dyslipidemia; BB, betablocker; ARB, angiotensin receptor blocker; Diur, diuretics; ASA, acetylsalicylic acid.
Template Graphics Software, France) was adapted for this purpose. Liquid-filled anatomical markers on the patient’s torso were used to couple the MRI geometric data with the data obtained from ECG mapping (Figure 1). The electrode positions were transformed by a rigid body transformation (rotation matrix and displacement vector) using a magnetic digitizer (Fastrak; Polhemus, Inc.) as previously described.7 – 10
Non-invasive imaging of cardiac electrophysiology
Prior to CRT-D implantation, patient-specific anatomical data were obtained by magnetic resonance imaging (MRI) with a Magnetom Vision Plus 1.5-T scanner (Siemens). Ventricular end-diastolic geometry and torso geometry were assessed in an ECG-gated cine mode during breath hold as previously described.7 – 9
Non-invasive imaging of cardiac electrophysiology was applied using a bidomain model-based boundary element formulation to relate step-like local activation at the endocardial and epicardial source points to the simulated potentials at the electrode locations. A Wilson terminal defined the reference for measured and computed unipolar signals. For inverse computation of the ventricular activation sequence, the target beats were selected using an automated signal processing algorithm. The model-based computation of the activation sequence of a single beat can be described as a continuous approximation approach. For an assumed model activation sequence, the ECG was simulated and compared with the measured ECG. The activation sequence was then systematically changed to minimize the difference between the simulated and the measured data. The initial estimation of the activation sequence was computed applying the critical point theorem. As this imaging problem was ill-posed, an additional regularization term of the Tikhonov second-order type was considered in the optimization process. This regularization imposes the constraint that the neighbouring source points have similar activation times (smooth activation pattern). The coupled regularization strategy was applied to solve the non-linear optimization problem. In a previous study, the spatial resolution of NICE respectively the localization accuracy was 18 + 5 mm.6 Details of the NICE method have been published previously.7 – 10
Electrocardiogram mapping
Data acquisition
A high-resolution 65-lead electrode array was applied 15 h after CRT-D implantation. The ECG recordings were performed using a batterypowered (6 V) high-precision ECG amplifier (Mark-8 system, Biosemi V.O.F.) with a sampling rate of 2048 Hz (0.3 – 400 Hz band pass filter) and an alternating current resolution of 500 nV/bit.
Non-invasive imaging of cardiac electrophysiology was performed during native sinus rhythm (intrinsic conduction) as well as during single RV, left (LV), and biventricular (biV) pacing. For each pacing mode atrial synchronous pacing was performed. The AV interval was optimized echocardiographically and the VV interval was programmed at 0 ms. For LV and biV pacing modes, different stimulation vectors via the different electrodes of the LV lead (e.g. prox1-dist4) were tested (Figure 2). For all pacing modes and the different pacing vectors right and LV total activation duration, earliest septal, endocardial and epicardial breakthrough sites, and the endocardial/epicardial activation sequences were analysed.
Patient Non-invasive imaging of cardiac electrophysiology (NICE) was performed in a 61-year-old male patient with a complete left bundle branch block who underwent CRT-D implantation (Promote Quadra, St. Jude Medical) due to ischaemic heart failure. The RV lead was implanted in an apical position. For LV pacing a quadripolar LV lead (Quartet, St. Jude Medical) was placed in a posterolateral cardiac vein. During implantation adequate pacing and sensing properties were tested. At all pacing sites, a safety margin of .2× pacing threshold was accepted as adequate. The pacing output was initially set to 3.5 V at 0.5 ms impulse duration at the time of implantation due to safety reasons. The patient’s characterisitics are shown in Table 1.
Magnetic resonance imaging
Model building Magnetic resonance imaging data were used to construct a patientspecific volume conductor model of the patient’s torso, including compartments of different conductivity (heart, lung, blood mass, and chest surface). A commercial software package (AMIRA Developer, TGS
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
studies and primarily endocardial ventricular activation was measured, as information of epicardial activation is limited to a small area of the left ventricle which can be obtained only by mapping in the coronary sinus.3 There have been only a few small studies regarding non-invasive imaging of ventricular electroanatomical activation with limited data on the effects of simultaneous RV endocardial and LV epicardial pacing (CRT) on biventricular endo/epicardial electrical activation.4 – 6 This is the first study investigating the impact of a quadripolar LV-lead on epicardial and endocardial ventricular activation using a novel non-invasive mapping technique.
M. Seger et al.
745
Non-invasive imaging of ventricular activation
2 Voltage leads (mV)
1 0 –1 –2 –3 –4 –5 –6 0
100
200
300 Time (ms)
400
500
600
0
100
200
300 Time (ms)
400
500
600
14
10 8 6 4 2 0 –2
Figure 1 Left panel: volume conductor model (VCM) comprising lungs, ventricles, and atria with two MRI-slices (long axis scan). Right upper panel: high-resolution body surface 65-lead ECG during of intrinsic activation. Right lower panel: standard 12-lead ECG during intrinsic rhythm.
The breakthrough sites were identified as the sites on the ventricular epicardium/endocardium where depolarization first appeared. Ventricular total activation duration was defined as the interval from the earliest breakthrough to the latest observed electrical activation of the RV and LV. The root-mean-square QRS duration was measured during native sinus rhythm as well as during different pacing modes as previously described.11
Results Intrinsic endocardial and epicardial ventricular activation The earliest ventricular activation was located in the RV endocardium. The earliest septal breakthrough site was located in the midseptum. The right septum was activated from apical/mid-septal to basal. In accordance with complete left bundle branch block, the LV endocardial and epicardial activation was markedly delayed. The earliest left septal endocardial breakthrough site was located in a mid-septal position. The LV activation wavefront turned around the apex and spread radially across the inferior/anterior wall towards the lateral wall of the left ventricle. The latest ventricular activation was located epicardially in the lateral wall of the left ventricle (Table 2, Figure 3). There was a marked correlation between the delayed electrical activation of the left lateral wall as obtained by NICE and the MRI-derived LV dysynchrony. The LV lead was
Proximal 4 Mid 3
Mid 2 RV Coil
Distal 1
Figure 2 Ten possible pacing vectors are programmable using the three ring electrodes and distal tip of the Quartet lead and the RV coil (image courtesy of St Jude Medical—already published in Europace Online ISSN 1532-2092 - Print ISSN 1099-5129).
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
Voltage RMS (mV)
12
746
M. Seger et al.
Table 2 Effect of different pacing modes on ventricular electrical properties as derived from NICE Intrinsic
RV
LVdist1-prox4
LVmid3-prox4
biVdist1-prox4
biVmid2-prox4
............................................................................................................................................................................... RVAD (ms)
0– 81
0 –130
74– 213
101–208
0 –165
0–150
LVAD (ms) Breakthrough site
49– 183 Left lat. Epi 183
33– 190 Left apical endo 197
0–94 Septal apical epi
0–142 Left lateral epi
224
222
0 –137 Right apical endo; left septal apical epi 171
0–133 Right apical endo; left lateral epi 164
184
188
198
219
150
148
184
191
214
209
166
151
QRS duration 12-lead (ms) QRS duration RMS (ms) TAD (ms)
Intrinsic, intrinsic activation; RV, right ventricular pacing; LVdist1-prox4, single left ventricular pacing from distal 1 to proximal 4; LVmid3-prox4, single left ventricular pacing from mid 3 to proximal 4; biVdist1-prox4, biventricular pacing from distal 1 to proximal 4; biVmid2-prox4, biventricular pacing from mid 2 to proximal 4; RVAD, right ventricular activation duration in relation to earliest activation onset; LVAD, left ventricular activation duration in relation to earliest activation onset; epi, epicardial; endo, endocardial; RMS, root mean square; TAD, total ventricular activation duration.
Right ventricular pacing Right ventricular pacing showed no significant effects on septal and LV activation when compared with native rhythm. Right ventricular pacing resulted in a decrease of the time course of RV endocardial breakthrough, a reversal of RV endocardial and epicardial activation, and a prolongation of the total activation duration of the LV and the RV.
Left ventricular pacing Left ventricular pacing resulted in a reversal of LV endocardial and epicardial activation of the left ventricle when compared with intrinsic activation. Moreover, LV pacing resulted in a marked shortening of epicardial and endocardial LV activation when compared with RV pacing and intrinsic activation. There were no significant effects on septal activation when compared with native rhythm. In addition, there was a marked prolongation of the total biventricular activation duration (TAD; Table 2, Figure 3).
Biventricular pacing Biventricular pacing showed a marked prolongation of RV activation time when compared with intrinsic activation. During biV pacing, the onset of LV endocardial as well as epicardial activation was markedly shortened when compared with native sinus rhythm. Moreover, during biV pacing the ventricular activation sequence markedly changed. In concordance to LV pacing, the earliest activation was located epicardially in the LV lateral free wall. This was followed by an early endocardial right ventricular apical breakthrough (Table 2, Figure 3).
Pacing vector-dependent changes of left ventricular and biventricular stimulation There was no marked effect on right ventricular activation duration (RVAD) and LV activation duration (LVAD) during biV pacing with a LV dist1-prox4 and a LVmid2-prox4 vector (Table 2, Figure 3).
Single-LV pacing with a dist1-prox4 vector resulted in an LVAD of 94 ms, whereas single-LV pacing with a mid3-prox4 vector resulted in an LVAD of 142 ms. Also the LV epicardial activation sequence changed. During dist1-prox4 pacing, the earliest LV lateral wall activation was located close to the apex as well as close to the LV lateral basis with a distinct multisite LV activation pattern. During mid3-prox4 LV pacing, this distinct multisite LV pacing pattern disappeared with a fused area of mid-lateral epicardial LV activation (Table 2, Figure 3).
Discussion This is the first study using an NICE for visualization of (bi)ventricular electroanatomical activation in a CRT patient with a novel quadripolar LV lead. This multipolar LV lead allows programming of several different pacing vectors. The spatial resolution of NICE also facilitated visualizing the impact of the different LV and biV pacing vectors on ventricular electroanatomical activation properties. There are only few studies on non-invasive imaging of cardiac electroanatomical activation from surface ECG mapping and most of these studies are either limited with respect to clinical applicability (conducted in animals) or limited by an incomplete depiction of cardiac electrical activation (e.g. visualization of either epi- or pericardial activation only).4,12 Non-invasive imaging of cardiac electrophysiology enables single-beat visualization of both endocardial and epicardial electroanatomical activation with sufficient spatial resolution as previously described.6,9 The model building for NICE can be performed using MRI or computed tomography scans prior to CRT implantation to guide the LV lead placement and the same anatomical model can be used after implantation to determine the most appropriate pacing vector to guide electronical repositioning and optimize electroanatomical activation. In this patient, RV pacing resulted in an increase of total RV activation duration when compared with intrinsic activation (Table 2). This increase may be due to the unphysiological activation of the RV as the activation wavefront propagates by cell-to-cell coupling until connecting to the intrinsic conduction system and spreading via the Purkinje system. As previously described, left bundle branch block
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
placed within the corresponding area of delayed electroanatomical activation as indicated by NICE.
747
Non-invasive imaging of ventricular activation
Anterior-Posterior view
Cranial view
Posterior oblique view
Intrinsic
91.5 183
0
95 190
0
91.5 183
0
95 190
0
106.5 213
0
104 208
0
91.5
183
RV
0
95
190
LVdist1 -prox4
0 106.5 213
0
106.5 213
LVmid3 -prox4
0 104 208
0
104
208
0
82.5
165
0
76
biVdist1 -prox4
0
82.5 165
0 82.5 165
biVmid2 -prox4
0 75 150
0 75 150
150
Figure 3 Colour-coded electroanatomic activation map of the RV and RV during intrinsic activation (intrinsic), RV pacing, LV pacing with cathode to anode stimulation from distal 1 to proximal 4 (LVdist1-prox4) and from mid 3 to proximal 4 (LVmid3-prox4) as well as during biV stimulation pacing with cathode to anode stimulation from distal 1 to proximal 4 (biVdist1-prox4) and from mid 2 to proximal 4 (biVmid2-prox4). Red colour illustrates earliest activation, blue (purple) illustrates the area of late (latest) ventricular activation. Time scale numbers reflect the activation duration in milliseconds. Head icon indicates the point of view.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
0
748
and often result in suboptimal pacing thresholds with suboptimal pacing response.13 Therefore, the clinical concept of multipolar pacing leads is the availability of different stimulation configurations due to different programmable pacing vectors. Using this quadripolar LV lead, 10 different biV and LV pacing vector options are available.2 If there are some local capture problems, individual electrical repositioning may help to find acceptable pacing thresholds by programming a different pacing vector. Moreover, implementation of information about scar tissue (as obtained by late enhancement MRI) into the electroanatomical model building may lead to a further refinement of non-invasive electroanatomical imaging. Novel approaches such as cellular automaton models for inverse computation of the electroanatomical activation may be suitable developments for this purpose. Also other limitations such as a suboptimal target vein anatomy may be overcome by the option of individual electrical repositioning.
Conclusions Non-invasive imaging of cardiac electrophysiology facilitates the identification of the target area for LV lead placement as it allows visualization of LV endo/epicardial activation. The combination of electrical and anatomical data may help to increase responders to CRT as this allows an individualized ‘patient-tailored’ pacing therapy. Assessment of this information prior to implantation may help to refine the implantation strategy. It may also help to identify if the electrical and anatomical properties are suitable for an endovascular approach or (e.g. if there is no suitable cardiac vein) if an epicardial surgical approach may be more promising. Moreover, in patients with a quadripolar LV lead NICE may even facilitate post-procedural CRT optimization by identification of the best electrical repositioning option. Conflict of interest: none declared.
Funding This study was financially supported by the Austrian Science Fund (FWF, Fonds zur Fo¨rderung der wissenschaftlichen Forschung), Grant START Y144 and by Oesterreichische Nationalbank Jubila¨umsfond (Jubila¨umsfondsprojekt Nr. 11183). TB received an one year EP-fellowship grant from the European Heart Rhythm Association (European Society of Cardiology). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References 1. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino J et al. Results of the Predictors of Response to CRT (PROSPECT) trial. Circulation 2008;117: 2608 –16. 2. Asbach S, Hartmann M, Wengenmayer T, Graf E, Bode C, Biermann J. Vector selection of a quadripolar left ventricular pacing lead affects acute hemodynamic response to cardiac resynchronization therapy: a randomized cross-over trial. PloS ONE 2013; 8:e67235. 3. Lambiase PD, Rinaldi A, Hauck J, Mobb M, Elliott D, Mohammad S et al. Non-contact left ventricular endocardial mapping in cardiac resynchronisation therapy. Heart 2004;90:44– 51. 4. Ramanathan C, Ghanem RN, Jia P, Ryu K, Rudy Y. Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia. Nat Med 2004; 10:422 – 8. 5. Jia P, Ramanathan C, Ghanem RN, Ryu K, Varma N, Rudy Y. Electrocardiographic imaging of cardiac resynchronization therapy in heart failure: observation of variable electrophysiologic responses. Heart Rhythm 2006;3:296 –310.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
results in a typical electroanatomical activation pattern with latest activation of the epicardial lateral free wall.6 Also in this patient, singleleft and biV pacing resulted in a marked change of the LV activation pattern with resynchronization of the delayed left lateral wall. During single-LV pacing the RV activation duration was decreased when compared with biV pacing. However, the TAD was even shorter during biV pacing when compared with single-LV pacing (Table 2). This indicates that the RV activation timing relative to the LV activation is of particular relevance for the TAD. This effect may also explain the varying impact of the different pacing modes on haemodynamic response. In this patient, epicardial pacing close to the area of latest LV activation as determined by NICE and simultaneous endocardial pacing at the right ventricular apex resulted in the shortest TAD. A limitation of this study is that the results are derived from only a single patient undergoing multipolar LV stimulation. Nevertheless, these results are in accordance with data from another study from our working group, showing similar effects of CRT on (bi)ventricular activation in patients with standard LV stimulation (no quadripolar leads).6 A major limitation of CRT is the high rate of non-responders. There have been several studies to find suitable parameters to predict the response to CRT. Despite some interesting results of smaller cohort studies, larger clinical trials failed to confirm these initially promising predictors such as some sophisticated echocardiographic parameters.1 The main goal of CRT is preceding the delayed LV by stimulation at the left lateral epicardium in patients with left bundle branch block. Thus, optimal placement of the LV lead is crucial. Stimulation as close as possible to the site of latest LV activation should be aimed at.13 Two recent trials nicely demonstrated that targeted LV lead placement close to sites of latest LV mechanical activation resulted in best CRT response.14,15 Both the TARGET and the STARTER trial used echocardiographic parameters for the determination of latest LV activation.14,15 Given the data from the PROSPECT trial, echocardiographic parameters of ventricular dyssynchrony show only a modest sensitivity and specificity for identification of responders to CRT.1 This may be due to the relatively high inter/intraobserver variabilities of these highly sophisticated analyses. Hence, identification of the area of latest LV activation by echocardiography may be difficult in clinical practice. Non-invasive imaging of cardiac electrophysiology may help to overcome these difficulties in the identification of the area of latest ventricular activation as the technical and interpretative factors are well defined. Up to now parameters obtained by standard 12-lead ECG mapping, such as QRS duration and left bundle branch block pattern, still seem to be one of the best predictors for identification of potential responders to CRT. Meta-analysis showed that these ECG criteria were still the strongest predictors for response to CRT. Therefore, the European Society of Cardiology revised in 2012 the guidelines for the management of heart failure and put even more focus on standard ECG criteria.16 Non-invasive imaging of cardiac electrophysiology may have further potential for improving responder identification rates as it works by fusing electrical data obtained from high-resolution body surface ECG with an individual model of the patient’s cardiac anatomy. In patients with severe heart failure, differences in local capture thresholds due to fibrotic changes in the myocardium are common
M. Seger et al.
Non-invasive imaging of ventricular activation
6. Berger T, Pfeifer B, Hanser FF, Hintringer F, Fischer G, Netzer M et al. Single-beat noninvasive imaging of ventricular endocardial and epicardial activation in patients undergoing CRT. PloS ONE 2011;6:e16255. 7. Modre R, Tilg B, Fischer G, Hanser F, Messnarz B, Seger M et al. Atrial noninvasive activation mapping of paced rhythm data. J Cardiovasc Electrophysiol 2003;14: 712 – 9. 8. Tilg B, Fischer G, Modre R, Hanser F, Messnarz B, Schocke M et al. Model-based imaging of cardiac electrical excitation in humans. IEEE Trans Med Imaging 2002;21: 1031– 9. 9. Berger T, Fischer G, Pfeifer B, Modre R, Hanser F, Trieb T et al. Single-beat noninvasive imaging of cardiac electrophysiology of ventricular pre-excitation. J Am Coll Cardiol 2006;48:2045 –52. 10. Fischer G, Hanser F, Pfeifer B, Seger M, Hintermuller C, Modre R et al. A signal processing pipeline for noninvasive imaging of ventricular preexcitation. Methods Inf Med 2005;44:508 –15. 11. Berger T, Hanser F, Hintringer F, Poelzl G, Fischer G, Modre R et al. Effects of cardiac resynchronization therapy on ventricular repolarization in patients with congestive heart failure. J Cardiovasc Electrophysiol 2005;16:611 –7.
749 12. Han C, Liu Z, Zhang X, Pogwizd S, He B. Noninvasive three-dimensional cardiac activation imaging from body surface potential maps: a computational and experimental study on a rabbit model. IEEE Trans Med Imaging 2008;27: 1622 –30. 13. Liu J, Adelstein E, Saba S. Targeting left ventricular lead placement to improve cardiac resynchronization therapy outcomes. Curr Cardiol Rep 2013;15:390. 14. Khan FZ, Virdee MS, Palmer CR, Pugh PJ, O’Halloran D, Elsik M et al. Targeted left ventricular lead placement to guide cardiac resynchronization therapy: the TARGET study: a randomized, controlled trial. J Am Coll Cardiol 2012;59: 1509 –18. 15. Saba S, Marek J, Schwartzman D, Jain S, Adelstein E, White P et al. Echocardiographyguided left ventricular lead placement for cardiac resynchronization therapy: results of the Speckle tracking assisted resynchronization therapy for electrode region trial. Circ Heart Fail 2013;6:427 –34. 16. Sipahi I, Carrigan TP, Rowland DY, Stambler BS, Fang JC. Impact of QRS duration on clinical event reduction with cardiac resynchronization therapy: meta-analysis of randomized controlled trials. Arch Intern Med 2011;171:1454 –62.
Downloaded from http://europace.oxfordjournals.org/ by guest on November 14, 2015
