Validation of electromechanical wave imaging in a caninemodel during pacing and sinus rhythmJulien Grondin, PhD,⁎ Alexandre Costet, MS,⁎ Ethan Bunting, MS,⁎ Alok Gambhir, MD, PhD,†

Hasan Garan, MD,† Elaine Wan, MD,† Elisa E. Konofagou, PhD⁎‡

From the ⁎Department of Biomedical Engineering, †Department of Medicine, College of Physicians andSurgeons, and ‡Department of Radiology, Columbia University, New York, New York.

BACKGROUND Accurate determination of regional areas ofarrhythmic triggers is of key interest to diagnose arrhythmias andoptimize their treatment. Electromechanical wave imaging (EWI) isan ultrasound technique that can image the transient deformationin the myocardium after electrical activation and therefore has thepotential to detect and characterize location of triggers ofarrhythmias.

OBJECTIVES The objectives of this study were to investigate therelationship between the electromechanical and the electricalactivation of the left ventricular (LV) endocardial surface duringepicardial and endocardial pacing and during sinus rhythm as wellas to map the distribution of electromechanical delays.

METHODS In this study, 6 canines were investigated. Two externalelectrodes were sutured onto the epicardial surface of the LV.A 64-electrode basket catheter was inserted through the apex ofthe LV. Ultrasound channel data were acquired at 2000 frames/sduring epicardial and endocardial pacing and during sinusrhythm. Electromechanical and electrical activation maps were

This work was supported by the National Institutes of Health (grantnos. R01-EB006042 and R01-HL114358). Present address of Dr AlokGambhir: Northeast Georgia Heart Center, Gainesville, Georgia. Addressreprint requests and correspondence:Dr Elisa E. Konofagou, Departmentof Biomedical Engineering, Columbia University, 1210 AmsterdamAvenue, ET 351, MC 8904, New York, NY 10027. E-mail address:ek2191@columbia.edu.

1547-5271/$-see front matter B 2016 Heart Rhythm Society. All rights reserved.

synchronously obtained from the ultrasound data and the basketcatheter, respectively.

RESULTS The mean correlation coefficient between electrome-chanical and electrical activation was 0.81 for epicardial anteriorpacing, 0.79 for epicardial lateral pacing, 0.69 for endocardialpacing, and 0.56 for sinus rhythm.

CONCLUSION The electromechanical activation sequence deter-mined by EWI follows the electrical activation sequence and morespecifically in the case of pacing. This finding is of key interest inthe role that EWI can play in the detection of the anatomical sourceof arrhythmias and the planning of pacing therapies such ascardiovascular resynchronization therapy.

KEYWORDS Arrhythmias; Electromechanical activation; Electricalmapping; Ultrasound; Noninvasive imaging; Pacing; Canine model

(Heart Rhythm 2016;13:2221–2227) I 2016 Heart Rhythm Society.All rights reserved.

IntroductionCardiac arrhythmia is experienced by approximately 5.8million people per year in the United Sates.1 Electrocardiog-raphy (ECG) is widely used to noninvasively detect anddiagnose cardiac arrhythmia. However, ECG has a lowspatial resolution and cannot always accurately determinethe location of the regions triggering arrhythmia. Surgicalablation of arrhythmias involves catheterization of the heartand detailed endocardial mapping to determine the source ofthe arrhythmia, which may be time-consuming, operator-dependent, and limited by accessibility of the tissue. Oncethe source of the arrhythmia is found, radiofrequency (RF)ablation can be used to terminate the arrhythmia. However,the procedure is highly dependent on the ability of theclinician to accurately determine the location of the

arrhythmogenic source. A time-effective and noninvasivemethod to identify the candidate regions for ablation couldsignificantly decrease the duration of the procedure. ECG hasbeen developed to map the electrical activity of the heart athigh spatial resolution noninvasively.2 However, this tech-nique requires special hardware as well as computedtomography and/or magnetic resonance imaging scans.

Electromechanical wave imaging (EWI) is an ultrasound-based technique that can noninvasively map the minutedeformation of the myocardium resulting from local electricalactivation at high temporal resolution. The terminology EWI ispreferred to mechanical wave imaging to avoid possibleconfusion with other mechanical waves propagating in themyocardium resulting from mitral or aortic valve closure3 orfrom blood-wall interaction and to be distinguished from timeto peak strain.4 Previous feasibility studies have shown thatEWI is capable of transmurally mapping the electromechan-ical activation sequence during sinus rhythm and pacing,localizing pacing sites,5,6 characterizing arrhythmias in car-diac resynchronization therapy (CRT) and in patients withatrial flutter,7 and mapping the electromechanical activationduring reentrant and focal arrhythmias.8 It has been reported

http://dx.doi.org/10.1016/j.hrthm.2016.08.010

Heart Rhythm, Vol 13, No 11, November 20162222

that 20%–40% of patients do not respond to CRT.9 Severalstudies reported a strong influence of the pacing lead place-ment on CRT response.10 Therefore, EWI could prove to be ofkey interest in providing assistance in optimizing lead place-ment by mapping the electromechanical activity of the heartduring the procedure.

In the myocardium, local electrical activation is followed bya localized minute muscle contraction after a delay in the orderof tens of milliseconds.11 It is hypothesized that EWI can detectthe onset of contraction because of its high axial and temporalresolution. This technique is different from the onset of short-ening obtained from the strain curves with speckle trackingechocardiography (STE) or tissue Doppler imaging.12,13 ForSTE, envelope-detected signals are used typically with a framerate of 50–70 Hz,14 although for both STE and tissue Dopplerimaging higher frame rates can be obtained at the cost of areduced field of view. In this study, EWI is performed withinterframe strain obtained from RF signals that preserve thephase information and at a frame rate of 2000 Hzwith a full fieldof view, which allows for capturing the onset of contraction inresponse to electrical activation. Previous studies from our groupreported a good correlation between electromechanical activa-tion determined by EWI and electrical activation measured byepicardial or endocardial electrodes sutured on the heart.5,6

These previous initial feasibility studies, however, were limitedto a low number of subjects and limited recording sites andimaging planes. A dedicated validation study with differentpacing configurations and a higher number of subjects, record-ing sites, and imaging planes has yet to be carried out. Multipleultrasound imaging planes and higher-density electrical mappingof the endocardial surface allow for better rendering of theelectromechanical and electrical activation of the endocardialsurface, respectively. This would also allow for mapping of thedistribution of electromechanical delays. The objectives of thisstudy were to investigate the relationship between the electro-mechanical and the electrical activation of the endocardialsurface of the canine left ventricle (LV) during epicardial andendocardial pacing and during sinus rhythm as well as to mapthe distribution of electromechanical delays.

MethodsAnimal preparationThe study was approved by the Institutional Animal Care andUse Committee of Columbia University and conforms to theGuide for the Care and Use of Laboratory Animals.Six canines (weighing 24.1 ± 0.4 kg) were premedicated withdiazepam (0.5–1 mg/kg) injected intravenously and subse-quently anesthetized with an intravenous injection of propofol(2–5 mg/kg). The canines were mechanically ventilated with arate- and volume-regulated ventilator on a mixture of oxygenand titrated 0.5%–5% isoflurane. Lidocaine (50 µg/(kg·min))was injected intravenously throughout the procedure and untilthe end of the experiment to minimize the occurrence ofventricular arrhythmia. A left lateral thoracotomy was performedwith electrocautery, and 1 rib was removed to expose the heart.Two external bipolar electrodes were sutured onto the epicardial

surface of the LV in the vicinity of the anterior and lateral regions(Figure 1). A 38-mm diameter 64-electrode basket catheter(Constellation, Boston Scientific, Marlborough, MA) with 8flexible splines and 8 equally spaced (3-mm) electrodes perspline was introduced into the LV through the apex. The basketcatheter was carefully positioned inside the LV to optimizeelectrode contact with the myocardium. A suture was attachedbeforehand to 1 spline of the basket catheter to serve as alandmark in order to determine the location of the splines usingB-mode imaging. The thoracic cavity was filled with saline foracoustic impedance matching during ultrasound data acquisition.

Pacing protocolThe epicardial electrodes were connected to a data acquis-ition system (DAQ 1) (NI USB-6259, National Instruments,Austin, TX), which could send a 10-V amplitude, 2-ms pulsewidth, and 500-ms cycle length pacing signal. The pacingsignal was also acquired by the data acquisition system.

Endocardial pacing was also achieved by pacing from anyelectrode on the basket catheter using an Arduino boardequipped with a microcontroller (Mega 2560, Arduino, Somer-ville, MA). An electrode with good contact with the endocardialwall was used for pacing. Only 1 electrode at a time was usedfor pacing. Epicardial or endocardial pacing was sustained untilall echocardiographic views were acquired.

Ultrasound data acquisitionUltrasound data were acquired either during sinus rhythm orduring epicardial or endocardial pacing. Four apical views wereacquired: the standard 4-, 2-, and 3-chamber views as well asthe 3.5-chamber view, which is located between the 4- and the2-chamber views. A 2.5-MHz center frequency phased array(P4-2, ATL/Philips, Andover, MA) operated by a Verasonicsultrasound system (V-1, Verasonics, Kirkland, WA) was usedto acquire RF channel data at 2000 frames/s using divergingwave imaging. The ECG signal was acquired synchronouslywith ultrasound data using an ECG unit (77804A, HP, PaloAlto, CA) connected to another data acquisition system (DAQ2) (NI USB-6210, National Instruments) and triggered by theVerasonics ultrasound system. A standard delay-and-summethod was used to reconstruct the entire image for each singlediverging beam transmitted. Conventional B-mode imageswere also acquired at 30 frames/s to assist myocardialsegmentation. Axial displacements in the myocardium wereestimated between consecutive RF frames using normalized 1-dimensional (1D) cross-correlation15 with a window length of6.2 mm and an overlap of 90%. Interframe axial strains wereestimated from the interframe axial displacements using a least-squares estimator implemented with a Savitzky-Golay filterwith a 5-mm kernel. Positive strains in the longitudinaldirection indicate longitudinal lengthening, whereas negativestrains indicate longitudinal shortening. During ventricularcontraction, there is longitudinal shortening. Therefore, theelectromechanical activation time point at a given location wasdefined when the temporal curve of interframe axial strainchanged sign, that is, as the first zero crossing,5 after the onset

electrodes

LV

Opening of the chest

Inser�on of the basket catheter inside the LV

Acquisi�on of ultrasound RF data in 4 different apical views

Suturing of electrodes onto the LV epicardial surface

Pacing from epicardial electrodes

Recording of endocardial electric poten�als

Pacing from endocardial

electrodes (basket)

Interframe axial displacements es�ma�on

Interframe axial strains es�ma�on

Evalua�on of the first maximum of each electrogram

Determina�on of the first zero-crossing of temporal strains

varia�on

Interpola�on of the electrical ac�va�on �mes onto an

ellipsoid

Interpola�on of the electromechanical ac�va�on

�mes onto an ellipsoid

Electrical ac�va�on map Electromechanical ac�va�on map

Ac�va�on �me (ms)

Ac�va�on �me (ms)

0

20

40

406080100120140

Figure 1 Flowchart of the methodology of synchronous acquisition of the electrical and electromechanical activation maps during epicardial and endocardialpacing and during sinus rhythm in the left ventricle of a canine. Regions in blue indicate early activation, and regions in red indicate late activation. LV ¼ leftventricle/ventricular; RF ¼ radiofrequency.

2223Grondin et al Electromechanical Wave Imaging in Canines

of electrical activation defined below. Isochrones of electro-mechanical activation were obtained for each echocardio-graphic view. Electromechanical activation times wereobtained on the endocardial surface in the vicinity of eachelectrode of the basket catheter. Conventional B-mode imageswere used to assist for the location of the electrodes that couldbe seen because of their high echogenicity (Figure 2). For eachultrasound view, we tried to angle the ultrasound transducer sothat the electrodes on the splines were visible on the ultrasoundimage. Manual selection was performed on the endocardialsurface of the LV at mid-distance between each pair ofelectrodes to determine electromechanical activation times tobe correlated with electrical activation times. Regions for whichno clear zero crossing was observed or for which the catheterwas pushing on the myocardium during contraction andtherefore giving nonphysiological deformation were excluded.

RV

Figure 2 B-mode image of the apical 4-chamber view with the basketcatheter inserted in the left ventricle. The electrodes of the basket cathetercan be distinguished as bright spots because of their high echogenicity andare indicated by blue arrows at the septal and lateral walls. LV ¼ leftventricle; RV ¼ right ventricle.

Electrical data acquisitionThe basket catheter was connected to 4 custom-builtacquisition boards (2 splines per acquisition board). Bipolarelectrograms were obtained from adjacent electrodes on thesame spline. Bipolar signals were preferred to unipolarsignals since they reduce far-field effects.16 Fifty-six bipolarelectrograms (7 bipolar electrograms per spline) weremultiplexed and output to DAQ 1, which was controlledusing MATLAB (MathWorks, Natick, MA). Each bipolarelectrogram was sampled at 1 kHz. The trigger of theVerasonics ultrasound system was also connected to DAQ

1 in order to acquire electrograms synchronously with theultrasound data. Electrograms were high-pass filtered with acutoff frequency of 0.5 Hz and then low-pass filtered with acutoff frequency of 100 Hz. The maximum peak of each ofthe 56 electrograms regardless of polarity was determinedusing parabolic interpolation and defined the electricalactivation times that were associated with the depolarizationof the myocardium. Some electrograms were excludedbecause of poor electrode contact with the endocardium or

Heart Rhythm, Vol 13, No 11, November 20162224

because of damaged electrodes and were thus discarded.Peaks within the range of ±0.5 mV were discarded.

Statistical analysisRegions where the electromechanical or the electricalactivation time could not be determined were linearlyinterpolated. Linear regression was performed betweenelectromechanical and electrical activation times; the corre-lation coefficient, the associated P value, the slope, and theintercept were obtained.

Ellipsoidal rendering of activation timeElectromechanical and electrical activation times on the endo-cardial surface of the LV were obtained for 4 echocardiographicviews. Electromechanical and electrical activation times werethen linearly interpolated onto an ellipsoid of the same geometryas the basket catheter to represent the isochrones on theendocardial surface. It was assumed that the spacing betweenneighboring splines remained identical and constant afterinsertion in the LV. The propagation of the electromechanicaland of the electrical activation was illustrated by depictingactivated regions as a function of time on the ellipsoid.

ResultsElectromechanical and electrical activation maps wereobtained in the LV of 6 canines during epicardial andendocardial pacing and during sinus rhythm.

Epicardial pacingElectrical and electromechanical activation times as well as thecorresponding linear regression during epicardial pacing aroundthe anterior region of the LV are shown in Figure 3A for 1canine. Early activated regions are displayed in blue, whereas lateactivated regions are displayed in red. Both the electrical andelectromechanical activation maps show that the first activatedregions are located in the anterior region and the last activatedregions are located approximately in the posterior region.

Electrical and electromechanical activation times as wellas the corresponding linear regressions were also performedduring epicardial pacing around the lateral region of the LVand are shown in Figure 3B for 1 canine. The electricaland electromechanical activation maps show that the firstactivated regions are located approximately in the lateralregion whereas the septal region is activated last. Goodagreement was observed between electrical and electro-mechanical activation times for both anterior and lateralepicardial pacing. The correlation coefficient, slope, andintercept for all 6 canines are listed in Table 1. The meancorrelation coefficient across all canines during epicardiallateral pacing was 0.79, the mean slope was 2.2, and themean intercept was 29 ms. During epicardial anterior pacing,the mean correlation coefficient was 0.81, the mean slopewas 1.8, and the mean intercept was 32 ms.

Endocardial pacingElectrical and electromechanical activation times as well as thecorresponding linear regressions were also performed duringendocardial pacing from 1 electrode on the basket catheter andare shown in Figure 3C for 1 canine. Electrical activation mapsindicate that activation starts in the paced region, subsequentlypropagating in the opposite direction. The location of the pacedregion differed across canines because of different choices ofelectrodes used to pace the ventricle. The electrical and electro-mechanical activation maps show a similar activation pattern.The mean correlation coefficient across all canines was 0.69, themean slope was 1.4, and the mean intercept was 41 ms.

Sinus rhythmElectrical and electromechanical activation times as well as thecorresponding linear regressions during baseline acquisition(sinus rhythm) are shown in Figure 3D for 1 canine. The firstactivated regions are located approximately in the septal region,whereas the last activated regions are located approximately inthe posterior-lateral region. Also, the apical region is activatedbefore the basal region. The electrical and electromechanicalactivation maps show a similar activation pattern. The meancorrelation coefficient across all canines was 0.56, the meanslope was 2.2, and the mean intercept was 39 ms.

Electromechanical delayThe distribution of electromechanical delays in the LV wasperformed during epicardial and endocardial pacing and duringbaseline acquisitions. Figure 4 shows the distribution of electro-mechanical delays during epicardial anterior, epicardial lateral,and endocardial pacing as well as during sinus rhythm for thesame canine described in Figure 3. During epicardial andendocardial pacing, the electromechanical delay is shorter nearthe pacing origin while it increases with distance. During sinusrhythm, the apical region presents shorter electromechanicaldelays whereas the basal region presents longer electromechan-ical delays. The regions of shortest and longest electromechanicaldelays for all 6 canines are indicated in Table 2.

DiscussionNoninvasive and efficient mapping of the electrical activity ofthe heart can improve the identification of the regionsresponsible for arrhythmia, which is of paramount interest tooptimize their associated therapies. EWI is a high temporalresolution, ultrasound-based technique that can map the electro-mechanical activity of the heart. The aims of this study were toshow how the electromechanical activation maps obtained withEWI can be associated with the electrical activation and to showthat EWI can map the electromechanical activation during sinusrhythm and during epicardial and endocardial pacing.

During epicardial and endocardial pacing, the electro-mechanical and electrical activations in the LV followeda similar pattern. It was also shown that the paced region(eg, anterior) had early activation times, whereas regionsfarther away from the pacing site (eg, posterior) wereelectrically and electromechanically activated later, which

Sept

Post

Lat

Ac�va�on �me (ms)

Electrical ac�va�on EWI ac�va�on

0

20

40

20406080100

Ac�va�on �me (ms)

Epicardial anterior pacing Epicardial lateral pacing

Post

Lat

Ac�va�on �me (ms)

Electrical ac�va�onAc�va�on �me (ms)

EWI ac�va�on

0

20

40

406080100120140

Sept

Post

Lat

Ac�va�on �me (ms)

Electrical ac�va�on EWI ac�va�onAc�va�on �me (ms)

0

20

40

60

80

100

120

140

Endocardial pacing

Sept

Post

Lat

Ac�va�on �me (ms)

Electrical ac�va�onAc�va�on �me (ms)

EWI ac�va�on

0

10

20

40

60

80

Sinus rhythm

A B

C D

Figure 3 Electrical and electromechanical activation times in a canine left ventricle during epicardial anterior (A), epicardial lateral (B), and endocardial (C)pacing and during sinus rhythm (D). The electrical and electromechanical maps show a similar pattern in each case. Linear regression between electromechanicaland electrical activation times is performed. The lightening cartoon indicates the paced region. EWI ¼ electromechanical imaging; Lat ¼ lateral; Post ¼posterior; R ¼ correlation coefficient; Sept ¼ septum.

2225Grondin et al Electromechanical Wave Imaging in Canines

is consistent with previous studies.17–19 The correlationcoefficient across all canines for epicardial pacing rangedfrom 0.47 to 0.94 and had a mean value of 0.79, the meanslope was 2.0, and the mean intercept was 30 ms. EWI is thusshown to be a good surrogate for electrical mapping,confirming reports of other studies17,19–21 although usingdifferent definitions of electrical and mechanical activation.For endocardial pacing, the correlation coefficient rangedfrom 0.61 to 0.84 and had a mean value of 0.69, a mean slopeof 1.4, and a mean intercept of 41 ms. Electrograms duringendocardial pacing were not as sharp as those duringepicardial pacing. Therefore, the activation time derived

from electrograms may be less accurate, which can partlyexplain the lower correlation coefficient obtained duringendocardial pacing compared with the one obtained duringepicardial pacing. The electromechanical delay distributionduring epicardial and endocardial pacing was not homoge-neous. As mentioned in a previous study, the distribution ofelectromechanical delays depends on the sequence ofactivation. The early activated regions have a shorter electro-mechanical delay than do the late activated regions.17,18,22

For baseline acquisitions, electromechanical and electricalactivation maps also showed a similar pattern. During sinusrhythm, the activation sequence was different from that during

Table 1 Linear regression results of electromechanical and electrical activation times in 6 canines during epicardial and endocardial pacingand during sinus rhythm

Variable Dog 1 Dog 2 Dog 3 Dog 4 Dog 5 Dog 6 Mean value

Epicardial lateral pacing R ¼ 0.87*** R ¼ 0.80*** R ¼ 0.91*** R ¼ 0.94*** R ¼ 0.47*** R ¼ 0.89*** R ¼ 0.79s ¼ 3.6 s ¼ 2.5 s ¼ 1.6 s ¼ 1.9 s ¼ 1.9 s ¼ 1.5 s ¼ 2.2i ¼ 19 ms i ¼ 12 ms i ¼ 24 ms i ¼ 47 ms i ¼ 39 ms i ¼ 34 ms i ¼ 29 ms

Epicardial anterior pacing NA R ¼ 0.76*** R ¼ 0.63*** R ¼ 0.89*** R ¼ 0.88*** R ¼ 0.80*** R ¼ 0.81s ¼ 1.5 s ¼ 1.7 s ¼ 2.3 s ¼ 2.3 s ¼ 1.4 s ¼ 1.8i ¼ 54 ms i ¼ 31 ms i ¼ 16 ms i ¼ 13 ms i ¼ 44 ms i ¼ 32 ms

Endocardial pacing R ¼ 0.84*** R ¼ 0.74*** R ¼ 0.63*** NA R ¼ 0.61*** R ¼ 0.61*** R ¼ 0.69s ¼ 1.9 s ¼ 1.8 s ¼ 1.6 s ¼ 0.9 s ¼ 0.9 s ¼ 1.4i ¼ 47 ms i ¼ 22 ms i ¼ 33 ms i ¼ 52 ms i ¼ 50 ms i ¼ 41 ms

Sinus rhythm R ¼ 0.61*** R ¼ 0.43*** R ¼ 0.76*** R ¼ 0.78*** R ¼ 0.39** R ¼ 0.41** R ¼ 0.56s ¼ 2.7 s ¼ 1.2 s ¼ 2.4 s ¼ 2.1 s ¼ 2.7 s ¼ 1.9 s ¼ 2.2i ¼ 34 ms i ¼ 23 ms i ¼ 60 ms i ¼ 40 ms i ¼ 33 ms i ¼ 46 ms i ¼ 39 ms

Values indicate the correlation coefficient (R), the slope (s), and the intercept (i) of the regression. In 2 cases, data were not available (NA).⁎⁎Po.01⁎⁎⁎Po0.001.

Heart Rhythm, Vol 13, No 11, November 20162226

pacing. The first activated regions were posterior-septal, septal,and anterior-septal, and the last activated regions were anterior-lateral, lateral, and posterior-lateral. Early activation in the septalregion of the LV has been previously reported.18,23 In addition,the apical region was activated before the basal region. This apexto base propagation is also consistent with previous stud-ies.18,23,24 The correlation coefficient across all canines rangedfrom 0.39 to 0.78 and had a mean value of 0.56. This indicatesthat the electromechanical activity depicted by EWI is associatedwith the electrical activity obtained by endocardial electricalmapping. The mean slope of the linear regression of electro-mechanical activation times vs electrical activation times was 2.2,which is41 and indicates that electrical activation was followedby electromechanical activation. More specifically, depolariza-tion was followed by the contraction with an increasing delay.

Sept

Post

Lat

EMD(ms)

20

40

60

80

Epicardial anterior pacing Epicardial lateral pacing

Sept

Post

Lat

40

60

80EMD(ms)

Sept

Post

Lat

Endocardial pacing

40

60

80

EMD(ms)Sept

Post

Lat

3040506070

EMD(ms)

Sinus rhythm

A B

C D

Figure 4 Electromechanical delays in a canine during epicardial anterior(A), epicardial lateral (B), and endocardial (C) pacing and during sinusrhythm (D). The electromechanical delays are shorter near the first activatedregions and longer for the last activated regions. The lightening cartoonindicates the paced region. EMD ¼ electromechanical delay; Lat ¼ lateral;Post ¼ posterior; Sept ¼ septum.

The mean intercept was 39 ms, indicating that the electro-mechanical delay in the first activated region in the ventricle was39 ms on average. The map of electromechanical delays wasobtained by representing the difference between electromechan-ical and electrical activation times. The distribution of electro-mechanical delays in the LV during sinus rhythm was differentfrom the paced distribution. The apical region had shorterelectromechanical delays than did the basal region. This obser-vation is consistent with a previous study by Gurev et al,18 whichis a simulation study using an electromechanical model.

Although similar electromechanical and electrical activationmaps were found, this study presents several limitations thatmay affect the quantitative aspects reported herein. Theorientation of the splines of the basket catheter inside the LVwas determined by the specular reflection from a suture attachedto one of the splines using conventional B-mode imaging.Although the reflection from the suture was not always easilyidentifiable, the activation maps obtained from the basketcatheter by pacing from epicardial electrodes sutured at a knownlocation were in good agreement with the expected earliestactivated spline. In addition, although significant efforts to align

Table 2 Regions of shortest and longest electromechanicaldelays in 6 canines during epicardial and endocardial pacing andduring sinus rhythm

Variable Dog 1 Dog 2 Dog 3 Dog 4 Dog 5 Dog 6

Epicardial lateral pacingShortest A-L P-L A-L P-L L LLongest P-S P-S P-S A-S S P-S

Epicardial anterior pacingShortest NA A-L A A-L A ALongest P-S P-L P-S P-L P-S

Endocardial pacingShortest P-S L P-L NA A-L PLongest A A A-L S S

Sinus rhythmShortest A-S S P-S S P-S SLongest P-L A A-L P L L

In 2 cases, data were not available (NA).A ¼ anterior; L ¼ lateral; P ¼ posterior; S ¼ septal.

2227Grondin et al Electromechanical Wave Imaging in Canines

the splines of the basket catheter with the echocardiographicviews were made, the alignment was imperfect. Therefore, theanatomical locations corresponding to electrical and electro-mechanical activation times used for correlation analysis maynot optimally match. The activation propagates throughout theentire volume of the myocardium. However, the electromechan-ical maps were obtained from 2D echocardiographic views andthe electrical measurements are obtained from 56 bipolarendocardial electrograms.

As shown in Table 1, linear regression could not beperformed in 2 cases: for one canine, during epicardialpacing, ultrasound data were not saved properly and couldnot be processed; and for another canine, endocardial pacingwas not performed.

For this basket catheter, the electrode spacing was 3mm. Considering a conduction velocity of 0.4–0.7 m/s atthe endocardium,25 the time delay between 2 neighboringelectrodes would be 4.3–7.5 ms for a propagation alongthe axis defined by the 2 electrodes. In that case, 2temporally separated depolarization peaks can bedetected on bipolar electrograms and this can yield anuncertainty in the order of the time delay between those 2peaks for the determination of electrical activation times.

Electromechanical activation times were defined at the zerocrossings on the strain curves. The zero crossings indicate thetime when the myocardium locally transitions from relaxation tocontraction. However, some regions may start to contract laterthan other regions because of distinct loading conditions orprestretched regions.22 Therefore, although a contraction forcemay start to develop locally after electrical excitation, theshortening of muscles may take additional time in prestretchedmyocardial regions. This may partly be explained by thedistribution of electromechanical delays and results in a slopeof41 for the regression analysis between electromechanical andelectrical activation times.

ConclusionThis study shows that the electromechanical activationsequence determined by EWI follows the electricalactivation sequence during epicardial and endocardial pacingand during sinus rhythm. A higher correlation betweenelectrical and electromechanical activation times wasobtained during pacing than during sinus rhythm. Thesefindings indicate that EWI has great potential to noninva-sively characterize and identify regions responsible forarrhythmias as well as to provide assistance during cardiacablation and pacing therapy. Future studies will include 3Dmapping of the electromechanical activation sequence.

References1. Tang DH, Gilligan AM, Romero K. Economic burden and disparities in

healthcare resource use among adult patients with cardiac arrhythmia. ApplHealth Econ Health Policy 2014;12:59–71.

2. Wang Y, Cuculich PS, Zhang J, Desouza KA, Vijayakumar R, Chen J, FaddisMN, Lindsay BD, Smith TW, Rudy Y. Noninvasive electroanatomic mapping ofhuman ventricular arrhythmias with electrocardiographic imaging. Sci TranslMed 2011;3:98ra84.

3. Remme EW, Lyseggen E, Helle-Valle T, Opdahl A, Pettersen E, Vartdal T,Ragnarsson A, Ljosland M, Ihlen H, Edvardsen T, Smiseth OA. Mechanisms ofpreejection and postejection velocity spikes in left ventricular myocardium: interactionbetween wall deformation and valve events. Circulation 2008;118:373–380.

4. Kanzaki H, Bazaz R, Schwartzman D, Dohi K, Sade LE, Gorcsan J III. Amechanism for immediate reduction in mitral regurgitation after cardiacresynchronization therapy: insights from mechanical activation strain mapping.J Am Coll Cardiol 2004;44:1619–1625.

5. Provost J, Lee WN, Fujikura K, Konofagou EE. Imaging the electromechanicalactivity of the heart in vivo. Proc Natl Acad Sci U S A 2011;108:8565–8570.

6. Provost J, Nguyen VT, Legrand D, Okrasinski S, Costet A, Gambhir A, Garan H,Konofagou EE. Electromechanical wave imaging for arrhythmias. Phys Med Biol2011;56:L1–L11.

7. Provost J, Gambhir A, Vest J, Garan H, Konofagou EE. A clinical feasibilitystudy of atrial and ventricular electromechanical wave imaging. Heart Rhythm2013;10:856–862.

8. Provost J, Costet A, Wan E, Gambhir A, Whang W, Garan H, Konofagou EE.Assessing the atrial electromechanical coupling during atrial focal tachycardia,flutter, and fibrillation using electromechanical wave imaging in humans. ComputBiol Med 2015;65:161–167.

9. Ypenburg C, Westenberg JJ, Bleeker GB, VAN de Veire N, Marsan NA,Henneman MM, van der Wall EE, Schalij MJ, Abraham TP, Barold SS, Bax JJ.Noninvasive imaging in cardiac resynchronization therapy. Part 1: selection ofpatients. Pacing Clin Electrophysiol 2008;31:1475–1499.

10. Ypenburg C, van Bommel RJ, Delgado V, Mollema SA, Bleeker GB, Boersma E,Schalij MJ, Bax JJ. Optimal left ventricular lead position predicts reverseremodeling and survival after cardiac resynchronization therapy. J Am CollCardiol 2008;52:1402–1409.

11. Cordeiro JM, Greene L, Heilmann C, Antzelevitch D, Antzelevitch C. Trans-mural heterogeneity of calcium activity and mechanical function in the canine leftventricle. Am J Physiol Heart Circ Physiol 2004;286:H1471–H1479.

12. Sengupta PP, Krishnamoorthy VK, Korinek J, Narula J, Vannan MA, Lester SJ,Tajik JA, Seward JB, Khandheria BK, Belohlavek M. Left ventricular form andfunction revisited: applied translational science to cardiovascular ultrasoundimaging. J Am Soc Echocardiogr 2007;20:539–551.

13. Voigt JU, Schneider TM, Korder S, Szulik M, Gurel E, Daniel WG, RademakersF, Flachskampf FA. Apical transverse motion as surrogate parameter to determineregional left ventricular function inhomogeneities: a new, integrative approach toleft ventricular asynchrony assessment. Eur Heart J 2009;30:959–968.

14. Teske AJ, De Boeck BWL, Melman PG, Sieswerda GT, Doevendans PA, CramerMJM. Echocardiographic quantification of myocardial function using tissuedeformation imaging, a guide to image acquisition and analysis using tissueDoppler and speckle tracking. Cardiovasc Ultrasound 2007;5:27–27.

15. Luo J, Konofagou E. A fast normalized cross-correlation calculation method formotion estimation. IEEE Trans Ultrason Ferroelectr Freq Control 2010;57:1347–1357.

16. Gepstein L, Evans SJ. Electroanatomical mapping of the heart: basic conceptsand implications for the treatment of cardiac arrhythmias. Pacing Clin Electro-physiol 1998;21:1268–1278.

17. Badke FR, Boinay P, Covell JW. Effects of ventricular pacing on regional leftventricular performance in the dog. Am J Physiol 1980;238:H858–H867.

18. Gurev V, Constantino J, Rice JJ, Trayanova NA. Distribution of electro-mechanical delay in the heart: insights from a three-dimensional electromechan-ical model. Biophys J 2010;99:745–754.

19. Wyman BT, Hunter WC, Prinzen FW, McVeigh ER. Mapping propagation ofmechanical activation in the paced heart with MRI tagging. Am J Physiol1999;276:H881–H891.

20. Faris OP, Evans FJ, Ennis DB, Helm PA, Taylor JL, Chesnick AS, Guttman MA,Ozturk C, McVeigh ER. Novel technique for cardiac electromechanical mappingwith magnetic resonance imaging tagging and an epicardial electrode sock. AnnBiomed Eng 2003;31:430–440.

21. Kroon W, Lumens J, Potse M, Suerder D, Klersy C, Regoli F, Murzilli R,Moccetti T, Delhaas T, Krause R, Prinzen FW, Auricchio A. In vivo electro-mechanical assessment of heart failure patients with prolonged QRS duration.Heart Rhythm 2015;12:1259–1267.

22. Prinzen FW, Augustijn CH, Allessie MA, Arts T, Delhaas T, Reneman RS. Thetime sequence of electrical and mechanical activation during spontaneous beatingand ectopic stimulation. Eur Heart J 1992;13:535–543.

23. Durrer D, van Dam RT, Freud GE, Janse MJ, Meijler FL, Arzbaecher RC. Totalexcitation of the isolated human heart. Circulation 1970;41:899–912.

24. Sengupta PP, Khandheria BK, Korinek J, Wang J, Jahangir A, Seward JB,Belohlavek M. Apex-to-base dispersion in regional timing of left ventricularshortening and lengthening. J Am Coll Cardiol 2006;47:163–172.

25. Akar FG, Spragg DD, Tunin RS, Kass DA, Tomaselli GF. Mechanismsunderlying conduction slowing and arrhythmogenesis in nonischemic dilatedcardiomyopathy. Circ Res 2004;95:717–725.