AJP - Heart Ad Instruments
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Am J Physiol Heart Circ Physiol 289: H2724-H2732, 2005. First published August 5, 2005; doi:10.1152/ajpheart.00639.2005
0363-6135/05 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Material
Right arrow All Versions of this Article:
289/6/H2724    most recent
00639.2005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, X.
Right arrow Articles by He, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zhang, X.
Right arrow Articles by He, B.

INNOVATIVE METHODOLOGY

Noninvasive three-dimensional electrocardiographic imaging of ventricular activation sequence

Xin Zhang,1 Indiresha Ramachandra,2 Zhongming Liu,1 Basharat Muneer,2 Steven M. Pogwizd,2 and Bin He1

1Department of Biomedical Engineering, University of Minnesota, Minneapolis, Minnesota; and 2Section of Cardiology, University of Illinois, Chicago, Illinois

Submitted 14 June 2005 ; accepted in final form 1 August 2005


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Imaging the myocardial activation sequence is critical for improved diagnosis and treatment of life-threatening cardiac arrhythmias. It is desirable to reveal the underlying cardiac electrical activity throughout the three-dimensional (3-D) myocardium (rather than just the endocardial or epicardial surface) from noninvasive body surface potential measurements. A new 3-D electrocardiographic imaging technique (3-DEIT) based on the boundary element method (BEM) and multiobjective nonlinear optimization has been applied to reconstruct the cardiac activation sequences from body surface potential maps. Ultrafast computerized tomography scanning was performed for subsequent construction of the torso and heart models. Experimental studies were then conducted, during left and right ventricular pacing, in which noninvasive assessment of ventricular activation sequence by means of 3-DEIT was performed simultaneously with 3-D intracardiac mapping (up to 200 intramural sites) using specially designed plunge-needle electrodes in closed-chest rabbits. Estimated activation sequences from 3-DEIT were in good agreement with those constructed from simultaneously recorded intracardiac electrograms in the same animals. Averaged over 100 paced beats (from a total of 10 pacing sites), total activation times were comparable (53.3 ± 8.1 vs. 49.8 ± 5.2 ms), the localization error of site of initiation of activation was 5.73 ± 1.77 mm, and the relative error between the estimated and measured activation sequences was 0.32 ± 0.06. The present experimental results demonstrate that the 3-D paced ventricular activation sequence can be reconstructed by using noninvasive multisite body surface electrocardiographic measurements and imaging of heart-torso geometry. This new 3-D electrocardiographic imaging modality has the potential to guide catheter-based ablative interventions for the treatment of life-threatening cardiac arrhythmias.

mapping; imaging; electrophysiology; arrhythmia; catheter ablation

VENTRICULAR ARRHYTHMIAS account for nearly 400,000 deaths per year in the United States alone (27). These lethal arrhythmias typically occur in patients with structural heart disease (such as ischemic heart disease or congestive heart failure). Pharmacological treatment of these arrhythmias has been difficult because of significant proarrhythmic effects of many antiarrhythmic agents, leading to utilization of more invasive approaches (10). Implantable defibrillators have been widely used in patients at high risk of sudden death (10). However, these devices do not function in a preventive manner but only treat ventricular tachycardia (VT) or ventricular fibrillation (VF) once it occurs. Alternatively, other invasive approaches, such as cardiac ablation techniques, are being used in a growing number of patients to eliminate critical sites in the heart that initiate the arrhythmias (41).

Catheter ablation of VT requires precise localization of the site of origin of the arrhythmias. This is usually done with catheter-based mapping procedures under fluoroscopic and/or magnetic guidance. Aside from the requirement of considerable time in the electrophysiology laboratory and the need to keep patients in prolonged periods of cardiac arrhythmias for catheter-based mapping, in some cases, it is impossible to adequately perform catheter-based mapping due to the hemodynamic instability of the patient's arrhythmias. Therefore, there is a need to develop noninvasive means of imaging cardiac activation, which could enable cardiologists to rapidly focus the intervention at the source of the arrhythmias without the need for lengthy intracardiac mapping.

Body surface potential mapping (1, 11, 42) has been used to attempt to reveal the underlying cardiac electrical activity noninvasively, in healthy subjects, patients with various cardiac abnormalities, and experimental animals (11). However, its ability to precisely predict cardiac activation is limited by the "smearing effect" of the torso volume conductor (1, 40). To overcome this smearing effect of the torso volume conductor, a number of investigators have studied ECG inverse solutions, which attempt to reconstruct the representation of cardiac electrical activity from thoracic ECG recordings. An early attempt to obtain unique ECG inverse solution was to use discrete point sources such as equivalent fixed or moving dipoles (26, 35), but numerous experimental studies suggested that distributed source models, such as epicardial potential distribution (2, 5, 7, 24, 25, 34, 37, 39) or heart surface isochrones (17), were needed to account for the distributed nature of cardiac electrical activity. Heart surface activation times have been previously estimated from body surface ECG recordings (9, 12, 17, 43). Despite the significant effort and progress made on heart-surface electrocardiographic imaging, these approaches are limited in that they only reveal information over the two-dimensional heart surface (epicardial or endocardial surface). Hypothetically, the intramural myocardial activation could be inferred from epicardial potentials/electrograms, but there have been no simultaneous intracardiac mapping data to confirm these predictions.

It is of paramount importance to image cardiac electrical activity throughout the three-dimensional (3-D) myocardium because cardiac arrhythmia may arise from myocardial regions far from heart surface. We have proposed a 3-D electrocardiographic imaging technique (3-DEIT), based on a novel imaging technique utilizing body surface potential maps (BSPMs) in which a 3-D electrophysiological heart model is incorporated, that can noninvasively localize sites of origin of myocardial activation (14, 20) and image ventricular activation sequence (13, 16), distributions of transmembrane potentials (15), and infarction (21). Work by our group (13–16, 20, 21) and others (28, 38) in computer simulations suggested the merits of the 3-DEIT approach in imaging and localizing cardiac electrical activity within the 3-D volume of myocardium. Although promising results have been obtained via computer simulation, there have been no in vivo experimental validation studies of the 3-DEIT approach.

In the present study, we experimentally validate our heart model-based 3-DEIT approach, by comparing the imaged activation sequence with the simultaneously recorded activation sequence obtained from 3-D cardiac mapping using plunge-needle electrodes throughout the ventricles. Here we demonstrate, for the first time, that the 3-D cardiac activation sequence can be noninvasively estimated from BSPMs by a novel imaging algorithm in a well-controlled experimental setting. This completely noninvasive approach could become a complementary means to image the 3-D cardiac activation sequence and locate the origin of cardiac arrhythmia, thereby providing valuable information for clinical diagnosis and treatment (e.g., ablative approaches) of ventricular arrhythmias.

This work has been presented in part at the Biophysical Society meeting in February 2005 and at the Computer in Cardiology meeting in September 2004 (45).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Animal model and in vivo mapping. Healthy New Zealand White rabbits were studied by using a protocol approved by the Institute Animal Care and Use Committees of the University of Minnesota and the University of Illinois at Chicago. The anatomical geometry information was obtained by ultrafast computerized tomography (UFCT) in anesthetized rabbits 1–4 days before the in vivo mapping experimentation. For each rabbit, two sets of CT images were obtained; one without intravenous contrast was used to construct the rabbit torso model, and another one with intravenous contrast (Omnipaque) was obtained for construction of a detailed ventricle model.

One to four days later, simultaneous body surface potential mapping and 3-D intracardiac bipolar mapping of cardiac electrical activity were performed. Forty to sixty breathable electrodes for infants (Vermed, VT) were placed uniformly to cover the anterolateral chest up to the midaxillary line (Fig. 1A). The heart was exposed via median sternotomy, and 20–25 transmural needle electrodes were inserted in the left and right ventricles of the rabbit (Fig. 1B). Each needle electrode contains eight bipolar electrode pairs with an interelectrode distance of 500 µm (8, 29–33). The chest and skin were closed in multiple layers by suture. Rapid ventricular pacing was then performed via bipolar electrode pairs on selected plunge-needle electrodes, while bipolar electrograms were continuously recorded from all plunge electrode pairs together with body surface potentials from surface electrodes. Figure 2A shows an example of body surface electrograms during ventricular pacing. Figure 2B illustrates the corresponding intraoperative electrograms recorded from eight bipolar electrode pairs at one plunge-needle electrode. At the completion of mapping, anesthetized rabbits underwent a (similar) postoperative UFCT scan. The rabbits were then euthanized, the chest was reopened, and the plunge needle electrodes were carefully localized as described in Ref. 33 by replacing each with a labeled pin. The heart was then excised and fixed in formalin. At a later date, the formalin-fixed heart with labeled pins was scanned in the UFCT to get precise 3-D localization of the transmural electrodes (see Fig. 1).



View larger version (68K):
[in this window]
[in a new window]
  		Fig. 1. A: determination of body surface recording electrodes using ultrafast computerized tomography (UFCT). B: reconstruction from UFCT images of a rabbit heart with labeled pins replacing the plunge-needle electrodes used during 3-dimensional intracardiac mapping.

 


View larger version (27K):
[in this window]
[in a new window]
  		Fig. 2. A: example of electrocardiograms recorded from anterior chest electrodes during ventricular pacing at midanterior endocardium. B: intraoperative electrograms corresponding to the pacing beat in the rectangle of A, recorded from 8 bipolar electrode pairs at 1 plunge needle electrode. Channels 1–8 (Ch 1–Ch 8) are from endocardium to epicardium. The red bar represents the time point of pacing stimuli, the blue dot represents the earliest activation time on endocardial electrode, and the red dot represents the latest activation on epicardial electrode. Difference of activation time between red dot and blue dot is 6 ms.

 
Heart model-based 3-DEIT approach. The heart model-based 3-D electrocardiographic imaging algorithm for estimating 3-D activation sequence has previously been detailed (16) (See Online Supplement1 ). In brief, a 3-D heart excitation model was constructed for each of the rabbits studied from geometric measurements via UFCT. A preliminary classification system (PCS) that approximately determines cardiac status based on a priori knowledge and the measured BSPMs was used to estimate the initial pattern of myocardial activation by means of an artificial neural network (ANN) (13, 16). The input of the PCS is the measured BSPM, and the output is the initially estimated myocardial segment whose corresponding regional current dipole gives the best match between simulated and measured BSPMs among all myocardial regional dipoles. The PCS provides an approximate estimate of the region of initiation of the paced activity and limits the searching space in our computer model for the optimization procedure. According to the output of the PCS, the parameters (origin of activation, activation propagation velocity, etc.) of the heart model were initialized, and the corresponding BSPM was calculated by using the boundary element method (BEM). Then a multiobjective nonlinear optimization procedure was employed to estimate the cardiac activation sequence.

With the objective function of the entire optimization system as E(x) = f[ECC(x),Eminp(x),ENPL(x)], the mathematical model of the desired optimization can then be represented as the following minimization problem

(1)

where X is the probable value region of the parameters in the heart model, and x is a parameter vector of the activation time in the heart model. ECC(x) is constructed with the average correlation coefficient (CC) between the measured and simulated BSPMs from instant T1 to instant T2 of the cardiac excitation after detection of initial activation. Eminp(x) is constructed with the deviation of the positions of minima of the measured and simulated BSPMs from instant T1 to instant T2. ENPL(x) is constructed with the relative error of the number of body surface recording leads, at which the potentials are less than a certain negative threshold (50% of maximum negative value) in the measured and simulated BSPMs from instant T1 to instant T2. In the present study, by setting Eminp and ENPL as constraints and keeping ECC as an objective function, Eq. 1 can be converted into a single-objective nonlinear optimization problem with two constrains, which was solved with the aid of the Simplex method. We used only the minima as one of the objective functions for the pacing protocol because there are reports that the minima in the BSPMs represent robust characteristics over subjects. Moreover, individual differences had little influence to the morphology and site of negativity (4, 40).

If the measured and model-generated BSPMs matched well, the optimization procedure stopped. If not, the heart model parameters were adjusted with the aid of the optimization algorithms described above and the simulation procedure proceeded until the objective functions satisfied the given convergent criteria. The activation sequence produced by the optimized heart model parameters was compared with the measured activation sequence (see below) to assess the performance of the noninvasive 3-D myocardial activation imaging. Figure 3 illustrates a schematic diagram of the present experimental validation study.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 3. Schematic diagram of the experimental study to validate the 3-dimensional electrocardiographic imaging technique by means of simultaneous intracardiac mapping in a rabbit model. 3-D AS, 3-dimensional activation sequence; BSPM, body surface potential map.

 
3-D intraoperative mapping. The electrograms recorded from transmural plunge-needle electrodes were analyzed according to a previously described approach (8, 29–33). Computer-assigned activation sequences were based on a peak criterion. Review and editing of electrograms were performed by the operator, and 3-D activation maps were constructed. First, as shown in Fig. 1B, the 3-D heart models with needle information were constructed from the UFCT scans of isolated heart with 27-gauge needles (replacing the pins at sites of transmural electrodes) by a specially designed software package that performs this semiautomatically. Each needle can be represented by several bright dot points on each slice, and a linear equation was used to approximately represent each needle. The intersection between each needle and heart surface was calculated on the basis of the fitted linear equation and the reconstructed 3-D heart model. Based on the structure information of each needle and coordinates of entry point, as well as the linear equation of the needle, the coordinates of eight transmural electrode pairs can be determined uniquely. To compare the activation sequence constructed from the 3-D intracardiac mapping with that estimated from noninvasive BSPMs, digital image processing was performed on the heart geometry information obtained from the postoperative UFCT scanning. Geometric transformation techniques using a third-order polynomial expression (6) was used to transform the coordinates of the postoperative scanned heart onto the preoperative scanned heart. The corresponding activation time of each recording electrode was assigned on the basis of peak criteria (30, 32), and then a weighted average interpolation algorithm (44) was applied to obtain the complete 3-D distribution of activation times throughout the ventricles. The 3-D activation sequence is color coded.

In addition to the above procedures, we also evaluated our 3-D activation mapping procedure with a previously developed 3-D activation mapping procedure (29–33), which used hand drawing on templates of four to five slices of the rabbit heart about 5–7 mm thick as derived from scaled drawing of pathological specimens of rabbit heart. Similar pattern of activation sequence maps were obtained for both sets, which confirmed the above-described computerized activation mapping from 3-D intracardiac electrophysiological recordings. Total activation time was defined as the interval between the earliest and latest activations for each beat.

Statistical analysis. All data are presented as means ± SD. Students t-test (unpaired) was used to evaluate the results, and P < 0.05 was considered significant. The general quality of 3-DEIT was quantitatively evaluated by comparing the estimated 3-D activation sequences with respect to those constructed from measurements, with the use of relative error (RE), error distance, and total activation time. The RE is a good indication of difference between estimated and measurement-constructed 3-D activation sequences and is defined as follows

(2)

where n is the number of myocardial cell units of the heart model, and AT and AT are the estimated activation time and measurement-constructed activation time at myocardial cell unit, respectively. The error distance is defined as the spatial distance between the earliest-activated myocardial cell unit in the 3-DEIT estimation and the pacing electrode. The total activation time for a paced beat is defined as the interval between the earliest and latest activation by either 3-DEIT imaged and 3-D-mapped activation sequences.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 GRANTS
 REFERENCES
 
In this in vivo experimental study, the heart model-based 3-DEIT analysis was conducted on four rabbits under isoflurane anesthesia. After insertion of plunge electrodes, closure of the chest did not alter total activation times of sinus beats (30 ± 3 vs. 29 ± 3 ms preclosure, P = NS), which were consistent with previously published data in control rabbits (31). The sinus rhythm signals recorded before and after open-chest surgery were compared. The averaged correlation coefficient (CC) of BSPMs is >85% during the ventricular depolarization period in two rabbits, which suggests that the patterns of BSPM were not significantly changed by open-chest surgery with careful chest closure in the rabbits.

The realistic geometry heart-torso models were built for all rabbits on the basis of their individual UFCT images. Some parameters of the heart models and experimental settings are shown in Table 1. The rabbit heart models, which contain a large number of myocardial cell units, were divided into approximately 30–40 myocardial segments. Cardiac electrical activity at each myocardial segment was represented by a regional current dipole for evaluating the BSPM during ventricular activation. The spatial resolution of the heart models was 1 or 0.75 mm, depending on the interslice distance of UFCT scan (1 additional slice was interpolated from 2 adjacent slices to increase the spatial resolution of heart model).


View this table:
[in this window]
[in a new window]
  		Table 1. Experimental parameters

 
By pacing selected pairs of bipolar electrodes of the plunge needle, the ventricular activation was induced. Figure 4 shows an example for rabbit 1, in which the pacing stimulation initiated at the right ventricle and propagated leftward. Figure 4A displays the measured BSPMs, with a time interval of 8 ms. The potential distribution is color coded from blue to red, corresponding to minimum and maximum values, respectively. It shows approximately a dipolar pattern over the anterior torso surface with positive activity at top left and negative activity at bottom right, with the amplitude being increased over time during the period shown. Figure 4B shows the measured 3-D activation sequence from the ventricular base to the ventricular apex derived from intracardiac electrograms of the same beat as in Fig. 4A. The distance interval between each cross-sectional layer of the ventricles is 1 mm, so the physical distance between the adjacent layers shown in Fig. 4 is 2 mm. The activation sequence after pacing at the right ventricular free wall is color coded from blue (early activation) to red (late activation), as shown by the color bar on the right. From Fig. 4B, the earliest activation (blue area) initiated in the midright ventricle close to the apex (layer 15), which then propagated to the left ventricle. This result was confirmed by pinpointing the plunge-needle electrode from which the stimulation was induced, and the pattern of activation sequence was consistent with the 3-D activation maps obtained by using a previously established qualitative procedure (32) over the templates of rabbit heart slices. Figure 4C shows the estimated 3-D activation sequence obtained by the 3-DEIT imaging approach, which also revealed early activation in the right ventricle, but just slightly more apical (layer 17). The quantitative comparison between the measured and imaged 3-D activation sequences shown in Fig. 4, B and C, gave a low RE of 0.29, suggesting the consistency of the measured and imaged activation sequence. The distance between the initiation site identified from the measured activation map and from the estimated origin of activation in the imaged activation sequence was 4.82 mm.



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 4. A: measured body-surface potential maps following ventricular pacing at a right ventricular site in rabbit 1. Time refers to the time instant since pacing. B: 3-D AS derived from intracardiac electrograms. Sections are arranged sequentially from base to midlevel (1–11, top row) and from midlevel to apex (13–23, bottom row). C: 3-D AS estimated noninvasively by means of 3-dimensional electrocardiographic imaging. See text for details.

 
Figure 5 shows an example for rabbit 2, in which the pacing stimulation was induced at a posterior-lateral left ventricular site. The measured 3-D activation sequence derived from intracardiac electrograms is shown in Fig. 5A, and the estimated 3-D activation sequence obtained by 3-DEIT imaging approach in Fig. 5B (the format of display is the same as in Fig. 4). In Fig. 5A, the earliest activation (blue area) initiated in the posterior lateral region of the left ventricle (layer 11), and a similar pattern is observed in Fig. 5B, in which the earliest activation was located in layer 9. Figure 5C illustrates the difference of measured and imaged activation sequence. Notably, the large error appeared at regions with late activation (right anterior base and right anterior apex), and little difference is observed in the regions of early activation. The quantitative comparison gave an RE of 0.322 and a distance between the initiation site identified from the measured activation map and the estimated origin of activation in the imaged activation sequence of 4.58 mm.



View larger version (38K):
[in this window]
[in a new window]
  		Fig. 5. A: 3-D AS derived from intracardiac electrograms after pacing at a posterior lateral left ventricular site in rabbit 2. B: 3-D AS estimated noninvasively by means of 3-dimensional electrocardiographic imaging. C: difference of measured and imaged AS. See text for details.

 
Table 2 shows summarized data of quantitative comparison of the activation sequences and the distance from 3-DEIT-estimated site of initial activation to the initial site determined from the measured 3-D activation sequence from a total of 100 paced beats (10 pacing beats for each of 10 pacing sites) in four rabbits. Also shown are summarized data of total activation times that were estimated by 3-DEIT and that were derived from intracardiac recorded electrograms. The RE between the 3-DEIT-estimated activation sequence and that derived from 3-D intracardiac mapping was 0.32 ± 0.06 (n = 100), suggesting good agreement between the 3-DEIT-estimated spatial distribution of activation time and that derived from intracardiac mapping. The 3-DEIT-estimated initiation site was within 5.73 ± 1.77 mm (n = 100) from the initiation site determined from intracardiac mapping, suggesting that the present 3-DEIT approach can reasonably localize the site of initiation of ventricular activation. Moreover, analysis of over 10,000 activation times (from over 160 electrodes, 100 paced beats) shows that 3-DEIT-estimated activation times (interval from earliest to latest activation) were comparable to those determined from 3-D intracardiac mapping (53.3 ± 8.1 vs. 49.8 ± 5.2 ms, P = NS, n = 100 paced beats), with an estimation error of 4.62 ± 2.29 ms.


View this table:
[in this window]
[in a new window]
  		Table 2. Quantitative comparison of estimated AS and measured AS

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
GRANTS
 REFERENCES
 
In the present study, we have demonstrated experimentally that 3-D cardiac activation sequence can be noninvasively estimated in vivo from BSPMs, UFCT images, and an imaging algorithm. These results suggest that the 3-DEIT can reasonably well estimate the activation sequence within the ventricles, localize the site of initial activation, and estimate the total activation time. Notably, all results derived from the intracardiac mapping presented above were confirmed by pinpointing the transmural electrode on which the stimulation was induced, and with the pattern of activation sequence using the previously established procedure of obtaining activation sequence from intracardiac mapping over the templates of rabbit heart slices (29, 31).

The 3-D electrocardiographic imaging approach represents an important improvement in electrocardiographic imaging. There has been considerable effort to solve the inverse problem of electrocardiography to localize and image cardiac electrical activity from noninvasive body surface electrocardiograms (1, 2, 5, 7, 9, 11–17, 20, 21, 24–26, 28, 34, 35, 37–40, 42, 43). While discrete cardiac electrical source models and heart surface two-dimensional (2-D) distributed source models have been used in the past decades to solve the imaging problem of electrocardiography, it is only recently that we and others have developed 3-D distributed source imaging techniques for solving the electrocardiography inverse problem over the 3-D volume of myocardium (13–16, 20, 21, 28, 38). The present study reports, for the first time, that such 3-D distributed electrocardiographic imaging is feasible in a well-controlled experimental setting.

Rigorous validation studies in biological systems are critical for any imaging approaches. The present study has used a novel experimental design to validate the 3-DEIT imaging approach by means of simultaneous 3-D intracardiac mapping. Despite numerous cardiac mapping studies, there have been no reports on validating electrocardiographic imaging approaches in an in vivo animal model using simultaneous high-density intracardiac mapping over the 3-D myocardium. Here mapping from up to 200 sites in the rabbit heart resulted in a resolution comparable to mapping from nearly 2,400 sites in the canine heart (based on cardiac size), which was sufficient to define 3-D activation sequence and electrophysiological mechanisms of VT in the feline and rabbit heart during various pathophysiological conditions (e.g., myocardial ischemia and heart failure) (30, 33). This degree of resolution, combined with the ability to record intracardiac electrograms in a closed-chest animal, provided a unique basis for in vivo assessment of 3-DEIT. Moreover, insertion of plunge-needle electrodes and closure of the chest was tolerated well with no significant effects on hemodynamics or conduction (see Refs. 19, 30, 33). Therefore, the myocardial activation sequences obtained in our closed-chest mapping studies are good representations of those in the intact heart under pacing conditions. The well-controlled pacing protocol enabled us to simulate physiologically realistic ectopic beats and test the 3-DEIT in a physiologically realistic setting.

The successful imaging of the activation sequence after ventricular pacing suggests the feasibility of noninvasively reconstructing ventricular activation sequence for physiologically realistic events (e.g., VT). To compare the measured activation sequence (from 3-D intracardiac mapping) and the 3-DEIT-estimated activation sequence, we used the measure of RE to assess the dissimilarity. It is well known that the electrocardiographic imaging procedures are sensitive to noise and have inherent error. While the RE for these in vivo experimental studies was somewhat larger than those reported in our computer simulation studies (16), the averaged RE of 0.32 over 100 paced beats represents a good match between what was measured and what was imaged. Note that the present RE results are reasonable compared with prior in vivo validation studies using other approaches. As previous 2-D electrocardiographic imaging studies indicate, the RE between the measured and imaged inverse solutions could be as large as 0.4–0.45 for epicardial potential inverse solutions in an animal study using dog's heart immersed within a human torso phantom (5) and as large as >1.0 for in vivo clinical studies (not simultaneous recordings) for epicardial potential inverse solutions (37). While the previous 2-D imaging results involved epicardial potentials instead of activation sequences, the present averaged RE of 0.32 is reasonable, which may be due to the unique characteristics of the 3-DEIT imaging algorithm in which the estimation was performed by minimizing the difference between measured and heart model-predicted BSPMs not only at a single time point but over a period of time after pacing (16). Therefore, the present 3-DEIT approach is not a spatial imaging technique but a spatiotemporal imaging approach that incorporates electrophysiological properties of cardiac activation, thus substantially improving the performance of the electrocardiographic inverse solutions. Also note that, in the present study, we focused on the imaging of paced ventricular activations, not including sinus rhythm. This is because the paced ventricular activity is a well-controlled and well-characterized process suitable for the goal of the present study, which is to validate the 3-DEIT technique. Cardiac electrical activity during sinus rhythm represents a diffused process that is beyond the scope of the present study. Likewise, imaging of reentrant arrhythmias is a complex problem that is beyond the scope of the present study.

The present 3-DEIT allows imaging of the total activation sequence of ventricular activation, and the total activation times measured by 3-DEIT was comparable to those measured by intracardiac mapping with an estimation error of 4.62 ± 2.29 ms (P = 100 paced beats). With a consideration of the localization errors of body-surface electrodes, the localization errors of plunge-needle electrodes, and volume conductor modeling errors in the in vivo experimental setup, the present localization error represents reasonable accuracy in assessing electrocardiographic imaging techniques using in vivo models. Note that the localization errors may be increased in a large-size biological system, such as in humans, because there are differences in size between rabbits and humans. However, in the inverse imaging problems, things do not scale linearly with size. There are inherited errors associated with certain procedures, such as the errors associated with electrode position determination and heart-torso geometrical modeling. These errors are not linearly scaled to the size of the subjects but are mainly determined by procedures, such as the precision of the positioning device for electrode location error. For these types of error, we believe the effective errors may not be significantly increased even as the size of the subjects increased. So it is anticipated that while the localization error would be increased in humans should the same experimental procedures be used, in a clinical setting the different procedures involved (e.g., there would be no sternotomy) may offset the effective localization errors. Moreover, in a separate clinical study we found the localization error for the site of pacemaker stimulation using 3-DEIT was 5.2 mm in a human subject (22). Further investigation will be needed to test the performance of the 3-DEIT in a clinical setting in intact humans.

We have constructed detailed heart-torso models for each of the rabbits we studied, consisting of ~10,000 cardiac units for 100 paced beats. This large number of experimental parameters is comparable to other experimental electrocardiographic imaging studies (34, 37, 43). These results demonstrate the novel concept of 3-DEIT approach in imaging 3-D activation sequence and localizing the site of initial activation.

In the present study, the median sternotomy may change the volume conductor property of thoracic volume and cause difference on conductivity distribution compared with intact torso. However, with careful multilayer-suture chest closing, as well as plastic-sheathed recording wires, we minimized such changes on the volume conductor. When comparing the sinus rhythm signals recorded before and after open-chest surgery, it was found that the averaged CC of BSPMs is >85% during the ventricular depolarization period in two rabbits, which suggests that the present open-chest surgery procedure did not significantly alter the patterns of BSPM. While the effect of open-chest surgery was neglected in the present analysis, it is likely contributing to the estimation errors.

In the present study, the anisotropy of myocardium was not taken into consideration in the heart modeling. This represents a limitation of the present study that may have contributed to the estimation error. As shown in Fig. 5C, the largest errors between measured and imaged activation sequences appeared in the regions of late activation, which may be associated with the isotropic assumption in the present study. However, with further development of techniques determining fiber orientations in an intact heart, the anisotropy of myocardium may be incorporated into the 3-DEIT approach in the future investigation and reduce the estimation errors.

What we have demonstrated in the present experimental study is the feasibility of imaging activation sequence within the 3-D ventricles using 3-DEIT in hearts without infarct/ischemia. For clinical situations with myocardial infarct scars, chronic ischemia, or patchy fibrosis, such information should be incorporated into the computer heart model as much as possible [e.g., incorporating a necrotic myocardial zone and altered conduction patterns in the computer heart model (23)], and further investigations will be needed.

The 3-DEIT represents a novel approach with which to locate the origin of cardiac activation (and potentially to locate the origin of cardiac arrhythmia) within the 3-D myocardium in a single beat. With further refinement, the 3-DEIT may be capable of localizing initiating sites from midmyocardial and epicardial regions far from the endocardial surface, which would help overcome an important limitation of clinical endocardial mapping device (3, 18, 36).

In summary, we have conducted controlled experimental studies in rabbits to test the feasibility of imaging 3-D activation sequence within the ventricles using the 3-DEIT approach. The present experimental study represents, to our knowledge, the first rigorous validation study of 3-D electrocardiographic imaging techniques, by means of simultaneous 3-D intracardiac mapping. The promising validation results indicate that the noninvasive 3-DEIT imaging approach can noninvasively estimate the 3-D ventricular activation and has the potential to advance electrocardiographic imaging from 2-D to 3-D, enhancing our ability to diagnose and treat life-threatening cardiac arrhythmias.


   GRANTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
REFERENCES
 
This work was supported in part by National Science Foundation Grant BES-0411480 (B. He), American Heart Association Grant 0140132N (B. He), National Institutes of Health Grant RO1-EB-00178 (B. He), and National Heart, Lung, and Blood Institute Grants R01-HL-46929 and R01-HL-73966 (S. M. Pogwizd).


   ACKNOWLEDGMENTS
 
This work has been presented in part at the Biophysical Society Meeting in February 2005 and at the Computer in Cardiology meeting in September 2004 (45).


   FOOTNOTES
 

Address for reprint requests and other correspondence: B. He, Dept. of Biomedical Engineering, Univ. of Minnesota, 7–105 BSBE, 312 Church St., Minneapolis, MN 55455 (e-mail: binhe{at}umn.edu)

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 Supplemental data for this article may be found at http://ajpheart.physiology.org/cgi/content/full/00639.2005/DC1. Back


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Abildskov JA, Burgess MJ, Lux RL, and Wyatt RF. Experimental evidence for regional cardiac influence in body surface isopotential maps of dogs. Circ Res 38: 386–391, 1976.[Abstract/Free Full Text]
  2. Barr RC and Spach MS. Inverse calculation of QRS-T epicardial potentials from normal and ectopic beats in the dog. Circ Res 42: 661–675.
  3. Ben-Haim SA, Osadchy D, Schuster I, Gepstein L, Hayam G, and Josephson ME. Nonfluoroscopic in vivo navigation and mapping technology. Nat Med 2: 1393–1395, 1996.[CrossRef][Web of Science][Medline]
  4. Benson DW, Sterba R, Gallagher JJ, Walston A II, and Spach MS. Localization of the site of ventricular preexcitation with body surface maps in patients with Wolff-Parkinson-White syndrome. Circulation 65: 1259–1268, 1982.[Abstract/Free Full Text]
  5. Burnes JE, Taccadi B, and Rudy Y. A noninvasive imaging modality for cardiac arrhythmias. Circulation 102: 2152–2158, 2000.[Abstract/Free Full Text]
  6. Castleman KR. Digital Image Processing. London, UK: Prentice-Hall International, 1996.
  7. Cheng LK, Bodley JM, and Pullan AJ. Comparison of potential- and activation-based formulations for the inverse problem of electrocardiology. IEEE Trans Biomed Eng 50: 11–22, 2003.[CrossRef][Web of Science][Medline]
  8. Chung MK, Pogwizd SM, Miller DP, and Cain ME. Three-dimensional mapping of the initiation of nonsustained ventricular tachycardia in the human heart. Circulation 95: 2517–2527, 1997.[Abstract/Free Full Text]
  9. Cuppen JJM and Van Oosterom A. Model studies with inversely calculated isochrones of ventricular depolarization. IEEE Trans Biomed Eng 31: 652–659, 1984.[Web of Science][Medline]
  10. Dimarco JP. Medical progress: implantable cardioverter-defibrillators. N Engl J Med 349: 1836–1847, 2003.[Free Full Text]
  11. Flowers NC and Horan LG. Body surface potential mapping. In: Cardiac Electrophysiology (2nd ed.), edited by Zipes DP and Jalife J. Philadelphia, PA: Saunders, 1995, p. 1049–1067.
  12. Greensite F. Remote reconstruction of confined wavefront propagation. Inverse Problems 11: 361–370, 1995.[CrossRef]
  13. He B and Li G. Noninvasive three-dimensional myocardial activation time imaging by means of a heart-excitation-model. Int J Bioelectromag 4: 87–88, 2002.
  14. He B and Wu D. Imaging and visualization of 3D cardiac electric activity. IEEE Trans Info Tech Biomed 5: 181–186, 2001.[CrossRef]
  15. He B, Li G, and Zhang X. Noninvasive imaging of ventricular transmembrane potentials within three-dimensional myocardium by means of a realistic geometry anisotropic heart model. IEEE Trans Biomed Eng 50: 1190–1202, 2003.[CrossRef][Web of Science][Medline]
  16. He B, Li G, and Zhang X. Noninvasive three-dimensional activation time imaging of ventricular excitation by means of a heart-excitation-model. Phys Med Biol 47: 4063–4078, 2002.[CrossRef][Web of Science][Medline]
  17. Huiskamp G and Greensite F. A new method for myocardial activation imaging. IEEE Trans Biomed Eng 44: 433–446, 1997.[CrossRef][Web of Science][Medline]
  18. Khoury DS, Berrier KL, Badruddin SM, and Zoghbi WA. Three-dimensional electrophysiological imaging of intact canine left ventricle using a noncontact multielectrode cavitary probe: study of sinus, paced, and spontaneous premature beats. Circulation 97: 399–409, 1998.[Abstract/Free Full Text]
  19. Kovoor P, Campbell C, Wallace E, Byth K, Dewsnap B, Epipper Uther JV, and Ross D. Effects of simultaneous insertion of 66 plunge needle electrodes on myocardial activation, function, and structure. Pacing Clin Electrophysiol 26: 1979–1985, 2003.[CrossRef][Medline]
  20. Li G and He B. Localization of sites of origins of cardiac activation by means of a new heart-model-based electrocardiographic imaging approach. IEEE Trans Biomed Eng 48: 660–669, 2001.[CrossRef][Web of Science][Medline]
  21. Li G and He B. Noninvasive estimation of myocardial infarction by means of a heart-model-based imaging approach—a simulation study. Med Biol Eng Comput 42: 128–136, 2004.[CrossRef][Web of Science][Medline]
  22. Li G, Zhang X, Lian J, and He B. Noninvasive localization of the origin of paced cardiac activation in a patient with pacemaker by means of a heart-excitation-model. IEEE Trans Biomed Eng 50: 1117–1120, 2003.[CrossRef][Web of Science][Medline]
  23. Liu C, Li G, and He B. Localization of the site of origin of reentrant arrhythmia from body surface potential maps: a model study. Phys Med Biol 50: 1421–1432, 2005.[CrossRef][Web of Science][Medline]
  24. MacLeod RS and Brooks DH. Recent progress in inverse problems in electrocardiology. IEEE Eng Med Biol Mag 17: 73–83, 1998.[Web of Science][Medline]
  25. Martyn Nash P, Chris Bradley P, and Paterson DJ. Imaging electrocardiographic dispersion of depolarization and repolarization during ischemia: simultaneous body surface and epicardial mapping. Circulation 107: 2257–2263, 2003.[Abstract/Free Full Text]
  26. Mirvis DM, Keller FW, Ideker RE, Cox JW, Dowdie RJ, and Zettergren DG. Detection and localization of multiple epicardial electrical generators by a two-dipole ranging technique. Circ Res 41: 551–557, 1977.[Abstract/Free Full Text]
  27. Myerburg RJ, Dessler KM, and Castellanos A. Sudden cardiac death: epidemiology, transient risk, and intervention assessment. Ann Intern Med 119: 1187–1197, 1993.[Abstract/Free Full Text]
  28. Ohyu S, Okamoto Y, and Kuriki S. Use of ventricular propagated excitation model in the magnetocardiographic inverse problem for reconstruction of electrophysiological properties. IEEE Trans Biomed Eng 49: 509–519, 2002.[CrossRef][Web of Science][Medline]
  29. Pogwizd SM, Chung MK, and Cain ME. Termination of ventricular tachycardia in the human heart: insights from three-dimensional mapping of nonsustained and sustained ventricular tachycardia. Circulation 95: 2528–2540, 1997.[Abstract/Free Full Text]
  30. Pogwizd SM and Corr PB. Reentrant and nonreentrant mechanisms contribute to arrhythmogenesis during early myocardial ischemia: results using three-dimensional mapping. Circ Res 61: 352–371, 1987.[Abstract/Free Full Text]
  31. Pogwizd SM, Hoyt RH, Saffitz JE, Corr PB, Cox JL, and Cain ME. Reentrant and focal mechanisms underlying ventricular tachycardia in the human heart. Circulation 86: 1872–1887, 1992.[Abstract/Free Full Text]
  32. Pogwizd SM, McKenzie JP, and Cain ME. Mechanisms underlying spontaneous and inducible ventricular arrhythmias in patients with idiopathic dilated cardiomyopathy. Circulation 98: 2404–2014, 1998.[Abstract/Free Full Text]
  33. Pogwizd SM. Nonreentrant mechanisms underlying spontaneous ventricular arrhythmias in a model of nonischemic heart failure in rabbits. Circulation 92: 1034–1048, 1995.[Abstract/Free Full Text]
  34. Ramanathan C, Raja NG, Jia P, Ryu K, and Rudy Y. Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia. Nat Med 10: 422–428, 2004.[CrossRef][Web of Science][Medline]
  35. Savard P, Roberge FA, Perry J, and Nadeau RA. Representation of cardiac electrical activity by a moving dipole for normal and ectopic beats in the intact dog. Circ Res 46: 415–425, 1980.[Free Full Text]
  36. Schalij MJ, van Rugge FP, Siezenga M, and van der Velde ET. Endocardial activation mapping of ventricular tachycardia in patients: first application of a 32-site bipolar mapping electrode catheter. Circulation 98: 2168–2179, 1998.[Abstract/Free Full Text]
  37. Shahidi AV, Savard P, and Nadeau R. Forward and inverse problems of electrocardiography: modeling and recovery of epicardial potentials in humans. IEEE Trans Biomed Eng 41: 249–256, 1994.[CrossRef][Web of Science][Medline]
  38. Skipa O, Sachse NF, Werner C, and Dossel O. Transmembrane potential reconstruction in anisotropic heart model. Int J Bioelectromag 17–18, 2002.
  39. Smith WM and Barr RC. The forward, and inverse problems: what are they, why are they important, and where do we stand? J Cardiovasc Electrophysiol 12: 253–255, 2001.[CrossRef][Web of Science][Medline]
  40. Spach MS, Barr RC, Lanning CF, and Tucek PC. Origin of body surface QRS and T-wave potentials from epicardial potential distributions in the intact chimpanzee. Circulation 55: 268–278, 1977.[Abstract/Free Full Text]
  41. Stevenson WG and Delacretaz E. Radiofrequency catheter ablation of ventricular tachycardia. Heart 84: 553–559, 2000.[Free Full Text]
  42. Taccardi B. Distribution of heart potentials on dog's thoracic surface. Circ Res 11: 862–869, 1962.[Abstract/Free Full Text]
  43. Tilg B, Fischer G, Modre R, Hanser F, Messnarz B, Schocke M, Kremser C, Berger T, Hintringer F, and Roithinger FX. Model-based imaging of cardiac electrical excitation in humans. IEEE Trans Med Imaging 21: 1031–1039, 2002.[CrossRef][Web of Science][Medline]
  44. Watson DF. Contouring—A Guide to the Analysis and Display of Spatial Data. Oxford, UK: Pergamon, 1992.
  45. Zhang X, Ramachandra I, Liu Z, Muneer B, Pogwizd SM, and He B 3-Dimensional activation sequence reconstruction from body surface potential maps by means of a heart-model-based imaging approach. Comput Cardiol 31: 1–4, 2004.



This article has been cited by other articles:


Home page
 J Am Coll CardiolHome page
T. Berger, G. Fischer, B. Pfeifer, R. Modre, F. Hanser, T. Trieb, F. X. Roithinger, M. Stuehlinger, O. Pachinger, B. Tilg, et al.
Single-Beat Noninvasive Imaging of Cardiac Electrophysiology of Ventricular Pre-Excitation
J. Am. Coll. Cardiol., November 21, 2006; 48(10): 2045 - 2052.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Material
Right arrow All Versions of this Article:
289/6/H2724    most recent
00639.2005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, X.
Right arrow Articles by He, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zhang, X.
Right arrow Articles by He, B.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2005 by the American Physiological Society.