Localization of septal pacing sites in the dog heart by epicardial mapping

Orientation of conduction velocity vectors on cardiac mapping surfaces

AIMS

Electroanatomical maps using automated conduction velocity (CV) algorithms are now being calculated using two-dimensional (2D) mapping tools. We studied the accuracy of mapping surface 2D CV, compared to the three-dimensional (3D) vectors, and the influence of mapping resolution in non-scarred animal and human heart models.

METHODS AND RESULTS

Two models were used: a healthy porcine Langendorff model with transmural needle electrodes and a computer stimulation model of the ventricles built from an MRI-segmented, excised human heart. Local activation times (LATs) within the 3D volume of the mesh were used to calculate true 3D CVs (direction and velocity) for different pixel resolutions ranging between 500 μm and 4 mm (3D CVs). CV was also calculated for endocardial surface-only LATs (2D CV). In the experimental model, surface (2D) CV was faster on the epicardium (0.509 m/s) compared to the endocardium (0.262 m/s). In stimulation models, 2D CV significantly exceeded 3D CVs across all mapping resolutions and increased as resolution decreased. Three-dimensional and 2D left ventricle CV at 500 μm resolution increased from 429.2 ± 189.3 to 527.7 ± 253.8 mm/s (P < 0.01), respectively, with modest correlation (R = 0.64). Decreasing the resolution to 4 mm significantly increased 2D CV and weakened the correlation (R = 0.46). The majority of CV vectors were not parallel (<30°) to the mapping surface providing a potential mechanistic explanation for erroneous LAT-based CV over-estimation.

CONCLUSION

Ventricular CV is overestimated when using 2D LAT-based CV calculation of the mapping surface and significantly compounded by mapping resolution. Three-dimensional electric field-based approaches are needed in mapping true CV on mapping surfaces.

Spectral analysis of electrograms during ventricular tachycardia in a canine model: relation with epicardial isochronal maps

The purpose of this study was to assess the capability of magnitude-squared coherence and bicoherence to differentiate monomorphic ventricular tachycardia (MVT) and polymorphic ventricular tachycardia (PVT) in a canine model and to relate these results to the epicardial isochronal maps on a beat-to-beat basis. Unipolar electrograms were simultaneously recorded from the surface of both ventricles with a 127-lead sock electrode array in 12 open-chest anesthetized dogs. The sampling frequency was 500 Hz. Atrioventricular block was induced by formaldehyde injection into the atrioventricular node. The left anterior descending coronary artery was occluded for 60 minutes under ventricular pacing (140 beats/min). During reperfusion, 12 MVT episodes lasting more than 42 seconds were recorded. Left stellate ganglion stimulation induced five PVT episodes lasting more than 42 seconds. Each of these recordings was divided into seven segments of 3,072 points (6.144 seconds). After visual selection, 104 segments were extracted and classified as 73 MVT and 31 PVT segments. Magnitude-squared coherence was estimated as the cross-spectrum from two epicardial signals (on the right and left ventricles, respectively), normalized with the respective autopower spectrum. Bicoherence was estimated as the bispectrum normalized with the autopower spectrum. Magnitude-squared coherence correctly identified 96% of MVT and 81% of PVT segments for a total accuracy of 91%. Bicoherence estimated with the left ventricular lead correctly identified 100% of MVT and 77% of PVT segments with an accuracy of 93%. Beat-to-beat epicardial maps of MVT displayed a cluster of sites of origin close to the reperfusion area, while the sites of origin from beats during PVT were much more dispersed over both ventricles. A strong and significant correlation was found between the number of electrodes with the earliest epicardial activation and coherence (r = .76, P < .0001) and bicoherence (r = .68, P < .0001), respectively. A high and significant correlation was also found between both spectral estimators (r = .74, P < .0001). Coherence and bicoherence discriminated accurately between MVT and PVT. Coherence achieved better results compared with bicoherence. Coherence and bicoherence measurements showed a quantitative relation with the spatial dispersion of the sites of origin. Both spectral techniques seemed powerful enough to be used in the development of implantable devices.

Mapping of septal ventricular tachycardia: clinical and experimental correlations

In patients with chronic myocardial infarction, ventricular tachycardia originating in the interventricular septum may account for a significant number of arrhythmia recurrences after direct ablative operations. We used total computer-assisted cardiac mapping (epicardial sock, left and right ventricular endocardial balloon electrode arrays) to assess whether tachycardia originating in deep or right-sided layers of the interventricular septum is associated with a specific pattern of epicardial activation sequence. We performed these studies during operations in 18 patients and during experiments in 12 dogs in which a septal myocardial infarction was produced by ligating the anterior septal coronary artery. Intraseptal needle electrodes were plunged into the septum of all animal preparations to generate pace-mapping data and to obtain intraseptal recordings (six preparations) during reentrant ventricular tachycardia induced by programmed stimulation. In addition, pace-mapping data of infarcted canine heart preparations were compared with those of nine healthy heart preparations. In the clinical study, 31 ventricular tachycardias with a septal site of origin were analyzed. Twenty tachycardias displayed an epicardial breakthrough in the area of the interventricular groove, whereas 11 had an epicardial breakthrough in the right ventricular free wall. Biventricular endocardial mapping revealed that left septal endocardial activation preceded right septal activation in the former and that right septal activation occurred earlier in the latter. In the experimental study, 14 ventricular tachycardias (cycle length 146 +/- 34 msec) were induced by programmed stimulation in 11 infarcted heart preparations. Eight tachycardias displaying an epicardial breakthrough on the right ventricle were found to originate in the right ventricular septal subendocardial layers, whereas six tachycardias in which the epicardial breakthrough occurred on the anterior interventricular groove originated in the left ventricular septal subendocardial layers. The epicardial breakthrough preceded the left ventricular endocardial breakthrough in six tachycardias (85.7%) originating in intermediate or right ventricular septal layers, but in only one of five tachycardias originating in the left ventricular septal layers. In the pace-mapping study, the epicardial breakthrough shifted progressively from the right ventricular free wall toward the interventricular groove area in response to pacing from the right, intermediate, and left ventricular thirds of the basal septum. This relationship was similar for infarcted and noninfarcted hearts, although transseptal conduction time was prolonged in infarcted hearts (45 +/- 10 msec vs 33 +/- 7 msec, p < 0.01). Therefore the information integrated from the localization of the epicardial breakthrough and the relative timing between the epicardial and the left ventricular endocardial breakthroughs can be used to estimate the depth of the site of origin of septal ventricular tachycardias. This study confirms that a three-dimensional view of the substratum of ventricular tachycardia can be derived from simultaneous epicardial and left ventricular endocardial mapping and can provide a superior basis for therapeutic interventions.

Distinct activation patterns of idioventricular rhythms and sympathetically-induced ventricular tachycardias in dogs with atrioventricular block

To investigate mechanisms of ventricular impulse formation in response to sympathetic stimulation in the healthy canine heart in situ, we compared the patterns of ventricular activation during the idioventricular rhythms arising after complete atrioventricular (AV) block and ventricular tachycardias induced by RSG or LSG stimulation. Isochronal maps were generated by computer from 116-127 unipolar electrograms recorded from the entire ventricular epicardium in 15 open chest, anesthetized dogs. In eight of these, bipolar electrograms were recorded with plunge electrodes from 11 selected endocardial sites located below epicardial breakthrough areas. Intracardiac recordings from the His-Purkinje system were made with electrode catheters. After electrograms were recorded during sinus rhythm, complete AV block was induced by injecting formaldehyde into the AV node and idioventricular rhythms occurred spontaneously at a rate of 37 +/- 12 beats/min (mean +/- SD, n = 25). During idioventricular rhythms, endocardial activation preceded the earliest epicardial breakthrough, which occurred in either the right anterior paraseptal region, antero-apical left ventricle, or postero-apical left ventricle. These sites were consistent with a focal origin in the subendocardial His-Purkinje system. Total epicardial activation times lasted for 47 +/- 13 msec (n = 40). Idioventricular rhythms were suppressed by overdrive pacing (intermittent trains of ten beats with decremental cycle length from 500 to 200 msec) or by intravenous calcium infusion (to plasma levels of 10.1-15.2 mM). Right or left stellate ganglion stimulation increased idioventricular rhythm rates (to 52 +/- 13 beats/min, n = 28) and also induced, in all preparations, ventricular tachycardias that had significantly faster rates (189 +/- 55 beats/min, n = 27, P less than 0.005). Ventricular fibrillation was induced after brief runs of ventricular tachycardia in five of the preparations. During ventricular tachycardias, epicardial activation occurred on the right ventricular outflow tract or the postero-lateral wall of the left ventricle, and preceded endocardial activation in 50% of cases. Total epicardial activation times (103 +/- 29 beats/min) were significantly longer than during idioventricular rhythms (P less than 0.005). Ventricular tachycardias displayed overdrive excitation at critical pacing cycle lengths (360-280 msec) and were not suppressed by calcium infusion. Thus, differential mechanisms of impulse formation with distinct localizations can be elicited from healthy ventricular myocardium.

Interpolating unipolar epicardial potentials from electrodes separated by increasing distances

In cardiac mapping, potentials for unexplored areas are estimated by interpolating values from nearest neighbor electrodes regardless of distances between these sites or wave front orientation. The effects of these variables on interpolated unipolar electrograms were analyzed two ways: with a computer model and with electrograms recorded 9.9 and 14.1 mm apart. For the model, wave fronts (n = 39) were generated from electrograms recorded during right ventricular (RV) activation in five dogs following the RV isolation procedure. Each wave front was assumed to propagate radially at 0.5 m/sec from a site 30 mm from the center of a square array with electrodes located at the center and corners. Each wave front crossed the array with its tangent at an angle of 0 degrees, 45 degrees, or 90 degrees to the diagonal line connecting opposite corner electrodes. Potentials for all five sites were generated from each wave front and were interpolated for the center site from the generated corner potentials. Generated and interpolated center site potentials were compared using correlation coefficients (r) and percent root mean square differences (%RMSD). Mean r values fell below 0.90 for interelectrode distances of 15.6 mm, 2.8 mm, and 1.4 mm at 0 degrees, 45 degrees, and 90 degrees wave front orientations, respectively. For experimentally measured potentials recorded 9.9 mm apart, results from interpolated electrograms were similar to results from the model at 0 degrees propagation. Electrograms interpolated from potentials measured 14.1 mm apart had poorer r and %RMS values than those from the computer model. Thus, with linear interpolation unipolar electrograms can be inaccurately interpolated from electrodes less than 3 mm apart or correctly interpolated from electrodes more than 14 mm apart depending upon wave front orientation.

The assumptions of isochronal cardiac mapping

Isochronal maps of cardiac activation are commonly used to study the mechanisms and to guide the ablative therapies of arrhythmias. Little has been written about the assumptions implicit in the construction and use of isochronal cardiac maps. These assumptions include the following: (1) the location of the recording electrodes is known with sufficient accuracy to determine the mechanism of an arrhythmia or to guide therapy; (2) a single, discrete activation time can be assigned to each recording electrode location; (3) the presence or absence of activation at an electrode site can be reliable ascertained, and when activation is present, the time of activation can be determined with sufficient accuracy to specify the mechanism of an arrhythmia or to guide therapy; and (4) the recording electrodes are close enough together that the activation sequence can be estimated with sufficient accuracy to determine the mechanism of an arrhythmia or to guide therapy. The manuscript reviews evidence that these assumptions may not always be true, and when they are not, the isochronal map may be misleading.

Transmural activations and stimulus potentials in three-dimensional anisotropic canine myocardium

Epicardial and endocardial pacing are widely used, yet little is known about the three-dimensional distribution of potentials generated by the pacing stimulus or the spread of activation from these pacing sites. In six open-chest dogs, simultaneous recordings were made from 120 transmural electrodes in 40 plunge electrodes within a 35 X 20 X 5-mm portion of the right ventricular outflow tract during epicardial and endocardial pacing at a strength of twice diastolic threshold and at 1 mA. The magnitude of extracellular potentials generated by the stimulus and the activation times were compared in regions proximal (less than 10-12 mm) and distal to the pacing site. Local fiber orientation was histologically determined at each recording electrode. For endocardial pacing, endocardial potentials were larger than epicardial potentials only in the proximal region (p less than 0.001); while in the distal region, epicardial potentials were larger (p less than 0.001), and endocardial activation occurred earlier than epicardial activation for both regions (p less than 0.001). For epicardial pacing, epicardial potentials were larger than endocardial potentials in both regions (p less than 0.001), and epicardial activation occurred earlier only in the proximal region (p less than 0.02), while endocardial activation occurred before epicardial activation in the distal region (p less than 0.01). In planes of recording electrodes parallel to the epicardium and endocardium, the initial isochrones were elliptical with the major axes of the ellipses along the mean fiber orientation between the pacing site and recording plane rather than along the local fiber orientation in the recording plane. Thus, the ellipses in each plane rotated with respect to each other so that in three dimensions the activation front was helicoid, yet the twist of the helix was less than that of the corresponding transmural rotation of fibers. For pacing from the right ventricular outflow tract, we conclude that beyond 10-12 mm from endocardial and epicardial pacing sites epicardial stimulus potentials in both cases are larger than endocardial potentials because of resistivity differences inside and outside the heart wall and activation in both cases is primarily endocardial to epicardial because of rapid endocardial conduction, and we conclude that the initial spread of activation is helicoid and determined by transmural fiber direction.

Effects of distant potentials on unipolar electrograms in an animal model utilizing the right ventricular isolation procedure

The effects of distant potentials on local epicardial unipolar electrograms were examined utilizing a model that enabled both ventricles to be paced independently in five dogs. The right ventricular isolation procedure electrically isolates the right from the left ventricle. Right ventricular electrograms were separated into their local (right ventricular) and distant (left ventricular) components by altering the left-right ventricular pacing interval. Waveform configuration, peak to peak amplitude, magnitude of the slope and timing of the fastest downstroke were carefully evaluated at each electrode site, both with and without the presence of distant left ventricular potentials. Except for the timing of the fastest downstroke, all of these variables were significantly altered by distant potentials. Although the slope of the fastest downstroke was significantly affected by distant potentials, it remained a sensitive indicator of local versus distant activation. All electrograms of local right ventricular activation had a slope magnitude greater than 2.5 mV/2 ms whereas none of the right ventricular electrograms containing only distant left ventricular activity had a magnitude greater than 2.5 mV/2 ms. Computer-generated electrograms were calculated by digitally summing the recorded local right and distant left ventricular components. The simulated electrograms correlated well with the recorded electrograms during synchronous ventricular pacing. Thus, the configuration, amplitude and slope of unipolar electrodes were profoundly influenced by distant potentials. The timing of the fastest downstroke is largely independent of the effect of distant potentials and most closely represents local activation. The magnitude of the slope of the recorded electrogram accurately distinguishes local from distant activation.

Mapping of ventricular tachycardia induced by programmed stimulation in canine preparations of myocardial infarction

To investigate the mechanism of uniform ventricular tachycardia induced by programmed stimulation, we recorded His bundle electrograms and unipolar electrograms from 64 subepicardial, subendocardial, and intramural sites in dogs. Isochronal maps were generated off-line by computer. Two groups of dogs were studied 3 days after occlusion of their left anterior descending coronary arteries; one group underwent reperfusion after 2 to 2.5 hr of occlusion and the other methylprednisolone treatment before permanent occlusion. In the former, subepicardial sequences presented either a pattern suggesting circus movement or a radial pattern in which excitation at intramural sites could precede earliest subepicardial excitation. In the latter preparations, subepicardial excitation patterns consistently suggested circus movement in the subepicardial muscle layer surviving over necrotic tissue. Assuming complete circus movement, the "missed" time interval, measured as the interval left unaccounted for by actual recording of local excitation between ventricular tachycardia cycles, ranged from 3% to 64% of the cycle length of ventricular tachycardia. While surviving subepicardial and intramural layers appeared to be involved in the mechanism of ventricular tachycardia, a late second breakthrough on the right ventricle, in conjunction with fixed-coupled H deflections on the His bundle electrograms, suggested the involvement of the conducting system in propagation of the impulse.

Identification of reproducible ventricular tachycardia in a canine model

Epicardial mapping was used as a standard to investigate how well the limb leads, both alone and in conjunction with 5 select epicardial electrodes, can verify reproducibility in a common, open-chest canine model of ventricular tachycardia (VT). Reproducible VT was defined as 2 or more episodes of monomorphic VT with similar rates, limb lead tracings and epicardial maps. In this study, 21 dogs underwent 2-hour occlusion of the left anterior descending coronary artery followed by reperfusion. Three days later, programmed stimulation was used to induce VT that was analyzed with limb leads I, II and III and 27 simultaneously recorded, bipolar epicardial electrodes. Thirteen dogs had VT of which 11 had polymorphic VT (varying QRS morphology). Twelve dogs yielded at least 1 form of monomorphic VT. Eight had 2 or more distinct forms of monomorphic VT (pleomorphism). Four of these 8 dogs had pleomorphic VT that was not apparent from the limb lead tracings, but was recognized from the epicardial activation maps constructed from the 27 epicardial recordings. To provide a method of distinguishing various VTs without the need of full epicardial mapping, 5 of the 27 epicardial electrodes were selected. These were positioned over the midanterior and midposterior right and left ventricles, and the left ventricular apex. By analyzing electrogram morphology and activation time, VT reproducibility could be as accurately identified with these 5 electrodes as with epicardial mapping derived from 27 electrodes. In conclusion, multiple VT morphologies are common in this open-chest canine model. Limb lead recordings alone are inadequate for analysis of VT reproducibility.(ABSTRACT TRUNCATED AT 250 WORDS)