- Email: [email protected]
Computerized potential distribution mapping: A new intraoperative mapping technique for ventricular tachycardia surgery
Computerized Potential Distribution Mapping: A New Intraoperative Mapping Technique for Ventricular Tachycardia Surgery Atsushi Harada, MD, James S. Tweddell, MD, Richard B. Schuessler, PhD, Barry H. Branham, MS, John P. Boineau, MD, and James L. Cox, MD Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, Barnes Hospital, St. Louis, Missouri
This study evaluated potential distribution mapping as a method for localizing the site of origin of ventricular tachycardia (VT). In contrast to conventional activation time maps, potential distribution maps require less editing and thus can be more automated and rapidly processed for interpretation of multiple beats of VT. As a series of potential distribution maps during VT is required for detailed analysis, an on-line computerized system was designed to display potential distribution maps sequentially at I-ms intervals as a color movie. Potential distribution maps and activation time maps were constructed from 182 epicardial and endocardial unipolar electrodes during 12 episodes of reproducible monomorphic VT in 9 dogs four to six days after experimental myocardial infarction (mean cycle length, 162 f 21 ms). At the onset of each depolarization during VT, a
I
ntraoperative mapping to localize the site of origin of ventricular tachycardia (VT) guides surgeons in determining the appropriate type of operative procedure and the region to which it should be applied. During the past two decades, activation mapping has been the exclusive method used to localize the site of origin of these arrhythmias [I, 21. However, the construction of activation time maps (ATMs)frequently requires editing to derive activation times from complex or questionable electrograms. This editing is time-consuming and requires considerable experience. In addition, ATMs only demonstrate local activity at the sites of the recording electrodes. Potential distribution mapping is another method to demonstrate the site of origin of VT [3-51. Studies involving VT simulated by rapid ventricular pacing have demonstrated that the intramural pacing site can be localized from the endocardial and epicardial surfaces with a high degree of accuracy [4, 51. This method has several advantages over ATMs. (1)Less editing is required to construct the isopotential map because measurement of potential is unambiguous. (2) Once the baseline is determined, no additional analysis time is required. Therefore, multiple beats of VT can be rapidly displayed. (3) Because unipolar Accepted for publication Dec 27, 1989. Address reprint requests to Dr Cox, Division of Cardiothoracic Surgery, Washington University School of Medicine, One Barnes Hospital Plaza, Suite 3108 Queeny Tower, St. Louis, MO 63110.
0 1990 by The Society of Thoracic Surgeons
potential minimum abruptly developed on the surviving epicardium and another on the surviving endocardium of the left ventricle, both immediately adjacent to the subendocardial infarct. These two minima on the initial potential distribution maps corresponded to the sites of earliest epicardial and endocardial activation breakthrough recorded on the activation time maps. These two minima subsequently expanded or moved into the adjacent area and coincided with the spread of activation fronts on the epicardial and endocardial surfaces. Thus, the rapid display of sequential, computerized potential distribution maps of multiple beats of VT provides a dynamic means of identifying the site of origin of VT, and therefore should facilitate intraoperative mapping. (Ann Thoruc Surg 2990;49:649-55)
potentials reflect distant activity, intramural sites of origin of VT may be detected from epicardial or endocardial recording electrodes. However, potential distribution mapping has not come into wide use because a series of maps is required to demonstrate the activation sequence. To overcome this disadvantage of potential distribution mapping and make it clinically feasible, an on-line computerized system was designed to display potential distribution maps sequentially as a movie. The purpose of this study was to determine if potential distribution mapping could provide a rapid and dynamic means of localizing the site of origin of VT in dogs with experimental myocardial infarction.
Material and Methods Definitions and Descriptions 1. Potential distribution maps (PDMs): A PDM constructed from unipolar electrograms at a specific instant in the cardiac cycle demonstrates the distribution of the potential fields. Reference baseline for calculation of the potential at each instant was obtained from the average of five consecutive samples (5 ms) in the flat portion of the late diastolic segment during normal sinus rhythm. A series of maps is required to demonstrate progressive changes of the potential distributions. 0003-4975/90/$3.50
650
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
Potential minimum: A potential minimum is the nadir of the negative potential area at an instant in the cardiac cycle. Primary potential minimum: A primary potential minimum is the earliest potential minimum reaching -3 mV in a series of PDMs [4, 51. Potential maximum: A potential maximum is the zenith of the potential area at an instant in the cardiac cycle. Activation time maps: An ATM is displayed as a sequence of different activation time zones. A computer program is used to determine local activation times from unipolar tracings. The peak negative derivative (-dV/dt,,,) of the major deflection of the unipolar complex is the time of local activation.
Animal Preparation Seventeen adult mongrel dogs of either sex weighing 18 to 25 kg were anesthetized with intravenous sodium pentobarbital (30 mg/kg), and mechanical ventilation was maintained by delivering 40% oxygen from a volume-cycled respirator (Bennett MA-1) through a cuffed endotracheal tube. After antibiotic prophylaxis (600,000 U of penicillin G administered intramuscularly), the chest was opened through a left thoracotomy in the fourth intercostal space and the heart was suspended in a pericardial cradle. The left anterior descending coronary artery was isolated immediately proximal to the first diagonal branch. A snare was lowered gradually onto a suture surrounding the artery, and partial occlusion was maintained for 20 minutes, followed by complete occlusion. The snare was released after two hours of total occlusion, and the myocardium was reperfused. Lidocaine hydrochloride (2 mg/ kg) was administered intravenously immediately before coronary occlusion and just before reperfusion. Then, the chest was closed in layers and the dogs were allowed to recuperate with only postoperative antibiotics given (600,000 U/day of penicillin G administered intramuscularly). Four to six days after myocardial infarction, the animals were reanesthetized and ventilated in the same fashion as described. A polyethylene catheter for monitoring arterial pressure was placed in the left femoral artery. A right thoracotomy was performed through the fifth intercostal space. Both venae cavae were encircled with tapes for control of these vessels, and the azygos vein was ligated and divided. The heart was exposed and suspended in a pericardial cradle. A sanguinous priming solution and pediatric oxygenator (Model H-400, William Harvey) were used for cardiopulmonary bypass. After heparinization (0.1 mg/kg) and direct cannulation of both venae cavae, normothermic cardiopulmonary bypass was instituted at a flow rate adequate for maintenance of a mean atrial pressure of 60 to 80 mm Hg. The interatrial groove was dissected to expose the septa1 portion of the left atrium. The heart was arrested with electrically induced ventricular fibrillation, and a standard left atriotomy was performed. A three-dimensional mold containing multiple endocardial electrodes was introduced into the left ventricle through the left atriotomy, thus avoiding the need for a left ventriculotomy. After a right atriotomy, a
Ann Thorac Surg 1990:49649-55
shaped, multielectrode endocardial mold was introduced into the right ventricular (RV) cavity. Sinus rhythm was restored by direct-current cardioversion. The entire epicardial surface of the heart was covered with a multielectrode epicardial mold. A bipolar electrode was sewn on the right ventricular outflow tract for both introducing ventricular tachycardias and recording reference electrograms. A common reference electrode for all cardiac unipolar electrograms was sewn on the right chest wall at a distance of at least 15 cm from the heart. Placement of the electrode in this position minimized electrical noise. Arterial blood gases and serum hematocrit levels were monitored at 15-minute intervals throughout each study. Sodium bicarbonate was administered as needed to maintain an acid-base equilibrium within normal limits at all times. Serum potassium levels were maintained between 3.7 and 4.5 mEqL.
Electrophysiological Studies Under normothermic total cardiopulmonary bypass, programmed electrical stimulation was performed to induce ventricular tachyarrhythmias using a programmable stimulator (model DTV-101, Bloom). Pacing was performed at a pulse width of 2 ms at twice late diastolic threshold current. A train of eight paced beats (S,s) at a basic cycle length of 300 ms was delivered, after which double premature stimuli ( S , and S,) were introduced at varying interval was lowered by coupling intervals. The ,S,S 5-ms decrements until noncapture of S, resulted, at which time the S,-S, coupling interval was narrowed by 5-ms decrements. This sequence was repeated until either the heart became refractory to the S, stimulus, ventricular fibrillation ensured (followed by direct-current cardioversion), or polymorphic or monomorphic ventricular tachycardia appeared. If the S, stimulus consistently failed to capture the ventricle, the pacing protocol was repeated at a basic cycle length of 250 ms. If ventricular tachycardia was not induced at one site, several alternative pacing sites were used.
Epicardial and Endocardia1 Electrodes and Mapping System Sixty-one silver bead unipolar electrodes, 2 mm in diameter, were positioned (15 mm apart) in 12 arrays on the inside of an epicardial mold (Fig 1A). Based on the epicardial structure of a normal canine heart, this mold was made of 5-mm thick foam rubber so that it could envelop the entire epicardial surface of the ventricles. The elasticity of the foam rubber maintained optimal contact between the unipolar electrodes and the epicardial surface during both systole and diastole. The endocardial molds for the left and right ventricles were also made of foam rubber. The RV mold, containing 60 unipolar electrodes, was designed to conform to the irregular internal anatomy of the RV cavity (Fig 1B). The right ventricular electrode was introduced into the RV cavity through the tricuspid valve annulus. The left ventricular mold (Fig 1C) was ellipsoid, 6 cm long and 3 cm across, with 61 unipolar electrodes evenly mounted (10 mm apart) in 12 arrays on its surface. Under normother-
HAR4DAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49649-55
mic cardiopulmonary bypass during electrically induced ventricular fibrillation, these electrode-containing molds were introduced into the RV and left ventricular (LV) cavities through the tricuspid and mitral valve annuli. The elasticity of these endocardial molds maintained optimal contact between the electrodes and the endocardial surface during both systole and diastole. These epicardial and endocardial molds standardized 182 unipolar electrode positions on the epicardial and endocardial surfaces (Fig 2). All signals from each unipolar electrode were connected to the differential amplifiers at an input impedance of lo'* ohms and a frequency response of 0.05 to 1,000 Hz. Data were recorded in three sets of 61 electrograms, in addition to the lead I1 electrocardiogram and bipolar reference electrode electogram. Because the current mapping system was capable of simultaneously recording 64 channels, three time-aligned sets of data using the bipolar reference electrogramand lead 'I were required to construct the entire epicardial and endocardial maps from 182 unipolar electrograms. Time-aligned analysis for
B
C
Fig 1. Epicardial and endocardial molds containing 182 2-rnm silver bead unipolar electrodes. These molds were made of foam rubber to conform to the epicardial and endocardial structures of the canine heart. (A) Epicardial mold containing 61 electrodes. (B)Right ventricular endocardial mold containing 60 electrodes. (C) Left ventricular endocardial mold containing 61 electrodes.
651
CHEST
pot
Fig 2. One hundred and eighty-two epicardial and endocardial electrode positions standardized by the epicardial and endocardial molds. (Ant. = anterior; END0 = endocardium; EPI = epicardiurn; FREE = free wall; LV = left ventricle; Post. = posterior; RV = right ventricle; SEP = septum,)
polymorphic VTs was impossible because of the changing rate and morphology of these arrhythmias. Therefore, the analysis of polymorphic VTs was excluded from this study. However, with a newer 250-channel data acquisition system, polymorphic VT can be analyzed using this technique without any additional processing beyond what is needed to analyze monomorphic tachycardias. A @-channel analog-to-digital converter digitized the unipolar data at a rate of 1,000 samplesk. A PDP 11/23PLUS computer stored the data and displayed the waveforms. The input range of the analog-to-digital converter was f 10 V. An amplifier gain of 500 was used for giving an input signal range of f 20 mV. Potential distribution maps for a selected window before and during early ventricular depolarization were automatically produced from the computer analysis of all 182 unipolar potentials referenced to the baseline T-P segments of each electrogram. If no baseline could be determined owing to the occurrence of a rapid VT, two other techniques for determining baseline were tested. Baseline was determined from the T-P segment during normal sinus rhythm for each electrogram, and the baseline was also determined for each electrogram by shorting the inputs of each amplifier to ground. Evaluation of PDMs during VT using all three techniques gave equivalent results. In addition, ATMs were similarly produced from the activation times derived from the same unipolar signals. Both ATMs and PDMs were displayed as smooth, computer-generated isochronous or isopotential maps. As construction of contour lines required considerable time, we used a more rapid "tile map" technique [ l l ] intraoperatively in which each of the electrode points divides the surface of the heart into polygonal regions. Each polygon tile color represented a potential level. More than 1,000 (1,000 ms) PDMs were generated with tile map technique within five minutes after acquisition of the data. SubsequentlYg a movie of the tile map was displayed by updating the color map for each electrode potential at each millisecond.
652
HARADAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49649-55
Fig 3. Epicardial and endocardial potential distribution maps during monomorphic ventricular tachycardia. Each panel (A-F) illustrates the computer-generated color isopotential maps of the epicardial and endocardial surfaces. The standard lead I1 echocardiograph (ECG) and bipolar reference electrogram (REF) are displayed in the left upper corner of each panel. The first, second, and third QRS complexes are s,, s,, and s, of a paced train, and the fourth QRS complex is the first beat of the induced rnonomorphic ventricular tachycardia. The vertical line and number on the ECG tracing indicate the time of each potential distribution map from the beginning of the ECG window. The scale is shown at the bottom of the figure. Negative potentials are idicated by blue areas, positive potentials by red areas, and near-neutral potentials by gray areas.
At the end of each experiment, the heart was excised with the endocardial and epicardial electrode molds still in place. The exact locations of the pacing and recording electrodes were verified. The heart was then sliced and stained with biphenyl tetrazolium chloride to define the extent of the infarction and the relationship to the recording electrodes. All animals received humane care in compliance with the ”Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the ”Guide for the Care and Use of Laboratory Animals” (NIH publication No. 85-23, revised 1985). In addition, the experimental protocol was reviewed and approved by the Washington University Committee for the Humane Care of Laboratory Animals.
Results Characteristics of Induced Ventricular Tachycardia Seventeen dogs were studied four to six days after myocardial iniarction. Monomorphic VT was induced in 9 of 17 dogs. Six of these dogs had one monomorphic VT and 3 dogs had two morphologically distinct VTs. The mean VT cycle length was 162 f 21 ms (range, 121 to 205 ms). These VTs terminated spontaneously or were terminated by premature ventricular pacing two to five minutes after onset. In 3 of the remaining 8 dogs, polymorphic VT was
induced and deteriorated into ventricular fibrillation. No VT was induced in 5 dogs. All 12 episodes of monomorphic tachycardia demonstrated an epicardial reentry pattern in regions overlying the infarct as determined by ATMs and PDMs.
Potential Distribution Maps During Monomorphic Ventricular Tachycardia Because of the similarity of results, a series of PDMs during monomorphic VT will be presented for a single dog (Fig 3). In the standard lead I1 electrocardiographic (ECG) tracing, the first, second, and third QRS complexes are S,, S,, and S, of the paced train; and the fourth QRS complex is the first beat of the induced monomorphic VT. The vertical line and number on the ECG tracing indicate the time of each PDM from the beginning of the ECG window. The potential scale is displayed at the bottom of the figure. Blue regions indicate negative potentials and red regions indicate positive potentials. Near-neutral potentials are indicated by the gray areas. Figure 3A shows the potential distributions at 820 ms after the beginning of the ECG window. At this instant, the majority of the epicardial and endocardial unipolar electrograms were in the repolarization phase (ie, T wave) of the second beat of the induced VT. Therefore, the potential minima are the results of the negative T waves present on the lateral LV epicardium and on the base of the LV and
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49:649-55
ECC
0 mr
EC(
Fig 4. Activation maps recorded from the same animal and during the same beat as in Figure 3. Zero milliseconds for reference time of the activation maps is selected at 800 ms from the beginning of the echocardiographic ECG) window. The earliest activation site is the paraseptal area and corresponds to the site of the prima y potential minimum on the potential distribution maps in Figure 3B. The shaded area on the left ventricular endocardium represents the infarction zone. (REF = bipolar reference electrogram.)
RV endocardium. The potential maxima are the results of the positive T waves observed on the epicardium of the RV and on the outflow tract of the RV endocardium. Later on, with the completion of local repolarization, these minima and maxima formed by the T waves slowly diminished the negative and positive areas. At 835 ms (Fig 3B), the beginning of the third QRS complex of the induced VT, a primary potential minimum abruptly developed in the paraseptal area of the epicardium just adjacent to the subendocardial infarction. At the same instant, there was little change in the epicardial and endocardial potential distributions except for the disappearance of a potential maximum on the RV epicardium. At 854 ms (Fig 3C), the potential minimum on the paraseptal epicardium expanded parallel to the left anterior descending artery. At this instant, the endocardial potential distribution suddenly changed, with a primary potential minimum of the RV endocardium developing anteriorly at the outflow tract. There was little change on the LV endocardium. At 860 ms (Fig 3D), while both epicardial and RV endocardial minima were expanding, another primary potential minimum of the LV endocardium abruptly developed on the LV septum adjacent to the preexisting potential maximum. This site of LV potential minimum was just adjacent to the subendocardial infarction. Subsequently, all epicardial and endocardial potential minima expanded or moved into the adjacent areas, with developing positive areas adjacent to these minima (Figs 3E, 3F). This pattern repeated for successive beats of VT.
Comparison Between Potential Distribution Maps and Activation Time Maps On the ATMs (Fig 4)recorded from the same animal and
653
during the same beat as that shown in Figure 3, the earliest activation breakthrough was observed on the paraseptal area of the epicardium including the site of epicardial primary potential minimum in the PDMs (Fig 3B). Right ventricular and LV endocardial activation breakthroughs also corresponded to the sites of endocardial primary potential minima following the epicardial primary potential minimum (Fig 4 vs Figs 3C, 3D). A comparison of ATMs and a series of PDMs showed that the potential minima during ventricular depolarization expanded or moved into the adjacent areas, concomitant with the spread of activation fronts at the epicardial and endocardial surfaces (Fig 4 vs Figs 3B-F). For example, a comparison of Figure 3E and Figure 4 shows that the demarcation lines of the negative potent'-l fields in the PDMs at 880 ms are in agreement with thc .
Comment The advent of direct surgical procedures for the treatment of refractory ischemic VT has led to a substantial improvement in the treatment of patients with these lifethreatening arrhythmias [6]. As electrophysiologically guided operation has resulted in a low postoperative inducibility and recurrence of VT, intraoperative electrophysiological mapping is essential for definitive surgical treatment [7, 81. The primary objective of intraoperative mapping is to localize the site of origin of VT, because ablation of the earliest activation site has been shown clinically to interrupt VT [7, 81. During the past two decades, epicardial, endocardial, and intramural activation time mapping have been used exclusively for identifying such sites of origin. Hand-held electrodes have been used to record multiple electrograms sequentially to derive the local activation time. However, the ATMs constructed by this technique are less accurate in demonstrating the activation sequence, because these are composite isochrones constructed from local activation times in different beats of VT. More accurate maps can be constructed using multiple simultaneous electrical recording systems. With the use of an increasing number of electrodes and amplifiers, a computer system is required to process the data rapidly [9, 101. However, the construction of ATMs frequently requires editing to derive activation times from complex or questionable electrograms. This editing is time-consuming and requires considerable experience. Additionally, intuitive judgment is used in interpreting the activation time data, whether they are displayed numerically or as isochrones. As a result of the time required to process the data, usually only one beat of the tachycardia is analyzed, precluding the use of map-
654
HARADAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49:649-55
A
\ Anterior
Rv
LV
Posterior B Fig 5. Photograph (A) and diagram ( B ) of a cross-section of the canine heart from which data are presented in Figures 3 and 4. (A) The section is stained with 2,3,5-triphenyl tetrazolium chloride solution. In (B), the shaded area represents necrotic tissue from myocardial infarction. The black arrow indicates the site of the epicardial primary potential minimum demonstrated in Figure 3B. The white arrow indicates the site of the left ventricular primary potential minimum in Figure 3 0 . (LV = left ventricle; RV = right ventricle.)
guided operation in polymorphic VT. In contrast to ATMs, PDMs require less editing to construct the isopotential contours, because the measurement of potential distributions is unambiguous, and require no additional analysis time to examine multiple beats of the tachycardia. Therefore, potential distribution mapping can be more automated and rapidly processed for interpretation. Another theoretical advantage is that unipolar potential maps can reflect distant activation events. Thus, epicardial and endocardial maps could theoretically detect early intramural wavefronts, whether in the ventricular free walls or septum, before the actual emergence of activation at these surfaces. Previous data on paced intramural
ventricular beats indicate that not only the site but also the intramural depth could be approximated from analysis of the PDMs from the ventricular surfaces bordering the site of stimulation [5]. With uniform conduction, the emergence of the wavefront would be expected to be at the site on the endocardial or epicardial surface closest to the point of origin. The presence of nonuniform activation due to the infarction or unidirectional block during reentrant ventricular rhythms could spatially dissociate the site of intramural depolarization and endocardial or epicardial breakthrough sites. Because PDMs reflect distant and local activity, electrode spacing is not as critical as it is for ATMs, which only represent local events. In addition, potential distributions recorded in areas of infarcted tissue (or areas with overlying thick fat pads) provide information to localize the origin of the VT.When constructing ATMs, these areas provide little information as no activation time can be assigned. In this situation, PDMs could prove to localize the intramural origin more accurately than endocardial or epicardial activation maps as the surface potentials would reflect intramural and distant activation. Epicardial, endocardial, and intramural potential distribution during ventricular pacing have been studied previously by several investigators [3-51. It has been well documented that the earliest potential minimum develops in an area close to the pacing site (earliest activation site) after the stimulation and that this minimum subsequently expands and moves to a distant area as it increases in negativity. Thus, the earliest potential minimum with low amplitude should represent the area closest to the earliest activation site [5]. We previously demonstrated that the earliest potential minimum reaching -3 mV is defined as a primary potential minimum that corresponds to the earliest activation site. Using the same definition for a primary potential minimum in this study, potential distribution mapping was evaluated as a method to localize the site of origin of VT. As described in the Results section, the epicardial primary potential minimum occurred above the subendocardial infarction and the LVendocardial primary potential minimum occurred adjacent to that infarction. These minima corresponded to the sites of activation breakthrough on the ATMs. Because an instant of -dV/dt,,, in the unipolar electrogram (activation time) usually did not coincide with the instant reaching -3 mV (primary potential minimum), activation breakthrough and primary potential minimum did not appear simultaneously. However, regardless of whether primary potential minimum precedes or follows activation breakthrough, it localizes the earliest activation site. Therefore, potential distribution mapping can serve the primary purpose of intraoperative mapping to localize the site of origin of VT. Furthermore, breakthrough on the ATMs appeared as a more widespread area than the primary potential minimum. One problem with viewing the PDMs statically is that there are often large areas of negativity. These areas are associated with the T waves, and when the maps are viewed dynamically in rapid succession, they are easily differentiated from the negativity associated with activation because they change much more slowly. However,
Ann Thorac Surg 1990;49:649-55
the potential distribution mapping technique could be improved if the T waves were filtered out of the signal. Spach and Barr [3] demonstrated that the demarcation line separating positive and negative potential (0 mV) shows good agreement with the location of the activation excitation wave. Our study demonstrated that the boundary of the negative potential field less than -3 mV at each instant coincides with the location of the activation wave. Rapid display of a series of PDMs could visually demonstrate the dynamic change of the activation fronts during VT. Therefore, we designed an on-line computer system to display PDMs sequentially as a movie. In conclusion, computerized potential distribution mapping provides a rapid means for the mapping of multiple beats of VT for the purpose of identifying the sites of origin of VT and also a dynamic means of demonstrating the activation sequence of these arrhythmias. Therefore, this intraoperative mapping technique should facilitate intraoperative mapping in patients undergoing operations for ventricular tachycardia. This work was supported by NIH grants RO1 HL33722 and RO1 HL32257. We thank George Quick, Michael M. Lischko, and Barbara E. Doerr for their excellent technical assistance, and Dawn Schuessler for the preparation of the manuscript.
References 1. Wittig JH, Boineau JP. Surgical treatment of ventricular arrhythmias using epicardial, transmural, and endocardia1 mapping. Ann Thorac Surg 1975;20:117-26.
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
655
2. Gallagher JJ, Kasell JH, Cox JL, Smith WM, Ideker RE, Smith WM. Techniques of intraoperative electrophysiologic mapping. Am J Cardiol 1982;49:221-40. 3. Spach MS, Barr RC. Analysis of ventricular activation and repolarization from intramural and epicardial potential distributions for ectopic beats in the intact dog. Circ Res 1975; 37830-43. 4. Taccardi B, Arisi G, Macchi E, Baruffi S, Spaggiari S. A new intracavitary probe for detecting the site of origin of ectopic ventricular beats during one cardiac cycle. Circulation 1987; 75:272-81. 5. Harada A, DAgostino HJ Jr, Schuessler RB, Boineau JP, Cox JL. Potential distribution mapping: a new method for precise localization of the intramural septa1 origin of ventricular tachycardia. Circulation 1988;78(Suppl 3):137-47. 6. Cox JL. Anatomic-electrophysologic basis of surgical treatment of refractory ischemic ventricular tachycardia. Ann Surg 1983;198:119-29. 7. Moran JM, Talano JV, Euler D, Moran JF, Montoya A, Pfiarrk R. Refractory ventricular arrhythmia: the role of intraoperative electrophysiological study. Surgery 1977;82:809-15. 8. Mason JW, Stinson EB, Winkle RA, et al. Surgery for ventricular tachycardia: efficacy of left ventricular aneurysm resection compared with operation guided by electrical activation mapping. Circulation 1982;65:114&55. 9. Witkowski FX, Corr PB. An automated simultaneous cardiac mapping system. Am J Physiol 1984;247:H661-8. 10. Downar E, Parson ID, Mickelborough LL, Cameron DA, Yao LC, Waxman MB. On-line epicardial mapping of intraoperative ventricular arrhythmias: initial clinical experience. JACC 1984;4:703-14. 11. Green PJ, Sibson R. Computing dirichlet tesselations in the plane. Comp J 1978;21:16&73.
This study evaluated potential distribution mapping as a method for localizing the site of origin of ventricular tachycardia (VT). In contrast to conventional activation time maps, potential distribution maps require less editing and thus can be more automated and rapidly processed for interpretation of multiple beats of VT. As a series of potential distribution maps during VT is required for detailed analysis, an on-line computerized system was designed to display potential distribution maps sequentially at I-ms intervals as a color movie. Potential distribution maps and activation time maps were constructed from 182 epicardial and endocardial unipolar electrodes during 12 episodes of reproducible monomorphic VT in 9 dogs four to six days after experimental myocardial infarction (mean cycle length, 162 f 21 ms). At the onset of each depolarization during VT, a
I
ntraoperative mapping to localize the site of origin of ventricular tachycardia (VT) guides surgeons in determining the appropriate type of operative procedure and the region to which it should be applied. During the past two decades, activation mapping has been the exclusive method used to localize the site of origin of these arrhythmias [I, 21. However, the construction of activation time maps (ATMs)frequently requires editing to derive activation times from complex or questionable electrograms. This editing is time-consuming and requires considerable experience. In addition, ATMs only demonstrate local activity at the sites of the recording electrodes. Potential distribution mapping is another method to demonstrate the site of origin of VT [3-51. Studies involving VT simulated by rapid ventricular pacing have demonstrated that the intramural pacing site can be localized from the endocardial and epicardial surfaces with a high degree of accuracy [4, 51. This method has several advantages over ATMs. (1)Less editing is required to construct the isopotential map because measurement of potential is unambiguous. (2) Once the baseline is determined, no additional analysis time is required. Therefore, multiple beats of VT can be rapidly displayed. (3) Because unipolar Accepted for publication Dec 27, 1989. Address reprint requests to Dr Cox, Division of Cardiothoracic Surgery, Washington University School of Medicine, One Barnes Hospital Plaza, Suite 3108 Queeny Tower, St. Louis, MO 63110.
0 1990 by The Society of Thoracic Surgeons
potential minimum abruptly developed on the surviving epicardium and another on the surviving endocardium of the left ventricle, both immediately adjacent to the subendocardial infarct. These two minima on the initial potential distribution maps corresponded to the sites of earliest epicardial and endocardial activation breakthrough recorded on the activation time maps. These two minima subsequently expanded or moved into the adjacent area and coincided with the spread of activation fronts on the epicardial and endocardial surfaces. Thus, the rapid display of sequential, computerized potential distribution maps of multiple beats of VT provides a dynamic means of identifying the site of origin of VT, and therefore should facilitate intraoperative mapping. (Ann Thoruc Surg 2990;49:649-55)
potentials reflect distant activity, intramural sites of origin of VT may be detected from epicardial or endocardial recording electrodes. However, potential distribution mapping has not come into wide use because a series of maps is required to demonstrate the activation sequence. To overcome this disadvantage of potential distribution mapping and make it clinically feasible, an on-line computerized system was designed to display potential distribution maps sequentially as a movie. The purpose of this study was to determine if potential distribution mapping could provide a rapid and dynamic means of localizing the site of origin of VT in dogs with experimental myocardial infarction.
Material and Methods Definitions and Descriptions 1. Potential distribution maps (PDMs): A PDM constructed from unipolar electrograms at a specific instant in the cardiac cycle demonstrates the distribution of the potential fields. Reference baseline for calculation of the potential at each instant was obtained from the average of five consecutive samples (5 ms) in the flat portion of the late diastolic segment during normal sinus rhythm. A series of maps is required to demonstrate progressive changes of the potential distributions. 0003-4975/90/$3.50
650
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
Potential minimum: A potential minimum is the nadir of the negative potential area at an instant in the cardiac cycle. Primary potential minimum: A primary potential minimum is the earliest potential minimum reaching -3 mV in a series of PDMs [4, 51. Potential maximum: A potential maximum is the zenith of the potential area at an instant in the cardiac cycle. Activation time maps: An ATM is displayed as a sequence of different activation time zones. A computer program is used to determine local activation times from unipolar tracings. The peak negative derivative (-dV/dt,,,) of the major deflection of the unipolar complex is the time of local activation.
Animal Preparation Seventeen adult mongrel dogs of either sex weighing 18 to 25 kg were anesthetized with intravenous sodium pentobarbital (30 mg/kg), and mechanical ventilation was maintained by delivering 40% oxygen from a volume-cycled respirator (Bennett MA-1) through a cuffed endotracheal tube. After antibiotic prophylaxis (600,000 U of penicillin G administered intramuscularly), the chest was opened through a left thoracotomy in the fourth intercostal space and the heart was suspended in a pericardial cradle. The left anterior descending coronary artery was isolated immediately proximal to the first diagonal branch. A snare was lowered gradually onto a suture surrounding the artery, and partial occlusion was maintained for 20 minutes, followed by complete occlusion. The snare was released after two hours of total occlusion, and the myocardium was reperfused. Lidocaine hydrochloride (2 mg/ kg) was administered intravenously immediately before coronary occlusion and just before reperfusion. Then, the chest was closed in layers and the dogs were allowed to recuperate with only postoperative antibiotics given (600,000 U/day of penicillin G administered intramuscularly). Four to six days after myocardial infarction, the animals were reanesthetized and ventilated in the same fashion as described. A polyethylene catheter for monitoring arterial pressure was placed in the left femoral artery. A right thoracotomy was performed through the fifth intercostal space. Both venae cavae were encircled with tapes for control of these vessels, and the azygos vein was ligated and divided. The heart was exposed and suspended in a pericardial cradle. A sanguinous priming solution and pediatric oxygenator (Model H-400, William Harvey) were used for cardiopulmonary bypass. After heparinization (0.1 mg/kg) and direct cannulation of both venae cavae, normothermic cardiopulmonary bypass was instituted at a flow rate adequate for maintenance of a mean atrial pressure of 60 to 80 mm Hg. The interatrial groove was dissected to expose the septa1 portion of the left atrium. The heart was arrested with electrically induced ventricular fibrillation, and a standard left atriotomy was performed. A three-dimensional mold containing multiple endocardial electrodes was introduced into the left ventricle through the left atriotomy, thus avoiding the need for a left ventriculotomy. After a right atriotomy, a
Ann Thorac Surg 1990:49649-55
shaped, multielectrode endocardial mold was introduced into the right ventricular (RV) cavity. Sinus rhythm was restored by direct-current cardioversion. The entire epicardial surface of the heart was covered with a multielectrode epicardial mold. A bipolar electrode was sewn on the right ventricular outflow tract for both introducing ventricular tachycardias and recording reference electrograms. A common reference electrode for all cardiac unipolar electrograms was sewn on the right chest wall at a distance of at least 15 cm from the heart. Placement of the electrode in this position minimized electrical noise. Arterial blood gases and serum hematocrit levels were monitored at 15-minute intervals throughout each study. Sodium bicarbonate was administered as needed to maintain an acid-base equilibrium within normal limits at all times. Serum potassium levels were maintained between 3.7 and 4.5 mEqL.
Electrophysiological Studies Under normothermic total cardiopulmonary bypass, programmed electrical stimulation was performed to induce ventricular tachyarrhythmias using a programmable stimulator (model DTV-101, Bloom). Pacing was performed at a pulse width of 2 ms at twice late diastolic threshold current. A train of eight paced beats (S,s) at a basic cycle length of 300 ms was delivered, after which double premature stimuli ( S , and S,) were introduced at varying interval was lowered by coupling intervals. The ,S,S 5-ms decrements until noncapture of S, resulted, at which time the S,-S, coupling interval was narrowed by 5-ms decrements. This sequence was repeated until either the heart became refractory to the S, stimulus, ventricular fibrillation ensured (followed by direct-current cardioversion), or polymorphic or monomorphic ventricular tachycardia appeared. If the S, stimulus consistently failed to capture the ventricle, the pacing protocol was repeated at a basic cycle length of 250 ms. If ventricular tachycardia was not induced at one site, several alternative pacing sites were used.
Epicardial and Endocardia1 Electrodes and Mapping System Sixty-one silver bead unipolar electrodes, 2 mm in diameter, were positioned (15 mm apart) in 12 arrays on the inside of an epicardial mold (Fig 1A). Based on the epicardial structure of a normal canine heart, this mold was made of 5-mm thick foam rubber so that it could envelop the entire epicardial surface of the ventricles. The elasticity of the foam rubber maintained optimal contact between the unipolar electrodes and the epicardial surface during both systole and diastole. The endocardial molds for the left and right ventricles were also made of foam rubber. The RV mold, containing 60 unipolar electrodes, was designed to conform to the irregular internal anatomy of the RV cavity (Fig 1B). The right ventricular electrode was introduced into the RV cavity through the tricuspid valve annulus. The left ventricular mold (Fig 1C) was ellipsoid, 6 cm long and 3 cm across, with 61 unipolar electrodes evenly mounted (10 mm apart) in 12 arrays on its surface. Under normother-
HAR4DAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49649-55
mic cardiopulmonary bypass during electrically induced ventricular fibrillation, these electrode-containing molds were introduced into the RV and left ventricular (LV) cavities through the tricuspid and mitral valve annuli. The elasticity of these endocardial molds maintained optimal contact between the electrodes and the endocardial surface during both systole and diastole. These epicardial and endocardial molds standardized 182 unipolar electrode positions on the epicardial and endocardial surfaces (Fig 2). All signals from each unipolar electrode were connected to the differential amplifiers at an input impedance of lo'* ohms and a frequency response of 0.05 to 1,000 Hz. Data were recorded in three sets of 61 electrograms, in addition to the lead I1 electrocardiogram and bipolar reference electrode electogram. Because the current mapping system was capable of simultaneously recording 64 channels, three time-aligned sets of data using the bipolar reference electrogramand lead 'I were required to construct the entire epicardial and endocardial maps from 182 unipolar electrograms. Time-aligned analysis for
B
C
Fig 1. Epicardial and endocardial molds containing 182 2-rnm silver bead unipolar electrodes. These molds were made of foam rubber to conform to the epicardial and endocardial structures of the canine heart. (A) Epicardial mold containing 61 electrodes. (B)Right ventricular endocardial mold containing 60 electrodes. (C) Left ventricular endocardial mold containing 61 electrodes.
651
CHEST
pot
Fig 2. One hundred and eighty-two epicardial and endocardial electrode positions standardized by the epicardial and endocardial molds. (Ant. = anterior; END0 = endocardium; EPI = epicardiurn; FREE = free wall; LV = left ventricle; Post. = posterior; RV = right ventricle; SEP = septum,)
polymorphic VTs was impossible because of the changing rate and morphology of these arrhythmias. Therefore, the analysis of polymorphic VTs was excluded from this study. However, with a newer 250-channel data acquisition system, polymorphic VT can be analyzed using this technique without any additional processing beyond what is needed to analyze monomorphic tachycardias. A @-channel analog-to-digital converter digitized the unipolar data at a rate of 1,000 samplesk. A PDP 11/23PLUS computer stored the data and displayed the waveforms. The input range of the analog-to-digital converter was f 10 V. An amplifier gain of 500 was used for giving an input signal range of f 20 mV. Potential distribution maps for a selected window before and during early ventricular depolarization were automatically produced from the computer analysis of all 182 unipolar potentials referenced to the baseline T-P segments of each electrogram. If no baseline could be determined owing to the occurrence of a rapid VT, two other techniques for determining baseline were tested. Baseline was determined from the T-P segment during normal sinus rhythm for each electrogram, and the baseline was also determined for each electrogram by shorting the inputs of each amplifier to ground. Evaluation of PDMs during VT using all three techniques gave equivalent results. In addition, ATMs were similarly produced from the activation times derived from the same unipolar signals. Both ATMs and PDMs were displayed as smooth, computer-generated isochronous or isopotential maps. As construction of contour lines required considerable time, we used a more rapid "tile map" technique [ l l ] intraoperatively in which each of the electrode points divides the surface of the heart into polygonal regions. Each polygon tile color represented a potential level. More than 1,000 (1,000 ms) PDMs were generated with tile map technique within five minutes after acquisition of the data. SubsequentlYg a movie of the tile map was displayed by updating the color map for each electrode potential at each millisecond.
652
HARADAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49649-55
Fig 3. Epicardial and endocardial potential distribution maps during monomorphic ventricular tachycardia. Each panel (A-F) illustrates the computer-generated color isopotential maps of the epicardial and endocardial surfaces. The standard lead I1 echocardiograph (ECG) and bipolar reference electrogram (REF) are displayed in the left upper corner of each panel. The first, second, and third QRS complexes are s,, s,, and s, of a paced train, and the fourth QRS complex is the first beat of the induced rnonomorphic ventricular tachycardia. The vertical line and number on the ECG tracing indicate the time of each potential distribution map from the beginning of the ECG window. The scale is shown at the bottom of the figure. Negative potentials are idicated by blue areas, positive potentials by red areas, and near-neutral potentials by gray areas.
At the end of each experiment, the heart was excised with the endocardial and epicardial electrode molds still in place. The exact locations of the pacing and recording electrodes were verified. The heart was then sliced and stained with biphenyl tetrazolium chloride to define the extent of the infarction and the relationship to the recording electrodes. All animals received humane care in compliance with the ”Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the ”Guide for the Care and Use of Laboratory Animals” (NIH publication No. 85-23, revised 1985). In addition, the experimental protocol was reviewed and approved by the Washington University Committee for the Humane Care of Laboratory Animals.
Results Characteristics of Induced Ventricular Tachycardia Seventeen dogs were studied four to six days after myocardial iniarction. Monomorphic VT was induced in 9 of 17 dogs. Six of these dogs had one monomorphic VT and 3 dogs had two morphologically distinct VTs. The mean VT cycle length was 162 f 21 ms (range, 121 to 205 ms). These VTs terminated spontaneously or were terminated by premature ventricular pacing two to five minutes after onset. In 3 of the remaining 8 dogs, polymorphic VT was
induced and deteriorated into ventricular fibrillation. No VT was induced in 5 dogs. All 12 episodes of monomorphic tachycardia demonstrated an epicardial reentry pattern in regions overlying the infarct as determined by ATMs and PDMs.
Potential Distribution Maps During Monomorphic Ventricular Tachycardia Because of the similarity of results, a series of PDMs during monomorphic VT will be presented for a single dog (Fig 3). In the standard lead I1 electrocardiographic (ECG) tracing, the first, second, and third QRS complexes are S,, S,, and S, of the paced train; and the fourth QRS complex is the first beat of the induced monomorphic VT. The vertical line and number on the ECG tracing indicate the time of each PDM from the beginning of the ECG window. The potential scale is displayed at the bottom of the figure. Blue regions indicate negative potentials and red regions indicate positive potentials. Near-neutral potentials are indicated by the gray areas. Figure 3A shows the potential distributions at 820 ms after the beginning of the ECG window. At this instant, the majority of the epicardial and endocardial unipolar electrograms were in the repolarization phase (ie, T wave) of the second beat of the induced VT. Therefore, the potential minima are the results of the negative T waves present on the lateral LV epicardium and on the base of the LV and
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49:649-55
ECC
0 mr
EC(
Fig 4. Activation maps recorded from the same animal and during the same beat as in Figure 3. Zero milliseconds for reference time of the activation maps is selected at 800 ms from the beginning of the echocardiographic ECG) window. The earliest activation site is the paraseptal area and corresponds to the site of the prima y potential minimum on the potential distribution maps in Figure 3B. The shaded area on the left ventricular endocardium represents the infarction zone. (REF = bipolar reference electrogram.)
RV endocardium. The potential maxima are the results of the positive T waves observed on the epicardium of the RV and on the outflow tract of the RV endocardium. Later on, with the completion of local repolarization, these minima and maxima formed by the T waves slowly diminished the negative and positive areas. At 835 ms (Fig 3B), the beginning of the third QRS complex of the induced VT, a primary potential minimum abruptly developed in the paraseptal area of the epicardium just adjacent to the subendocardial infarction. At the same instant, there was little change in the epicardial and endocardial potential distributions except for the disappearance of a potential maximum on the RV epicardium. At 854 ms (Fig 3C), the potential minimum on the paraseptal epicardium expanded parallel to the left anterior descending artery. At this instant, the endocardial potential distribution suddenly changed, with a primary potential minimum of the RV endocardium developing anteriorly at the outflow tract. There was little change on the LV endocardium. At 860 ms (Fig 3D), while both epicardial and RV endocardial minima were expanding, another primary potential minimum of the LV endocardium abruptly developed on the LV septum adjacent to the preexisting potential maximum. This site of LV potential minimum was just adjacent to the subendocardial infarction. Subsequently, all epicardial and endocardial potential minima expanded or moved into the adjacent areas, with developing positive areas adjacent to these minima (Figs 3E, 3F). This pattern repeated for successive beats of VT.
Comparison Between Potential Distribution Maps and Activation Time Maps On the ATMs (Fig 4)recorded from the same animal and
653
during the same beat as that shown in Figure 3, the earliest activation breakthrough was observed on the paraseptal area of the epicardium including the site of epicardial primary potential minimum in the PDMs (Fig 3B). Right ventricular and LV endocardial activation breakthroughs also corresponded to the sites of endocardial primary potential minima following the epicardial primary potential minimum (Fig 4 vs Figs 3C, 3D). A comparison of ATMs and a series of PDMs showed that the potential minima during ventricular depolarization expanded or moved into the adjacent areas, concomitant with the spread of activation fronts at the epicardial and endocardial surfaces (Fig 4 vs Figs 3B-F). For example, a comparison of Figure 3E and Figure 4 shows that the demarcation lines of the negative potent'-l fields in the PDMs at 880 ms are in agreement with thc .
Comment The advent of direct surgical procedures for the treatment of refractory ischemic VT has led to a substantial improvement in the treatment of patients with these lifethreatening arrhythmias [6]. As electrophysiologically guided operation has resulted in a low postoperative inducibility and recurrence of VT, intraoperative electrophysiological mapping is essential for definitive surgical treatment [7, 81. The primary objective of intraoperative mapping is to localize the site of origin of VT, because ablation of the earliest activation site has been shown clinically to interrupt VT [7, 81. During the past two decades, epicardial, endocardial, and intramural activation time mapping have been used exclusively for identifying such sites of origin. Hand-held electrodes have been used to record multiple electrograms sequentially to derive the local activation time. However, the ATMs constructed by this technique are less accurate in demonstrating the activation sequence, because these are composite isochrones constructed from local activation times in different beats of VT. More accurate maps can be constructed using multiple simultaneous electrical recording systems. With the use of an increasing number of electrodes and amplifiers, a computer system is required to process the data rapidly [9, 101. However, the construction of ATMs frequently requires editing to derive activation times from complex or questionable electrograms. This editing is time-consuming and requires considerable experience. Additionally, intuitive judgment is used in interpreting the activation time data, whether they are displayed numerically or as isochrones. As a result of the time required to process the data, usually only one beat of the tachycardia is analyzed, precluding the use of map-
654
HARADAETAL POTENTIAL DISTRIBUTION MAPPING
Ann Thorac Surg 1990;49:649-55
A
\ Anterior
Rv
LV
Posterior B Fig 5. Photograph (A) and diagram ( B ) of a cross-section of the canine heart from which data are presented in Figures 3 and 4. (A) The section is stained with 2,3,5-triphenyl tetrazolium chloride solution. In (B), the shaded area represents necrotic tissue from myocardial infarction. The black arrow indicates the site of the epicardial primary potential minimum demonstrated in Figure 3B. The white arrow indicates the site of the left ventricular primary potential minimum in Figure 3 0 . (LV = left ventricle; RV = right ventricle.)
guided operation in polymorphic VT. In contrast to ATMs, PDMs require less editing to construct the isopotential contours, because the measurement of potential distributions is unambiguous, and require no additional analysis time to examine multiple beats of the tachycardia. Therefore, potential distribution mapping can be more automated and rapidly processed for interpretation. Another theoretical advantage is that unipolar potential maps can reflect distant activation events. Thus, epicardial and endocardial maps could theoretically detect early intramural wavefronts, whether in the ventricular free walls or septum, before the actual emergence of activation at these surfaces. Previous data on paced intramural
ventricular beats indicate that not only the site but also the intramural depth could be approximated from analysis of the PDMs from the ventricular surfaces bordering the site of stimulation [5]. With uniform conduction, the emergence of the wavefront would be expected to be at the site on the endocardial or epicardial surface closest to the point of origin. The presence of nonuniform activation due to the infarction or unidirectional block during reentrant ventricular rhythms could spatially dissociate the site of intramural depolarization and endocardial or epicardial breakthrough sites. Because PDMs reflect distant and local activity, electrode spacing is not as critical as it is for ATMs, which only represent local events. In addition, potential distributions recorded in areas of infarcted tissue (or areas with overlying thick fat pads) provide information to localize the origin of the VT.When constructing ATMs, these areas provide little information as no activation time can be assigned. In this situation, PDMs could prove to localize the intramural origin more accurately than endocardial or epicardial activation maps as the surface potentials would reflect intramural and distant activation. Epicardial, endocardial, and intramural potential distribution during ventricular pacing have been studied previously by several investigators [3-51. It has been well documented that the earliest potential minimum develops in an area close to the pacing site (earliest activation site) after the stimulation and that this minimum subsequently expands and moves to a distant area as it increases in negativity. Thus, the earliest potential minimum with low amplitude should represent the area closest to the earliest activation site [5]. We previously demonstrated that the earliest potential minimum reaching -3 mV is defined as a primary potential minimum that corresponds to the earliest activation site. Using the same definition for a primary potential minimum in this study, potential distribution mapping was evaluated as a method to localize the site of origin of VT. As described in the Results section, the epicardial primary potential minimum occurred above the subendocardial infarction and the LVendocardial primary potential minimum occurred adjacent to that infarction. These minima corresponded to the sites of activation breakthrough on the ATMs. Because an instant of -dV/dt,,, in the unipolar electrogram (activation time) usually did not coincide with the instant reaching -3 mV (primary potential minimum), activation breakthrough and primary potential minimum did not appear simultaneously. However, regardless of whether primary potential minimum precedes or follows activation breakthrough, it localizes the earliest activation site. Therefore, potential distribution mapping can serve the primary purpose of intraoperative mapping to localize the site of origin of VT. Furthermore, breakthrough on the ATMs appeared as a more widespread area than the primary potential minimum. One problem with viewing the PDMs statically is that there are often large areas of negativity. These areas are associated with the T waves, and when the maps are viewed dynamically in rapid succession, they are easily differentiated from the negativity associated with activation because they change much more slowly. However,
Ann Thorac Surg 1990;49:649-55
the potential distribution mapping technique could be improved if the T waves were filtered out of the signal. Spach and Barr [3] demonstrated that the demarcation line separating positive and negative potential (0 mV) shows good agreement with the location of the activation excitation wave. Our study demonstrated that the boundary of the negative potential field less than -3 mV at each instant coincides with the location of the activation wave. Rapid display of a series of PDMs could visually demonstrate the dynamic change of the activation fronts during VT. Therefore, we designed an on-line computer system to display PDMs sequentially as a movie. In conclusion, computerized potential distribution mapping provides a rapid means for the mapping of multiple beats of VT for the purpose of identifying the sites of origin of VT and also a dynamic means of demonstrating the activation sequence of these arrhythmias. Therefore, this intraoperative mapping technique should facilitate intraoperative mapping in patients undergoing operations for ventricular tachycardia. This work was supported by NIH grants RO1 HL33722 and RO1 HL32257. We thank George Quick, Michael M. Lischko, and Barbara E. Doerr for their excellent technical assistance, and Dawn Schuessler for the preparation of the manuscript.
References 1. Wittig JH, Boineau JP. Surgical treatment of ventricular arrhythmias using epicardial, transmural, and endocardia1 mapping. Ann Thorac Surg 1975;20:117-26.
HARADA ET AL POTENTIAL DISTRIBUTION MAPPING
655
2. Gallagher JJ, Kasell JH, Cox JL, Smith WM, Ideker RE, Smith WM. Techniques of intraoperative electrophysiologic mapping. Am J Cardiol 1982;49:221-40. 3. Spach MS, Barr RC. Analysis of ventricular activation and repolarization from intramural and epicardial potential distributions for ectopic beats in the intact dog. Circ Res 1975; 37830-43. 4. Taccardi B, Arisi G, Macchi E, Baruffi S, Spaggiari S. A new intracavitary probe for detecting the site of origin of ectopic ventricular beats during one cardiac cycle. Circulation 1987; 75:272-81. 5. Harada A, DAgostino HJ Jr, Schuessler RB, Boineau JP, Cox JL. Potential distribution mapping: a new method for precise localization of the intramural septa1 origin of ventricular tachycardia. Circulation 1988;78(Suppl 3):137-47. 6. Cox JL. Anatomic-electrophysologic basis of surgical treatment of refractory ischemic ventricular tachycardia. Ann Surg 1983;198:119-29. 7. Moran JM, Talano JV, Euler D, Moran JF, Montoya A, Pfiarrk R. Refractory ventricular arrhythmia: the role of intraoperative electrophysiological study. Surgery 1977;82:809-15. 8. Mason JW, Stinson EB, Winkle RA, et al. Surgery for ventricular tachycardia: efficacy of left ventricular aneurysm resection compared with operation guided by electrical activation mapping. Circulation 1982;65:114&55. 9. Witkowski FX, Corr PB. An automated simultaneous cardiac mapping system. Am J Physiol 1984;247:H661-8. 10. Downar E, Parson ID, Mickelborough LL, Cameron DA, Yao LC, Waxman MB. On-line epicardial mapping of intraoperative ventricular arrhythmias: initial clinical experience. JACC 1984;4:703-14. 11. Green PJ, Sibson R. Computing dirichlet tesselations in the plane. Comp J 1978;21:16&73.