Equilibrium radionuclide angiography (ERNA) accurately evaluates regional and global ventricular function (1- 2). The ERNA “phase” analysis applies a first harmonic fit of the blood pool time versus radioactivity curve to measure the magnitude and sequence of ventricular contraction in each pixel of the image. The generated symmetric cosine curve is characterized by its depth or magnitude, the amplitude, and its location during the cardiac cycle, or phase angle, Ø. The sequence of contraction, assessed by Ø, generally parallels the conduction sequence of the myocardium (3).

The use of ERNA with phase image analysis is the only noninvasive method that can accurately, objectively, and reproducibly determine ventricular volume, function, contraction pattern, and synchrony. Its application to ventricular tachycardia (VT) has been limited to localization of the VT exit site as the earliest contracting segment in a few patients (4- 6). Not only can the use of ERNA potentially determine the localization of the VT exit site but also the subsequent contraction sequence and its functional effects (1- 3). Similar outcomes have been evaluated in other conditions of altered conduction (3,7- 9). In a unique and challenging protocol, we took advantage of the full potential of ERNA at the bedside to determine the clinical effects of VT rhythms induced during electrophysiologic study. Information regarding function, conduction, and synchrony is important in the evaluation of cardiac resynchronization therapy, where ERNA may also be useful.

Over the course of 2 years, 32 monomorphic VT rhythms were imaged in 26 patients (24 men, 2 women, 15 to 68 years of age, mean age 54 years), presenting with a history of symptomatic, sustained VT and/or sudden cardiac death. Patients were selected randomly based on camera availability, and they gave written informed consent approved by The UCSF Human Research Committee. We performed ERNA during VT that was induced under electrophysiologic study in 21 patients or that occurred spontaneously in 5 patients. Twenty patients had coronary disease, 5 had a noncoronary cardiomyopathy, and 1 had right ventricular (RV) dysplasia.

Twenty-two patients were imaged in normal sinus rhythm, within minutes of VT imaging with imaging in sinus rhythm preceding VT in 19 patients. There were no events, volume, or medication changes between studies.

Rate of VT, axis, conduction pattern, and exit site were derived from leads I, II, III, and V₁ measured continuously throughout electrophysiologic study and from the 12-lead ECG recorded intermittently with VT during electrophysiologic study and with spontaneous VT. Electrocardiogram criteria were used to determine VT exit sites according to the following criteria of Josephson et al. (10): left bundle branch block = RV or left ventricular (LV) septal origin; left bundle branch block with inferior axis = RV outflow tract origin; right bundle branch block with q waves in V₅,V₆, II, III,or AVF = LV apical origin; right bundle branch block with r in V₅, V₆, II, III, or AVF = LV base origin. Unlike electrophysiologic analysis, ECG correlation did not permit more refined segmental correlation.

Planar ERNAs in VT were acquired with a LEM portable camera (Siemens Inc., Des Plaines, Illinois) with the use of an all-purpose, 20° slant hole collimator with modified in vivo red cell labeling (11). The “best septal” left anterior oblique (LAO), anterior, and 70° LAO projections were acquired in that order, each for 3 min, in sinus rhythm and with VT unless limited by spontaneous VT conversion or patient intolerance.

Images were acquired in 28 frames with a 9-point spatial smooth and displayed in a 128 × 128 format, viewed in 256 gray shades. Automated or manually drawn LV edges were used to calculate LV ejection fraction (EF) in the “best septal” projection (1,9). One-minute acquisition was adequate to extract phase and relative volume data. The estimated radiation exposure for the acquired ERNA is 0.3 rems.

In-house software-corrected images for counts decay between sinus rhythm and VT to calculate relative changes of end-diastolic volume (EDV), end-systolic volume, stroke volume (SV), and the cardiac output (CO), in those studied in both rhythms. With multiple induced VT morphologies, comparisons in sinus rhythm were made to VT having the most likely clinical pattern.

Regional LV wall motion was assessed by a researcher who was blinded to the procedure, in the projections available and graded 0 to 4, as normal, mildly hypokinetic, severely hypokinetic, akinetic, and dyskinetic, respectively, according to established criteria (1,9) and supported by the amplitude image. The VT exit site was localized in 15 segments (Figure 1).

The fundamental Fourier harmonic was applied to the first 25 frames of each “best-septal” ERNA to generate phase and amplitude images (3). Regional phase angle, Ø, was coded in color or 256 gray shades as previously presented (Figures 2, 3, 4) (3).

The distribution of phase angles in the RV and LV regions of interest in the “best-septal” LAO projection was measured from RV and LV phase histograms, which related Ø on the abscissa to the number of pixels at each Ø on the ordinate. To exclude background noise, an empirical 5% threshold was applied. Moveable cursors set histogram limits, highlighting related pixels and permitting localization of the site of earliest Ø, the VT exit site, as well as the progression of Ø, the contraction sequence, in each ventricle. The mean and standard deviation (SD) of phase angles, LVØ and RVØ, indicators of intraventricular synchrony, were derived from their respective histograms. The difference (mean LVØ −mean RVØ), reflects interventricular synchrony.

By using phase image analysis in all projections, the reader, who was blinded to wall motion, ECG, or electrophysiologic study, triangulated the VT exit site to a specific ventricular location in 1 of the 15 segments. With limited projections, the site of earliest Ø was localized to a projected line of possible sites. The relationship of the phase VT exit site to the electrocardiographic and electrophysiologic VT exit sites was noted.

Patients were studied in the fasting, nonsedated state, off antiarrhythmic drugs according to the standard procedure (12). Catheters were positioned against the RV apex, the outflow tract, and within the LV with ECG monitoring and with serial recording of the intracardiac electrogram and arterial pressure. Programmed ventricular stimulation induced sustained VT, lasting >30 s. Endocardial activation time was measured by a researcher who was blinded to the procedure. Initially, 15 segments through both ventricles were sampled. Then, based on ECG findings and any known regional scar, dense sampling of multiple electrograms was made in relation to initially sampled sights. The VT exit site was localized to 1 of the 15 segments in most cases, but some studies identified similarly early activation in more than one adjacent segment (Figure 1). Full electrophysiologic study required mapping of 10 LV sites. Multiple VT morphologies were not mapped.

Intravascular blood pressure measured at electrophysiologic study or with cuff readings were recorded. Systolic pressure ≤90 mm Hg, related to symptoms including chest pain, lightheadedness, and near syncope, required cardioversion and indicated rhythm intolerance. Ventricular tachycardia rhythms with systolic pressure >90 mm Hg were tolerated for their duration.

The site of earliest Ø, the VT exit site, and pattern of Ø progression, the contraction/conduction patterns, were compared with the ECG exit site and that mapped at full electrophysiologic study. Where electrophysiologic study localized the exit site to several segments, scintigraphic agreement was considered present when the site of earliest Ø was localized to one of these segments. Left ventricular volumes, SD RVØ, SD LVØ, and mean LVØ-RVØ in VT, were compared with those in sinus rhythm and related to VT tolerance. Serial change in LVEF was expressed in absolute EF units and, like CO and all LV volumes, as the percentage change from sinus rhythm and analyzed with paired t tests. An independent means test was applied to continuous variables, and the chi-square test was applied to discrete variables. Independent significance of variables was established by multivariate analysis.

Images were acquired during 32 different VT rhythms in 26 patients, 4 of whom demonstrated multiple VT rhythms. Three projections were acquired in 24 patients. In 2 patients with unstable VT, only a “best-septal” LAO projection was acquired. Of 22 VT rhythms imaged also in sinus rhythm, 13 were tolerated clinically (Group I), whereas 9 were not tolerated (Group II). The clinical diagnosis could not differentiate between groups.

With VT, the QRS widened from 0.11 ± 0.01 s to 0.16 ± 0.02 s, p = NS. Neither the QRS duration in sinus rhythm, nor in VT, nor their difference could differentiate groups. There was no correlation between QRS duration and SD LVØ, SD RVØ, or mean LVØ-mean RVØ. The ECG pattern localized 6 exit sites to the RV, 26 to the LV, and 13 to the septum.

One VT pattern was fully mapped at electrophysiologic study in each of 19 patients. Brief VT duration in 7 patients led to incomplete electrophysiologic mapping. The specific VT exit site could not differentiate groups.

Phase imaging correctly localized 16 of 19 (84%) VT exit sites that were fully mapped at electrophysiologic study (p < 0.05), 13 to the LV, and 3 to the RV (Figures 1, 2, 3, 4). The phase pattern and mean LVØ-mean RVØ, paralleled the VT ECG conduction delay measured during the electrophysiologic study.

Electrocardiogram localization of the VT exit site agreed with the electrophysiologic site in 14 of 19 fully mapped VT rhythms and paralleled the findings on phase image analysis in 15 of these 19. The scintigraphic VT exit site and contraction pattern agreed with the ECG results in 11 additional VT rhythms that were not fully mapped at electrophysiologic study. Hence, for a total of 32 VT exit sites localized by ERNA, 26 (81%) correlated with electrophysiologic or ECG localization (p < 0.05).

Two patients demonstrated 2, and 2 others had 3 VT patterns. The localization obtained by ERNA agreed with the single VT exit site fully mapped at electrophysiologic study in these 4 patients. The use of ERNA revealed 6 of these 10 exit sites to originate in separate anatomic locations. Two VT rhythms in each of 2 patients appeared to originate from similar anatomic exit sites, with varying patterns of contraction and phase sequence that paralleled ECG conduction patterns (Figures 3, 4). In these patients, ventricular function and tolerance varied at a similar VT rate (Figures 3, 4), suggesting an influence of the VT exit site, or other factor on patient tolerance.

Rate in sinus rhythm did not differ between Groups I (70.46 ± 8.75) and II (72.0 ± 7.76). Intolerance to VT related to a greater rate, 157.70 ± 29.18 versus 130.38 ± 23.25, p < 0.0001, and a greater absolute increase in rate, 85.70 versus 59.92, p < 0.05 in Group II compared with Group I, but relative rate augmentation with VT, 2.21 ± 0.48 versus 1.89 ± 0.44, did not differ between groups (p = 0.18).

Left ventricular ejection fraction in sinus rhythm did not differ between groups and decreased with VT in Groups I (28.38 ± 7.75 to 19.00 ± 7.12) and II (31.22 ± 6.77 to 15.11 ± 5.22), both p < 0.03. We found that VT LVEF was lower, with a greater decrease of LVEF in absolute EF units (16.11 ± 10.5 vs. 9.38 ± 7.9), and in relative percentage (51 ± 13% ± vs. 33 ± 12%), in Group II compared with Group I (all p < 0.05). There was a greater relative decrease in LVSV (65 ± 15% vs. 45 ± 16%, p < 0.01), CO (30 ± 14% vs. 2 ± 15%, p < 0.001), and LVEDV (36 ± 18% vs. 27 ± 14%, p < 0.001) in Group II compared with group I but not in LV end-systolic volume (19 ± 19% vs. 9 ± 14%, p = NS).

The SD LVØ was the only functional parameter in sinus rhythm that differed significantly between Groups I (35.38 ± 8.4) and II (45.0 ± 9.4), p < 0.01. We found that LV SDØ and RV SDØ increased in all VT episodes. Wall motion score increased during VT, from 1.8 to 2.7, with greater dyssynchrony during VT, expressed by the increase in both SD LVØ (35.38 ± 8.4 to 62.30 ± 13.5) in Group I and (45.00 ± 9.4 to 70.55 ± 17.3) in Group II, and SD RVØ (31.92 ± 16.7 to 48.46 ± 13.5) in Group I and (40.55 ± 16.3 to 63.88 ± 15.4) in Group II (all p < 0.01), with greater SD LVØ and SD RVØ in VT in Group II than in Group I (all p < 0.05).

In 16 VT episodes, the ERNA-mapped VT exit site was adjacent to a discrete aneurysm (Figures 2, 3), a paradoxically moving wall segment generally apical in location, which appeared de novo during VT in 14 cases with an increase in mean wall motion score from 1.4 to 2.2 (p < 0.05). The VT phase pattern and the mean LVØ-RVØ difference paralleled the ECG conduction delay (Figures 2, 3, 4). The specific location of the VT exit site could not differentiate between groups. Significant changes in LVEF, LVEDV, LVSV, and CO from sinus rhythm to VT, as well as SD LVØ in sinus rhythm, differentiated Groups I and II.

The VT exit site determined by ERNA mirrored the clinical VT exit site determined by ECG and electrophysiologic study. Two groups of VT rhythms with a difference in tolerance were identified. The changes in rate, ventricular volumes, and CO, which previously were related to VT intolerance, were confirmed in the present study by the use of ERNA analysis. When combined with related images, the use of ERNA provided us with a unique and complete noninvasive evaluation of both functional and electrophysiologic aspects of VT.

The hemodynamic and clinical outcomes of VT have been shown to be determined by the functional effects of VT, with related autonomic responses and other factors, imposed on baseline ventricular function (13- 17). This study demonstrates that ERNA can be used to analyze both the mechanical and “electrical” effects of VT. Such an analysis presents important parallels for the evaluation of the effects of cardiac resynchronization therapy (14- 15,18- 19) in which accurate, reproducible quantitation of ventricular function, conduction sequence, and synchrony are important. The SDØ and other new parameters have been developed to measure ventricular synchrony based on phase image data (20- 21). This study supports further application of the ERNA phase method to assess resynchronization therapy.

Sequential phase progression is a surrogate for the regional contraction sequence. It has been related to the conduction sequence in bundle branch block (8), applied to characterize conduction abnormalities (3,7- 9), and used to determine the functional effects of pacemakers (9). In limited VT studies (4- 6,14), its use has led to the identification of the VT exit site. In the current study, ERNA presented functional changes with a graphic image assessment of the conduction pattern in VT patients with electrophysiologic correlation.

The first harmonic sinusoid provides only a rough fit of the time versus radioactivity curve. Although it can be related with great resolution to the timing of ventricular contraction, only its sequence has been confirmed. Sinusoids are symmetric as are volume curves at high rates, as in VT, where the first harmonic fit is particularly appropriate. A specific temporal estimate of the contraction sequence would require a multiharmonic curve fit.

The relationship between phase contraction sequence and conduction has been confirmed by electrophysiologic mapping or indirectly by the surface ECG (3,14,22). In previous studies, ECG patterns correlated grossly with the electrophysiologic VT exit site (10). The current study shows that ERNA is capable of more-refined localization of the VT exit site than the ECG and correlated well with the exit sites localized on ECG and electrophysiologic study and their related conduction pattern.

The importance of conduction, related contraction and the VT exit site, as determinants of LV systolic function, is supported by several cases with multiple VT exit sites and varying tolerance, in the presence of similar VT rates. Results are presented of ERNA acquisition and phase image analysis of multiple VT exit sites and multiple VT morphologies (Figures 3, 4), which had never before been attempted.

With normal baseline LV function, VT rate, incomplete relaxation, and reduced filling influence VT tolerance (15- 16,21). With abnormal LV function, the VT exit site, the resultant activation and contraction patterns, and the level of induced dyssynchrony appear important (18,20- 21). Our study showed that decremental function and VT tolerance related significantly to the level of baseline dyssynchrony but related imperfectly to baseline rate, wall motion, and LVEF.

Aneurysms have previously been shown to appear with pacing of the peri-infarction area in animals (14) and to resolve with resynchronization therapy (14,21), based likely on altered conduction (21). When using electrophysiologic studies, researchers often have localized the VT exit site to the perianeurysmal border region (5), which also was confirmed in our study, as ERNA often demonstrated new aneurysms with VT.

The study population was small in this challenging protocol but adequately served the purpose. The brief patient observation period likely resulted in an underestimation of VT intolerance. Ventricular tachycardia-induced dysfunction could relate to induced ischemia. However, ischemia was not evident before VT, rhythm tolerance was unrelated to coronary disease or age, no patient had overt ischemia during electrophysiologic study, and ischemia occurs uncommonly with induced VT (17). Both exit site and dyssynchrony alter regional perfusion (5,19). Yet, even with ischemia, the functional and electrophysiologic effects of VT intolerance remain valid.

Autonomic changes, retrograde atrial activation, mitral regurgitation, hypovolemia, and hypertrophic obstructive cardiomyopathy may reduce VT tolerance. Patients were kept euvolemic, whereas other conditions were not evident or were not specifically measured.

The ECG localization of the VT exit site was determined and known before electrophysiologic localization. However, electrophysiologic localization was made on the basis of objective timing information. As noted in the literature, ECG correlation could do no better than gross localization.

Acquisition time, data density, and image smoothing influence phase values but not its sequence. The 5% phase threshold applied here is empirical and set to separate cardiac from extracardiac regions by removing the low-amplitude extracardiac noise. It has been widely and successfully applied (3,9,20), but a varied threshold may be more optimal.

The planar ERNA method mapped the site of earliest phase angle. Although this site successfully paralleled exit site localization, it did so only within the anatomic limits presented by that planar methodology. Although providing differentiation regarding the localization of multiple VT exit sites, it does not have the spatial resolution of the electrophysiologic method. Phase imaging with SPECT ERNA (23) could add resolution to exit site localization and assessment of the contraction pattern. However, given the unique nature of this protocol and the requirement for a portable camera, SPECT could not be applied.

The use of ERNA provided a complete noninvasive method to determine the functional and electrophysiologic effects of VT. This unique study provides an objective analysis of the effect of the rhythm, rate, conduction pattern, and synchrony on VT tolerance. Graphic images displayed the VT exit site and related contraction and conduction patterns, providing further insight into the effects of the rhythm. Phase image analysis of multiple VT exit sites was presented. Such ERNA analysis, applied to patients undergoing resynchronization therapy, could potentially provide insight into the factors that promote benefit from that intervention.

For accompanying figures and videos of phase images of an RV VT exit site and multiple exit sites, please see the online version of this article.