Outflow tract ventricular tachycardia (OTVT) represents the most common subgroup of idiopathic ventricular tachycardia (VT). It typically occurs in healthy patients of young to middle age without structural heart disease and might be provoked by emotional stress, exercise, or dietary stimulants (1). The clinical presentation of OTVT is heterogeneous, ranging from isolated premature ventricular contractions (PVCs) to repetitive nonsustained VT to sustained VT. Although this diagnosis can be malignant, it generally carries an excellent prognosis and can be effectively treated by drugs or radiofrequency (RF) catheter ablation (2- 5).

Detailed intracardiac electrical mapping has demonstrated that the vast majority of OTVTs originate from the anterior and superior septal aspect of the right ventricular outflow tract (RVOT), just inferior to the pulmonic valve (6- 7). Less commonly, the site of origin can be localized to the right ventricular (RV) infundibulum, RV free wall, and posterior aspect of the interventricular septum. In approximately 10% to 15% of cases, the arrhythmia originates from the left ventricular outflow tract (LVOT) and can be mapped to the region of the aortic cusps (1,8- 10). Rarely, OTVTs can be ablated from within the anterior interventricular vein, aorto-mitral continuity, or the root of the pulmonary artery.

Because OTVT has a focal origin and occurs in patients with structurally normal hearts, it is an arrhythmia that is particularly conducive to localization with the 12-lead electrocardiography (ECG). Typically, OTVT originating in the RV manifests an inferior axis in the frontal plane and left bundle branch block (LBBB) configuration with precordial R/S transition at or after lead V₃ (11- 13). By contrast, LVOT VT usually manifests either a right bundle branch block (RBBB)/inferior axis or a LBBB/inferior axis with a precordial R/S-wave transition at or before lead V₃ (1,14). The ECG characteristics have also been described for idiopathic VT foci originating from the aorto-mitral continuity, anterior interventricular vein, and the left coronary cusp (LCC)–right coronary cusp (RCC) junction (6,10- 11,13- 15).

Criteria to distinguish RVOT from LVOT origin for patients with precordial transition occurring at lead V₃ have been lacking. This is of particular clinical importance, because a substantial number of OTVTs demonstrate lead V₃ transition and the need for left-sided ablation significantly alters patient counseling with regard to procedural time and risk (16). The aim of this study was to develop an ECG algorithm for reliably predicting the site of origin of OTVTs with lead V₃ precordial R/S transition. Existing ECG algorithms currently do not take into account cardiac rotation, respiratory variation, or the position of ECG leads on the chest, which might vary depending on body habitus, breast size, and technician expertise when placing leads. We hypothesized that comparison of the PVC/VT with the sinus rhythm (SR) QRS morphology would be an effective means of distinguishing LVOT from RVOT VT.

This study was designed in 2 parts: 1) a retrospective review of OTVT ablation cases in order to develop the ECG algorithm; and 2) a prospective assessment of the algorithm on a second group of patients.

We reviewed records on all patients who underwent mapping and ablation of idiopathic PVCs or VT at the Hospital of the University of Pennsylvania between January 2002 and December 2009. Patients with PVCs/VT manifesting a LBBB/inferior axis and a precordial transition (from R/S <1 to R/S >1) at lead V₃ were included. The majority of patients had normal left ventricular function by echocardiography, although patients with a presumed cardiomyopathy due to frequent ventricular ectopy were not excluded. Patients with ECG evidence of prior myocardial infarction, RBBB during SR, or whose clinical arrhythmia could not be abolished with catheter ablation were excluded.

A standard quadripolar catheter was positioned in the RV apical position, and a 4-mm nonirrigated catheter (Navistar or Celcius, Biosense-Webster, Diamond Bar, California) was initially positioned in the RVOT for mapping. In patients with sufficient ectopy, activation mapping was performed, recording the earliest local bipolar activation time compared with surface QRS of the clinical PVC. Pace-mapping at a threshold just above local capture was performed in all cases with careful comparison of the paced surface QRS morphology with that of the clinical PVC. All idiopathic PVCs/VT in this series originated from the septal side of the RVOT just beneath the pulmonic valve. The site of RVOT origin was classified by fluoroscopy and electroanatomic mapping as originating from either the anterior or posterior septal aspect of the RVOT for simplicity (12).

The decision to extend mapping to a LVOT site was made if no adequate RVOT sites were identified or ablation in the RVOT was unsuccessful in abolishing the arrhythmia. The LVOT sites were mapped via a retrograde aortic approach. All mapping was performed after heparin bolus, maintaining an activated clotting time >250 s. In addition to standard fluoroscopy, a 3-dimensional electroanatomic mapping system (CARTO, Biosense-Webster, or NavX, St. Jude Medical, Minnetonka, Minnesota) and intracardiac echocardiography (Acuson, Siemens Medical, Mountain View, California) were used to localize the anatomic position of the ablation catheter within the outflow tract (Figure 1). For most of our experience, a nonirrigated 4-mm tip catheter with power delivered up to 50 W and temperature up to 50°C was used for RVOT and LVOT ablation. Beginning in 2008, irrigated 3.5-mm tip catheters were used almost exclusively in the LVOT and on occasion by some operators in the RVOT. In the RVOT, a maximum power setting of 30 to 35 W was used. For the LVOT/aortic cusp region, the typical maximal power setting was 30 W; however, in select cases with late termination of PVCs or suppression and recurrence of ectopy, powers of up to 50 W were used.

Acute ablation success was defined as the absence of the clinical PVC at 30 min after the last RF delivery, both with and without isoproterenol, and confirmed by continuous full disclosure cardiac telemetry in the subsequent 24 h of inpatient care.

Sinus rhythm and VT ECG morphology were measured on the same 12-lead ECG with electronic calipers on the Prucka CardioLab (GE Healthcare, Waukesha, Wisconsin) recording system. Standard 12-lead ECG electrode placement was used. Lead gain was uniform with paper speed of 100 mm/s. During the clinical arrhythmia, the following measurements were obtained during both SR and the PVC/VT: 1) R- and S-wave amplitudes in lead I, II, III, aVF, and V₁ to V₃; 2) R-wave duration in leads V₁ to V₃; 3) QRS duration; and 4) QRS frontal axis (Figure 2). The T-P segment was considered the isoelectric baseline for measurement of R- and S-wave amplitudes. The QRS duration was measured from the site of earliest initial deflection from the isoelectric line in any lead to the time of latest activation in any lead. The R-wave duration was measured from the site of earliest initial deflection from the isoelectric line to the time at which the R-wave intersected the isoelectric line. For all cases, QRS measurements were performed on isolated PVCs representative of the clinical VT before the induction of sustained VT and compared with the SR QRS complex. As a means of accounting for respiratory variation, we standardized our SR measurements by measuring the largest R- and S-wave over a 10-s window at 25-mm/s sweep speed (Figure 3). The transition ratio was calculated in leads V₂ and V₃ by computing the percentage R-wave during VT (R/R+S)_VT divided by the percentage R-wave in SR (R/R+S)_SR.

We also assessed a simple, easily used, qualitative measure of comparing the precordial transition of the PVC/VT with SR. We hypothesized that a PVC/VT precordial transition to R>S at an interspace equal to or earlier than the location of the SR transition would suggest an LVOT origin, whereas a transition at an interspace later than the SR transition would suggest an RVOT origin. The sensitivity and specificity for this measure was assessed.

We then performed a prospective evaluation of a second cohort of consecutive patients with LBBB/inferior PVCs/VT with lead V₃ R/S transition who met the same inclusion criteria used in the retrospective cohort and who presented for catheter ablation between January 2010 and August 2010. Quantitative and qualitative measurements in this group were performed by 2 blinded observers, before any knowledge of the site of VT origin. The site of successful ablation was identified as the one that led to abolishment of the clinical PVC/VT. The utility of the precordial transition ratio for predicting the site of PVC/VT origin was then determined.

Continuous variables are presented as mean ± 1 SD. Continuous variables were compared with Student t test and categorical variables with Fisher exact test. Predictors with a p value <0.1 were entered into a stepwise forward multivariate logistic regression to determine independent predictors of PVC origin (LVOT vs. RVOT). Agreement between the ECG measurements of the blinded observers was assessed with the intraclass correlation coefficient, assuming a 2-way random model with absolute agreement. The significance of the intraclass correlation coefficient was assessed with an F test to determine whether the correlation was >0. The kappa statistic was also computed for agreement in classification between the 2 observers. A receiver operating characteristic curve was performed for sensitivity and specificity analysis. Statistics were performed on SPSS statistical software (release 18.0, SPSS, Chicago, Illinois). A p value ≤0.05 was considered statistically significant.

We reviewed data from 293 patients who underwent catheter ablation of OTVT between January 2002 and December 2009 at our institution. After excluding patients with precordial transition V₁ or V₂ (n = 49), precordial transition V₄ to V₆ (n = 35), structural or ischemic heart disease (n = 55), RBBB (n = 8), paced rhythm (n = 2), unable to acutely abolish PVC/VT (n = 5), and incomplete records (n = 99), we identified 40 cases of successful OTVT ablation with lead-V₃ transition that met our inclusion criteria. Patient demographic data between RVOT and LVOT groups were similar at baseline (Table 1). The RVOT cohort consisted of 80% anteroseptal sites and 20% posteroseptal sites. The LVOT cohort consisted of 35% LCC, 15% RCC, 40% left-right coronary cusp junction, and 10% great cardiac vein near the anterior interventricular vein.

Absolute measurements of PVC and SR QRS indexes are listed in (Table 2). The R-wave amplitude was greater for LVOT compared with RVOT PVCs in leads V₂ (p < 0.001) and V₃ (p < 0.001). The R-wave duration ratio of the PVC to SR in lead V₂ (p = 0.002) and lead V₃ (p = 0.026) was also significantly greater for LVOT compared with RVOT PVCs. The V₂ transition ratio was significantly greater for LVOT (range 0.42 to 2.89) compared with RVOT origin (range 0.02 to 0.57) (p < 0.001); however, the V₃ transition ratio was not significantly different between PVCs of LVOT versus RVOT origin (p = 0.093) (Figure 4). The overall QRS duration was longer for LVOT compared with RVOT PVCs (p = 0.048).

In a multivariate logistic regression, including R-wave duration in leads V₁ and V₂, R-wave amplitude in leads V₂ and V₃, and the R-wave transition ratio in lead V₂, the V₂ R-wave transition ratio was the only independent predictor of PVC origin (p < 0.001, 95% confidence interval: 0.01 to 0.41).

A receiver operating characteristic curve for the retrospective data is illustrated in (Figure 5). A V₂ transition ratio ≥0.6 predicted an LVOT origin with a sensitivity of 95% and specificity of 100% (Figure 4). This cutoff yielded a positive predictive value of 100% and a negative predictive value of 95%. A more practical cutoff of ≥0.5 yields a sensitivity of 95% and specificity of 95%.

The simple qualitative measure of PVC precordial transition (R>S) occurring at or before the SR transition (R>S) had a sensitivity of 47% and specificity of 64% for identifying an LVOT origin. However, a PVC precordial transition occurring later than the SR transition had a 19% sensitivity and 100% specificity for RVOT origin. Therefore, a PVC that transitions later than SR effectively rules out an LVOT origin. The surface ECGs of representative PVCs with lead V₃ R/S transition are illustrated adjacent to their corresponding sinus beats in (Figure 6).

The ECG measurements of the ensuing 21 cases of OTVT who underwent successful RF ablation at our institution were performed. The algorithm was able to correctly predict the site of successful ablation (LVOT vs. RVOT) in 91% (19 of 21) of cases. The interobserver agreement in measurement of R- and S-wave amplitudes yielded an intraclass correlation coefficient of 0.966 (95% confidence interval: 0.937 to 0.982; p < 0.001), and the kappa statistic for classification agreement was 0.877 (p < 0.001). With the simple qualitative measure alone, where LVOT origin is suggested by a PVC transition at or earlier than the SR transition, there was 71% accuracy in diagnosing the PVC origin (100% interobserver agreement). On the other hand, when the PVC transition occurred later than the sinus rhythm transition, an LVOT origin could be excluded with 100% accuracy. A proposed algorithm combining the qualitative and quantitative measures is shown in (Figure 7).

We present a novel surface ECG measurement, the V₂ transition ratio, for distinguishing LVOT from RVOT origin in patients with left bundle branch pattern idiopathic VT/PVCs with lead V₃ precordial transition. This measure accounts for variations in body habitus, cardiac rotation, respiratory variation, and ECG lead positioning by measuring precordial transition during the PVC/VT relative to the SR precordial transition. A V₂ transition ratio ≥0.6 predicted an LVOT origin with 95% sensitivity and 100% specificity. For patients referred for catheter ablation of OTVT, this simple ECG measurement might be performed in the office both to help plan an ablation strategy and to enhance patient counseling with regard to procedural time, potential outcome, and risks associated with arterial access, mapping, and ablation.

One might argue that the V₂ transition ratio is cumbersome for everyday use in clinical practice. In the electrophysiology lab, this measurement is easily made with digital calipers available on any clinical electrophysiologic recording system. For more practical clinical use, however, we also found that a precordial transition during the PVC/VT that occurs later than the SR transition excludes an LVOT origin with 100% accuracy. This simple measure can be easily used by any cardiologist or electrophysiologist when counseling patients about PVC ablation.

There have been many prior studies evaluating the surface ECG for localization of OTVTs. Dixit et al. (12) in our laboratory established the typical ECG features for RVOT septal and free wall PVC origin. Ouyang et al. (13) found that a greater R-wave duration and R/S-wave amplitude ratio in leads V₁ or V₂ reliably predicted an aortic sinus cusp compared with RVOT origin. Lin et al. (11) described site-specific ECG features of PVCs originating from the aortic cusps with electroanatomic and ICE-guided pacemapping. They concluded that the LCC typically produces a precordial transition by V₂, whereas the RCC demonstrated a precordial transition by lead V₃. Furthermore, studies by Yamada et al. (9) and Bala et al. (10) characterized the left-right coronary cusp junction as often possessing QRS in leads V₁ to V₃ or QS in lead V₁. One previous study of patients undergoing RF catheter ablation for idiopathic VT or symptomatic PVCs implicated the R-wave amplitude proportion as a means of discriminating between RVOT and extra-RVOT foci; however, the study was limited by small size, lack of prospective evaluation, or an assessment of the frontal plane (17). It has previously been shown that lead V₃ R/S transition is highly variable, with only 55% sensitivity and 38% specificity for predicting an RVOT origin. (16).

The sample size, although comparable to other studies of OTVT, is small and should be validated further in larger populations. We used the location of the ablation catheter in the RVOT or LVOT during abolition of the clinical arrhythmia as the PVC/VT site of origin. It is possible that ablation lesions from the RVOT or LVOT extended into adjacent structures and that a PVC with epicardial or LVOT origin could be abolished from the RVOT. This might explain the few cases that did not match our criteria. Detailed measurements with digital calipers were used to develop the algorithm. It is possible that such measurements will prove more difficult in clinical practice, although as stated in the preceding text, the theory of incorporating the SR precordial transition into the decision algorithm for predicting a site of PVC origin still applies.

We present a novel electrocardiographic measure of the ratio of the VT and SR precordial transition, “the V₂ transition ratio,” which can reliably distinguish left from right outflow tract PVC/VT origin in patients with OTVT and lead V₃ precordial R/S transition. A V₂ transition ratio ≥0.6 predicts an LVOT origin with high sensitivity and specificity. A precordial transition later than the SR transition excludes an LVOT VT origin. This algorithm might lead to improved patient counseling and planning of medical therapy or ablation procedures in patients referred for ablation of outflow tract tachycardias.