Study population

From July 2006 to July 2014, 1199 consecutive patients (497 male; mean age 48.2 ± 17.2 [7-86] years), who were proven not to have structural heart diseases by routine biochemistry examination, CT scan, echocardiography, and even coronary angiography and cardiac magnetic resonance in some special cases, received RFCA for VAs at our cardiac electrophysiology center. Of these, 45 patients (3.75%; 24 males; mean age, 56.6 ± 16.4 [17-80] years) with ECG suggestive of VAs arising from the left ventricular summit, which was further confirmed by the effective target on fluoroscopy and three-dimension mapping system (Carto 3 or Ensite-NavX system), were enrolled in the present study. Transesophageal echocardiography was only examined on older patients to exclude thrombus in left atrial appendage because of the potential risk of asymptomatic atrial fibrillation. All patients signed the informed consent form and agreed to the electrophysiological examination and radiofrequency ablation treatment. All these treatments patients received were considered standard care for their conditions.

ECG measurements

Twelve-lead ECGs of the clinical arrhythmias were obtained and analyzed before RFCA. No patients had T inversion, pathological Q wave in precordial leads or epsilon wave. ECG pattern analysis focused on (1)QRS morphology of the clinical VAs in all 12 leads; (2) the axis deviation; (3) Transitional zone index in precordial leads (defined as transition zone score of VAs minus transition zone score of sinus beat); (4)QRS duration; (5)The pseudo delta wave time (PdW, the interval from the beginning of the QRS complex to the earliest fast deflection in any precordial lead) [11]; (6) The intrinsicoid deflection time (IDT, the interval from the beginning of the QRS complex to the peak of the R wave in V2) [11]; (7)The maximum deflection index (MDI, IDT divided by the QRS duration) [11]; (8) QS wave ratio of aVL/aVR. ECG of each patient was measured by 2 cardiologists unknown the site of origin; any discrepancy was judged by a third cardiologist. Capital letters (Q,R,S) represented waves of relatively high amplitude(>0.5 mV), while lowercase letters (q,r,s) represented waves of relatively low amplitude(<0.5 mV).

Electrophysiological study and radiofrequency ablation

Electrophysiological study was performed after discontinuation of anti-arrhythmic drugs for at least five half-lives. Catheters were delivered to the distal coronary sinus (CS) via subclavian or femoral veins and to the right ventricle via femoral veins. 12-lead ECG and intracardiac electrogram were monitored and recorded on a multichannel oscilloscopic recorder. If the clinical arrhythmias didn’t appear at the beginning of the electrophysiologic study, induction of VAs was attempted by programmed electrical stimulation from the right ventricular apex, delivering a maximum of three extrastimulis with basic drive cycle lengths of 600, 500, and 430 ms. If VAs was still not provoked at baseline, intravenous isoproterenol (2–4 μg/min) was administered. In all case, activation mapping and pace mapping were both performed to identify the VA origin site. The activation time was measured during VA from the onset of the ventricular activation of the distal bipole of the mapping catheter to the earliest onset of the QRS complex in any of the 12 ECG leads. Pace mapping was performed during sinus rhythm at a pacing cycle length of 500 ms with stimulus amplitude of 1 mA greater than the late-diastolic threshold, the pace match was assessed by whether the paced QRS morphology match the clinical VAs in the 12-lead ECG (a perfect pace map was defined as 12 leads complete match, an excellent pace map referred to more than 10 leads match). When a site was mapped with a local ventricular activation preceding the QRS onset by more than 20 ms and pacing from the site produced an excellent match to the QRS complex of the clinical VAs, catheter ablation was attempted.

For these VAs, complete LVOT mapping and ablation by retrograde transaortic approach and antegrade trans-DGCV approach was performed in every patient. When endocardial catheter ablation was unsuccessful with a transaortic approach or a much earlier local ventricular activation was mapped in the DGCV than at any other endocardial site, epicardial catheter ablation by trans-DGCV approach was attempted. If both transaortic and trans-DGCV RFCA didn’t abolish the VAs, mapping and ablation below aortic root via a transseptal approach was finally attempted. Additional LVOT mapping by antegrade transseptal approach was performed for the reason that only by retrograde transaortic approach, the anatomic area just below the left coronary cusp (LCC), which was the endocardial side of left ventricular summit, was sometimes insufficiently mapped. Antegrade transseptal approach was achieved under the guidance of fluoroscopy with transseptal puncture through the anteroinferior fossa ovalis, slightly posterior to the mapping catheter placed in the noncoronary sinus cusp. After transseptal puncture, the reversed S curved ablation catheters was delivered to the area just below the LCC in the left ventricular outflow tract. With the assistance from an Agllis NxT sheath, the ablation catheters were manipulated with ease in this area. During the whole procedure, activated clotting times was monitored and maintained between 250 and 300 s by unfractionated heparin to prevent thrombus.

If the overall site of earliest ventricular activation was found within the DGCV or below the LCC, coronary angigraphy was performed to assess the distance between the tip of ablation catheter and left coronary arteries and their ostia, identify the exact anatomic location and avoid the inadvertent damage to coronary artery.

Once a target site was located, RF energy was released by the irrigated-tip catheter in a temperature-controlled mode with a target temperature of 43 °C, maximum power output of 35 W, and flow rate of 17 ml/min, to achieve a impedance decline of 10-15Ω. If ablation in DCGV was required, the irrigated-tip catheter mode was switched to maximum power output of 25 W and flow rate 30 ml/min. During the catheter ablation within DGCV or below LCC, left coronary angiography was repeated every 15 s to evaluate the distance of the ablation catheter relative to the left coronary arteries and their ostia, the energy delivery should be immediately stopped once the catheter slightly dislodged. Energy delivery was forbidden within 5 mm of a coronary artery. Coronary angiography was routinely performed after successful ablation to assess coronary blood supply. During the first 10 s of the energy release on the target site(20 s for antegrade transseptal approach), if the frequency of VT or PVCs accelerated or decelerated, the energy was applied for another 60 s, if not, the energy application was stopped, and another target site was searched for. Programmed electrical stimulation and isoproterenol infusion were both performed before withdraw of all catheters and sheaths. The RFCA endpoint was defined as absence of spontaneous or inducible VT or PVCs.

Post ablation follow-up

After RFCA, all patients underwent 72-h ECG monitoring deprived of any antiarrhythmic drugs. After discharge, all patients were followed up at our outpatient department. To assess long-term outcome, 24-h Holter monitoring and echocardiography were carried out at 1, 3, 6 and months later. Whenever the patients had symptoms suggesting recurrence of PVCs/IVT, ECG and 24-h Holter were performed immediately. Successful catheter ablation was defined as no recurrence of VAs during follow-up.

Statistical analysis

All the authors did not have access to identifying information during or after data collection. Continuous data were manifested as mean ± standard deviation (X ± s), the analysis of variance (ANOVA) and the student’s t-test were applied to compare the continuous data, while categorical data were manifested as number and percentage, the X²-test or Fisher method were applied to compare the categorical data. P < 0.05 indicated a significant difference. A Receiver operating characteristic (ROC) curve analysis was used to test the predictive value of PdW, IDT, MDI. A two-tailed value of P < 0.05 indicated a statistical significance.

Study population

Results of electrophysiological study, mapping and ablation

Electrocardiographic characteristics

Applying the PdW to predict trans-DGCV approach presented with the sensitivity, specificity, positive predictive value, negative predictive value of 100, 78, 61, 100%, applying the IDT to predict trans-DGCV approach presented with the sensitivity, specificity, positive predictive value, negative predictive value of 100, 78, 63, 100%, applying the MDI to predict trans-DGCV approach presented with the sensitivity, specificity, positive predictive value, negative predictive value of 100, 65, 50, 100%. Receiver Operating characterisitcs (ROC) curve showed that PdW, IDT, MDI of VAs had a good ability to differentiate trans-DGCV approach and non-trans-DGCV approach (transaortic and transseptal approach) with area under curve (AUC) of 0.9098, 0.8988, 0.8806 respectively, PdW > 53 ms, IDT > 74 ms, MDI > 0.45 strongly indicated the trans-DGCV approach could firstly be applied for ablation of VAs arising from LV summit, see Fig. 1. The typical electrocardiographic characteristics of VAs from each group were shown in Fig. 2, and a typical case of RFCA procedure from each group was presented in Figs. 3, 4 and 5.

Anatomy of the LV summit

Epicardially, the LV summit is the area on the epicardium of the LV beneath the bifurcation of the left main coronary artery and superior to the horizon line of the first septal perforating branch, bounded by the left anterior descending artery (LAD) anteriorly and the left circumflex artery (LCX) laterally, the DGCV transversely bisects the LV summit into a superior portion, which is in close proximity to the proximal coronary arteries with epicardial fat overlaying it and inaccessible to catheter ablation, and an inferior portion that can be accessible by trans-DGCV or transpericardial approach [9]. Endocardially, it is an area within LVOT just below LCC, located anterosuperior of the AMC, and is accessible to endocardial catheter ablation [10].

The approaches for RFCA of VAs arising from LV summit

Previous studies described different ablation approaches for VAs arising from left ventricular summit. Yamada et al. reported that VAs origin in left ventricular endocardial summit just below LCC can be ablated by a retrograde transaortic approach with a catheter inversion technique [8], meanwhile, he also demonstrated that VA origin in the accessible area of left ventricular epicardial summit can be ablated by a trans-DGCV approach or a pericardial approach via suboxiphoid puncture [9]. Ouyang et al. recently reported that the left ventricular endocardial summit VAs can be better mapped and ablated via a transseptal approach with the ablation catheter reversed S curved [10]. Nonetheless, no matter which ablation approach mentioned above was chosen for left ventricular summit VAs ablation, its advantages and limitations should be taken into full consideration. The transaortic approach is usually firstly attempted for its relatively safe for there is no coronary artery endocardially, however, because of the relatively rigid aortic valve leaflets limiting catheter manipulation and vigorous movements of the aortic valves rendering a contact of the mapping catheter on the tissue in this region unstable, the area below the ASC is sometime insufficiently mapped, leading to unsuccessful ablation. Trans-DGCV approach should be attempted when an epicardial origin within DGCV was supposed, however, it is sometime difficult to place the tip of an ablation catheter in the narrow vascular lumen, meanwhile, high impedance produced during ablation limits the release of energy, and potentially fatal risks, like pericardial tamponade and coronary artery injury, may occur. Transseptal approach was recently proposed for left ventricular summit VAs ablation, by which access into the left ventricular endocardial summit below the ASCs and catheter manipulation in this area are thought easier, however, the transseptal procedure with a Brockenbrough needle is not always safe, catastrophic complications, including atrial rupture and acute mitral valve damage, may be induced by this approach. In addition, Yamada et al. demonstrated by catheter inversion technique via the retrograde transaortic approach, the looped mapping catheter looked the same as the distal half part of the reverse S curve with the antegrade transseptal approach [8], thus they thought inversion with the retrograde transaortic approach should be attempted before the transseptal approach. Pericardial approach via subxiphoid puncture was rarely adopted for its high risk of severe complications, such as heart rupture and pericardial tamponade, and it was forbidden in inaccessible area where lay the proximal coronary arteries and energy application can lead to fatal myocardial infarction. Thus, when VAs were suspected of a LV summit origin, taking full consideration about the ablation approach prior to the procedure was important, it may save the RFCA time and avoid serious complications.

ECG characteristics of LV summit ECG

Previous studies has described electrocardiographic characteristics criteria to identify the left ventricular summit VAs. Because the LV summit is located most superior in the LV beneath the LCC, VAs originating from this region exhibit high inferior lead R-wave amplitudes and no S wave in V5/V6. Left ventricular summit VAs of left bundle branch block morphology are a rare phenomenon. In our study, all VAs, no matter ablated endocardially and epicardially, had high inferior lead R-wave amplitudes, no S wave in V5/V6, most VAs had right bundle branch block pattern, which is consistent with previous reports [9].

PdW, IDT, MDI in approach selection

Although a series of ECG characteristics of LV summit VAs have been concluded to differentiate itself from the other sites of PVCs/IVTs origin, the ECG characteristics predicting which ablation approach adopted for LV summit VAs have not been studied. A series of electrocardiographic criteria were once put forward to help recognizing the epicardial origin of VAs and consistently suggested that epicardial VAs usually show prolonged PdW, IDT, MDI [1, 9]. The potential explanation for this electrocardiographic phenomenon is that the Purkinje network is primarily located in the subendocardium, thus excitation of an epicardial ectopy exactly needs more time to reach the Purkinje network to fire arrhythmias than that of an endocardial ectopy, resulting in a relative prolonged PdW, IDT, MDI in epicardial VAs than endocardial VAs. In the present study, we also found that for LV summit VAs, the prolonged PdW time and IDT time and MDI indicated an epicardial origin and trans-DGCV approach could be adopted firstly.

Catheter ablation of LV summit VAs by transseptal approach

Ouyang et al. described that LV summit VAs ablated by transseptal approach [10], for these VAs, earliest ventricular activation of target sites preceding the QRS onset by 39.5 ± 7.7 ms, pace mapping at the target site showed mismatch of QRS morphology to clinical VAs, radiofrequency resulted in rapid disappearance of VAs within 10 s. We also found the LV summit VAs ablated by transseptal approach showed poor pace match, however, opposed to Ouyang et al., we found the earliest ventricular activation of target sites preceding the QRS onset just by 28.61 ± 2.71 ms, and these VAs usually need more than 10 s’ energy discharging (less than 20 s). The relative delayed ventricular activation time and prolonged energy application, and the poor pace match of target sites by transseptal approach from endocardium, to some extent, suggested that the origin of VAs may be located subepicardially or intramurally. However, complete epicardial mapping via suboxiphoid approach was not performed in any of our patients, whether a much earlier local ventricular activation can be recorded epicardially is unknown.

Study limitations

In our study, complete mapping of LV summit before ablation by transaortic approach, by trans-DGCV approach, by transseptal approach and by pericardial approach respectively was not performed. The RFCA procedure in the present study was proceeded by transaortic approach and trans-DGCV approach firstly, followed by a transseptal approach. Pericardial approach was abandoned for its high risk of catastrophic complications. Thus where the exactly earliest local ventricular activation of PVCs/IVTs located was not fully investigated. However, for mapping in LV summit VAs, it is something different. The LV summit is the most common area where intramural VAs are located [12]. In the area, even complete mapping from both the endocardium and epicardium performed before ablation, we can not precisely locate the origins of VAs. In our study, as we were more concerned about safety and effectivity of RFCA procedure of LV summit VAs, we performed complete mapping and ablation by transaortic approach and trans-DGCV approach firstly, when failed, the more aggressive transseptal approach was proceeded, by our ablation strategy, a part patients can get away from severe RFCA complications.