Differentiation between epicardial and endocardial conduction gaps for left atrial posterior wall isolation using peak frequency annotation algorithm

Key Teaching Points.

• •In the left atrial posterior wall, atrial potentials originating from endocardial and epicardial sides are often observed simultaneously. The Omnipolar Technology Near Field (Abbott, St. Paul, MN) algorithm can differentiate between endocardial and epicardial electrograms recorded in the left atrial posterior wall.
• •The emphasis map, which can superimpose the peak frequency map and voltage map, can accurately identify the location of the residual conduction gap by adjusting the threshold of the peak frequency value.
• •Both the Omnipolar Technology Near Field algorithm and emphasis map may be useful for achieving left atrial posterior wall isolation.

Introduction

A multicenter randomized clinical trial has found that the addition of left atrial posterior wall (LAPW) isolation to pulmonary vein isolation (PVI) did not significantly improve freedom from atrial arrhythmias compared with using PVI alone in patients with persistent atrial fibrillation (AF).^1,2 However, LAPW isolation added to PVI is a therapeutic option that could be considered based on individual characteristics in patients with persistent AF. When performing LAPW isolation, a transmural conduction block by left atrial (LA) roof and floor linear ablations remains challenging, mainly due to epicardial muscular fibers that bridge epicardial by endocardial conduction, such as the septopulmonary bundle.³ A residual conduction through epicardial fibers was observed in 40.2% of cases following first-pass linear ablation procedures of the LA roof and floor.⁴

Recently, a novel EnSite Omnipolar Technology Near Field (OTNF; Abbott, St. Paul, MN) algorithm, which can automatically measure peak frequency (PF) value in local electrograms, has become available. Previous studies have found that the OTNF algorithm can differentiate between far- and near-field signals in complex electrograms.^5,6 However, the utility of the OTNF algorithm in identifying the localization of residual conduction gaps after LAPW ablation has not been elucidated. Herein, we present a case of persistent AF, wherein residual endocardial conduction after LAPW isolation was identified using the OTNF algorithm by differentiating between the epicardial and endocardial electrograms.

Case report

A 68-year-old man with persistent symptomatic AF refractory to medical therapy was referred for radiofrequency catheter ablation (RFCA). The AF lasted approximately 3 years, and the LA diameter was 48.3 mm. Therefore, we decided to perform PVI with LAPW isolation as our ablation strategy. After converting AF to sinus rhythm by electrical cardioversion, an extensive encircling PVI was completed using a 3-dimensional mapping system (EnSite X, Abbott) and a TactiFlex (Abbott) ablation catheter. Regarding PVI, the radiofrequency (RF) application was set as follows: power, 50 W; contact force, 10–20 g; and duration, 20 s, 15 s, and 15 s for anterior, roof, and posterior segments, respectively. After PVI was completed, linear LA roof and floor ablations for LAPW isolation were performed using a point-by-point application. The RF application was set as follows: power, 50 W; contact force, 10–20 g; and duration, 20 s for the roof line or 15 s for the floor line. The target interlesion distance was 4 mm for both linear ablations. The esophageal temperature was monitored and RF delivery was interrupted once the temperature reached 39°C.

After linear ablation, the local electrograms remained within the LAPW box, indicating incomplete LAPW isolation. Therefore, to identify the location of the residual conduction gap, high-density maps were created using the Advisor HD Grid (Abbott) catheter during atrial pacing from the distal coronary sinus at 600 milliseconds, using the following 2 distinct annotation algorithms: absolute dV/dt and OTNF. Activation vectors obtained via the absolute dV/dt algorithm on an omnipolar (HD Wave Solution algorithm; Abbott) voltage map suggested a residual conduction gap on the left side of the floor line (Supplemental Video 1). In contrast, the activation vectors obtained via the OTNF algorithm showed that conduction propagated through the center of the floor line into the LAPW box slightly later than that observed on the left side (Supplementary Video 2). The electrograms annotated using the absolute dV/dt and OTNF algorithms inside the LAPW box are shown in Figures 1A and 1B, respectively. The electrograms recorded at the lower left of the LAPW were formed with a dull component, whereas those recorded at the center of the LAPW were formed with a dull component, followed by a sharp component. In the central LAPW, the OTNF algorithm annotated the sharp components of the electrograms (red line and white arrow in Figure 1B). In contrast, the absolute dV/dt algorithm annotated the dull components (yellow line and white arrow in Figure 1A). The emphasis map with a fusion of the omnipolar voltage and PF maps revealed the residual conduction gap at the center of the floor line by increasing the threshold PF value (Figure 2). At the presumed gap site on the central floor line, the absolute dV/dt algorithm annotated the dull component with the high and steep slope of the electrogram, whereas the OTNF algorithm annotated the sharp component slightly later compared with the dull component (Figure 3A). Subsequently, RF applications were added to the presumed gap site (Figure 3A). Immediately after application at 50 W, the sharp components of the electrograms recorded on the grid catheter (electrodes C1–2 and D1–3) positioned at the LAPW box were gradually delayed and finally eliminated (Figure 3B). However, low-frequency electrograms remained inside the LAPW box, indicating that the LAPW was isolated only on the endocardial side. Although high-output pacing at 10 V inside the LAPW box did not result in atrial capture outside the LAPW box, the pacing failed to capture residual electrograms within the LAPW (Supplemental Figure 1). To complete the transmural LAPW isolation, an additional RF was applied to the left side of the floor line. Ultimately, however, we abandoned transmural LAPW isolation because the esophageal temperature increased easily. Atrial burst pacing did not induce atrial arrhythmias after PVI and the selective endocardial LAPW isolation. The patient had no perioperative complications and no recurrences of atrial arrhythmias during the follow-up of 4 months.

Figure 1.

Electroanatomic map and residual electrograms in the left atrial posterior wall after linear left atrial roof and floor ablations. A: An Omnipolar voltage map generated using an annotation algorithm employing absolute dV/dt. The yellow lines and white arrows indicate annotated sites in the local electrograms. B: Map constructed using Omnipolar Technology Near Field (OTNF, Abbott, St. Paul, MN) annotation algorithm. The red lines and white arrows indicate annotated sites. The OTNF algorithm annotated sharp potentials slightly later than those annotated using the absolute dV/dt algorithm. In the voltage map, purple indicates regions with an amplitude of 0.5 mV or more, and gray indicates regions with <0.05 mV.

Figure 2.

Emphasis maps varying a threshold in peak frequency value. Emphasis maps were constructed using the peak frequency and Omnipolar voltage maps. The area with a peak frequency value above the threshold is highlighted. The yellow arrow indicates the residual conduction, where left atrial posterior wall isolation was achieved after additional radiofrequency applications.

Figure 3.

The electrograms recorded in the left atrial posterior wall (LAPW) during the additional radiofrequency (RF) application at the center of the left atrial (LA) floor line. A:Left panel: Emphasis map with a threshold at a peak frequency of >200 Hz. The yellow arrow on the central LA floor indicates the successful radiofrequency catheter ablation (RFCA) site for LAPW isolation. The grid catheter was placed in the LAPW during the RF application. The red tags represent RFCA points on the bilateral extended-encircling pulmonary vein isolation, LA roof, and floor. Right panel: The difference in annotation timing (yellow lines and white arrows) between using the absolute dV/dt and the Omnipolar Technology Near Field (OTNF, Abbott, St. Paul, MN) algorithms at the successful ablation site. B: Electrograms were recorded in the LAPW using a grid catheter during the pacing of the distal coronary sinus. After RF application at the presumed gap site for 7 seconds, the sharp components of the electrograms were delayed (red arrows) and finally disappeared after 12 seconds. Electrograms with low peak frequency retained in the LAPW box thereafter (blue arrowheads). CS = coronary sinus.

Discussion

To the best of our knowledge, this is the first case of LAPW isolation showing that an OTNF algorithm could differentiate between epicardial and endocardial signals, resulting in the successful localization of an endocardial conduction gap. Localizing residual conduction gaps after LA linear ablation for LAPW isolation is sometimes challenging. One reason is that atrial potentials originating from the endocardial and epicardial sides can be recorded simultaneously.⁷ Conventional absolute dV/dt annotation algorithm commonly annotates high-amplitude electrograms. Thus, the dV/dt algorithm probably fails to accurately identify the localization of an endocardial residual conduction gap in situations when both endocardial and epicardial electrograms are recorded. In contrast, using the PF value, the novel OTNF annotation algorithm can identify whether the obtained electrograms are near- or far-field signals. Hence, when performing LAPW isolation, the OTNF algorithm may differentiate between endocardial and epicardial signals. In this study, the absolute dV/dt and OTNF annotation algorithms showed different activation vector patterns in the LAPW box. Based on the annotated electrograms, various patterns of activation vectors via the 2 algorithms might have resulted from the differences in the timing of epicardial and endocardial conduction into the LAPW box. The conduction gaps at the central LAPW presumably remained on both the endocardial and epicardial sides, whereas the conduction gap at the left LAPW was located only on the epicardial side. The OTNF algorithm preferentially identified the location on the central floor line where the endocardial conduction gap was present. However, post-mapping could not be obtained after the selective endocardial LAPW isolation due to the limited procedural time. Therefore, the present study could not demonstrate whether the PF value near the gap in the central LAPW had changed after the selective endocardial LAPW isolation.

In this case, although only selective endocardial LAPW isolation was successful, transmural LAPW isolation was not ultimately achieved. The LA wall consists of 2 layers—endocardial and epicardial layers—and the gap with fat tissues between the septoatrial bundle (endocardial side) and septopulmonary bundle (epicardial side) mostly occurs in the LAPW.³ A previous study that evaluated the localization of the septopulmonary bundle using magnetic resonance imaging demonstrated that the septopulmonary bundle extended from the LA roof through the LAPW and could exceed beyond the floor line,⁸ suggesting that the gap between the epicardial and endocardial sides could occur in the LA floor. Accordingly, the selective endocardial LA floor block in the present study might have resulted from the gap between endocardial and epicardial bundles. In addition, high-output pacing from the LAPW failed to capture the residual atrial electrograms after the selective endocardial LAPW isolation, probably due to the gap between endocardial and epicardial bundles by structures with low electrical conductivity. However, because we did not perform pacing at an output of >10 V, we could not conclude whether the inability to capture residual atrial signals was due to a unidirectional block in the epicardial side or insufficient pacing output to capture epicardial potentials.

The emphasis map, available in the EnSite X system, allows for the visualization of other maps superimposed on a selected base map. Herein, the PF map was superimposed on the voltage map, which allowed us to simultaneously assess the amplitude of atrial potentials and localization of near-field potentials based on the PF values. In addition, the emphasis map helped determine the target site for RFCA, and the conduction gap site where RFCA completed LAPW isolation was highlighted by varying the threshold of the PF value. A previous study of patients with atrial flutter demonstrated that PF values recorded at isthmus sites were significantly higher with slower conduction velocities than those at the nonisthmus sites.⁵ In the present case, a possible examination for the higher PF value indicating successful ablation site is that conduction through the conduction velocity decreased through the residual gap.⁹ Given these findings, narrowing down the area with high PF values using the emphasis map may be useful for localizing the residual conduction gap. In our case, a threshold PF value of >200 Hz was the optimal cutoff for the localization of the residual conduction gap. However, the optimal cutoff PF value may differ depending on various factors, such as patient characteristics and the propagation direction of activation. Further investigations are required to determine the optimal cutoff value for identifying the localization of the conduction gap in LAPW isolation.

Conclusion

The OTNF annotation algorithm could differentiate between epicardial and endocardial potentials, resulting in the successful localization of an endocardial gap for LAPW isolation. Assessing the PF values of local electrograms may be useful for localizing residual conduction gaps after LA linear ablations.

Disclosures

The authors have no conflicts of interest to disclose.