Impact of cavotricuspid isthmus depth on the ablation index for successful first-pass typical atrial flutter ablation

Cavotricuspid isthmus (CTI) linear ablation has been established as the treatment for typical atrial flutter. Recently, ablation index (AI) has emerged as a novel marker for estimating ablation lesions. We investigated the relationship between CTI depth and ablation parameters on the procedural results of typical atrial flutter ablation. A total of 107 patients who underwent CTI ablation were retrospectively enrolled in this study. All patients underwent computed tomography before catheter ablation. From the receiver-operating curve, the best cut-off value of CTI depth was < 4.1 mm to predict first-pass success. Although the average AI was not different between deep CTI (DC; CTI depth ≥ 4.1) and shallow CTI (SC; CTI depth < 4.1), DC required a longer ablation time and showed a lower first-pass success rate (p < 0.01). In addition, the catheter inversion technique was more frequently required in the DC (p < 0.01). The lowest AI sites of the first-pass CTI line were determined in both the ventricular (2/3 segment of CTI) and inferior vena cava (IVC, 1/3 segment of CTI) sides. The best cut-off values of the weakest AIs at the ventricular and IVC sides for predicting first-pass success were > 420 and > 386, respectively. Among patients with these cut-off values, the first-pass success rate was 89% in the SC and 50% in the DC (p < 0.01). Although ablation parameters were not significantly different, the first-pass success rate was lower in the DC than in the SC. Further investigation might be required for better outcomes in deep CTIs.


Material and methods
Study design. All consecutive patients who underwent CTI linear ablation at Wakayama Medical University Hospital from January 2018 to December 2020 were retrospectively included in this study. Of these, a 3D electroanatomical mapping system (CARTO, Biosense Webstar, Irvine, CA) was not applied in three patients, computed tomography was not performed prior to CTI ablation in two patients, and ablations for recurrent AFL were performed in three patients. Therefore, these patients were excluded from the analysis. This study was carried out in accordance with the Declaration of Helsinki. This study was approved by the local ethics committee (Research Ethics Committee of Wakayama Medical University, 3053) and the requirement for written informed consent was waived because of the retrospective nature of the study.
Ablation method. The patients were mildly sedated with hydroxyzine pamoate and/or dexmedetomidine hydrochloride. Right internal jugular vein access was obtained, and a multipolar deflectable catheter was inserted into the coronary sinus. Another multipolar catheter (Snake, Japan Life Line, Tokyo) was placed close to the tricuspid annulus via the right femoral vein. A deflectable sheath (Agilis, Abbott, St. Paul, MN) was also introduced from the right femoral vein for an ablation catheter (SmartTouch SF, Biosence Webster, Irvine, CA). Radiofrequency ablation was performed starting from the tricuspid valve to the inferior vena cava (IVC). An irrigated ablation catheter was used with point-by-point application. In patients with AFL during the procedure, CTI-dependent AFL was diagnosed by entrainment pacing from the CTI. If the patient presented with sinus rhythm, continuous pacing at 500-700 ms cycle from the proximal coronary sinus was performed during CTI ablation.   www.nature.com/scientificreports/ Both the VisiTag module and the fluoroscopy provided each radiofrequency (RF) location during the procedures. In this study, our setting of VisiTag is as follows: minimum time of 5 s, maximum range of 2.5 mm, minimum contact force (CF) of 5 g, and force over time of 25%. RF energy was 25-35 W, with saline irrigation at a flow rate of 8 mL/min or 15 mL/min, duration at each RF site was between 15 and 35 s, aiming for bipolar electrogram voltage attenuation of ≥ 80%, and distance between two neighboring RF sites did not exceed 6 mm. In cases with strong pain, 25 W of RF energy was applied, especially near IVC. Although AI data was displayed to operator in last 25 patients from April to December 2020, operators did not refer to AI during ablation and each AI data was retrospectively analyzed after ablation. The bidirectional conduction block was confirmed using differential pacing from the proximal coronary sinus and inferior lateral wall of the right atrium. The widely spaced double potential was also confirmed. If incomplete conduction block was suspected near the IVC, the ablation catheter would be deflected by more than 90° to ablate the interior of the pouch (Fig. 1) 13 . The complication was defined as vascular complication (pseudoaneurysm, fistula and bleeding requiring transfusion), cardiac tamponade, valvular damage, pneumothorax and hemothorax.
Ablation data analysis. The RF sites of the CTI were divided into two parts: the ventricular side (2/3 segment of CTI) and the IVC side (1/3 segment of CTI) 12 . First-pass success was defined as requirement of no additional ablation after performing the first CTI line. Each VisiTag data, including ablation duration, mean CF, FTI, and AI, were retrospectively analyzed. The lowest AI sites of the first-pass CTI line were determined in both the ventricular and IVC sides (Supplementary Figure 1). The interrupted ablation sites due to catheter instability were defined as ablation duration of less than 15 s and excluded from the lowest AI site of the first-pass line determination.
Computed tomography protocol and analysis. All patients underwent multi-detector computed tomography (MDCT) before catheter ablation. MDCT was performed using a 320-slice MDCT (Aquilion ONE, Canon Medical Systems, Otawara, Japan), with retrospective ECG-gated scans. In the routine protocol of atrial fibrillation, the total iodine dose was 160 mg/kg. On occasion, the application of contrast media was avoided because of chronic kidney disease and/or on the operator's judgment. The images during the atrial diastolic phase of the cardiac cycle (30-40% of the interbeat interval) were reconstructed in different planes on a workstation (Intuition Thin Client, TeraRecon, Durham, NC, USA). Anatomical information was analyzed in the sagittal plane across both the center of the tricuspid valve and the IVC.
As shown in Fig. 2, we analyzed the following: a) the length the CTI, b) the depth of the CTI, c) the length from the right coronary artery to the right atrium, d) the height of the Eustachian ridge, and e) the angle between the CTI and the IVC. The measurement of Eustachian ridge was made from its basis to its distal extremity in the plane passing through the center of the IVC 8 . In MDCT image with contrast media, the thickness of CTI was also analyzed. Measurement data of CTI was blinded to operators before ablation.

CTI depth for first-pass success of ablation.
From the ROC, the best cut-off value of CTI depth to predict first-pass success of CTI ablation was < 4.1 mm with a sensitivity of 67% and specificity of 81% (Fig. 3). In both ventricular and IVC sides, averages of lowest AI were lower in unsuccessful cases than in successful cases (p < 0.01 and p < 0.01, respectively) (Fig. 4A). The conduction gaps in ventricular side (n = 40) were more frequently observed than in IVC side (n = 15) (37% vs. 14%, p < 0.01). From the ROC, the cut-off values of AI for www.nature.com/scientificreports/ predicting first-pass success of conduction block in ventricular and IVC sides were > 420 and > 386, respectively, with area under the curve of 0.68 and 0.75, respectively (Fig. 4B, C). Among patients with anterior (> 420) and posterior (> 386) AIs, first pass success was achieved in 16 of total 18 patients with SC and in 6 of total 12 patients with DC. Limited in patients with these cut off values, the first-pass success rate was higher in the SC than in the DC (89% vs. 50%, p < 0.05). The sensitivity and specificity, positive predictive value and negative predictive value of lowest AI cut off values to predict first-pass success were 40%, 78%, 89% and 23% in SC and 30%, 84%, 50% and 70% in DC, respectively.

Discussion
Major findings. In the present study, to best our knowledge, we firstly investigated the association of CTI depth and AI with first-pass success of radiofrequency catheter ablation for typical AFL. A CTI depth of < 4.1 mm was a best cut off value for first-pass success of CTI ablation. Although the individual ablation parameters were not different, the first-pass success rate was higher in the SC than in the DC.

CTI depth and first-pass success. CTI ablation is an established strategy for typical atrial flutter and
can provide durable success and a low recurrence rate. However, we sometimes experience laborious cases with prolonged procedural time. Considering several reports, the anatomy of the CTI can reflect procedural difficulty, and concave-shaped and/or pouch-like CTI, are especially associated with laborious interventions 6,7 . In this study, CTI of less than 4.1 mm depth was a best cut off for the first-pass success of CTI ablation. This cut-off value is similar to the < 3.9 mm CTI depth necessary for knuckle curve ablation reported by Shimizu et al. 13 . Consistent with a previous report, longer procedural time was observed in the deep CTI cases 7 .

Ablation parameters in shallow and deep CTI.
It has been acknowledged that CTI ablation tends to be more difficult in concave CTIs. However, it was unclear whether the CF was adequate, there was any gap between each ablation lesion, and ablation power was enough to achieve block line in concave CTIs. Currently, the CF and VisiTag module have been proven as effective tools, and the AI can evaluate lesion formation more precisely than FTI regardless of ablation power [14][15][16][17] . Although each ablation parameter was not different between the SC and the DC, the first-pass success rate of the SC was higher than that of the DC. In the present study, we conducted CTI ablation with a 3D mapping system, and there was no gap between each ablation point on the CARTO system. However, in deep CTIs, complete block line was not occasionally achieved regardless of the lesions having adequate AI. Several potential mechanisms have been proposed. First, there might be a micro pouch, which was not assessed by CT, especially in the DC (Supplementary Fig. 2A). Second, the atrial wall in some patients may be www.nature.com/scientificreports/ thick. Knecht et al. reported that the thickness of the CTI is an independent predictor of difficulty in CTI ablation, but this could not be fully evaluated in this study because of non-contrast computed tomography (CT) imaging used in some patients 8 . Third, in patients with a high Eustachian ridge, a catheter inversion technique was required. Therefore, in this type of patients, it is difficult to achieve complete ablation of the lesions near the IVC (Supplementary Fig. 2B). Lastly, there might be a gap in the atrial myocardium regardless of the absence or presence of gaps in the 3D mapping system. The locations are result of the ablation catheter pushing the myocardium wall. Therefore, the curve shaped atrial wall in the DC might cause the gap (Supplementary Fig. 2C).
To improve the first-pass success rate, one of the strategies is to raise the target AI. In a previous report, the anterior and posterior AI targets were set at 500/400 12 . A higher first-pass rate of 93% was achieved without any complications. The appropriate target AI was decided according to the arterial wall thickness. Another solution . AI and first-pass success. (A) Box plot of lowest AI in comparison with first-pass success and no firstpass success cases. The lowest AI was higher in first-pass success cases both in ventricular side (p < 0.01) and IVC side (p < 0.01). Data are presented as box plots with median and 25th and 75th percentiles (box) and 5th to 95th percentiles (whiskers). IVC; inferior vena cava. (B) ROC for predicting first-pass success of CTI ablation according to the lowest AI at the ventricular side. The area under the curve was 0.68 with a cut-off value of > 420. The sensitivity, specificity, positive predictive value and negative predictive value were 55%, 78%, 80% and 51%, respectively. (C) ROC for predicting first-pass success of CTI ablation according to the lowest AI at the IVC side. The area under the curve was 0.75 with a cut-off value of > 386. The sensitivity and specificity were 60% and 87%. The sensitivity, specificity, positive predictive value and negative predictive value were 60%, 87%, 96% and 26%, respectively. AI, ablation index; CTI, cavotricuspid isthmus; ROC, receiver operating curve. www.nature.com/scientificreports/ is to modulate the catheter shape in the DC according to preoperative MDCT in formation although radiation exposure increases. Kajihara et al. adapted the procedure to anatomical characteristics by inverting the catheter near the IVC 18 . Combining the above two strategies, a tailor-made ablation method would be preferable by assessment of CT prior ablation. Further investigation is needed to improve AI guided CTI ablation, especially in deep CTI case.

Study limitations.
There are some limitations to this study. Because patients were retrospectively enrolled and derived from a single center, a selection bias may exist. Furthermore, the sample volume was relatively small. In addition, the CTI lines between actual ablation lesions and CT imaging might be different. Because this study focused only on the first-pass success rate of CTI ablation, its long-term efficacy remains uncertain. The data of catheter angulation was lacking. The thickness of the CTI was analyzed in limited populations, which could influence the success rate. In this study we divided CTI into ventricular and IVC side. Because we evaluate RF location based on VisiTag, the site might be different from anatomical assessment. Lastly, the area under the curve of ROC to predict first-pass success of conduction block in ventricular and IVC sides by AI was low.

Conclusions
Although the ablation parameters were not different, the first-pass success rate was lower in the deep CTI than in the shallow CTI. Further investigation, including target AI, distance interval of RF sites and appropriate catheter position, would be required for better outcomes in deep CTIs.