Identification of critical isthmus using coherent mapping in patients with scar‐related atrial tachycardia

Abstract Introduction Accurate identification of slow conducting regions in patients with scar‐related atrial tachycardia (AT) is difficult using conventional electrogram annotation for cardiac electroanatomic mapping (EAM). Estimating delays between neighboring mapping sites is a potential option for activation map computation. We describe our initial experience with CARTO 3 Coherent Mapping (Biosense Webster Inc,) in the ablation of complex ATs. Methods Twenty patients (58 ± 10 y/o, 15 males) with complex ATs were included. We created three‐dimensional EAMs using CARTO 3 system with CONFIDENSE and a high‐resolution mapping catheter (Biosense Webster Inc). Local activation time and coherent maps were used to aid in the identification of conduction isthmus (CI) and focal origin sites. System‐defined slow or nonconducting zones and CI, defined by concealed entrainment (postpacing interval < 20 ms), CV < 0.3 m/s and local fractionated electrograms were evaluated. Results Twenty‐six complex ATs were mapped (mean: 1.3 ± 0.7 maps/pt; 4 focal, 22 isthmus‐dependent). Coherent mapping was better in identifying CI/breakout sites where ablation terminated the tachycardia (96.2% vs 69.2%; P = .010) and identified significantly more CI (mean/chamber 2.0 ± 1.1 vs 1.0 ± 0.7; P < .001) with narrower width (19.8 ± 10.5 vs 43.0 ± 23.9 mm; P < .001) than conventional mapping. Ablation at origin and CI sites was successful in 25 (96.2%) with long‐term recurrence in 25%. Conclusions Coherent mapping with conduction velocity vectors derived from adjacent mapping sites significantly improved the identification of CI sites in scar‐related ATs with isthmus‐dependent re‐entry better than conventional mapping. It may be used in conjunction with conventional mapping strategies to facilitate recognition of slow conduction areas and critical sites that are important targets of ablation.


| INTRODUCTION
Conventional mapping of complex scar-related atrial tachycardias (ATs) can be challenging. 1 Three-dimensional (3D) electroanatomic mapping (EAM) systems facilitate such difficult interventional ablation procedures and allow accurate navigation to a predefined site with favorable spatial resolution and visualization of the activation sequence (activation mapping) and voltage information (voltage mapping). 2 An important factor that influences the accuracy of the activation map includes the consistency of electrogram annotation which is dependent on numerous algorithms (eg, peak amplitude or rapid downstroke of the unipolar signal) that can be selected for automatic signal annotation. 3 The presence of scarred tissue from prior surgery or ablation can cause a variable delay between the onset of the local electrogram and the time to the peak amplitude. 4 This limits the accuracy of the activation mapping which relies on the correct annotation of the impulse timing different from that of a reference electrode. Other limitations of activation mapping include choosing the mapping window of interest based on an arbitrarily defined early and late activation and difficulty in differentiating active diastolic activity that is part of the re-entrant circuit from passive diastolic activity recorded in scar unrelated to the tachycardia. 5 An algorithm incorporating conduction velocity vectors computed taking into account a global best-fit solution allows the management of such complex electrograms, identifies abnormal substrate with slow or nonconducting (SNO) zones and facilitates interpretation of such complex AT mechanisms. 5 In this pilot study, we eval-  focus had either a transesophageal echocardiography or cardiac computed tomography scan done to exclude LA appendage thrombus within 2 days before procedure.

| Electrophysiological study
The details of mapping have been described previously. In brief, each patient underwent an electrophysiological study and catheter ablation in the postabsorptive, nonsedated state, 6 were placed through sheaths into the atrial chamber of interest.
Activation mapping was directly performed for all patients who were in incessant AT on presentation and AT induction was done for those in sinus rhythm before mapping. Induction of AT was done with stimulation from the proximal and distal coronary sinus. 8,9 If AT is not induced with programmed electrical stimulation, intravenous isoproterenol (at graded dosages from 1 to 4 μg/min) was infused until AT developed or the sinus rate increased to 20% above the resting value. 8,10,11 After inducing sustained AT, EAM was performed and the suspected circuits were confirmed with entrainment maneuvers, and postpacing interval (PPI) analyses from multiple sites to identify the mechanism of the AT. Pacing sites with a PPI less than or equal to 20 ms of the CL were considered as VICERA ET AL.

| 1437
part of the circuit. Macroreentrant AT accounted for at least 80% of the tachycardia CL on activation mapping. 12 The absence of entrainment from multiple pacing sites and a centrifugal activation from a focal site indicate focal AT as the mechanism. 9,11,13,14 Transseptal puncture under right atriography guidance was performed to access the LA, if needed, and intravenous heparin was administered to maintain an activated clotting time between 250 and 350 seconds.

| Data collection settings
All maps were acquired using a multielectrode mapping catheter (Lasso Nav or PentaRay Nav; Biosense Webster Inc,) and CARTO 3 system CONFIDENSE module with the continuous acquisition of electroanatomical (EA) points that meet a set of predefined filters. Points were accepted only if they pass all the selected filters including cycle length stability within a 5% range of the tachycardia cycle length, local activation time (LAT) stability within 4 ms and an electrode position stability within 4 mm. A tissue proximity filter based on catheter location and impedance measurements was used to determine electrode proximity to cardiac tissue. The greatest negative deflection (-dV/dt) in each distal unipolar signal was used for the LAT calculation. Map consistency display was used to search for, and filter points whose LAT values are not consistent with the LAT values of neighboring points to highlight these outliers. The interpolation threshold of the color fill-in of the reconstruction was set at 5% during map acquisition to permit a relatively uniform density of mapping points. 15 Bipolar electrograms were filtered between 16 and 500 Hz, unipolar electrograms between 2 and 240 Hz and recorded digitally. 7 SNO zones were defined by the absence of recordable activity in collected points having bipolar voltage amplitude ≤ 0.03 mV (baseline noise in the Biosense system). SNO zones were the result of the coherent mapping algorithm, exhibit absence of conduction velocity vectors on coherent map, and appear as brown on the 3D maps ( Figure 1A). 15 For optimal performance, the following were likewise adjusted.
• Scar settings were defined with a voltage cutoff of less than 0.05 and less than 0.5 mV used to define scar and low voltage zones, respectively. 16 F I G U R E 1 Effect of different scar settings on activation and coherent maps for identification of isthmus site top panels correspond to coherent maps. The bottom panels correspond to LAT maps. The scar threshold setting was adjusted with 0.01, 0.03, and 0.05 mV thresholds from left to right, respectively. The figure demonstrates increasing SNO areas on the recalculated coherent maps and increasing patches of gray areas on the LAT maps as the scar threshold setting is increased. The changes are more apparent on the coherent maps than the LAT maps. The 0.01 mV is shown only for demonstration purpose but was not applied during the actual cases because a threshold less than 0.03 mV is below the noise threshold for the CARTO system. LAT, local activation time; SNO, slow or nonconduction areas • Signals that were identified as double-potential and fractionated were tagged.
• The interpolation threshold of the color fill-in of the reconstruction was set at 5% during map acquisition to permit a relatively uniform density of mapping points.
• Anatomical structures were marked and cut before coherent map calculation.
All the maps were interpreted by two operators (JJBV and PTL), if there were disagreements in the map interpretation, a third operator (YJL) was consulted. The standard LAT map was evaluated first with adjustments made with results of entrainment mapping and the use of the extended early-meets-late (EEML) feature of the software. This was followed by calculation of the coherent map. The ablation strategy was performed based on the interpretation of the coherent map.

| Standard activation and propagation
The details of standard EAM has been described previously. 7 In brief, a sharp and stable signal from a decapolar catheter inserted into the coronary sinus was used to provide the timing reference signal during the mapping procedure. 11 After completing an intact right atrial (RA) and/or LA geometry reconstruction, depending on suspected chamber site of origin based on surface ECG P wave morphology, 12,17 a multielectrode mapping catheter was selected as the roving catheter that was used to collect the LAT (relative to the reference signal) and voltages. The signal from the roving catheter was used to build a sequential map. Total activation time was defined as the time interval from the earliest to the latest activation point in the atrium. 7 The extended early-meets-late feature of the software was turned on during map review to aid map interpretation. This feature highlights areas of potential conduction block thereby providing a basis for a better interpretation of the LAT and propagation map. 18  The main conditions used by the algorithm are:

| Coherent mapping
• low (bipolar voltage) potential points such as scar, • double-potential points, and

| Coherent map interpretation and identification of critical isthmus sites
Maps were reviewed and evaluated for the presence of focal activation or macroreentry. In cases of macroreentry, critical isthmus sites, defined as the narrowest region of orthodromic conduction in the isthmus bounded on both sides by SNO regions (brown areas on the coherent map) due to functional or permanent block as determined from the coherent map or by block on one side and an anatomic edge on the other. 19 The critical isthmus was identified and measured by adjusting the displayed scar threshold setting between 0.03 and 0.05 mV, only including points above the background noise setting.

| Catheter ablation
A planned ablation lesion set was defined based on the studied coherent map targeting the earliest activation sites for focal AT and slow con- AT. For isthmus-dependent ATs, the ablation lesions were performed continuously while repositioning the catheter tip every 40 seconds between nonconducting sites that allowed crossing the critical isthmus. 8,9,12 The common isthmus was not always targeted and depended on the preference of the operator such as the decision to ablate the anatomical isthmus or cavotricuspid isthmus for peritricuspid reentrant tachycardia. Ablation was considered successful if the tachycardia terminated or changed to a different activation pattern with an associated change in cycle length and negative inducibility of clinical AT by programmed extra stimuli from the CS catheter with intravenous isoproterenol (1-4 μg/min) infused to achieve at least a 20% heart rate increment. 13,14,23 If tachycardia transitioned to another AT, the operators were encouraged to remap the atrial chamber of interest.

| Follow-up
Patients were followed-up at our cardiology clinic or with the referring physicians initially at 2 weeks after discharge, then every Holter monitoring and/or cardiac event recording for 1 week was performed. Atrial tachyarrhythmia recurrence was defined as an episode lasting more than 1 minute and confirmed by an ECG occurring after 3 months of the blanking period. If more than one episode of recurrent symptomatic atrial arrhythmia was documented, the patients were encouraged to receive a second ablation procedure or were prescribed antiarrhythmic drugs to control recurrence. 7

| Statistical methods
Parametric data are reported as mean ± SD. A χ 2 with a Fisher's exact test was used for categorical data. The student independent t test with Levene's test for homogeneity was used for continuous data.

| Patient characteristics
Twenty patients (15 males, mean age: 58.2 ± 10.3 years old) met the inclusion criteria and were included in the study; nine with history of cardiac surgery (five for valvular heart disease, two with concurrent coronary arterial bypass graft [CABG], and another two with concurrent MAZE procedure), two with the previous repair of congenital atrial septal defect, two with previous CABG, 16 with history of previous atrial ablation and one that developed AT after AF ablation. Hypertension was present in 35%, diabetes in 15%, heart failure in 5%, coronary artery disease in 30%, hyperlipidemia in 20%, valvular heart disease in 25%, previous atrial arrhythmia ablation in 80% and prior cardiac surgery in 45%. The mean left ventricular ejection fraction (LVEF) was preserved (mean: 59.6% ± 9.4%) with only two patients with LVEF less than 55%.
The mean left atrial diameter (LAD) was 45.7 ± 9.8 mm with only five patients with LAD of less than 40 mm. Tables 1 and S1 provides a summary of the clinical cases.    The LAT maps were extended to 80% to 100% during map review, however, there was no identifiable potential isthmus in 30.8% of the LAT maps.    Figure 3B demonstrates the critical isthmus site of a mitral flutter bordered by anatomical and functional nonconducting areas with slow conduction, concealed entrainment, and verified by termination with ablation. Coherent mapping significantly identified more ATs with conduction isthmus ≥ 3 compared with LAT mapping which was not able to identify conduction isthmus ≥ 3 (Table 3).

| Comparison with previous studies
Our study confirms the outcome of the recent study by Anter

| Diagnosis and ablation of scar re-entrant AT
A change in cycle length and activation pattern during ablation occurred in seven of the ATs included in our study likely due to the presence of codominant circuits. Although, the most intuitive ablation strategy is to target the common isthmus as ablation of one loop may produce a sudden transformation to a new re-entrant tachycardia formed by the remaining loop that requires ablation at a second isthmus. 25 However, sometimes radiofrequency ablation of two distinct isthmuses may be a better option than transection of the common isthmus due to location or a wider diameter of the common isthmus than separate multiple isthmuses. 26  open-heart surgery and should be suspected in case of tachycardia change during radiofrequency ablation. 25 Our study showed that the coherent mapping module can clearly identify multiple circuits through the projected conduction velocity vectors and SNO sites.
However, the complexity of the AT mechanism and circuits can still cause a change in the propagation and transfer to another existing circuit after the ablation of the first circuit. The identification of all the loops is critical to the ablation strategy which targets the common isthmus when at least one nonanatomic loop is involved and is reported by Takigawa et al 27 to achieve acute success in terminating the tachycardia in 80%. It is therefore essential to determine whether one loop is a "functional" circuit that will keep rotating once the first loop is abolished by radiofrequency ablation of the other loop, or whether it is an "innocent" bystander that will not require additional treatment. 26 Recently, Takigawa et al 27 reported that when dual-loop ATs included a nonanatomic loop, the common isthmus was generally short, and the ablation was successful. When the arrhythmia combined two nonanatomic loops, a complete anatomic linear ablation was usually not needed and accounted for the better outcome observed in their study. On the contrary, when the dual-loop AT included two anatomical multifocal AT circuits, a complete anatomic isthmus block was required. 27 Verma et al, 28  Similar to our previous findings, 16 this study showed that focal ATs tended to originate from areas of low voltage (0.39 ± 0.21 mV).