Magnetic resonance imaging–guided cryoballoon ablation for left atrial substrate modification in patients with atrial fibrillation

Cryoballoon ablation (CBA) for pulmonary vein isolation (PVI) is an established modality for the treatment of atrial fibrillation (AF). We report feasibility of left atrial (LA) substrate modification in addition to PVI both using the cryoballoon.

Conclusion: LA substrate modification in addition to PVI using LGE-MRI-guided CBA is feasible but still experimental. The efficacy and safety have to be investigated in a prospective randomized trial. In patients with significant extrapulmonary LA substrates, the utility of CBA is less than straightforward. In such patients, the choices are to either use CBA in a PVI-only approach or alternatively, use CBA for PVI followed by the use of radiofrequency (RF) ablation for additional substrate modification. The ability to use CBA for LA substrate modification in addition to PVI would therefore provide a promising advantage.
Three-dimensional (3D) late gadolinium enhancement magnetic resonance imaging (LGE-MRI) has emerged as a unique noninvasive approach to quantify LA substrates as well as ablation-induced scar formation. 3 Substrate modification by targeting MRI-detected LA- LGE is an increasingly popular approach that might be effective to decrease the risk of AF recurrence. 4 In this study, we report our initial clinical experience of LGE-MRI-guided LA substrate modification in addition to PVI using CBA.

| METHODS
This was a retrospective observational study of consecutive patients who underwent first-time CBA for AF at the University of Utah between June, 2014 to August, 2016. The study was approved by the University of Utah Institutional Review Board.

| Patient selection
Patients were included in the study if they were more than 18-years old, had symptomatic paroxysmal or persistent AF, and were referred for first-time AF ablation. All patients underwent a preablation LGE-MRI followed by scans within 24 hours after ablation and at 3 months. The 24-hour LGE-MRI assessed acute postablation complications including esophageal thermal injury (ETI) while the preablation and 3-month scans assessed LA substrate and CBA-induced scar formation, respectively. All patients were admitted for overnight observation and discharged the next day after their 24-hour LGE-MRI scan. The admission was extended if follow-up studies such as repeat MRI or an esophagogastroduodenoscopy (EGD) was required based on the extent of esophageal enhancement on the acute LGE-MRI. Patients who had a contraindication for LGE-MRI were not included. 4 Other exclusion criteria were recent (within last 3 months) coronary artery syndrome, cardiac surgery, or transient ischemic attack (TIA)/stroke, presence of known LA thrombus, or the inability to remain on continuous anticoagulation postoperatively. All patients were followed up within 8 weeks in clinic.

| Magnetic resonance imaging procedures
Patients underwent 3D LGE-MRI evaluation before ablation obtained on a 1.5-Tesla Avanto or 3-Tesla Verio clinical scanner (Siemens Medical Solutions, Erlangen, Germany) acquired 15 minutes after injection of 0.01 mmol/kg of gadolinium contrast (Multihance; Bracco Diagnostics Inc, Princeton, NJ) using a 3D ECG-gated, respiratorynavigated, inversion-recovery prepared gradient-recalled echo pulse sequences as previously described. 3 Quantification of LA-LGE and scarring were defined using dedicated software (Corview; Marrek, Salt Lake City, UT). Patients were assigned to one of four groups of increasing LA-LGE stages: Utah stage 1 (minimal: <10%), Utah stage 2 (mild: 10%-20%), Utah stage 3 (moderate: 20%-30%), and Utah stage 4 (extensive: >40%). 4 A substrate map displaying areas of LA-LGE was generated and used during the ablation procedure. Similar images were acquired 3 months later to determine the levels of postablation scarring created by CBA. 5 Immediately 24 hours after ablation, all patients underwent MRI to assess ETI. Any delayed enhancement in this area was considered to be esophageal enhancement secondary to ETI and categorized as none (no detectable enhancement), mild (focal enhancement), moderate (transmural or near transmural enhancement), and severe (transmural enhancement involving more than 5 mm of esophageal tissue or present in more than one location. 6 Presence of atrioesophageal fistula was also assessed. Patients with severe enhancement had to repeat an MRI study within 24 hours. Patients with moderate esophageal enhancement had a repeated study within 7 days. All patients with persistent enhancement on their repeat study 1588 | were referred for EGD. A repeat MRI was performed at 3 months after ablation in all patients. Ablation scar size as well as presence of PV stenosis were assessed and correlated to regions of LA substrate measured before ablation.

| Ablation procedure
All patients were continued on their home oral anticoagulation (OAC) agent periprocedurally. Patients who had AF as their presenting rhythm at the time of ablation underwent transesophageal echocardiography to exclude presence of LA thrombus. A weight-based bolus of heparin was given immediately before transseptal puncture and intermittently every 10 minutes thereafter to achieve a target A freezing temperature of −30°C and application time of 80 to 120 seconds was applied when targeting areas of atrial LGE. We used ICE to ensure the alignment of the cooling zone of the balloon with the LA tissue. In cases where extensive LGE was present in the anterior wall, the catheter-tissue alignment was ensured by anchoring the Achieve catheter in the LA appendage.
Information from MRI with preablation LGE was integrated into a 3D mapping system. Alignment with landmarks from fluoroscopy or ICE was extensively used to navigate to extrapulmonary areas of LA-LGE and to guide catheter positioning as well as to ensure overlapping of lesions without any gaps. Esophageal temperature was monitored at all times by maneuvering a temperature probe in the esophagus and stopped freezing if the temperature was decreased bellow 20°C. Ablation ended if PVI was achieved by entrance block and noninducibility by rapid atrial pacing as well as homogenization of LA-LGE outside the PVs by CBA. During and at the end of the procedure, the LA was surveyed with ICE and presence of any pericardial effusion was checked.

| Data analysis and statistical methodology
Data are expressed as number (percentage) or median with interquartile range (IQR). Continuous variables were analyzed using the two-sample t-test when distribution was normal or Mann-Whitney test with a non-normal distribution. The χ 2 test or Fisher exact test was used to analyze categorical variables. Time to first recurrence of atrial arrhythmias was analyzed using Kaplan-Meier estimates. A two-sided probability of P ≤ .05 was considered significant. Statistical analyses were performed using R software (R Core Team, Vienna, Austria; URL: http://www.R-project.org/).

| Evaluation of LA substrate by LGE-MRI
Homogenization of arrhythmogenic tissue outside the PVs has been suggested as a means to achieve better rhythm outcome after AF ablation. 7

| Feasibility of extrapulmonary cryoballoon ablation
Cryoablation outside the PVs has been reported for isolation of the LA roof, 10 In a multicenter nonrandomized study, PVI + posterior wall isolation (PWI) vs standard PVI using CBA (390 patients, persistent AF) achieved significantly more freedom from AF recurrence at 12 months (80% vs 51%; P = .001). 16 However, one-third of patients required additional RF ablation to achieve PWI and every second patient needed additional ablation procedures. In contrast to our approach using LGE-MRI guidance, atrial substrates have been evaluated by high-density voltage mapping. The invasive nature of this approach might have hindered prior appropriate patient selection and ablation procedure planning with potential avoidance to switch to an additional RF approach. It excluded reporting of followup data on permanent lesion characteristics. Data regarding the incidence of atrioesophageal injury, PV stenosis, or TIA/stroke were not reported. In concordance with our findings, procedure and fluoroscopy times were prolonged, but PWI by CBA was feasible without major periprocedural complications.   27 However, cryo-lesions were wider and more continuous than RF-lesions. 28 In our cohort of LA substrate-positive patients, who are prone to worse rhythm outcome, freedom from AF recurrence was 74.5% and in line with the findings from multisite-wide success rates as well as recently published experiences of PVI+PWI using CBA. 16,19,29 However, not all AF patients benefit from additional substrate ablation outside the PVs as reported in the STAR-AF II trial. 30 Evaluation of the individual LA substrate and selection of a tailored ablation approach seem to be crucial for rhythm outcome. 8,9 In light of the current results, the value of CBA for LA substrate modification is still experimental and might be promising in wellselected patients. However, this needs critical appraisal in terms of safety and efficacy in a larger number of patients also investigating potential proarrhythmogenic effects of cryo-lesions outside the PVs.
Addressing these questions, the presented concept will be in-

| Assessment of esophageal thermal injury by LGE-MRI
Esophageal thermal injury with deterioration into atrioesophageal fistula remain life-threatening complications after AF ablation.
LGE-MRI has been reported a useful screening modality for early detection and further monitoring of ETI. 6 Esophageal enhancement immediately after ablation is frequently observed in 50 to 80% of cases but did not predict severe ETI on EGD at such an early stage. [31][32][33] While overall in line with previous reports, the majority (54%) of our patients presented with mild esophageal enhancement acutely after cryoablation. Only one patient had moderate enhancement without ETI on EGD. As seen before, pathological imaging commonly appeared in the acute postablation period with almost all cases exhibiting complete resolution within 3 months. 6,33 It seems possible that acute enhancement dissolving within 1 to 3 days, indicates transient procedure-related inflammatory response with edema rather than clinically relevant esophageal tissue necrosis leading to severe complications.
In a recent analysis, only moderate and severe esophageal enhancement exhibited high sensitivity and negative predictive value (100% each, 78% specificity) for the detection of ETI confirmed by EGD. 33 Especially the combination of lasting symptoms and persistent esophageal enhancement of the anterior wall without posterior wall involvement beyond 24-hour postablation were associated with esophageal ulcer. 32 Identification of these patients could potentially lead to early treatment and minimization of severe ETI-related complications. Further research is needed to determine the optimal time point of imaging and other predictors within patient's demographics, mediastinal anatomy, and ablation techniques leading to or preventing of ETI.

| Limitations
We report about a single-center retrospective initial experience that investigated a small cohort of preselected patients with mainly paroxysmal AF type and low LA-LGE burden. Due to the initial experiences with this new technique, procedural and fluoroscopy times were longer than known from standard CBA procedures but are subject to a steep learning curve. There was no randomization against substrate-positive patients treated with standard PVI by CBA or comparison to an RF approach. Whether LGE on MRI can only be interpreted as arrhythmogenic tissue predicting better rhythm outcome when addressed by CBA is debatable. Challenging catheter maneuverability could have resulted in incomplete substrate modification and PV reconnection as potential reasons for AF recurrence or even increased arrhythmogenicity. The presented techniques may only be reproducible by experienced operators and hold an increased risk of harming sensible cardiac structures during extrapulmonary ablation. Therefore, it is of critical note, that overall statements on efficacy and safety cannot be made. Validation in multicenter randomized trials with review of complication rates and long-term rhythm outcome in a larger number of patients is needed.

| CONCLUSION
In preselected AF patients, LA substrate modification in addition to PVI using LGE-MRI-guided cryoablation is feasible but still experimental. Freedom from AF recurrence after 12 months was achieved in three of four LA substrate-positive patients. Efficacy and safety have to be further investigated in a randomized controlled trial and a larger number of patients.