Relation of left atrial appendage closure devices to topographic neighboring structures using standardized imaging by cardiac computed tomography angiography

Background Although left atrial appendage (LAA) anatomy and topographic relations are well understood, little is known about the impairment of neighboring structures (NBS) by an implanted left atrial appendage closure (LAAC) device. This prospective longitudinal observational study for the first time describes distances of implanted LAA closure (LAAC) devices to NBS using a standardized imaging protocol of cardiac computed tomography angiography (cCTA). Hypothesis cCTA imaging is an eligible tool for post‐implantation evaluation of LAAC devices and their relation to neighboring structures. Methods cCTA data sets of consecutive patients 6 months after successful LAAC were acquired on a third generation dual‐source CT system and reconstructed with a slice thickness of 0.5 mm. The standardized multi‐planar reconstruction LAA occluder view for post‐implantation evaluation (LOVE) algorithm was used to measure the distances to NBS in relation to LAA morphology and implanted LAAC devices. Results A total of 48 patients (median age 80 years, 25% female) were included. Left upper pulmonary vein and circumflex artery were generally closest to occlusion devices (median 2.9 and 2.8 mm, respectively). AMPLATZER AMULET devices were closer to the mitral valve annulus than WATCHMAN devices (6.6 mm (inter quartile range [IQR] 4.9‐8.6) vs 10.9 mm (IQR 7.4‐14.0), P = 0.001). Distances to the left upper pulmonary vein were affected by LAA morphology, with cauliflower type having the closest proximity (1.7 mm [IQR 1.0‐3.4], P = 0.048). Conclusion A standardized cCTA imaging protocol is an eligible tool to accurately measure distances to NBS. Left upper pulmonary vein and circumflex artery are closest to LAAC devices and could thus be most prone to impairment.


| INTRODUCTION
In non-valvular atrial fibrillation (AF), the left atrial appendage (LAA) has been described as the main site of thrombus formation. 1,2 This provides the rationale for direct targeting of this site by interventional LAA closure (LAAC) to reduce the risk of ischemic stroke. Especially, as the effectiveness of this procedure has been demonstrated by follow-up of the PROTECT-AF study compared to conventional anticoagulation therapy, 3  reports of occlusion of Cx, 8 perforation of LPA 9,10 and pulmonary vein compression 11 have been published. We therefore see a need for further assessment of the relations of implanted LAAC devices to NBS.
Assessment of NBS can be performed through transesophageal echocardiogram (TEE). 12 However, cardiac computed tomography angiography (cCTA) may reveal the potential as an even more accurate imaging modality, 7,13 as it offers higher resolution images for measuring distance to and impairment of LAA NBS. Recently, a standardized cCTA imaging protocol has been proposed by our working group, 13,14 which could prove useful in the post-implantation evaluation of implanted LAAC devices. Higher resolution and non-invasiveness of cCTA evaluation of LAAC compared with TEE, paired with higher reliability and objectivity of a standardized protocol are arguments in favor of cCTA. 13,14 Therefore, the present study evaluates the distances of NBS to the implanted device using the standardized cCTA protocol-LAA occluder view for post-implantation evaluation (LOVE) -6 months after successful LAAC. Distances to NBS are assessed in relation to types of occlusion devices as well as the LAA morphologies.

| Study population
This is a prospective, non-randomized, observational, longitudinal single-center study. Consecutive patients with non-valvular AF and indication for oral anticoagulation due to a CHA 2 DS 2 -VAsc score ≥ 2 undergoing LAAC device implantation between June 2014 and December 2017 were included. Accordant to the 2016 guidelines of the European Society of Cardiology (ESC), 15 sex category as independent risk factor was excluded from the original CHA 2 DS 2 VASc score.
Inclusion criteria were a relative or absolute contraindication for oral anticoagulation, which was major or recurrent bleeding, HAS-BLED score ≥ 3 or intolerance to oral anticoagulation and age ≥ 18. Exclusion criteria were a treatable cause or a single episode of AF, planned catheter ablation of AF or electrical cardioversion within 30 days prior or after LAAC, myocardial infarction within the last 3 months, congestive heart failure of New York Heart Association (NYHA) stage IV, atrial septum defect (ASD) or interventional or surgical occlusion of ASD, mechanical heart valve, status after heart transplant, intracere-  catheter. Injection rate was 5 mL/second followed by a 50 mL saline flush. The systematic approach to evaluate implanted LAAC devices that has been recently described by the so called LOVE views, revealing optimal device-related angulation allowing optimal evaluation of the device post implantation 13 was applied.

| Definitions
For measurement of the distances of the implanted occlusion device to NBS, the recently described LOVE axial and sagittal views 13 were applied in all patients. Relevant NBS were defined: the MVA, the LPA, the LUPV and the Cx. LAA morphology was assessed for each patient and classified into one of four types suggested by Wang et al, 16 namely windsock, chicken wing, cauliflower, and cactus.

| Study endpoints
At cCTA visits, distances to neighboring structures were analyzed for each morphological type of LAA individually, as well as for both types of implanted LAAC devices (WATCHMAN and AMPLATZER AMULET).
All patients were followed-up regarding anticoagulant therapies and adverse clinical events for 12 months. This especially included arterial or venous thromboembolism, stroke or transient ischemic attack (TIA).    Table 1].

| Distance to NBS
Distances to LPA, LUPV, and Cx were best visualized and measured in LOVE axial view, while the distance to MVA was demonstrated best in LOVE sagittal view (Figures 1 and 2

| LAA morphology and NBS
The prevalence of different LAA morphologies was 38% windsock, 33% chicken wing, 23% cauliflower, and 6% cactus type. Distances to MVA and LPA showed no relevant difference among the morphological types.
In cauliflower type, devices were closest to the LUPV (1.7 mm) compared to windsock (2.6 mm), cactus (3.8 mm) and chicken wing (4.3 mm) types. A P value of 0.048 indicates statistical significance of those
AMPLATZER AMULET devices were also closer to LPA (3.9 mm vs

mm) and LUPV (2.5 mm vs 3.4 mm), although both differences
were not statistically significant. In contrast, median distance of WATCHMAN devices to Cx was slightly smaller than in AMPLATZER AMULET devices (2.7 vs 3.1 mm, P = 0.323) [ Table 3].

| Adverse events
In the present study cohort, only one patient had a vascular adverse     Other uses of cCTA have been proposed in the planning of LAAC and assessment of early and mid-term outcomes of device implantation, 6 three-dimensional geometric CT analysis of the LAA could prove to be effective for prediction of PDL. 20 Aspects in favor of the LOVE cCTA protocol are non-invasiveness and higher resolution measurements of CT compared to TEE evaluation of LAAC, paired with higher objectivity and reliability of a standardized protocol.