Abstract
Background
Left atrial appendage occlusion (LAAO) is an established alternative to long-term anticoagulation in patients with atrial fibrillation. Predicting successful LAAO device implantation with preplanning imaging studies, including cardiac computed tomography (CCT), remains a challenge.
Objectives
The purpose of this study was to determine whether specific preprocedural and postprocedural CCT measurements are associated with successful LAAO implantation and can predict early postimplant complications.
Methods
This was a single-center retrospective study including 165 patients who underwent LAAO with Watchman FLX between November 2022 and January 2025 and had both preimplant and postimplant CCT. Adverse events including peri-device leak, device-related thrombus, or cerebrovascular events (transient ischemic attack/stroke) within 90 days were followed.
Results
Adverse events occurred in 40 patients (24.2%), including 32 peri-device leaks, 9 device-related thrombi, and 1 transient ischemic attack. Those with events had significantly higher mean left atrial pressure (15.0 vs 13.5 mm Hg; P = 0.03), larger pre-CCT LAA ostium maximum diameter (26.6 vs 23.5; P < 0.001), larger pre-CCT minimum ostial diameter (21.8 vs 18.5 mm; P < 0.001), and smaller postimplant area change (66.5 vs 136.9 mm2; P < 0.001).
Conclusions
Smaller LAA ostial dimensions and greater device compression (postimplant area change) on CCT were associated with fewer complications post-LAAO. Favorable preimplant LAA ostial morphology seems to be oval shaped and of smaller size. These characteristics could be helpful in predicting better postimplant outcomes.
Key words: cardiac CT, device-related thrombus, left atrial appendage closure, peri-device leak
Central Illustration
Left atrial appendage occlusion (LAAO) is a reasonable alternative for mitigating the risk of cardioembolic events in selected patients.1 In order to secure favorable outcomes and lower complication rates, periprocedural imaging is an important step in every procedure planning. The challenges surrounding the placement of LAAO device lie in the inherent variability of left atrial appendage (LAA) anatomy and its neighboring structures. For a successful and safe percutaneous approach, a comprehensive understanding of critical anatomic landmarks and the diverse endocardial morphologic variants of the LAA is essential.2 Due to its high resolution and multiplanar reconstruction capabilities, cardiac computed tomography (CCT) is increasingly becoming the primary imaging tool for periprocedural planning.3 If used in preplanning, CCT is associated with higher procedural success rate (100% vs 82% compared to transesophageal echocardiography [TEE]), higher procedural efficiency, lower average number of devices used, shorter procedural time, lower radiation exposure, and lower contrast doses.4 Inadequate implantation may lead to peri-device leaks (PDLs), higher risk for device-related thrombus (DRT), and increased risk of cerebrovascular events. To ensure adequate sealing of the LAAO device and the absence of device-related complications, postprocedural imaging between days 45 and 90 is recommended.1 Predicting successful LAAO device implantation with preplanning imaging studies, including cardiac CT, remains a challenge. Therefore, we aimed to investigate whether the use of CCT before and after procedure could differentiate patients with and without adverse outcomes following LAAO.
Methods
Study design and patient population
This retrospective study was approved by the Lee Health Institutional Review Board. This was a single-center retrospective analysis involving 165 consecutive patients who underwent percutaneous LAA closure with Watchman FLX device (Boston Scientific) and had preprocedural and postprocedural CCT done between November 2022 and January 2025. Preprocedural CCT scans, as well as postprocedural assessments 45 to 90 days following procedure, were performed. Of these patients, 160 had their measurements analyzed using the Vitrea Workstation (Canon Medical Systems), while data for the remaining 5 were obtained from the preplanning records from the outside institution. Patients who did not undergo preprocedural and/or postprocedural CCT scan and patients lacking documented 90-day follow-up in their charts were excluded from this study. Baseline characteristics were collected. Pre-CCT measurements included LAA ostium diameters (minimum and maximum), area, and perimeter. Post-CCT measurements focused on parameters related to the LAAO device, specifically the LAAO area, perimeter, and intershoulder diameter of the device (Figure 1). Postimplant change in area or ΔA (CT post-LAAO device intershoulder area−CT pre-LAAO ostium area) to characterize the device compression was calculated. Echocardiographic evaluations of left atrial volume index (LAVI) and left ventricular ejection fraction were conducted prior to implantation. Furthermore, procedural notes were reviewed, and device size, mean left atrial pressure (LAP) following transseptal puncture, device recapture, and the use of intracardiac echocardiography were reported.
Figure 1.
Calculation of ΔA
(A) Multiplanar reconstruction (MPR) with cross-hairs demonstrating LAA ostium area prior implant using the 2-chamber plane for orientation and the circumflex coronary artery as anatomical landmark (like transesophageal echo; red arrow). (B) MPR format demonstrating how to assess LAAO device intershoulder area (red arrow) by aligning planes parallel to the widest shoulder distance (ΔA = CT post-LAAO device intershoulder area − CT pre-LAAO ostium area; in example shown below: ΔA = 560 mm2 - 383 mm2 = 177 mm2). ΔA = postimplant change in area; CT = computed tomography; LAA = left atrial appendage; LAAO = left atrial appendage occlusion.
Image acquisition
Images were acquired by using either a dual-source FORCE CT scanner (Siemens Healthineers) or a single-source 128-row detector SOMATOM go. Top CT scanner (Siemens Healthineers), with prospective gating acquisition mode predominantly in systolic phases. Delayed gated acquisition was performed following the same acquisition mode 45 seconds after initial iodinated contrast injection to rule out thrombus in LAA in initial scans or to identify DRT and PDL in follow-up scans. Images were reconstructed using iterative reconstruction at isotropic 0.65 mm voxels and a Bv36 smooth kernel.
Study defined outcomes
Adverse events were defined as any PDL, DRT, and cerebrovascular events (stroke and/or transient ischemic attack). At the end of the follow-up period, the patients were categorized into 2 groups based on the presence or absence of an adverse outcome.
Statistical analysis
Categorical variables are reported as count (percentage), and continuous variables are reported as mean ± SD or median (IQR), as appropriate. Statistical analyses involved the comparison of continuous variables using the Mann-Whitney test and categorical variables using Fisher exact test. A P value of <0.05 was considered statistically significant. Statistical analyses were performed with SPSS (Version 21.0, SPSS Inc) and R software (R Core Team 2021).
Results
Among 165 patients included in the comparative analysis, 40 (24.2%) experienced adverse events. The mean age was 79.3 ± 7.2 and 77.0 ± 7.6 in the adverse event vs no adverse event groups, respectively. The baseline characteristics of 2 groups are presented in Table 1. There was no statistical difference in age, sex, hypertension, presence of diabetes, body mass index, heart failure, coronary artery disease, history of myocardial infarction or stroke, CHADS2-VASc and HAS-BLED scores, and left ventricular ejection fraction. The most common LAA morphology was windsocks and was present in 68% and 71% of patients in adverse and nonadverse event group, respectively, with no statistical difference among groups (Figure 2). No statistical difference between groups was found when placement was guided by intracardiac echocardiography (Table 1).
Table 1.
Baseline Characteristics of the Study Groups
| Subjects With Adverse Events (n = 40) | Subjects Without Adverse Events (n = 125) | P Valuea | |
|---|---|---|---|
| Age (y) | 79.3 (60-93) | 77.0 (57-94) | 0.15 |
| Female | 13 (32%) | 45 (36%) | 0.84 |
| Essential hypertension | 38 (95%) | 116 (93%) | 1.00 |
| BMI (kg/m2) | 28.6 (20.3-47.2) | 29.3 (19.8-45.0) | 0.35 |
| Diabetes | 12 (30%) | 33 (26%) | 0.68 |
| History of systolic HF | 7 (18%) | 11 (8%) | 0.15 |
| History of diastolic HF | 5 (13%) | 31 (25%) | 0.25 |
| CAD | 28 (70%) | 81 (65%) | 0.57 |
| History of MI | 6 (15%) | 26 (21%) | 0.50 |
| History of stroke/TIA | 5 (13%) | 32 (26%) | 0.13 |
| CHADS2-VASc score | 4.5 (2-7) | 4.4 (2-8) | 0.83 |
| HAS-BLED score | 2.7 (1-5) | 2.5 (1-6) | 0.48 |
| Echo LVEF (%) | 54.3 (23-72) | 56.7 (25-78) | 0.55 |
| Prior catheter ablation | 11 (28%) | 20 (16%) | 0.28 |
| Echo LAVI (mL/m2) | 48.0 (21-173) | 38.3 (15-81) | 0.04 |
| Intracardiac echocardiography | 23 (58%) | 61 (49%) | 0.52 |
Values are median (IQR) or n (%).
BMI = body mass index; CAD = coronary artery disease; HF = heart failure; LAVI = left atrial volume index; LVEF = left ventricular ejection fraction; MI = myocardial infarction; TIA = transient ischemic attack.
P < 0.05 for statistical significance.
Figure 2.
Number of Left Atrial Appendage Morphologies in Each Group
Abbreviation as in Figure 1.
PASS (Position, Anchor, Size, and Seal) criteria were achieved in 100% of patients by TEE standards. The median device size used was 27 (20-40) and 27 (20-35) in the adverse event vs no adverse event groups, respectively. Device recapture was not different between 2 groups. The LAVI was statistically higher in the adverse event group (48.0 vs 38.3; P < 0.04). The difference between the LAA ostial area preimplant and the device inter shoulder area (ΔA) in subjects with adverse events was 66.5 (−124.8 to 309.2) mm2 vs no event ΔA of 136.9 (−180.0 to 781.5) mm2 (P < 0.001). There were no other significant differences in post-CCT measurements for either group (Table 2). Lastly, there were total of 42 adverse events in 40 patients: 32 PDLs, 9 DRTs, 1 transient ischemic attack, and 0 postimplant stroke (Table 3).
Table 2.
Periprocedural and Procedural Characteristics
| Subjects With Adverse Events (n = 40) | Subjects Without Adverse Events (n = 125) | P Valuea | |
|---|---|---|---|
| LAA morphology | 0.58 | ||
| Windsock | 27 (68%) | 71 (57%) | |
| Chicken-wing | 7 (18%) | 38 (30%) | |
| Broccoli | 1 (2.5%) | 2 (2%) | |
| Cauliflower | 2 (5%) | 6 (5%) | |
| Cactus | 2 (5%) | 6 (5%) | |
| Not reported | 1 (2.5%) | 2 (2%) | |
| CT pre-LAAO ostium average diameter (mm) | 24.3 (18.0-31.7) | 21.3 (11.0-29.7) | <0.001 |
| CT pre-LAAO ostium minimum diameter (mm) | 21.8 (14-31.0) | 18.5 (7.4-29.0) | <0.001 |
| CT pre-LAAO ostium maximum diameter (mm) | 26.6 (19.9-35.5) | 23.5 (12.0-33.6) | <0.001 |
| CT pre-LAAO ostium area (mm2) | 471.8 (250.5-785.8) | 374.2 (102.1-757.0) | <0.001 |
| CT pre-LAA perimeter | 78.2 (57.0-101.0) | 69.7 (39.8-99.0) | 0.001 |
| CT pre-LAA depth (mm) | 18.4 (12.4-26) | 17.9 (10.3-33.0) | 0.69 |
| LAAO device size | 27 (20-40)a | 27 (20-35)a | 0.19 |
| 20 mm | 1 (3%) | 13 (10%) | |
| 24 mm | 18 (45%) | 42 (34%) | |
| 27 mm | 8 (20%) | 39 (31%) | |
| 31 mm | 7 (18%) | 19 (15%) | |
| 35 mm | 5 (13%) | 12 (10%) | |
| 40 mm | 1 (3%) | 0 (0%) | |
| Device recapture ≥1 | 1 (3%) | 9 (7%) | 0.59 |
| Mean LAP (mm Hg) | 15.0 (8-30) | 13.5 (4-60) | 0.03 |
| CT post-LAAO device average diameter (mm) | 26.0 (19.0-34.0) | 25.2 (16.4-35.3) | 0.37 |
| CT post-LAAO device minimum diameter (mm) | 25.4 (19.0-34.0) | 24.5 (13.4-34.8) | 0.30 |
| CT post-LAAO device maximum diameter (mm) | 26.3 (19.0-34.1) | 25.7 (17-35.9) | 0.56 |
| CT post-LAAO device intershoulder area (mm2) | 538.2 (274.0-910.0) | 512.5 (211.0-978.8) | 0.50 |
| CT post-LAAO device perimeter (mm) | 81.4 (59.0-107.0) | 79.0 (27.7-111.1) | 0.52 |
| CT post-LAAO ΔA(mm2) | 66.5 (−124.8 to 309.2) | 136.9 (−180.0 to 781.5) | <0.001 |
Values are n (%) or median (IQR).
ΔA = postimplant change in area; CT = computed tomography; LAA = left atrial appendage; LAAO = left atrial appendage occlusion; LAP = left atrial pressure.
P < 0.05 for statistical significance.
Table 3.
Follow-Up Imaging Findings
| Subjects With Composite Adverse Events (n = 40) | Subjects Without Composite Adverse Events (n = 125) | P Valuea | |
|---|---|---|---|
| Average (d) | 55.9 (18-242) | 59.6 (14-375) | 0.85 |
| Endothelization, reported | 34 | 119 | 0.004 |
| No | 5 (15%) | 7 (6%) | |
| Predominantly | 16 (47%) | 30 (26%) | |
| Completely | 13 (38%) | 82 (68%) | |
| Peri-device leak (any) | 32 (80%) | / | / |
| Major leak (>5 mm) | 5 | ||
| Minor leak (<5 mm) | 24 | ||
| Not specified | 3 | ||
| Device-related thrombus | 9 (23%) | / | / |
| TIA | 1 (2.5%) | / | / |
| Stroke | 0 | / | / |
Values are median (IQR) or n (%).
Abbreviation as in Table 1.
P < 0.05 for statistical significance.
The adverse outcome group had larger mean LAP (15.0 [8-30] mm Hg vs 13.5 [4-60] mm Hg; P = 0.03), larger pre-CT LAA ostium average diameter (24.3 [18-31.7] mm vs 21.3 [11.0-29.7] mm; P < 0.001), larger pre-CT LAA ostium minimum diameter (21.8 [14.0-31.0] mm vs 18.5 [7.4-29.0] mm; P < 0.001), and larger pre-CCT LAA ostium maximum diameter (26.6 [19.9-35.5] mm vs 23.5 [12.0-33.6] mm; P < 0.001) (Table 2). Similarly, patients with PDL only had larger pre-CT LAA ostium average diameter (P = 0.001), pre-CT LAA ostium minimum diameter (P = 0.002), larger pre-CCT LAA ostium maximum diameter (P = 0.002), CT pre-LAAO ostium area (P = 0.009), CT pre-LAA perimeter (P = 0.01), and smaller ΔA (P = 0.06) (Table 4).
Table 4.
PDL Outcome
| Peri-Device Leak (n = 32) | No Peri-Device Leak (n = 125) | P Valuea | |
|---|---|---|---|
| CT pre-LAAO ostium average diameter (mm) | 24.1 (18.0-31.7) | 21.4 (11.0-31.0) | 0.001 |
| CT pre-LAAO ostium minimum diameter (mm) | 21.4 (14.0-29.1) | 18.7 (7.4-31.0) | 0.002 |
| CT pre-LAAO ostium maximum diameter (mm) | 26.5 (19.9-35.5) | 23.7 (14.0-34.5) | 0.002 |
| CT pre-LAAO ostium area (mm2) | 460.4 (250.5-785.8) | 383.2 (102.1-776.0) | 0.009 |
| CT pre-LAA perimeter | 77.4 (57.0-100.8) | 70.4 (39.8-101.0) | 0.01 |
| Mean LAP (mm Hg) | 14.8 (8-30) | 13.6 (4-60) | 0.18 |
| CT post-LAAO ΔA (mm2) | 74.7 (−124.8 to 309.2) | 130.7 (−180.0 to 781.0) | 0.06 |
Values are median (IQR).
PDL = peri-device leak; other abbreviations as in Table 2.
P ≤ 0.05 for statistical significance.
Discussion
The current study sought to investigate the use of CT-derived LAA measurements as predictors of outcomes following transcatheter LAAO. The principal finding is that a CT phenotype characterized by a smaller pre-CT minimum diameter, smaller pre-CT maximum diameter, smaller perimeter, and smaller diameter area was associated with better outcomes. Secondly, a larger ΔA showed statistical significance in reducing adverse outcomes in the studied population. Lastly, a higher invasive mean LAP value following transeptal puncture and larger LAVI were more prevalent in subjects with worse outcomes.
Numerous studies have explored a variety of imaging modalities, initially favoring TEE for the implantation of Watchman devices.5 In more intricate cases, the utilization of CT-based three-dimentional (3D) printing models has emerged as a beneficial approach for assisting in device selection and predicting compression during interventions for LAA closure.6 Among these modalities, cardiac CT has emerged as a great tool for LAAO periprocedural planning, given its higher spatial resolution, clearer depiction of the appendage morphology and surrounding structures, and better procedural outcomes compared to implant preplanning with TEE.7, 8, 9 Whether there are imaging phenotypes by CT that may be utilized in preplanning imaging to predict implant success remains a question.
Achieving optimal sizing and perfect radial compression between the device and the LAA is a complex issue. More favorable outcomes are probably not related to a single measurement or morphological assessment alone, rather to the combination of multiple factors. As per clinical trial data, the largest ostial dimensions based on the preimplant TEE has been used by implanters to guide device sizing and this is the current practice extended to CT. This practice has shown a robust correlation between the maximum diameter of the LAA ostium measured by CT and the eventual size of the deployed Watchman device, correlating with favorable procedural outcomes during subsequent follow-ups.10 Similarly, in our study, the maximal diameter of the LAA ostium differed between the 2 groups. However, we also found that a CT phenotype characterized by a lower pre-CT minimum diameter, smaller perimeter, and smaller diameter area was associated with less adverse outcomes. Furthermore, the patients with less adverse outcome were found to have smaller left atrium. Given smaller ostial dimensions with lower minimal diameter, we suspect that there is a better radial compression of the device toward ostium. Additionally, our hypothesis is that a smaller minimal diameter configures differently the LAA ostium shape, making it less circular and more oval-shaped. This variant could be favorable in avoiding less PDLs and DRT due to a better distribution of compressive forces. Essentially, a larger minimal diameter alters the ostial shape geometry, leading to a more circular configuration and reducing the margin for error in device sizing by TEE, particularly in the presence of variations in volume status, left atrial pressure, and technical imaging limitations (Figure 3). The larger ΔA could be an indicator of good device compression, and in our cohort, it seems to be larger in the group without adverse outcomes. We cannot establish or discard that some of these phenotypical ostial morphologies in CT prior LAAO device implant may be a result of dynamic phenomenon related to increases in LAP or just intrinsic anatomical variants (Central Illustration).
Figure 3.
Schematic of LAA Ostium Morphology and Relationship to Events
Larger minimum dimensions in left atrial appendage ostium (LAAo) were found to be associated with adverse outcomes. In green box, several examples of cases with no events illustrate the oval morphologies associated with less events when compared to higher incidence of adverse events in subjects with LAAo having larger minimum diameters. DRT = device-related thrombus; PDL = peri-device leak; other abbreviation as in Figure 1.
Central Illustration.
Multimodality Imaging: Predicting Adverse Events After LAAO
CCT = cardiac computed tomography; TIA = transient ischemic attack; other abbreviations as in Figures 1 and 3.
Study limitations
Being hypothesis-generating, this study has several limitations, including a relatively small sample size and variability in CT scanner types and imaging techniques. As a single-center analysis limited to patients treated with the most commonly used LAAO device, the findings should be interpreted with caution and warrant validation in larger, prospective cohorts with longer follow-up period. Additionally, these results cannot be extrapolated to patients who underwent echocardiographic assessment before and after LAA closure. Lastly, we recognize that a complete description of PDL location and mechanism is lacking.
Conclusions
Although the maximal LAA ostial diameter is traditionally used to guide Watchman FLX sizing, successful CT-based LAAO planning appears to be more complex than relying solely on this measurement. Both larger maximal and minimal ostial diameters were associated with higher rates of adverse events. Conversely, a larger ΔA, a surrogate for higher compression, appears to be associated with fewer adverse outcomes. Smaller, more oval preimplant ostial morphology also appeared favorable and may help predict better postimplant outcomes.
Perspectives.
COMPETENCY IN MEDICAL KNOWLEDGE: Inadequate Watchman implantation may lead to PDLs, a higher risk for DRT, and an increased risk of cerebrovascular events.
TRANSLATIONAL OUTLOOK: This study’s findings should inform the field and could be helpful in predicting better postimplant outcomes.
Funding support and author disclosures
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Footnotes
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.
References
- 1.Saw J., Holmes D.R., Cavalcante J.L., et al. SCAI/HRS expert consensus statement on transcatheter left atrial appendage closure. Heart Rhythm. 2023;20(5):e1–e16. doi: 10.1016/j.hrthm.2023.01.007. [DOI] [PubMed] [Google Scholar]
- 2.Gonzalez-Casal D., Datino T., Soto N., et al. Anatomy of the left atrial appendage for the interventional cardiologist. Herzschrittmacherther Elektrophysiol. 2022;33(2):195–202. doi: 10.1007/s00399-022-00866-8. [DOI] [PubMed] [Google Scholar]
- 3.Dieker W., Behnes M., Fastner C., et al. Impact of left atrial appendage morphology on thrombus formation after successful left atrial appendage occlusion: assessment with cardiac-computed-tomography. Sci Rep. 2018;8(1):1670. doi: 10.1038/s41598-018-19385-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Reddy V.Y., Holmes D., Doshi S.K., Neuzil P., Kar S. Safety of percutaneous left atrial appendage closure: results from the watchman left atrial appendage system for embolic protection in patients with AF (PROTECT AF) clinical trial and the continued access registry. Circulation. 2011;123(4):417–424. doi: 10.1161/CIRCULATIONAHA.110.976449. [DOI] [PubMed] [Google Scholar]
- 5.Möbius-Winkler S. Percutaneous left atrial appendage closure: technical aspects and prevention of periprocedural complications with the watchman device. World J Cardiol. 2015;7(2):65. doi: 10.4330/wjc.v7.i2.65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hell M., Achenbach S., Yoo I., et al. 3D printing for sizing left atrial appendage closure device: head-to-head comparison with computed tomography and transoesophageal echocardiography. EuroIntervention. 2017;13(10):1234–1241. doi: 10.4244/EIJ-D-17-00359. [DOI] [PubMed] [Google Scholar]
- 7.Korsholm K., Jensen J.M., Nielsen-Kudsk J.E. Cardiac computed tomography for left atrial appendage occlusion. Interv Cardiol Clin. 2018;7(2):229–242. doi: 10.1016/j.iccl.2017.12.004. [DOI] [PubMed] [Google Scholar]
- 8.Eng M.H., Wang D.D., Greenbaum A.B., et al. Prospective, randomized comparison of 3-dimensional computed tomography guidance versus TEE data for left atrial appendage occlusion (PRO3DLAAO) Catheter Cardiovasc Interv. 2018;92(2):401–407. doi: 10.1002/ccd.27514. [DOI] [PubMed] [Google Scholar]
- 9.Korsholm K., Berti S., Iriart X., et al. Expert recommendations on cardiac computed tomography for planning transcatheter left atrial appendage occlusion. JACC Cardiovasc Interv. 2020;13(3):277–292. doi: 10.1016/j.jcin.2019.08.054. [DOI] [PubMed] [Google Scholar]
- 10.Xu B., Betancor J., Sato K., et al. Computed tomography measurement of the left atrial appendage for optimal sizing of the watchman device. J Cardiovasc Comput Tomogr. 2018;12(1):50–55. doi: 10.1016/j.jcct.2017.11.012. [DOI] [PubMed] [Google Scholar]