Multimodality Imaging for Optimal Device-size Selection During Percutaneous Left Atrial Appendage Closure in Thai Population

Background: Optimal device size selection is crucial for percutaneous left atrial appendage (LAA) closure. Transesophageal echocardiography (TEE) is the standard imaging technique for LAA assessment, however there are discrepancies among different imaging modalities. We aimed to evaluate the agreement between device size and LAA size measured by three methods: multi-detector cardiac computed tomography (MDCT), TEE, and angiography. Methods: Patients who underwent percutaneous LAA closure at King Hospital from 2012 to 2020 were included in this study. MDCT, TEE and angiography were reviewed. LAA ostial diameter, landing zone diameter and maximal depth from each imaging modality was measured and analyzed. Agreement between landing zone diameter and implanted device size was assessed. Results: We reported on 61 consecutive patients who underwent percutaneous LAA closure. The mean age of patients was 74.0 ± 8.4 years. The mean CHA 2 DS 2 score, CHA 2 DS 2 -VASc score and HAS-BLED score were 2.8 ± 1.4, 4.6 ± 1.8 and 2.6 ± 1.0, respectively. Device implantation was successful in all patients (100%). Two different LAA closure devices were used: Watchman (n = 43, 70.5%) and Omega (n = 18, 29.5%). Maximum landing zone diameter measured by MDCT scan, TEE and angiography were 23.4 ± 3.9 mm, 22.2 ± 4.8 mm and 22.7 ± 3.5 mm, respectively. MDCT measurement was signicantly larger than TEE measurement (p = 0.015) and closer to implanted device size compared with TEE and angiography. The difference between landing zone diameter measured by CT scan and device size was -1.65 ± 2.0 mm compared with -4.8 ± 4.6 mm for TEE and -4.3 ± 3.3 mm for angiography. Conclusion: MDCT sizing of LAA results in larger measurement than TEE. Routine implementation of MDCT sizing may improve procedural success with more accurate device size selection.


Introduction
Percutaneous left atrial appendage (LAA) closure is an emerging alternative treatment for stroke prevention in patients with atrial brillation who are at increased risk of bleeding or contraindicated to oral anticoagulation. [1][2][3][4][5] Transesophageal echocardiography (TEE) is the standard imaging for LAA assessment for procedural planning and device size selection.[6-8] Appropriate device size is crucial in percutaneous LAA closure as under-sizing can lead to peri-device leakage and device embolization.
Device over-sizing may also cause LAA perforation, pericardial effusion and cardiac tamponade. [8] Evaluation of LAA by TEE is challenging due to complex anatomy and technical di culties. Recent studies reported that LAA measurement using multi-detector cardiac computed tomography (MDCT) is more accurate compared to TEE and LAA angiography,[9-12] however, manufacturer recommendations for device size selection is currently based on TEE measurement. We aimed to evaluate the LAA size measured by TEE, MDCT and angiography, and study the agreement between LAA size from each imaging modality and the size of the successfully implanted LAA closure device.

Materials And Methods
All patients who had atrial brillation and underwent percutaneous LAA closure at King Chulalongkorn Memorial Hospital between 2012 to 2020 were included in this study. Baseline patient characteristics, procedural records and post-procedural patient follow-up were collected from electronic medical records. Pre-procedural and intra-procedural TEE, pre-procedural CT and intra-procedural LAA angiography were reviewed and analyzed retrospectively by investigators. The study was conducted in accordance with the Institutional Human Subjects Committee guidelines and approved by the Institutional Review Board at Faculty of Medicine, Chulalongkorn University.
Transesophageal echocardiography sizing of LAA Pre-procedural and intra-procedural two-dimensional (2D) TEE images were meticulously reviewed by investigators. TEE was performed with standard TEE transducer (Vivid, Philips, the Netherlands). TEE images of four standard planes (0 0 , 45 0 ,90 0 and 135 0 ) were identi ed for LAA sizing ( Figure 1). The LAA ostium and landing zone were measured from all 4 TEE views: 1) maximal diameter of the LAA ostium, 2) maximal LAA depth measured from the LAA ostial plane to the LAA apex, 3) maximal diameter of the LAA landing zone which is measured from the circum ex coronary artery to a point 2 cm below the tip of the left upper pulmonary vein limbus for Watchman device and measured at 10 to 12 mm distal to the ostium at an angle perpendicular to the neck axis for Omega device, and 4) landing zone depth measured from the landing zone plane to the LAA apex for Watchman device and measured perpendicularly toward the back of LAA for Omega device. LAA morphology and the presence or absence of thrombus were also assessed by pre-procedural TEE.
CT imaging and sizing of LAA Pre-procedure contrast-enhanced multi-detector CT (320-multidetector scanner, Aquilion One Vision Edition, Toshiba Medical Systems, Japan) was used to evaluate LAA. Retrospective ECG-gate for cardiac CT was performed. A volume of 50 ml of contrast (Iopomira 370 mgI/mL) was delivered via the antecubital vein with 50 ml saline ush at 5 ml/s. Multi-planar reconstruction was used to obtain orthogonal views and oblique planes of the LAA. LAA size was measured with 4 dimensions: 1) maximal and minimal diameter of the LAA ostium, 2) maximal LAA depth measured from the LAA ostial plane to the LAA apex, 3) maximal and minimal diameter of the conventional landing zone ~10 mm distal to the LAA ostium, and 4) maximal and minimal diameter of LAA at 12 mm and 15 mm distal to the LAA ostium ( Figure 1).

LAA angiography and sizing
LAA angiography was performed at the time of procedure using pigtailed catheter. Two standard angiography views of LAA were included (RAO 30 0 Cranial 10-20 0 and RAO 30 0 Caudal 10-20 0 ). LAA sizing by angiography was measured with 4 dimensions: 1) maximal diameter of the LAA ostium; 2) maximal diameter at the landing zone 10-20 mm distal to the LAA ostium; 3) maximal LAA depth measured from the LAA ostial plane to the LAA apex, and 4) landing zone depth measured from the landing zone plane to the LAA apex.
LAA closure procedure All LAA closure device implantations were performed by experienced interventional cardiologists via femoral venous access and local anesthesia with light sedation. Trans-septal puncture was achieved under TEE guidance. Unfractionated intravenous heparin was administered at a dose of 70 IU/kg, target activated clotting time > 250 seconds. The decision regarding nal device size was based on preprocedural TEE, intra-procedural TEE, preprocedural CT and intra-procedural LAA angiography with reference to manufacturer's sizing guidelines. The position of the device, the presence of peri-device leak, and the degree of compression were assessed by TEE before device release and immediately after deployment. Left ventricular systolic function, pulmonary venous in ow, pericardium condition, inter-atrial septum and degree of mitral regurgitation were carefully assessed prior to procedural conclusion. Transthoracic echocardiography was performed before hospital discharge and TEE was conducted at 45 to 60 days post-procedure to assess peri-device leakage, device-related thrombosis, device position, pericardial effusion and other complications. Mild peri-device leakage was de ned as a jet < 3 mm in diameter. Moderate and severe leakages was de ned as jet diameters of 3 to 5 mm and >5 mm, respectively. The decision to maintain oral anticoagulation or antiplatelets was determined by the attending cardiologists. Patients were routinely followed for complications and cardiovascular events at 6 weeks, 3 months, 6 months, 1 year, and then annually or as appropriate.
Comparison of LAA sizing among different imaging modalities LAA size measurements from pre-procedural TEE, pre-procedural CT and LAA angiography were compared in four ways: 1) maximal diameter of the LAA ostium, 2) maximal diameter at the landing zone, 3) maximal LAA depth, and 4) landing zone depth. Maximal diameter at the landing zone of each imaging modalities were also compared to the nal device size that was successfully implanted and is considered to be the optimal device size for the individual patient in our study.

Statistical analysis
Continuous variables are reported as mean ± SD and categorical variables are reported as counts and percentages. Continuous variables were compared using independent t-test and Mann-Whitney U test. Analysis of categorical variables employed the Chi-square test and Fisher's exact test. We investigated the agreement between maximal LAA landing zone diameter from each imaging modality and nal implanted device size using adapted Bland-Altman plot into the LAA. The 95% limits of agreement (mean difference ± 2 x SD) were calculated for each imaging modality. A two-sided p value < 0.05 was considered statistically signi cant. Statistical analyses were performed with SPSS (SPSS version 23).

Patients characteristics
A total of 61 patients were included in this study. The mean age was 74.0 ± 8.4 years with 26 (42.6%) female patients. The mean CHA 2 DS 2 score, CHA 2 DS 2 -VASc score and HAS-BLED score were 2.8 ± 1.4, 4.6 ± 1.8 and 2.6 ± 1.0, respectively. Important predictors of stroke including age >75 years, congestive heart failure, and previous stroke presented in 45.9%, 16.4%, and 47.5% of patients, respectively. There were 31 patients (50.8%) with a previous history of bleeding. Paroxysmal AF was found in 30 (49.2%) patients and permanent AF in 31 (50.8%) patients. Oral anticoagulation had been prescribed in 75.4% of patients, mainly warfarin (62.3%). Baseline characteristics are summarized in Table 1.
LAA closure procedural characteristics Device implantation was successful in all patients. Two different LAA closure devices were used: Watchman (n = 43, 70.5%) and Omega (n = 18, 29.5%). Median hospital length of stay was 1 day (interquartile range 1 to 2). All procedures were performed under local anesthesia with conscious sedation and TEE guidance. Mean numbers of attempts was 1.7 ± 1.6 times with 7 patients (11.5%) needed to change device size. Mean procedural time was 51.1 ± 24.4 minutes. Mean uoroscopy time was 7.7 ± 6.0 minutes.
The procedure was successful without peri-device leakage identi ed by TEE and angiogram in 57 patients (93.4%) and successful with mild leakage in 4 patients (6.6%). Final implanted device size was within manufacturer's recommendations in 42.1%, oversized in 54.4% and undersized in 3.5% of all patients. There was no major leakage with any patient. Immediate complications were air embolism into right coronary artery (n = 1, 1.6%) and pericardial effusion (n = 2, 3.3%). Post procedural medications were oral anticoagulant for 6 weeks followed by dual antiplatelets for 6 months in 36.1%, dual antiplatelets for 6 months followed by single aspirin therapy in 47.5% and extended dual antiplatelets in 13.1% with 3.3% not needing medications. LAA closure procedural characteristics were summarized in Table 2.
Forty-ve to sixty days TEE follow-up and long-term outcome At routine 45 to 60 day patient follow-up TEE, there were no leakage in 70.5%, mild leakage in 26.2% and moderate leakage in 3.3%. There was no difference in the severity of peri-device leakage among patients who had implanted device size within manufacturer's recommendation, oversized and undersized (p =0.898), as shown in Table 3. There was pericardial effusion in 1 patient (1.6%), device related thrombosis in 2 patients (3.3%), LA thrombus in 1 patient (1.6%) and device protrusion in 2 patients (3.3%). Median duration of follow up was 14.7 months (IQR: 5.3-45.9 months). At long-term follow up, three patients (4.9%) had recurrent stroke, two patients (3.3%) had cardiovascular death and one patient (1.6%) had pericardial effusion requiring pericardiocentesis.
Comparison of LAA size and device size selection by different imaging modalities Pre-procedural, periprocedural and post-procedural TEE were completed in 57 (93.4%) cases. All patients had angiographic assessment for LAA sizing. Only 18 patients (29.5%) had undergone a pre-procedural CT scan. Maximum diameter of LAA ostium measured by TEE, CT scan and angiography were 24.1 ± 6.1 mm, 26.6 ± 4.4 mm, and 22.5 ± 3.8 mm, respectively. Maximum landing zone diameter measured by TEE, CT scan and angiography were 22.2 ± 4.8 mm, 23.4 ± 3.9 mm and 22.7 ± 3.5 mm, respectively. Both maximum diameter of LAA ostium and landing zone diameter were larger as measured by CT scan compared to TEE (p = 0.025 and p =0.015).
Using the nal implanted device size as an optimal size for comparison, the CT scan was closer to device size compared to TEE and angiography with a difference of -1.65 ± 2.0 mm between landing zone diameter as measured by CT scan and device size compared to -4.8 ± 4.6 mm and -4.3 ± 3.3 mm, respectively). Agreement and accuracy of landing zone measurement for each imaging modality compared to implanted device size were studied using adapted Bland-Altman plots as shown in Figure 2

Discussion
We reported on a single-center retrospective study of 61 consecutive patients who underwent percutaneous LAA closure. Our study aimed to investigate LAA measurement with three different imaging modalities (TEE, MDCT and LAA angiography) and compare the accuracy of LAA sizing with the implanted device size and study the safety and e cacy of percutaneous LAA closure.
Our study showed high procedural success with only 4 patients (6.6%) experiencing mild peri-device leakage. Only 4.9% of cases had in-hospital complications which is in the range of other previous large studies and registries complication rates of 2.2-6.7%[7,8,5,13-15]. Our center chose a higher acceptable device compression rate of 10-30% resulting in a higher number of cases with oversizing (54.4%), based on manufacturer's recommendations [16,7]. However, there was no difference in peri-device leakage between the correct sized group and the oversized group at the 45 to 60-day follow-up assessment which may imply broader safety compression rate than currently recommended.
One of the main ndings from our study is that all parameters of LAA measurement including LAA ostial diameter, maximal depth and landing zone diameter are larger with MDCT sizing in comparison with TEE sizing. Landing zone diameter was 1.7 ± 2.6 mm larger with MDCT sizing than TEE sizing which may lead to device upsizing in a few cases. Using the nal implanted device size as the optimal LAA size, MDCT sizing was closer to the optimal size than TEE sizing with the difference between MDCT sizing and device size of only -1.7 ± 2.0 mm compared to -4.8 ± 4.6 mm for the difference between TEE sizing and optimal device size. Previous studies reported larger discrepancies between MDCT and TEE of up to 2-3 mm which led to sizing errors of up to 3.4% of cases [10,9,12,17,18]. We hypothesize that because our center uses a higher compression rate than manufacturer's recommendation, larger device size may correct the smaller landing zone diameter measured by TEE and have similar procedural success as using MDCT sizing.
Our study has some limitations. We conducted a single center retrospective study resulting in a relatively small number of patients. There were two LAA closure devices with different sizing recommendations and device characteristics. Not all the included patients had MDCT images with only patients in the Omega device group having MDCT sizing. TEE, MDCT and angiography were retrospectively reviewed and measured by investigators who were not? the same group of clinicians who performed the procedure and may have introduced measurement bias.

Conclusion
MDCT sizing of LAA was more accurate compared to TEE sizing and LAA angiography. Routine implementation of MDCT sizing may improve procedural success with more accurate device size selection. Sizing discrepancies using standard TEE images and LAA angiography may be overcome by increased device size and higher acceptable compression rate.

Declarations
Funding: The study did not receive funding.

Con ict of interest:
There are no con icts of interest to declare.