Comparison of multiple imaging modalities for measuring orifice diameter and selecting occluder size in patients undergoing left atrial appendage closure

Abstract Background Left atrial appendage (LAA) closure (LAAC) can safely and effectively prevent stroke events caused by atrial fibrillation. However, the structure of the LAA is highly variable among individuals, and the optimal method for obtaining measurements remains unknown. Hypothesis We aimed to study the accuracy of left atrial computed tomography angiography (CTA), three‐dimensional (3D) reconstruction using CTA, two‐dimensional transesophageal echocardiography (2D‐TEE), and digital subtraction angiography (DSA) for measuring the diameter of the LAA and compare their value for selecting occluder size. Methods We retrospectively evaluated data for 148 patients with nonvalvular atrial fibrillation who underwent successful LAAC. CTA and 2D‐TEE of the left atrium and pulmonary vein were performed before LAAC. We performed 3D reconstruction of the left atrium and LAA using Mimics and 3‐matics software. DSA of the LAA was performed during surgery. Results Values measured via CTA 3D reconstruction were significantly higher than those measured using other methods. DSA‐measured values were significantly lower than those measured via CTA and CTA 3D reconstruction. Occluder size was positively correlated with LAA orifice diameter. The differences between occluder size and DSA, 2D‐TEE, CTA, CTA 3D reconstruction measurements were 4.96 ± 2.58, 4.64 ± 2.50, 4.04 ± 1.37, and 2.92 ± 1.38 mm, respectively. Intraclass correlation coefficients for these methods were −.067, .006, .241, and .519, respectively. Conclusion CTA 3D reconstruction provides the best correlation and consistency between the measured LAA orifice diameter and occluder size. Adding 2–4 mm to the maximum LAA orifice diameter based on 3D‐CTA may aid in selecting the appropriate WATCHMAN device.

the maximum LAA orifice diameter based on 3D-CTA may aid in selecting the appropriate WATCHMAN device.

K E Y W O R D S
3D reconstruction, atrial fibrillation, left atrial appendage occlusion, occluder size

| INTRODUCTION
Atrial fibrillation is a common symptom of rapid arrhythmia in middleaged and older individuals, and its incidence increases with age.
During atrial fibrillation, a thrombus tends to form in the left atrial appendage (LAA), originating from this area in 90% of cardiogenic thrombus. A recent study reported that this rate increases to almost 100% in patients with nonvalvular atrial fibrillation (NVAF). 1 Emboli that develop following thrombus shedding represent the main cause of death and disability in patients with atrial fibrillation. 2 Left atrial appendage closure (LAAC) has been recommended to reduce the risk of stroke following atrial fibrillation. LAAC aims to prevent thrombus formation in the LAA during atrial fibrillation via an occluder implanted at the LAA orifice, thereby reducing the risk of long-term disability or death caused by thromboembolism. The size and morphology of the LAA vary significantly among patients.
Preoperative computed tomography angiography (CTA) threedimensional (3D) reconstruction, two-dimensional transesophageal echocardiography (2D-TEE), and intraoperative digital subtraction angiography (DSA) are often used to evaluate the size, depth, and axial direction of the LAA and provide adequate guidance for operations. However, values have differed substantially among studies. The early prospective WATCHMAN trials suggested that an average of 1.8 occluders were required for each occlusion operation, reported successful occlusion in 82% of cases. 3 One of the most common reasons for reselecting the occluder in this study was that the originally selected device was too small, indicating that selecting occluder sizer based on the size of LAA on DSA or TEE may not be the most accurate method. This study aimed to compare the value of different imaging modalities for measuring the LAA orifice diameter and selecting occluder size. principle and the residual shunt was ≤5mm. The immediate compression ratio was 22.25 ± 3.33 mm as measured via TEE. There were nine cases of residual shunting, which were distributed as follows: 1 mm, n = 1 case; 2 mm, n = 5 cases; 3 mm, n = 3 cases. Thus, all cases met standards for occluder release.

| General information
All patients were followed-up via telephone. Three patients died    Figure 1A).

| Statistical analyses
Statistical analysis was performed by SPSS 22.0 (IBM). Measurement data conforming to a normal distribution were described as the mean and standard deviation, while those not conforming to a normal distribution were described as the median and quartile. The Shapiro-Wilk test was used to test for normality. Normally distributed measurement data were compared between two groups using t-tests and among multiple groups using analyses of variance.  (Figure 3).
The ICCs of occluder size and LAA orifice diameter measured using DSA and 2D-TEE were −0.07 and 0.01, respectively; however, the difference was not statistically significant (Table 4).

| DISCUSSION
Atrial fibrillation is one of the most common arrhythmias in older adults, and the risks of stroke and heart failure are significantly higher among patients with atrial fibrillation than among those without. The The size and effective depth of the LAA have important reference values that allow for the selection of the appropriate occluder type and size. However, the structure of the LAA is highly variable among individuals, and the optimal method for obtaining these measurements remains to be determined.
In the DSA method, contrast agent is injected into the LAA during operation to make the blood flow in the LAA visible under  value. In addition, the surgeon's experience and equipment will also Note: Since the 2D-TEE, CTA, and occlude size did not meet the normal distribution, the Spearman rank correlation analysis and Wilcoxon signedrank test were used.
F I G U R E 2 Correlation between LAA orifice diameter as measured by different imaging methods and occluder size. 2D-TEE, two-dimensional transesophageal echocardiography; 3D, three-dimensional; CTA, computed tomography angiography; DSA, digital subtraction angiography; LAA, left atrial appendage.
operator's experience and equipment-related errors, such as artifact interference; (2) TEE is a semi-invasive examination that can be highly  selection. A small-sample single-center study in China reported on the use of the Mimics simulated 3D imaging system for visualizing the LAA in patients with atrial fibrillation before surgery, demonstrating that the system provided the necessary information and allowed for the success of LAAC. 9 One study indicated that LAA values measured using DSA, 2D-TEE, and coronary CTA differed significantly. 10  Our study has some limitations, including its single-center retrospective design, highlighting the need for a prospective multicenter study with a larger sample size for further verification.
The occluders used in this study were all "internal plug" WATCHMAN occluders, while cases with "external cover" occluders were excluded.
Thus, our conclusions may be not applicable to LAAC procedures utilizing "external cover" occluders such as ACP, LACbes, and LAmbre occluders. Further, we only utilized 2D-TEE in the current study.
However, real-time 3D-TEE exhibits greater functionality and better image quality, which may provide more structural information regarding the LAA.
Specifically, 3D-TEE overcomes the insufficient spatial resolution of 2D-TEE. It can slice the image from any angle or plane stereoscopically and intuitively without the interference of spontaneous development, making the hit highly applicable for real-time evaluation of LAA morphology. Particularly, compared with 2D-TEE, 3D-TEE can better distinguish thrombi and comb muscle/trabecula 12 ; allow for dynamic observation of systolic and diastolic movement as well as volume changes in the LAA 13 ; and ensure better consistency between the measured value and the size of the occluder. 11 Compared with DSA and 2D-TEE, CTA 3D reconstruction is advantageous in that it can be performed preoperatively and has been associated with good patient tolerance. It is an excellent, accurate, and intuitive technical means for the preoperative evalua-