Evaluating the morphology of the left atrial appendage by a transesophageal echocardiographic 3-dimensional printed model

Abstract

The novel 3-dimensional printing (3DP) technique has shown its ability to assist personalized cardiac intervention therapy. This study aimed to determine the feasibility of 3D-printed left atrial appendage (LAA) models based on 3D transesophageal echocardiography (3D TEE) data and their application value in treating LAA occlusions.Eighteen patients with transcatheter LAA occlusion, and preprocedure 3D TEE and cardiac computed tomography were enrolled. 3D TEE volumetric data of the LAA were acquired and postprocessed for 3DP. Two types of 3D models of the LAA (ie, hard chamber model and flexible wall model) were printed by a 3D printer. The morphological classification and lobe identification of the LAA were assessed by the 3D chamber model, and LAA dimensions were measured via the 3D wall model. Additionally, a simulation operative rehearsal was performed on the 3D models in cases of challenging LAA morphology for the purpose of understanding the interactions between the device and the model.Three-dimensional TEE volumetric data of the LAA were successfully reprocessed and printed as 3D LAA chamber models and 3D LAA wall models in all patients. The consistency of the morphological classifications of the LAA based on 3D models and cardiac computed tomography was 0.92 (P < .01). The differences between the LAA ostium dimensions and depth measured using the 3D models were not significant from those measured on 3D TEE (P > .05). A simulation occlusion was successfully performed on the 3D model of the 2 challenging cases and compared with the real procedure.The echocardiographic 3DP technique is feasible and accurate in reflecting the spatial morphology of the LAA, which may be promising for the personalized planning of transcatheter LAA occlusion.

Figure 1

The working flow of echocardiographic 3-dimensional (3D) replica generation for 3D printing. The whole data postprocessing for 3D printing was demonstrated, from the original raw DICOM image acquisition to the STL file. (A) The full volumetric raw data acquired from 3D TEE in a DICOM format. (B and C) Raw data view before and after postprocessing by the GVI mode. The chamber of the LAA was displayed as black in the original image (B), whereas it was white after being processed by GVI (C). (D) LAA chamber after threshold segmentation. (E) LAA chamber model STL file after the whole segmentation and other postprocesses. (F) Step that generated the wall model, which was needed to extend the chamber mask 1 to 2 mm to span the thickness of the LAA wall, and then a wall model replica was finally obtained (G). DICOM = Digital Imaging and Communications in Medicine, GVI = gray value inverted, LAA = left atrial appendage, STL = Standard Tessellation Language, TEE = transesophageal echocardiography.

Figure 2

LAA measurement on 3D TEE images and on the 3-dimensional (3D)-printed wall model. LAA measurement comprising ostial dimensions and the depth on 3D TEE (A–C), and on the 3D-printed wall model (D–F). In 3D TEE, the left circumflex coronary artery was considered to be the anatomic mark to reconstruct the axial view of the LAA ostium (A). D1 represented the maximal diameter of the ostium, and D2 was the minimal dimension that was perpendicular to D1 (B). (C) Depth of the LAA. Regarding the ostium plane on the 3D model, the internal transition point was first decided and then the other point was set at a position lower than the model margin (1–2 cm), which was similar to the 3D TEE measurement method (D). The Vernier caliper was placed on the ostium plane to measure D1 and D2 (E), and then the measured value was recorded in the model as the depth of the LAA model (F). LAA = left atrial appendage, TEE = transesophageal echocardiography.

Figure 3

The 3D-printed models of LAA. (A–D) Solid cardiac chamber models used to display the various morphologies of the LAA. (A) Windsock type; (B) chicken wing type; (C) cauliflower type; and (D) cactus-type LAAs. (E–H) Corresponding 3D wall models for A–D, which were hollow and flexible, and used for the LAA dimension measurements and device implantation practices. LAA = left atrial appendage.

Figure 4

Bland-Altman plots showing agreement in the measurements obtained in the 3-dimensional (3D) print model and 3D TEE methods. (A) Maximal dimension of the long axis; (B) Short-axis dimension of the LAA ostium; (C) Depth of the LAA. LAA = left atrial appendage, TEE = transesophageal echocardiography.

Figure 5

The device implantation rehearsal mimicked the real procedure during a challenging case with a bi-lobed LAA. The 2 ostia could be visualized by 3-dimensional (3D) TEE (A), and detailed anatomic structures were displayed by the 3D-printed chamber model demonstrating the similar size, the orientation, and relationship of the 2 lobes (B). A simulation operative practice demonstrated that two device occlusion could be easily performed on the 3D model (C), which was similar to the actual procedure in this specific case (D). The status of these 2 devices positioned in the model, and the relationships between the devices and with the LAA model were well-displayed and compared with the actual occlusion image, which indicates that the performance of the model can be considered for future preprocedural planning. LAA = left atrial appendage, TEE = transesophageal echocardiography.