Preliminary report on use of 3-dimensional computed tomographic images in a disease-based transesophageal echocardiographic simulation system

Abstract

We used 3-dimensional computed tomographic images to create a disease-based transesophageal echocardiographic simulation system for complex congenital heart defects. We enrolled 7 pediatric patients with complex congenital heart defects in this proof-of-concept study. Preoperative computed tomographic images and intraoperative transesophageal echocardiographic images were acquired for all patients. Two- and 3-dimensional computed tomographic cross-sectional images were created to simulate the process of transesophageal echocardiographic image acquisition. Computed tomographic images simulating the midesophageal 4- and 5-chamber views, aortic valve short-axis views, long-axis views, and ascending aortic short-axis views were created to correspond with the actual transesophageal echocardiographic images from each patient. Four reviewers then evaluated the image quality of the computed tomographic images, the agreement between the echocardiographic and tomographic images, and the ability of the 3-dimensional computed tomographic full-volume and cross-sectional images to yield the spatial and temporal congruence of transesophageal echocardiograms. In most of the patients, computed tomography yielded images of good-to-excellent quality. Strong agreement was noted between the computed tomographic and transesophageal echocardiographic images acquired in the same patients. The ability of 3-dimensional computed tomography to yield the spatial and temporal congruence of transesophageal echocardiography in selected planes was also good to excellent. We found that 3-dimensional computed tomographic images can simulate the process of transesophageal echocardiography in acquiring the echocardiographic image clearly. This imaging method has the potential to be applied successfully to a disease-based transesophageal echocardiographic simulation system.

Keywords: Adult health personnel/education; computer graphics; echocardiography, transesophageal/instrumentation; heart defects, congenital/ultrasonography; image processing, computer-assisted/methods; imaging, three-dimensional/methods; tomography, x-ray computed/methods.

Fig. 1 Images show the process of using 3-dimensional computed tomography to simulate that of transesophageal echocardiographic (TEE) image acquisition in a patient with tetralogy of Fallot. A 2-dimensional image of the heart in 3 orthogonal planes—A) cross-sectional, B) coronal, and C) sagittal views—can be produced simultaneously. Each planar image can be adjusted by repositioning the orthogonal position lines in the other 2 planes. D) In this example, the esophagus is first identified in cross-sectional view by putting the crossover point of the 2 orthogonal position lines on it; then, retroflex and omniplane angulation motion is simulated in the E) sagittal and F) coronal images. Next, a simulated TEE image is acquired in cross-sectional view. Three simulated images from different angles at the midesophageal level are shown: G) midesophageal 4-chamber view (0°), H) aortic valve short-axis view (40°), and I) long-axis view (120°).

Fig. 2 In the heart of a patient with tetralogy of Fallot, reconstructed 3-dimensional computed tomographic full-volume and cross-sectional images simulate transesophageal echocardiographic results and congruence in midesophageal 4-chamber view (ME 4-chamber), aortic valve short-axis view (ME AV-SAX), long-axis view (ME LAX), and ascending aortic short-axis view (ME Aortic SAX).

Fig. 3 In a patient with tetralogy of Fallot, transesophageal echocardiographic images (without and with color-flow Doppler), compared with corresponding computed tomographic 3-dimensional and 2-dimensional cross-sectional images, show the spatial and temporal congruence of the 3 imaging modes in midesophageal A), B), C) aortic valve short-axis view and D), E), F) long-axis view. Ao = aorta; LV = left ventricle; RV = right ventricle