1 | INTRODUCTION
Cardiac computed tomography (CT), fluoroscopy and trans-esophageal echocardiography (TEE) are three major imaging modalities, which support the increasing complexity of trans-catheter interventions in congenital and structural heart diseases. 1–3 These modalities are based on different mathematical coordinate systems, and are often demonstrated at different visual perspectives. Multi-planar reconstruction (MPR) of cardiac CT, commonly used as pre-operative planning tool, is based on the standard Cartesian coordinate system (Figure 1A). Although manual adjustment is allowed for plane rotation and movement of the origin of the coordinate system into the region of interest (ROI), the re-sliced 2D images can still be intuitively understood because CT datasets include abundant isotropic details in cardiac anatomy. Instead of multi-slice imaging, fluoroscopy provides with the X-ray projection from the visual perspective of the image intensifier, whose trajectory is clinically described by the spherical coordinate system (Figure 1B) as right anterior oblique (RAO)/left anterior oblique (LAO) and cranial (CRA)/caudal (CAU). The coordinate system of TEE is the most complex so the same structure scanned at different esophageal levels requires mental co-registration in understanding (Figure 1C, right). The TEE is basically a cylindrical coordinate system (Figure 1C, left) and its monitor produces 2D anatomical information on the polar coordinate reference plane, similar to the axial and the coronal CT MPR planes for trans-esophageal and trans-gastric windows, respectively. However, the longitudinal axis (similar to z-axis in Cartesian system) of such cylindrical coordinate system is not a straight line, not to mention the anteflex and retroflex probe manipulation; rather, the longitudinal axis actually means the esophageal tract. Furthermore, the reference plane is allowed to rotate around its polar axis (the mid-line of the scanning sector) and the geometry of flex to the right/left maneuver is even more complex. As a result, both the acquisition and the understanding of TEE images need learning curve.
Mental co-registration of these modalities is the foundation to facilitate heart-team communication. 4,5 For quick switch between imaging modalities with different coordinate systems, assistive theory and technology have been developed. Echocardiography-fluoroscopic fusion imaging transforms the coordinate system of TEE probe to that of the image intensifier so that 2D and 3D TEE images can be calibrated and projected onto the 2D fluoroscopy.6,7 As a result, images of these two modalities can be viewed in the same visual perspective 4,5,8 with acceptable mean target registration error. Piazza and colleagues9,10 introduced the concept of optimal projection curve that comprises all the projection angles of which the X-ray tube-to-image intensifier direction is orthogonal to the normal vector of the interested anatomical targets. Their method is applied to minimize the parallax during the deployment of a quasicylindrical device into an anatomical structure of variable geometry. However, the bridge between cardiac CT and TEE is lacking. Adjusting the MPR crosshairs into the center of mitral valve plane and crossing the third MPR line through the LV apex are the traditional methods to obtain LV four-chamber, two-chamber and long-axis CT re-slices without foreshortening.11 These views precisely describe the LV segments and mitral apparatus, but are usually at a different reference plane from that of TEE imaging. Although TEE simulator based on CT images has been developed as a valuable teaching tool, 12 it sacrifices the wide visual field of CT scan and loses image quality during datasets transformation. Due to rapid development of new trans-catheter devices, more and more distinct imaging requirements are called for. As a result, it is mandatory to bridge the gap between the pre-operative planning tool and the intra-operative imaging guidance.
For this purpose, we developed the methods of stepwise cardiac CT MPR manipulation to mimic trans-esophageal echocardiography. On the other hand, from the patient-specific information of cardiac CT, we can also plan the TEE imaging in a preoperative rehearsal to make the desired standardized TEE views more reproducible.
2 | METHOD (Figure 2)
Since the motion of TEE probe is confined to the esophagus-stomach tract, it is difficult for TEE sonographers to reproduce the cardiac CT MPR views used for preoperative plan. By contrast, it is easier to use CT datasets to simulate TEE imaging. There are three basic MPR planes – axial, coronal and sagittal – that are perpendicular to each other. One can adjust the cross-sectional orientation of the other two planes by moving and rotating their MPR lines on the selected MPR plane (i.e., we can adjust coronal and sagittal MPR lines on axial plane and vice versa). To represent the TEE scanning sector, we imagine a circular sector with the origin in the MPR crosshairs on the axial plane and the sector has a line of symmetry as the sagittal MPR line. Once we move the MPR crosshairs into the esophagus with the following sequential steps, we can translate basic TEE probe movements into corresponding CT MPR manipulation with the following steps.
Step 1: Manipulation on the sagittal plane: advance, withdrawal, anteflex and retroflex
The first step to simulate TEE on CT axial MPR plane is to invert the z-axis (the coronal MPR line on the sagittal MPR plane), because CT offers a perspective from a caudal direction on the axial MPR plane but TEE, the cranial perspective on the reference plane. Next, we simulate the advance/withdrawal by moving the crosshairs along the esophageal tract into the gastric space. Moreover, we can simulate the anteflex by clockwise rotation of the axial MPR line on the sagittal MPR plane and the retroflex by counterclockwise rotation.
Step 2: Manipulation on the axial plane: turn to the left/right
Because the sector width of TEE is limited, sonographers have to turn the TEE probe rightward to observe the right-sided structures and vice versa. We can simulate this maneuver by rotating the sagittal MPR line on the axial MPR plane. Step1 and Step 2 can be repeated randomly until satisfactory image is obtained on the axial MPR plane before moving on to the step 3.
Step 3: Manipulation on the coronal plane: rotation, flex to the left/right
Since the z-axis has been inverted at the very beginning, the anterior posterior direction on the coronal MPR plane is reversed, too. As a result, the counterclockwise rotation of TEE scanning plane will be simulated by a clockwise rotation of the axial MPR line on the coronal MPR plane. TEE probe is steerable for delicate adjustment of imaging orientation. The flexible joint is around 5cm above the scanning center. To simulate the “flex to the right,” two steps have to be taken. Firstly, we rotate the crosshairs about the flexible joint of TEE probe. Secondly, a negative rotation with the same degree that the crosshairs are rotated should be added. Examples of flex to the right at mid-esophagus level and flex to the left at transgastric level are illustrated in Figure 3 and Figure 4 respectively.