Aortic atherosclerosis is a risk factor for cerebrovascular events. Two-dimensional transesophageal echocardiographic quantification of descending aortic plaques is time-consuming and underestimates plaque burden. The aim of this study was to assess the feasibility and accuracy of a novel semiautomated program that uses three-dimensional (3D) transesophageal echocardiography to identify and quantify aortic plaque severity as determined by plaque thickness, volume, and number. The relationship between maximum plaque thickness and volume was also examined.
Descending aortic 3D transesophageal echocardiographic images from 58 patients were analyzed for plaque thickness, volume, and number using semiautomated custom software. The reference standard was manual assessment by an expert reader using 3D multiplanar reconstructions. Agreement and κ values were calculated to determine the program’s accuracy against the reference standard. Correlation and bias were examined using linear regression and Bland-Altman statistics. Pearson’s correlation was used to examine the relationship between maximum plaque thickness and volume.
Analysis was possible in all patients. Overall agreement for the absolute presence or absence of plaque per patient was 95%. Agreement regarding the number of plaques per patient and plaque severity was high at 95% and 85%, respectively. Plaque volume was slightly underestimated by the program compared with manual measurements. The correlation between plaque thickness and volume was 0.56.
The results of this study demonstrate that semiautomated plaque analysis of 3D transesophageal echocardiographic descending aortic data sets is feasible and accurate in determining plaque severity as measured by plaque thickness, volume, and number. This methodology allows the standardization of plaque quantification, which will improve its utility in clinical trials. A greater understanding of the importance of plaque thickness versus volume is needed.
Strokes are the fourth leading cause of death in the United States, and transesophageal echocardiographic (TEE) imaging is routinely performed in patients with stroke when an extracardiac source of embolism has not been identified. Typically, at the end of a TEE examination, the echocardiographer will examine the aorta for complex atherosclerotic plaques, which when present significantly increase stroke risk. To be considered complex, these plaques must be ≥4 mm in thickness, be ulcerated, or contain mobile thrombi. However, determination of plaque complexity from two-dimensional (2D) TEE images may be inaccurate, because only one cross-sectional plane can be visualized or measured at a time, ignoring the complex three-dimensional (3D) shape of the plaque.
Three-dimensional TEE imaging of the aorta results in the acquisition of 3D data sets that can be used to accurately quantify the plaque’s 3D dimensions and volume. However, this quantification is a tedious and time-consuming process because it requires manual analysis of multiple 2D images from the 3D data set. We hypothesized that this process would benefit from automation and accordingly developed semiautomated software that detects and quantifies plaques from 3D descending aortic TEE images. The aim of this study was to assess the feasibility and accuracy of the program to (1) identify the presence or absence of plaque, (2) identify the locations and number of plaques, (3) determine plaque severity, (4) determine plaque volume, and (5) determine the relationship, if any, between maximum plaque thickness and plaque volume.
Clinically indicated TEE studies were performed in 58 patients (35 men; mean age, 60 ± 16 years). Forty-one patients were studied at the University of Chicago and 17 at the Lenox Hill Heart and Vascular Institute. Three-dimensional TEE (X-7t; Philips Medical Imaging, Best, The Netherlands) data sets of the descending aorta were acquired, using single-beat, narrow-angle acquisition mode. These images were acquired at 0°, with the TEE probe rotated toward the descending aorta ( Figure 1 ). First, the 2D short-axis image of the aorta was optimized, allowing maximum contrast between blood-tissue planes. Then, multiplane mode was used to ensure that the short axis and long axis of the aorta were centered within the 3D pyramidal volume. Finally, narrow-angle, single-beat data sets, with the acquisition pyramid minimized to include only the aorta, were acquired. Gain and rendering settings were left up to the discretion of the operators, as both sites have extensive experience acquiring 3D data sets. No postprocessing was required after acquisition.
If no atherosclerotic plaque was present, a representative region of the descending aorta was imaged. If atherosclerotic plaque was present, overlapping 3D data sets were obtained from the level of the diaphragm to the aortic arch. The descending aorta was defined as the portion that could first be imaged during withdrawal from the stomach. The aortic arch was defined where the short-axis image of the aorta could no longer be kept centered and circular during withdrawal of the probe. These data sets were then exported in Digital Imaging and Communications in Medicine format for semiautomated analysis on a dedicated computer system. The institutional review boards of both institutions approved the protocol. The software was written in Chicago and Milan.
Semiautomated Plaque Analysis
An algorithm that requires minimal user interaction for plaque segmentation and quantification was developed (see the Appendix for further details). No patent application has been made.
Three-Dimensional Data Set Orientation
The first step of the analysis was to orient the 3D volume such that one axis corresponds to the longitudinal dimension of the aorta. This was achieved by having the user define a volume of interest by selecting two slices that mark the ends of the aorta ( Figure 2 A) within the 3D data set. The user then selected a point in each of these slices that visually corresponded to the center of the aortic lumen. A line passing between these two points represented the long axis of the aorta. The 3D data set was then rotated so that the short-axis aortic lumen was presented, which was perpendicular to the calculated long axis. This presented view was required for the segmentation process.