Fully Automatic 3-D-TEE Segmentation for the Planning of Transcatheter Aortic Valve Implantation

Abstract

A novel fully automatic framework for aortic valve (AV) trunk segmentation in three-dimensional (3-D) transesophageal echocardiography (TEE) datasets is proposed. The methodology combines a previously presented semiautomatic segmentation strategy by using shape-based B-spline Explicit Active Surfaces with two novel algorithms to automate the quantification of relevant AV measures. The first combines a fast rotation-invariant 3-D generalized Hough transform with a vessel-like dark tube detector to initialize the segmentation. After segmenting the AV wall, the second algorithm focuses on aligning this surface with the reference ones in order to estimate the short-axis (SAx) planes (at the left ventricular outflow tract, annulus, sinuses of Valsalva, and sinotubular junction) in which to perform the measurements. The framework has been tested in 20 3-D-TEE datasets with both stenotic and nonstenotic AVs. The initialization algorithm presented a median error of around 3 mm for the AV axis endpoints, with an overall feasibility of 90%. In its turn, the SAx detection algorithm showed to be highly reproducible, with indistinguishable results compared with the variability found between the experts' defined planes. Automatically extracted measures at the four levels showed a good agreement with the experts' ones, with limits of agreement similar to the interobserver variability. Moreover, a validation set of 20 additional stenotic AV datasets corroborated the method's applicability and accuracy. The proposed approach mitigates the variability associated with the manual quantification while significantly reducing the required analysis time (12 s versus 5 to 10 min), which shows its appeal for automatic dimensioning of the AV morphology in 3-D-TEE for the planning of transcatheter AV implantation.

Automated segmentation of the atrial region and fossa ovalis towards computer-aided planning of inter-atrial wall interventions

“…6 A). As such and as previously observed in other studies [13,35] , small differences between central axis will result in high P2S errors. When evaluating the common AO region, i.e.…”

“…The atlas-based technique allows a coarse identification of the AO axis, presenting a sub-optimal performance in the left ventricular outflow tract (LVOT) region, due to the lack of walls. Inspired by [13] , we propose an AO axis correction technique. The method starts by estimating the leaflets positions through a Canny edge filter ( Fig.…”

“…The success of the aforementioned image-fusion strategies is directly related to the accuracy of the image alignment approach [7] . While some researchers focused on strategies to align the preand intra-procedural data using 2D/3D alignment approaches [10] , landmark identification [11] and/or image registration [12] , other researchers optimized the planning stages suggesting automated algorithm versions [2,13] . These automated methods showed a performance comparable to the manual analysis [2,14,15] , requiring none or little user input, removing or minimizing the variability between observers and reducing the time spent during the planning [2] .…”

Artificial intelligence and echocardiography

“…Automated assessments of 3D transoesophageal echocardiograms of the mitral valve provided more reproducible and consistent quantitative assessment of the mitral valve annulus size and its morphology than human interpretation ( 6 , 47 ). An extensive work also has been done in the field of aortic valve segmentation for planning transthoracic aortic valve implantation procedure ( 48 , 49 , 50 ).…”

Automatic aortic root segmentation and anatomical landmarks detection for TAVI procedure planning

“…Grbic et al [19] employed robust machine learning algorithms to estimate the valve model parameters from non-contrast CT including information on valve leaflets and calcium. Segmentation of aortic valve from TEE using an improved probability estimation and continuous max-flow approach has also been proposed [20], or using a combination of shape-based B-Spline explicit active surface and generalized Hough transform [21].…”