Semi-automatic aortic valve tract segmentation in 3D cardiac magnetic resonance images using shape-based B-spline Explicit Active Surfaces

摘要

Accurate preoperative sizing of the aortic annulus (AoA) is crucial to determine the best fitting prosthesis to beimplanted during transcatheter aortic valve (AV) implantation (TAVI). Although multidetector row computedtomography is currently the standard imaging modality for such assessment, 3D cardiac magnetic resonance (CMR) is afeasible radiation-free alternative. However, automatic AV segmentation and sizing in 3D CMR images is so farunderexplored. In this sense, this study proposes a novel semi-automatic algorithm for AV tract segmentation and sizingin 3D CMR images using the recently presented shape-based B-spline Explicit Active Surfaces (BEAS) framework.Upon initializing the AV tract surface using two user-defined points, a dual-stage shape-based BEAS evolution isperformed to segment the patient-specific AV wall. The obtained surface is then aligned with multiple reference AV tractsurfaces to estimate the location of the aortic annulus, allowing to extract the relevant clinical measurements. Theframework was validated in thirty datasets from a publicly available CMR benchmark, assessing the segmentationaccuracy and the measurements’ agreement against manual sizing. The automated segmentation showed an averageabsolute distance error of 0.54 mm against manually delineated surfaces, while demonstrating to be robust against thealgorithm’s parameters. In its turn, automated AoA area-derived diameters showed an excellent agreement againstmanual-based ones (-0.30±0.77 mm), being comparable to the interobserver agreement. Overall, the proposed frameworkproved to be accurate, robust and computationally efficient (around 1 sec) for AV tract segmentation and sizing in 3DCMR images, thus showing its potential for preoperative TAVI planning.