Aortic valve landmarks detection in Computed Tomography (CT) images is essential for planning Transcatheter Aortic Valve Implantation (TAVI) and establishing predictive factors of cardiac conduction disturbance. The fully automatic detection facilitates the measurement and planning while reducing interobserver variability. Although the emerged Convolutional Neural Networks (CNNs) have been applied to fully automatic landmark detection, it is challenging for CNNs to directly process large-scale CT volumes due to the limited computational resources. Some common preprocessing to deal with the limitation of resources, such as down-sampling or center cropping, can cause the volume to lose detailed features or global information. To address these issues, we propose a two-stage detection method based on CNN. The proposed method initially detects the approximate positions of the landmarks in the global view and then obtains refined results in local regions, without overdependence on prior knowledge or labor-intensive preprocessing. Eight important aortic valve landmarks, including three hinge points, three commissure points, the middle point of the cusps, and the point of the lower part of the membranous septum are automatically detected from our network. An overall result of Mean Radial Error (MRE) of 2.23 mm is yielded from our data set containing 150 individual cardiac CT volumes. The method takes 0.15 seconds per stage to process one volume, showing high efficiency.
KEYWORDS: Aorta, Image segmentation, 3D modeling, Education and training, Data modeling, Visualization, Feature fusion, Contrast transfer function, Aneurysms, Tissues
PurposeSegmentation of vascular structures in preoperative computed tomography (CT) is a preliminary step for computer-assisted endovascular navigation. It is a challenging issue when contrast medium enhancement is reduced or impossible, as in the case of endovascular abdominal aneurysm repair for patients with severe renal impairment. In non-contrast–enhanced CTs, the segmentation tasks are currently hampered by the problems of low contrast, similar topological form, and size imbalance. To tackle these problems, we propose a novel fully automatic approach based on convolutional neural network.ApproachThe proposed method is implemented by fusing the features from different dimensions by three kinds of mechanisms, i.e., channel concatenation, dense connection, and spatial interpolation. The fusion mechanisms are regarded as the enhancement of features in non-contrast CTs where the boundary of aorta is ambiguous.ResultsAll of the networks are validated by three-fold cross-validation on our dataset of non-contrast CTs, which contains 5749 slices in total from 30 individual patients. Our methods achieve a Dice score of 88.7% as the overall performance, which is better than the results reported in the related works.ConclusionsThe analysis indicates that our methods yield a competitive performance by overcoming the above-mentioned problems in most general cases. Further, experiments on our non-contrast CTs demonstrate the superiority of the proposed methods, especially in low-contrast, similar-shaped, and extreme-sized cases.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.