Advanced Transcatheter Aortic Valve Implantation (TAVI) Planning from CT with ShapeForest

Abstract

Transcatheter aortic valve implantation (TAVI) is becoming a standard treatment for non-operable and high-risk patients with symptomatic severe aortic valve stenosis. As there is no direct view or access to the affected anatomy, comprehensive preoperative planning is crucial for a successful outcome, with the most important decisions made during planning being the selection of the proper implant size, and determining the correct C-arm angulations. While geometric models extracted from 3D images are often used to derive these measurements, the complex shape variation of the AV anatomy found in these patients causes many of the shape representations used to estimate such geometric models to fail in capturing morphological characteristics in sufficient detail. In addition, most current approaches only model the aortic valve (AV), omitting modeling the left ventricle outflow tract (LVOT) entirely despite its high correlation with severe complications such as annulus ruptures, paravalvular leaks and myocardial infarction. We propose a fully automated method to extract patient specific models of the AV and the LVOT, and derive comprehensive biomarkers for accurate TAVI planning. We utilize a novel shape representation -the ShapeForest -which is able to model complex shape variation, preserves local shape information, and incorporates prior knowledge during shape space inference. Extensive quantitative and qualitative experiments performed on 630 volumetric data sets demonstrate an accuracy of 0.69 mm for the AV and 0.83 mm for the LVOT, an improvement of over 16% and 18% respectively when compared against state of the art methods.

Automatic 3D aortic annulus sizing by computed tomography in the planning of transcatheter aortic valve implantation

“…11,14e17,21,23,24 In fact, the LOAs against expert measurements were lower than many reported in the literature for automated solutions. 17,21,23,24 Interestingly, when assessing interobserver variability for the SA approach, LOAs for the proposed software were also excellent (BA:0.02 ± 0.70 in Ref. 15, and ICC: 0.98 in Ref.…”

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

“…To help clinicians assess the AV morphology, several methods have already been proposed to (semi-) automatically segment it in MDCT [12]- [17], 2-D-TEE [18]- [20] and, more recently, 3-D-TEE datasets [12], [21]- [24]. Besides being possible to separate these studies based on their target modality, they can also be divided according to the methodology used (machine learning [12], [15], [17], deformable models [13], [21], [22], [24], 0018-9294 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.…”

“…See http://www.ieee.org/publications standards/publications/rights/index.html for more information. among others) or the required level of user interaction (fully automatic [12], [15]- [17], [21], [22], semi-automatic [18]- [20], [23], [24] or interactive [14]). Among 3-D-TEE methodologies, Ionasec et al [12] proposed the first fully automatic algorithm for AV segmentation.…”

Artificial intelligence and automation in valvular heart diseases

“…A generic physiological model which can represent aortic valve and its pathological variations was constructed first, patient-specific parameters of the model were estimated from volumetric sequences by trajectory spectrum learning, marginal space learning and discriminative learning. Their study was further improved with shape forest to constrain the classical statistical shape model [66]. Waechter et al [67] extracted the aortic valve geometry from CT images using model-based segmentation and applies pattern search method to detect the coronary ostia.…”