Automatic Identification of Bioprostheses on X-ray Angiographic Sequences of Transcatheter Aortic Valve Implantation Procedures Using Deep Learning.

Transcatheter aortic valve implantation (TAVI) has become the treatment of choice for patients with severe aortic stenosis and high surgical risk. Angiography has been established as an essential tool in TAVI, as this modality provides real-time images required to support the intervention. The automatic interpretation and parameter extraction on such images can lead to significative improvements and new applications in the procedure that, in most cases, rely on a prior identification of the transcatheter heart valve (THV). In this paper, U-Net architecture is proposed for the automatic segmentation of THV on angiographies, studying the role of its hyperparameters in the quality of the segmentations. Several experiments have been conducted, testing the methodology using multiple configurations and evaluating the results on different types of frames captured during the procedure. The evaluation has been performed in terms of conventional classification metrics, complemented with two new metrics, specifically defined for this problem. Those new metrics provide a more appropriate assessment of the quality of the results, given the class imbalance in the dataset. From an analysis of the evaluation results, it can be concluded that the method provides appropriate segmentation results for this dataset.

Predicting TAVI paravalvular regurgitation outcomes based on numerical simulation of the aortic annulus eccentricity and perivalvular areas.

Trans-catheter aortic valve implantation (TAVI) is an increasingly adopted technique which provides a minimal invasive solution for patients who suffer from severe aortic stenosis. Some complications of the procedure could be annular rupture or paravalvular leakage, both related with adverse outcome. In TAVI with balloon expandable devices, a mismatch between those two factors leads to a conflict situation, where improving one worsens the other. The presented research proposes a methodology that uses numerical simulation to obtain certain TAVI outcomes related with aortic regurgitation due to paravalvular leakage, such as perivalvular area, aortic eccentricity or annular pressure. The application of the methodology for two patients shows the possibility of predicting those quantities. The highest stress values are distributed along the contact area. Results also show that a great deformation on the aortic annulus does not necessarily imply a higher stress; pressure can either be converted into root reshape or into root stretching. Validation of the results was done using scientific publications, clinical guidelines and clinical reports. Numerical simulation provides a suitable tool that could possibly contribute to optimize the planification procedure adjusting the mismatch between size and pressure.

A biomechanical model of the pathological aortic valve: simulation of aortic stenosis.

Aortic stenosis (AS) disease is a narrowing of the aortic valve (AV) opening which reduces blood flow from the heart causing several health complications. Although a lot of work has been done in AV simulations, most of the efforts have been conducted regarding healthy valves. In this article, a new three-dimensional patient-specific biomechanical model of the valve, based on a parametric formulation of the stenosis that permits the simulation of different degrees of pathology, is presented. The formulation is based on a double approach: the first one is done from the geometric point of view, reducing the effective ejection area of the AV by joining leaflets using a zipper effect to sew them; the second one, in terms of functionality, is based on the modification of AV tissue properties due to the effect of calcifications. Both healthy and stenotic valves were created using patient-specific data and results of the numerical simulation of the valve function are provided. Analysis of the results shows a variation in the first principal stress, geometric orifice area, and blood velocity which were validated against clinical data. Thus, the possibility to create a pipeline which allows the integration of patient-specific data from echocardiographic images and iFR studies to perform finite elements analysis is proved.