True Subject-Specific Computational Models of the Human Tricuspid Valve

Abstract

Objective: 1.6 million Americans currently suffer from clinically significant tricuspid valve regurgitation. A first critical step toward successfully treating these patients is to fully understand and characterize the complex mechanics of the valve. This goal has eluded us thus far for historic and technical reasons. Computer models may be critical tools in this endeavor. However, current models are based either on averaged valve geometries and properties, or worse yet, on fictitious geometries and properties. The objective of our work is to overcome these shortcomings of the past and build the first and fully subject-specific computer models of eight tricuspid valves. Methods: In beating, healthy, human hearts, maintained after explantation from donors in an organ preservation system, we recorded the dynamics of the tricuspid annulus using sonomicrometry crystals, in addition to hemodynamic data on a loaded right ventricle. Upon arresting the hearts, we excised the tricuspid valve complex, imaged the leaflets and digitized their geometry. We then recreated the 3D annulus onto which we non-rigidly transformed the leaflet shapes. Subsequently, we used imaging data to assign chordal insertions and rebuild the valve in our computer model. Additionally, we characterized material properties and microstructural information of the leaflets and chordae through in-vitro planar biaxial and uniaxial testing, and 2-Photon microscopy, respectively. Finally, we informed our model with annular displacements and transvalvular pressure gradients measured ex-vivo and simulated the dynamics of the valve over the cardiac cycle in Abaqus/Explicit. Results: The valve models faithfully capture leaflet coaptation, as validated against echo scans of the same beating hearts. We observe peak leaflet stresses in the belly regions and dynamic strains consistent with our previous in-vivo measurements in sheep. Conclusions: Utilizing these high-fidelity computer models, we may virtually implant repair devices on the tricuspid valve and optimize surgical procedures at a patient-specific level.