Search

It was recently submitted that the rupture risk of an ascending thoracic aortic aneurysm (ATAA) is strongly correlated with the aortic stiffness. To validate this assumption, we propose a non-invasive inverse method to identify the patient-specific local extensional stiffness of aortic walls based on gated CT scans. Using these images, the local strain distribution is reconstructed throughout the cardiac cycle. Subsequently, obtained strains are related to tensions, through local equilibrium equations, to estimate the local extensional stiffness at every position. We apply the approach on 11 patients who underwent a gated CT scan before surgical ATAA repair and whose ATAA tissue was tested after the surgical procedure to estimate the rupture risk criterion. We find a very good correlation between the rupture risk criterion and the local extensional stiffness. Finally it is shown that patients can be separated in two groups: a group of stiff and brittle ATAA with a rupture risk criterion above 0.9, and a group of relatively compliant ATAA with a rupture risk below 0.9. Although these results need to be repeated on larger cohorts to impact the clinical practice, they support the paradigm that local aortic stiffness is an important determinant of ATAA rupture risk.

We recently developed an approach to characterize local nonlinear, anisotropic mechanical properties of murine arteries by combining biaxial extension-distension testing, panoramic digital image correlation, and an inverse method based on the principle of virtual power. This experimental-computational approach was illustrated for the normal murine abdominal aorta assuming uniform wall thickness. Here, however, we extend our prior approach by adding an optical coherence tomography (OCT) imaging system that permits local reconstructions of wall thickness. This multimodality approach is then used to characterize spatial variations of material and structural properties in ascending thoracic aortic aneurysms (aTAA) from two genetically modified mouse models (fibrillin-1 and fibulin-4 deficient) and to compare them with those from angiotensin II-infused apolipoprotein E-deficient and wild-type control ascending aortas. Local values of stored elastic energy and biaxial material stiffness, computed from spatial distributions of the best fit material parameters, varied significantly with circumferential position (inner vs. outer curvature, ventral vs. dorsal sides) across genotypes and treatments. Importantly, these data reveal an inverse relationship between material stiffness and wall thickness that underlies a general linear relationship between stiffness and wall stress across aTAAs. OCT images also revealed sites of advanced medial degeneration, which were captured by the inverse material characterization. Quantification of histological data further provided high-resolution local correlations among multiple mechanical metrics and wall microstructure. This is the first time that such structural defects and local properties have been characterized mechanically, which can better inform computational models of aortopathy that seek to predict where dissection or rupture may initiate.

It is now a rather common approach to perform patient-specific stress analyses of arterial walls using finite-element models reconstructed from gated medical images. However, this requires to compute for every Gauss point the deformation gradient between the current configuration and a stress-free reference configuration. It is technically difficult to define such a reference configuration, and there is actually no guarantee that a stress-free configuration is physically attainable due to the presence of internal stresses in unloaded soft tissues. An alternative framework was proposed by Bellini et al. (Ann Biomed Eng 42(3):488-502, 2014). It consists of computing the deformation gradients between the current configuration and a prestressed reference configuration. We present here the first finite-element results based on this concept using the Abaqus software. The reference configuration is set arbitrarily to the in vivo average geometry of the artery, which is obtained from gated medical images and is assumed to be mechanobiologically homeostatic. For every Gauss point, the stress is split additively into the contributions of each individual load-bearing constituent of the tissue, namely elastin, collagen, smooth muscle cells. Each constituent is assigned an independent prestretch in the reference configuration, named the deposition stretch. The outstanding advantage of the present approach is that it simultaneously computes the in situ stresses existing in the reference configuration and predicts the residual stresses that occur after removing the different loadings applied onto the artery (pressure and axial load). As a proof of concept, we applied it on an ideal thick-wall cylinder and showed that the obtained results were consistent with corresponding experimental and analytical results of the well-known literature. In addition, we developed a patient-specific model of a human ascending thoracic aneurysmal aorta and demonstrated the utility in predicting the wall stress distribution in vivo under the effects of physiological pressure. Finally, we simulated the whole process preceding traditional in vitro uniaxial tensile testing of arteries, including excision from the body, radial cutting, flattening and subsequent tensile loading, showing how this process may impact the final mechanical properties derived from these in vitro tests.