The goal of the LEAD project Aortic Dissection is to develop computational tools and advanced algorithms on the basis of noninvasive medical imaging, and to quantify the underlying cardiovascular mechanics in patient-specific anatomical and fluid-structure interaction (FSI) models. That computational framework is capable of investigating wall stresses, the hemodynamics, false lumen propagation, exchange of blood between true and false lumina, thrombus formation and growth, basically at any stage of the disease. This helps to better understand the mechanobiological events and to finally assist clinicians with the diagnosis, treatment and management of aortic dissection patients.
Phase I of the LEAD project Aortic Dissection started in January 2018 with one aim to strengthen the cooperation and scientific exchange between 10 different institutes and 5 faculties at Graz University of Technology, and to broaden national and international visibility. We have developed and implemented advanced models to better understand events that occur during and after aortic dissections following an ambitious research program within basic science. Phase I of the LEAD project included four work packages (WPs): WP 1 – Characterization and Constitutive Modeling of Dissected Tissues; WP 2 – Failure Criterion, Initiation, and Propagation; WP 3 – Modeling of Blood Flow and Thrombus Growth; WP 4 – FSI Simulation of Aortic Dissection. Currently there are 10 PhD students and 1 postdoctoral researcher involved with different scientific backgrounds, from 7 different nations, and hence from different social and cultural backgrounds. A common office space for all PhD students has led to a very close research collaboration among the PhD students across different disciplines, now visible in many joint publications. Here some research results of the four WPs are outlined.
WP 1: Characterization and Constitutive Modeling of Dissected Tissues.
The overall results of WP 1 can be divided into computational and experimental results, which are equally important for the success of this WP. The computational results are based on the development of constitutive models capturing pathological changes in the dissected aorta including disease-dependent parameters. We identified the degradation of elastic fibers and the apoptosis and dysfunction of smooth muscle cells as key factors resulting in a dissection. On that basis we developed a validated constitutive model of the layered aorta to determine the stress distribution in a patient-data motivated geometry of a dissected aorta under physiological boundary conditions. The experimental results clearly demonstrated that the failure stress of dissected tissues is significantly lower under different loading conditions in comparison to control tissue. In addition, dissected tissues haven often shown a distinct softening behavior, which can be explained by the altered elastin architecture.
WP 2: Failure Criterion, Initiation, and Propagation.
Based on failure experiments of dissected tissues, an anisotropic failure criterion for soft biological tissues has been developed. The initiation of rupture, starting from an intimal tear into which blood can entry and form a false lumen, was simulated considering for the significant interaction of the fluid (blood) and the structure (aortic tissue). In patient-specific geometries of fully established aortic dissections, coupled models and simulations of the fluid-structure interaction (FSI) have been conducted, giving important information on the flow fields and tissue deformations. Impedance cardiography was used to represent the variation of the blood-flow-induced conductivity changes in different stages of an aortic dissection in terms of changes in the amplitude of the obtained signals. Sensitivity analyses helped to reduce model uncertainties and improved the understanding and quantification of the relative importance of modeling aspects and parameter choices.
WP 3: Modeling of Blood Flow and Thrombus Growth.
This WP modeled blood as a non-Newtonian liquid and developed a viscosity model with shear thinning and yield stress. Two thrombus growth models were formulated – based on hemodynamics and on the theory of porous media, relying on different approaches for transport-based models. A generalized Reynolds number approach allows the flow to be efficiently simulated as Newtonian. The hemodynamic and hematocrit influences on thrombus growth for varying false-lumen geometry were quantified. The modeling of thrombus growth showed that high hematocrit hinders thrombus formation and growth. Parameters with strong influence on thrombus formation and growth were identified, allowing for model reduction. The sensitivity of the model outcome to input parameters was quantified, showing the need for nonlinear material modeling.
WP 4: FSI Simulation of Aortic Dissection.
This WP developed stable space-time finite elements for incompressible flow problems and appropriate domain decomposition methods for an efficient parallel solution of the overall system. For the use of computational hemodynamics we developed a new stabilization method for non-Newtonian fluid problems, which removes common limitations of traditional low-order finite element solvers. As a byproduct, this resulted also in a new method for accurately estimating local blood pressure from measured flow velocity data. Together with uncertainty quantification, this allowed clear hemodynamics modeling recommendations from multi-sensor data. Methods based on deep neural networks were also established not only to extract patient-specific information from medical images, but also as part of a framework to solve high-dimensional stochastic PDEs for near real-time uncertainty quantification.
We have now published more than 60 peer-reviewed papers, and additional 23 are currently under review. We have produced 40 completed Bachelor and Master Theses, while 7 are ongoing. In addition, we have provided 67 invited talks, received 10 awards and applied for 13 follow-up projects (7 are under review). Within the LEAD project Aortic Dissection, we have organized two international project workshops outside of Graz, three days each, and we have arranged an International Advisory Board consisting of 8 renowned professors.
To build a talent base for future research projects and for Phase II of the LEAD project we tried to train our PhD students, Master and Bachelor students comprehensively and individually. Apart from the two international workshops, we held a Summer School in Graz with six internationally recognized lecturers over one week; another Summer School will take place from August 30 to September 3, 2021. We followed incoming and outgoing mobility programs, and have invited recognized researchers from abroad to teach in form of courses, seminars or lectures – from incoming guests, who stayed short-terms up to several months and cooperated intensively with us. Outgoing PhD students could experience various academic cultures, and strengthen international and inter-institutional collaborators within the LEAD project.
As a visible result of the joint activities, Aortic Dissection at Graz University of Technology has grown to an internationally recognized research center for modeling and simulation of aortic dissection, a serious condition which may lead to aortic rupture or decreased blood flow to organs. Over the past three years, the LEAD consortium has endeavored to gain a better understanding of this disease on the one hand, but on the other hand to open up a new perspective that will enable more accurate predictions with engineering tools.
For related publications of the LEAD Project click here.