Oct. 2 — Abdominal aortic aneurysm (AAA) is a vascular disease that affects tens of thousands of people around the world each year. The disease occurs most commonly in men over 60, and is characterized by a dilation of the abdominal aortic wall and often a persistent blood clot.

The aorta is the largest artery in the body; it allows oxygenated blood to be distributed to organs and tissues throughout the body. Aneurysms, or localized dilatations, can form in the ascending or descending portions of the aorta within the chest or within the descending portion in the abdomen. Abdominal aneurysms are more common and when they rupture, they are often lethal.

Because aneurysms are stressed by the continuous action of changing blood pressures, methods of engineering that were developed to study the flow of viscous fluids (e.g., blood) and the failure of structures (rupture of the aorta) can be used to help understand the complex history of these deadly lesions. Once an AAA ruptures, death is nearly inevitable due to extreme internal bleeding, signifying the critical need to predict which aneurysms are most vulnerable.

To understand why some AAAs rupture and others do not, researchers must first understand better how they form and enlarge, digging deep into both the micro and macro levels to understand the complex interactions that define the disease.

One group attempting to accomplish this task is the Continuum Biomechanics Lab in the Department of Biomedical Engineering at Yale University. Using computational models in combination with biological experiments, researchers in the lab can study AAA development at both the micro and macro scales.

Jay Humphrey, professor of biomedical engineering, emphasizes that “the biological and mechanical complexity of AAAs demands a multidisciplinary team to advance our understanding.” George Tellides, professor of surgery and Andrew Sherman at the Yale Center for Research Computing, teamed up with Humphrey to enable the development of patient-specific computational models for multiple aspects of AAAs. They, in combination with post-doctoral fellows and Ph.D. students have recently put forth new hypotheses on factors that drive intraluminal, or interior areas of the artery, clotting as well as the mechanobiological stability of the diseased aortas.

To protect the identity of patients, associated medical images are de-identified and provided to the research team for analysis.

According to Paolo Di Achille, a researcher in the group, “the medical images we have can be processed with semi-automatic algorithms to reconstruct the geometry of the vasculature. Our goal is to get accurate representations of the blood flow throughout the cardiac cycle. Then we look more specifically at the interactions between blood flow and both the cells that line the vessel wall and the cells within the blood that are responsible for clotting.”

To explore the hemodynamics in AAAs, the fluid dynamics of blood flows, the researchers rely on a mix of computational resources from Yale and the Extreme Science and Engineering Discovery Environment (XSEDE), which is funded by the National Science Foundation.

“Access to XSEDE resources is a key aspect of our research. The code we started working with was already equipped to work on supercomputers, but having access to high performance computers enables us to explore additional parallel capabilities,” said Di Achille.

To take full advantage of XSEDE’s capabilities, Di Achille and Humphrey applied to XSEDE’s Extended Collaboration and Support Service (ECSS), a program that pairs researchers with expert staff members in advanced cyberinfrastructure. The researchers at Yale partnered with researcher Yifeng Cui at the San Diego Supercomputing Center (SDSC) to optimize their code and improve simulations. Although the researchers’ ECSS allocation expired in August, they found the program very helpful to adapting their new codes to a supercomputing environment.

“I would say that’s the biggest advantage of ECSS, talking with experts to get real insight into how our codes can be improved. It’s the best way for us to optimize our codes, because we don’t have time to read all the literature on these new architectures,” said Di Achille.

Working with Cui, the team ran their code on TACC’s most powerful supercomputer, Stampede, and used its visualization nodes to create realistic depictions of AAAs. The team also used the supercomputer Trestles, which was recently transferred from SDSC to the Arkansas High Performance Computing Center.

The researcher’s sophisticated models allowed them to investigate another major issue in AAA — thrombus formation.

Thrombus, or a blood clot, is estimated to develop in three-fourths of all AAAs. Its presence usually indicates a more dangerous form of the disease. Platelets are blood-borne cells that control the formation and dissolution of thrombus.

“Normally blood clots form when we have wounds or some laceration of the vessel wall,” said Di Achille. “For some reason, at some point of AAA evolution, blood clots form and we want to find out why.”

The group’s study of how thrombus arise in AAAs, their progression, and why certain aneurysms develop them and others don’t, led to a paper published in late 2014 in the Proceedings of Royal Society. Di Achille describes the group’s most interesting finding as a possible explanation for the unique biomechanical factors that lead to thrombus formation.

“The enlargement of the vessel in this disease seems to create a unique fluid dynamic situation that both activates and localizes platelets within AAAs,” said Di Achille.

To advance their research, the team is using this information to develop more comprehensive models to track the progression of the disease. Ultimately, it is hoped that this research will not only increase our basic understanding of AAAs, but it will also help clinicians make more informed decisions when treating patients.

Said Di Achille: “Clinical decisions are extremely complex. If our modeling results could help in this decision process and improve outcomes, that would be very satisfying.”

—

Source: Makeda Easter, TACC