October 13, 2011

Thomas Hughes in Vis Lab
In the Visualization Lab, Dr. Hughes and his research assistants can view large scale models of their work, such as the coronary arteries on the screen behind him that have been modeled using data from Dr. Hughes' own arteries. The two three-dimensional models Dr. Hughes holds are modeled from his computations as well: the white one is of a thoracic aorta modeled using the data from the aorta of one of his graduate students and the red one is an abdominal aorta modeled after Dr. Hughes' own aorta. Photo courtesy of Pratt Institute

For one professor in the Department of Aerospace Engineering and Engineering Mechanics, spending time crunching numbers is leading to technologies that could save lives. Dr. Thomas Hughes and his colleagues have pioneered patient-specific 3-D models of blood flow through the heart and blood vessels that could help guide best practices for cardiologists.

Rather than relying on earlier computer models — where simple two-dimensional geometry shared little resemblance to actual anatomy — medical doctors can now use the work of Hughes to better understand how various medical interventions in the heart and vessels affect blood flow. As a result, crucial information can be provided about the safety and effectiveness of commonly used devices like stents, angioplasties and bypass grafts.

"What we introduced in the mid-1990s was a new paradigm of modeling using different computer science technologies," Hughes said. These technologies included familiar diagnostic tests like ultrasounds, CT scans and MRIs, but the revolutionary part of Hughes' research was to extract the data gathered from medical imaging in the form of a DICOM file.

By extracting the information, medical images could then be manipulated in computer programs to build 3-D volumetric models of a patient's circulatory system, organs, blood vessels and bloodstream. From this model, blood flow can be computed, including the direction of flow and the friction it creates on the surface of the vessel walls, called "shear stress."

The impact of this blood flow modeling helps determine best practices for cardiologists who plan life-saving interventions.

"Let's say you have a left ventricle that's not functioning properly," Hughes proposed in a hypothetical example of how his method could be applied. "It's not pumping enough blood into the aorta and you want to put in a device that will aid in pumping." That device — a miniature pump — is called a left ventricular assist device (LVAD). After a surgeon bores a hole in the left ventricle to insert the LVAD, he attaches its tube to the aorta, either on the descending branch of the aorta, where it's most commonly placed, or on the ascending branch. However, the difference between these options is crucial.

"The flow pattern is very different," Hughes said. "What you want to see is a physiological shear rate, a clean flow that's more or less unidirectional and signals to the cells to line up like shingles along the blood vessel wall."

Just as rain flows off the shingles of a roof, blood should flow past the shingles of the vessel wall cleanly.

"If the blood flow is going back and forth in a chaotic way, the cells don't know what to do and instead create a cobblestone structure that creates interstitial spaces," Hughes said. "That's how lipids get into the wall and form plaques."

So what does this mean for Hughes' 3-D modeling? It shows that common placement of the LVAD tube — on the descending branch of the aorta — creates blood flow that doesn't adequately mimic what a healthy body's flow would do on its own. In fact, a descending aortic placement can lead to variety of complications such as heart attacks, blood clots, stroke and more.

In contrast, Hughes' 3-D models show the best place to insert the LVAD tube is on the ascending branch of the aorta; a graft that doesn't have negative side effects.

"If you just look at the overall amount of blood flowing through the branches, it seems to be adequate," he said.

However, what matters is not just the amount of blood flow, but the way the blood is moving through those vessels — something only the computations can reveal.

"The important message here is the three-dimensional details are important," Hughes said. "They really need to be taken into consideration when you're designing an intervention."

Using blood flow analysis techniques to improve the safety and efficacy of cardiovascular surgery is just one application of Hughes' research into the computational methods for understanding solid, structural and fluid mechanics. Another area Hughes is studying is aneurysm growth and adaptation. He wants to know what mechanisms of aneurysm creation in the body help determine aneurysms that may be prone to failure.

This research is one reason Hughes was recently elected to the Royal Society, the oldest known scientific society in the world, whose members have included Isaac Newton, Charles Darwin and Albert Einstein.

Hughes develops his models working with physicians and mechanical engineers through the interdisciplinary Institute for Computational Engineering and Sciences at the university.