A stroke of genius

3 min read

University of Houston engineers are working with neurology specialists on new technology to help identify which brain aneurysms are at highest risk of rupture and could cause a stroke.

Cerebral aneurysms are ballooning weak spots in the wall of a blood vessel in the brain. The goal of the study is to develop a fully integrated computational medical tool that will be useful in helping to select patients for treatment whose aneurysms are most likely to rupture.

Ralph Metcalfe, a mechanical engineering professor at UH and deputy director of the UH biomedical engineering programme, and his graduate student, Aishwarya Mantha, are working on this project with a Methodist Neurological Institute team consisting of interventional neuroradiologists Drs. Charles Strother and Goetz Benndorf, and Christof Karmonik, a researcher.

Using computer simulations of blood flow in realistic geometric models of aneurysms, some blood flow characteristics have been identified that may contribute to aneurysm formation.

Metcalfe said, "Most aneurysms don't rupture, but if they do, the results are fatal in about 50 percent of the cases. The question is how to predict who is most at risk."

Since treatment of aneurysms is associated with some risk, Metcalfe's group and his Methodist colleagues are trying to develop a better method of identifying which aneurysms are most vulnerable for rupture. Once these patients are identified, physicians can then determine the best course of medical treatment.

"One of the key points is that aneurysms don't seem to form randomly," Metcalfe said. "They do seem to form at locations that are associated with the fluctuations in the flow of blood, leading to the question of what it is about the flow of blood that tends to correlate with the formation of aneurysms."

The Methodist researchers acquire 3D images of the intracranial vascular system by injecting dye into the vessels and rotating an X-ray tube around the patient's head, a technique that has become a standard for high-quality vascular imaging in this institution.

Using this geometric and blood flow data taken from a specific patient's clinical profile, Metcalfe's team can perform computer simulations of blood flow in that patient's arteries using existing computational fluid dynamics models.

"We can't look at a person and tell the likelihood that an aneurysm will rupture," Strother said. "But we do know that force and stresses created by blood flow produces aneurysms. Our hope is that this study will help us learn enough to predict which ones are at high risk of rupture so that treatment can be offered before they become harmful."

This work has two potential applications. The first is as a research tool, with Metcalfe's team performing simulations of specific aneurysms. Using a technique employed by Karmonik to simulate removal of an aneurysm on the computer, they analyse how the blood behaves as it flows near the aneurysm site and determine if that can be correlated to a certain type of behaviour of the blood at potential sites where aneurysms form. Very accurate simulations are done for a complete description of the flow fields, studying all the fluid dynamic variables in great detail, such as the wall shear stresses, the pressures and the velocity.

"The second application is as a potential clinical tool," Metcalfe said. "Once we have a reasonable idea of the fluid dynamic variables needed to study and identify a potential problem, we then use a program that provides a detailed, 3D description of the aneurysms of the real patients."

Benndorf added that the potential clinical importance of these computer simulations lies in the future possibility of directly predicting blood flow so that patient-specific medical devices can be used in aneurysm treatment. He is studying how stents can be tailored to the patient's individual anatomy and blood flow in order to optimise their therapeutic effect and maximise the possibility of a successful outcome.

When Metcalfe's group imports a patient's images into a computer program, they remove some geometric glitches and generate a computational map that involves the charting of hundreds of thousands of tiny elements that represent the area being studied. That map is then entered into a program that calculates the fluid dynamics.

"It takes a lot of computer time to perform these simulations," Metcalfe said. "There are several hundred thousand elements that are discrete zones within a geometric mesh, and then there are 700 steps representing intervals of time over the cycle of each heart beat."

"The critical step here is to make these complicated flows much more accessible to people like medical researchers and physicians," Metcalfe said. "We're developing 3D visualisations so doctors can go inside the virtual artery and actually see what's happening as the blood cells flow through."