George Mason University
George Mason University Mason
George Mason University

Predicting Aneurysm Rupture

March 7, 2018   /   by Martha Bushong

Detmer's research analyzes the difference between low and high risk aneurysms using statistical models and brain imaging.

Felicitas Detmer says that when she meets new people, they often ask about her research topic. When she says it’s about brain aneurysms, they often tell her something like “my friend had an aneurysm” or once somebody said to her, “you know, my father died of an aneurysm.” These comments show how common aneurysms are.

Detmer describes a brain aneurysm as a dilation of a brain artery “that looks like to a berry-shaped outpouching.” She says, “Aneurysms occur quite frequently–– overall two to five percent of the population have aneurysms–– but usually they don’t cause any symptoms. If an aneurysm ruptures, or bursts and starts bleeding, it can cause a stroke, which often has fatal consequences.”

There are different treatment options to prevent unruptured aneurysms from bleeding. But the risk associated with these treatment options and their complications is much higher than the natural aneurysm rupture risk. That is why when an aneurysm is diagnosed, it can be quite challenging to decide whether to treat it.

“My research involves developing a statistical model which can to identify those aneurysms that are likely to rupture in the future and thus require treatment,” she says. “To do so, I used patient and image data of about 2,000 aneurysms. As input parameters for my model, I consider patient information, parameters that describe the shape of the aneurysm, and blood flow information.”

Detmer uses blood flow information because she knows that the blood flow acts on the vessel wall, which can be sensed by the cells in the vessel wall and by that influences the mechanisms that are involved in aneurysm growth and rupture. Therefore, the research simulates the blood flow in the brain using computational models and once all the simulations are done, she uses the results to train her statistical model.

“So far, I have developed a model which is able to identify most of the ruptured aneurysms in our database,” she says. “Particularly we found that high-risk aneurysms have a more complex shape and are exposed to a higher force acting on the vessel wall.”

“She is constructing statistical models based on flow and geometric characteristics of the aneurysms in addition to anatomical and patient information that is typically used to evaluate rupture risk,” says her advisor Juan Cebral, a professor in the Department of Bioengineering. “These models have the potential of improving diagnosis by a more objective and quantitative evidence data-based decisions.”

Detmer’s hope is that in the future this model can be applied in clinical practice to identify high-risk aneurysms, treat them and by that prevent these patients from suffering a stroke, while at the same time not exposing patients with low-risk aneurysms to the unnecessary risk of treatment.

“My research involves developing a statistical model which can to identify those aneurysms that are likely to rupture in the future and thus require treatment."

Felicitas Detmer