UNC Health Talk

Imagining a New Kind of Image

Imagining a New Kind of Image

Graduate student Sarah Shelton reinvented herself from a dancer to a scientist. Now, she’s creating a new ultrasound technique to improve cancer diagnostics.

When Sarah Shelton was a kid, she saw a performance featuring flamenco music and dance. Entranced by the performance, Shelton wanted to pursue flamenco but couldn’t find classes until she was a freshman at UNC. Finally, she took up flamenco, and after graduating from UNC she joined an intensive training program in Spain to become a professional dancer and teacher.

Shelton excelled to the point where she could give inspired performances and teach the complex dance to others.

Flamenco is not like other types of dancing. There’s a complex interplay between the singer, guitarist, and the dancer. They feed off each other. Each one improvises and each one responds. Shelton excelled to the point where she could give inspired performances and teach the complex dance to others.

But being a professional flamenco dancer is no waltz in the park. It’s a tough way to make a living, and she missed science.

When she returned to the United States, she reimagined her future, deciding to earn a master’s degree in environmental engineering. By chance, though, the environment she wound up studying was the human body – in particular, the blood vessels that feed cancerous tumors.

Now, she’s pursuing her doctorate in biomedical engineering, a joint department between the UNC School of Medicine and the NC State College of Engineering. She’s part of a team conducting a clinical trial to see if a new kind of ultrasound can provide doctors with valuable information they could use in their diagnoses of patients. One day, the method could become the standard method of diagnosis and might even help patients avoid months of unhelpful and toxic treatment.

We sat down with Sarah for a student profile to discuss her path to UNC, her research project, and the science behind acoustic angiography.

Name: Sarah Shelton

Birthdate: Feb. 2, 1984

Hometown: Cary, North Carolina

Education: B.S. in environmental science / UNC; master’s in environmental science and engineering / UNC Gillings School of Global Public Health

Dissertation: Using acoustic angiography ultrasound to reveal the vascular fingerprint of malignant tumors

Mentor: Paul A. Dayton, PhD

Awards: Integrative vascular biology training program (NIH T-32), outstanding presentation at IVB symposium, IEEE student travel award, NCSU Dean’s Doctoral Fellowship, NCSU Graduate Merit Award

Extracurriculars: Flamenco dancing, cooking, hiking

Why biomedical science?

“When I started at UNC as an undergrad, the biomedical engineering department didn’t exist, and back then I was interested in environmental science from the perspective of public health. So I majored in that. Then I went off and became a flamenco dancer, but I missed the science so I came back and got my graduate degree [at the UNC Gillings School of Global Public Health].

“When I was in Spain, I missed the creative and technical side of research – the challenging part of science. So I thought about working in industry and went to graduate school, where I did ground water modeling and some mathematical modeling. But my thesis focused on creating a model of tumor growth.”

How did that happen?

“It just kind of happened. My advisor was Bill Gray, whose work is all about multiscale mathematical modeling of porous media. At some point, he showed me a paper from some other lab that applied porous media modeling to biological systems. He asked what I thought about it. We discussed it. I had to look up all these biology terms to understand it. And after we discussed it a few more times, we decided that we could do it, and maybe we could do it better – the idea was that we could devise a model to show how tumors grow using the framework established for multiscale porous media modeling. I really liked this idea. This became my project.

“I liked the biology side of it. I liked learning about cancer and really wanted to do cancer research – to understand how tumors grow and how that biology works.”

“Turned out that this was really interesting to me. I liked the biology side of it. I liked learning about cancer and really wanted to do cancer research – to understand how tumors grow and how that biology works.”

Why UNC?

“I did my undergraduate and graduate work here, and so I actually tried to get away to do my PhD somewhere else, but UNC keeps bringing me back. It just felt right. It wasn’t a plan. I looked at other places, but it seemed that I needed to be here and I’m glad I’m here. The fact that it’s close to home had nothing to do with it. I was packed and ready to move across the country and just didn’t.

“I interviewed elsewhere but this was the best fit for what I wanted to pursue.”

Why the Dayton lab?

“During my last semester at the school of public health, I was thinking about wanting to do cancer research, so I Googled “if I want to study mathematical models and biology, what field is it?” And I got all these hits for biomedical engineering.

“There’s a lot of data to get through in biomedical engineering but also experiments that have to be done. There’s a little bit of everything instead of doing just one thing all the time. You have to put a bunch of different things together to make a whole picture.

“I didn’t know where exactly I wanted to get my PhD, but after talking to Paul [Dayton] and his students, his lab seemed perfect for me. They do imaging, which seemed like the natural way to understand what’s going on during tumor growth. I didn’t know anything about ultrasound or imaging, but I thought his research was really interesting. I got into it and I loved it.

“Also, I remember talking to Paul about wanting to work on things that were translational, things relevant to health and disease in people.

“He helped me figure out how to shadow people in the hospital. And now we’re starting to work on a clinical trial. He’s been really supportive and has helped me get to the point where I’m achieving the things I wanted to.”

What is your research project?

“Paul’s whole lab does ultrasound research. The overall goal is to make existing ultrasound technology better and use it for something relevant to disease. I’m focusing on cancer angiogenesis – the abnormal process of blood vessel growth in tumors.

“Malignant tumors have these twisted blood vessels. Benign tumors are not like that. So we use a contrast agent that harmlessly flows through the blood vessels so that we can use ultrasound to image the vasculature of tumors.

“We have a unique approach for imaging these microbubbles which is called acoustic angiography.”

“The contrast agents are microbubbles, which are micron-sized spheres just slightly smaller than red blood cells. We inject the microbubbles intravenously, and, they travel through the vasculature. The microbubbles reflect ultrasound so that we can get an image of blood vessels. We have a unique approach for imaging these microbubbles which is called acoustic angiography.”

How are the bubbles imaged, exactly?

“We use the frequency of the ultrasound. A low frequency excites the bubbles, and then we use a high frequency transducer to receive signals that come back from the excited microbubbles. This way, we don’t see any tissue; we isolate the signal from the microbubbles so that we’re only getting images of the microbubbles. And because the microbubbles are only in the vasculature, by proxy, we are only imaging the vasculature.

“The high frequency gives us high-resolution images. This allows us to see smaller structures. The frequency separation also lets us reject background signal from structures other than microbubbles. This is important because there’s actually a lot of stuff in the background – tissue, fat, other parts of the tumor. But in our images of vasculature, the background is black.”

How could this help patients?

“We hope it will add valuable information to the diagnostic process. For example, to diagnose breast cancer, people first get a mammogram.”

“We hope it will add valuable information to the diagnostic process. For example, to diagnose breast cancer, people first get a mammogram. If something looks suspicious, they go back in for a second look – often with ultrasound imaging. What doctors actually look at is a gray-scale image. There are clinical grades – if the mass looks like this, then it looks like cancer. There are metrics to determine if it’s a benign or malignant mass.

“Our hope is that our acoustic angiography method can do a better job at distinguishing what is and is not cancer. Our method might be able to help patients avoid getting biopsied. That’s long term. We would need many more trials and more sophisticated transducers – the equipment that picks up the signals so that we can image the vasculature.”

Is this in clinical trials?

“The trial just got underway. We’re adding an ultrasound before people get biopsied. So we’ll have our images and we’ll have biopsy results to see what we were actually looking at. Then we’ll do a study to see whether the radiologists think our images add information that they’re not getting from the other ultrasound images, regarding vascular structure.

“In animal models, we’ve shown that if you look at our acoustic angiography images of a tumor, you don’t have to know anything about imaging or ultrasound or even tumor biology to notice that cancerous tumors are very different from healthy tissue.

“If someone simply tells you that cancer makes twisted blood vessels, then you could look at two images and say, ‘this one is cancer and this one is not.’”

How else could acoustic angiography be used?

“Down the road, we could possibly track response to treatment. We could see if the tumor vasculature is changing in response to therapy. We hypothesize that the vasculature may indicate whether someone is responding to treatment. We think this vascular change would appear before the tumor would actually shrink. Right now, that’s how a treatment is judged – if a tumor is smaller after a certain amount of treatment.

“If we could produce an earlier indication of whether a treatment is working, then we could potentially know earlier whether doctors should switch treatment. This, obviously, is way down the road. But it’s something we’re thinking about now.”

Your future?

“I definitely want to stay in research and stay in academia to become a PI. And I want to continue to study cancer angiogenesis. I will probably continue to research imaging modalities, possibly including ultrasound. I have a few years to figure it out before I graduate. But I love what I’m doing.”