Interview with Dr. Stacey Finley

Dr. Stacey Finley is a pioneering researcher and professor in the field of biomedical engineering at the University of Southern California. She is widely recognized for her work in computational modeling of cancer, using advanced systems biology techniques to better understand and treat disease. Dr. Finley holds the prestigious title of Nichole A. and Thuan Q. Pham Professor and serves as the Director of USC's Center for Computational Modeling of Cancer. Her work has earned numerous honors, including recognition as an outstanding young innovator and a National Science Foundation CAREER Award. Dr. Finley’s research focuses on unraveling the complex biological processes underlying cancer, with the ultimate goal of improving treatment strategies. She is also deeply committed to mentoring the next generation of scientists and engineers.

Isaac

Hi, Dr. Finley. Your background is in chemical engineering. During your early career, what inspired you to shift towards biomedical engineering and cancer research?

Dr. Finley

Yes. So my undergraduate degree as well as my PhD are both in chemical engineering. In that context, I studied a lot of chemical reactions and looked at the dynamics of those reactions—how fast they occur, and how the levels of species involved are changing. I got really excited about building models of chemical reactions, even as an undergraduate student.

As I went into my PhD, I wanted to apply that knowledge in the context of human health and disease. In my PhD, I also worked on metabolic reactions, mainly to see how we might predict new reactions that could occur to break down toxic compounds—getting a little bit closer to human health and the relevance of that research.

Then I decided to do a postdoctoral fellowship. It was during that transition that I said, “Well, let me look specifically for opportunities that are focused on health and disease.” I stumbled across a fantastic research group at Johns Hopkins that was using mathematical modeling to study tumor angiogenesis—the formation of new blood vessels inside tumors. I learned that the same kinds of models I was using to study chemical reactions could be applied to study biochemical reactions and reactions happening in biology. So that’s really where the transition from chemical engineering to more biomedical engineering occurred.

Isaac

What were some of the more challenging aspects of moving from engineering to cancer research?

Dr. Finley

I think the biggest difficulty in moving from chemical engineering to cancer research is just not knowing the biology. Cancer is such a complex and heterogeneous disease—even within one type of cancer, every patient’s tumor is very different. So first, just trying to grasp the complexity of the disease, and then learning about the proteins, signaling species, and transcription factors that are important in driving tumor growth and influencing the response to treatment. I think the biggest hurdle is comprehending the complexity and learning about the biology.

Isaac

In my research with Dr. Rockne, he emphasized that it’s really important to take the most important elements of a model and sort of simplify the rest—because otherwise, there are too many factors to keep track of. So in your research, how do you simplify very complex models into something that accurately reflects how cancers respond to treatments?

Dr. Finley

I think the approach we take to building a model that is predictive, but still manageable and not too complex, starts with the biological question. Really, the question should drive what components you add to the model and also how complex the model is. You should only be developing a model that is complex enough to answer the biological question you set out to answer. It’s really important to keep that question at the top of your mind so it informs your decisions about how to develop the mathematical model.

As an example, if you're interested in a particular signaling pathway and how that pathway influences the proliferation of cancer cells, then maybe you don’t look at the 20 other pathways also influencing cancer cells. Instead, you might say, “I'm going to focus on this one,” and examine how that pathway behaves under different tumor growth conditions. So it’s still capturing what’s happening, but in a way that’s specific to the question you’re interested in answering.

Isaac

I'd like to ask about a specific factor in cancer treatment—cancer-associated fibroblasts in colon cancer. Could you speak a bit about how those impact the treatment, as well as the model?

Dr. Finley

Yes. One project in the lab is looking at colorectal cancer. We know this is an important type of cancer to study—it’s the second leading cause of cancer-related deaths in men, and the five-year survival rate is pretty low, maybe just around 10%. Colorectal cancer patients often develop resistance to standard therapies, so there’s a lot of motivation for studying it.

This type of tumor has a large stromal compartment. Tumors aren’t just made up of cancer cells; they consist of many different cell types. The stromal compartment is a mix of various cells that also influence tumor growth. One of those cell types is called cancer-associated fibroblasts. They influence the structure of the tumor and how cancer cells respond to treatment.

Cancer-associated fibroblasts also take up nutrients, so they’re competing with cancer cells and other cells in the tumor microenvironment for nutrients and oxygen needed for growth. We've focused on colorectal cancer cells and their interactions with these fibroblasts because there’s evidence in the literature and clinical observations suggesting that fibroblasts interact with colorectal cancer cells—possibly even outcompeting them for nutrients. So we’re studying that kind of metabolic interaction between those two cell types.

Isaac

Does this relate to angiogenesis at all?

Dr. Finley

Not directly. Fibroblasts don’t directly influence angiogenesis, so I don’t think there’s a strong connection there.

Isaac

Thank you. In my research, I’ve been learning about different new treatments, like extracellular vesicles, oncolytic viruses, and CAR T cells. How can computational models be used to improve the effectiveness of these newer therapies?

Dr. Finley

There’s definitely a lot of excitement about CAR-engineered cells—whether it’s T cells or other engineered cells that now express CARs. This is immunotherapy, essentially using the body’s own immune system to target and kill cancer cells.

Modeling can help the field make better progress in a few ways. One is by helping us understand the dynamics—what’s the sequence of events? How fast does a CAR T cell recognize a cancer cell? How quickly is it activated? And how does that activation affect the cancer cell? Understanding those dynamics is very useful and well-suited to mathematical modeling.

Another important aspect is heterogeneity. Not every interaction between a CAR T cell and a cancer cell will result in the cancer cell being killed. There’s a probability associated with that, and it could depend on many different factors in the tumor environment. Modeling can help us ask, “What happens if that probability drops to 40% instead of 100%? How many T cells would you need to kill off the cancer cells in that case?” These kinds of questions—where there’s a lot of complexity or variability—are areas where modeling can provide useful insights.

Another useful area is in studying the distribution of CAR T cells throughout the body. When you administer a population of CAR T cells—say a million cells—not all of them go directly to the tumor. They’re distributed across different tissues. Most will likely end up in the liver, which is a major organ for filtering and removing cells and molecules from the body. So then you can ask, “What does that distribution look like? How fast does it happen? And what’s the effect of having only 10% of the CAR T cells go to the tumor versus 100%?”

So again, wherever there’s uncertainty about the rate or probability of something happening, modeling can help by offering predictions—showing what the possible outcomes could be based on different assumptions. That’s where it can be a very powerful tool.

Isaac

Could you speak a bit about the importance of treating each component as a probability rather than using an average? For example, instead of saying 40% or 10% of cells will go to the tumor site, what’s the value in modeling each individual one as a separate path?

Dr. Finley

Yes, I think when you do that—when you account for individual probabilities instead of just relying on an average—you get a more accurate prediction. Rather than saying, "On average, 40% of the cells will reach the tumor," you’re working with a distribution of outcomes. That allows you to better understand the effect of variability.

It’s not always going to be 100% certain that something will happen. There's a probability that an interaction will occur, or that a cancer cell will be killed. So it's useful to acknowledge that uncertainty and incorporate it into the model. Doing so helps you make predictions that are more sound and rigorous, and that better reflect what's likely to happen in a real, biological system.

Isaac

I was also wondering about biomarkers in cancer treatment. How can computational models help predict which therapies are most likely to work for individual patients?

Dr. Finley

That’s a great question. Take CAR T cells, for example. The CAR is designed to recognize a specific antigen that’s expressed on tumor cells. So a natural biomarker in that case would be whether a patient’s tumor cells express that particular antigen.

If we can profile a cancer patient’s tumor and determine whether it expresses that antigen, then we can decide whether to use a CAR T cell therapy that targets it. So if the cancer cells express antigen X, then yes—we would give the CAR T cells that recognize antigen X.

That’s a pretty straightforward use of biomarkers. Another example would involve looking at the overall composition of the tumor. Some immune cells actually suppress the activity of CAR T cells—like T regulatory cells or other suppressive immune cells. So if we profile a patient’s tumor and find a high concentration of those suppressive cells, that could suggest the CAR T cell therapy might be less effective.

In both cases, biomarkers help us classify which patients are most likely to benefit from a certain therapy, and they can also help predict how effective that therapy might be.

Isaac

One thing I have observed is that science is becoming more interdisciplinary in recent years. How do you bring together engineers, biologists, and mathematicians in your research, and why is that so crucial for advancing cancer treatment?

Dr. Finley

Yes, absolutely. I think it goes back to the point about the complexity of cancer. Just having one viewpoint—like that of a cancer biologist—isn’t going to be sufficient to make real progress in treating cancer patients.

Instead, you need people who approach the system from different perspectives. That’s where interdisciplinary collaboration comes in. The way I see a problem as an engineer may be very different from how a mathematician, a physicist, or a biologist sees it. When you have people coming from these different backgrounds, they can complement each other, and that leads to better, faster progress toward improving outcomes for cancer patients.

So how is this done? For me, it’s often through giving presentations about the kind of modeling work we do to audiences made up of biologists or clinicians. They’ll say, “Maybe you’re missing this part of the tumor microenvironment that’s really important,” or, “Actually, you don’t need to worry too much about that piece.” That kind of feedback—coming from a totally different perspective—is so valuable. One of the best ways to promote interdisciplinary research is for people to regularly share their work outside of their immediate field, where they can gain insights they might not have otherwise considered.

Isaac

Have there been any recent advances in technology—like computing power, data collection, or even AI—that have had an impact on your cancer therapy research?

Dr. Finley

Yes, definitely. One major advance has been the development of single-cell technologies. That’s been incredibly useful over the past decade. We no longer have to treat the T cells inside a tumor as one homogeneous population. Now, we can see that there’s a distribution—some are subtype A, some are subtype B, and others are subtype C. That kind of resolution really helps inform the models we build.

Another big development is the ability to profile a tumor across space. Instead of treating the tumor as a single lump, we can now study how the expression of different antigens or the presence of different cell populations varies in different regions of the tumor. These spatial technologies have been very helpful.

On the computational side, high-performance computing has definitely been a game-changer. And while we don’t use AI heavily in my lab yet, it's becoming increasingly valuable. For example, AI can learn features from clinical images and help predict a patient’s response to treatment—or even generate predictions of what future imaging might show based on earlier scans.

As more data becomes available and AI and machine learning tools continue to improve, I think those technologies will play an even bigger role in pushing the field forward.

Isaac

With all these new tools, what do you hope to see as the next big breakthrough in cancer treatment?

Dr. Finley

That’s a good question. I think the breakthrough the field has been aiming for—and still is—is personalized medicine. Being able to treat a specific patient’s tumor, not just the "average" tumor, is incredibly valuable.

To get there, we’ll need things like single-cell profiling, spatial omics, transcriptomics, and proteomics. In our work, we’re also very interested in metabolism, so I think spatial metabolomics will be really important, too. So yes, personalized medicine—but informed by all these different types of data, and guided by models that can make reliable predictions. That’s the direction I hope we’re heading toward.

Isaac

I’m interested in joining the field of mathematical oncology as I head into college and beyond. What advice would you give about getting started—not just in mathematical oncology, but in science and cancer research in general?

Dr. Finley

I’d say: ask questions and seek out experiences in research labs. Even if you're not directly working on a big project, just sitting in on group meetings and research presentations is valuable. Especially those times when you’re sitting in a talk and feel a little out of your depth—that's where the real learning happens. That discomfort can push you to learn more and grow faster. So immerse yourself, ask questions, and try to be around cancer biology and research as much as you can.

Isaac

As I go to college, what majors or areas of study would you recommend for someone interested in the intersection of oncology and mathematics?

Dr. Finley

I might be a little biased, but I’d recommend something quantitative—like engineering, computational biology, or even physics. Those give you a strong mathematical foundation, which is incredibly helpful when you're looking at biological questions through a quantitative lens.

From my experience—and from what I’ve seen with colleagues and students—it tends to be easier to start with the math and then learn the biology, rather than the other way around. So starting in a quantitative discipline and then exploring which areas of cancer biology interest you could be a great path forward.

Isaac

Again, thank you so much for speaking with me today. One last thing—if there’s one thing you hope the next generation of scientists carries forward in the fight against cancer, what would it be?

Dr. Finley

I think it’s really important to keep the patient in mind. That can be harder to do when you’re working in basic science or in areas that aren’t immediately translational. But even when building a mathematical model, try to remember that it’s meant to represent a real person—someone with a life, a history, a family, and an environment that shapes their disease.

It’s easy to get caught up in the science and leave the patient out of the picture, especially in areas like mathematical oncology. But keeping that human element in mind helps you stay grounded and reminds you of the potential impact your work can have.

Isaac

Thank you so much. Your advice today—especially about majors and getting started—has been really helpful. I’m applying to college soon, so this really means a lot.

Dr. Finley

It was my pleasure. I'm always happy to talk science and careers—anytime.