City of Hope Researchers Explore a New Way to Predict Breast Cancer Risk

For decades, breast cancer risk assessment has largely focused on inherited genetic mutations, leaving many women without clear answers about their personal risk. While genetic testing can help identify some individuals with inherited susceptibility, researchers estimate that the vast majority of breast cancer cases occur in women without those known genetic markers. Now, new research from scientists at City of Hope is exploring a potentially groundbreaking way to detect breast cancer risk earlier — before cancer even develops — by examining the physical and mechanical properties of breast cells themselves.

In an interview with The Educated Patient, Mark LaBarge, Ph.D., professor and stem cell biologist in the Department of Population Sciences at the Beckman Research Institute at City of Hope National Medical Center, discusses how his team developed a novel technology that uses mechanical cell measurements and machine learning to identify signs of accelerated biological aging in breast tissue. The research could eventually help personalize screening strategies, reduce unnecessary anxiety for some women, and open the door to earlier prevention and lifestyle interventions aimed at slowing biological aging linked to cancer risk.

What makes this new research exciting, and how could it potentially change the way we think about breast cancer risk before cancer even develops?

Mark LaBarge, Ph.D.: Right now, the strongest tools we have for breast cancer risk assessment are focused on inherited genetic risk through germline gene sequencing. But inherited risk only accounts for about 6% to 10% of women who develop breast cancer. For the other roughly 90%, we do not currently have a truly individualized way to assess risk based on their own biology.

Most current methods rely on population-based models that compare one person’s medical history to others who developed breast cancer. While those models can be helpful, they often overestimate or underestimate a person’s actual risk.

What’s exciting about this research is that we identified a biological signal linked to susceptibility that can be detected through a simple mechanical measurement of cells. This was unexpected and represents an entirely new approach. The technology is scalable and opens the possibility of more personalized breast cancer risk assessment in the future.

Can you explain in simple terms what researchers were measuring and why these cellular changes may reveal early signs of cancer risk or accelerated biological aging?

ML: In simple terms, we pass cells through a tiny tube that narrows at one point, which squeezes the cells. We measure what the cells look like before squeezing, during squeezing and how long it takes them to return to their original shape afterward.

Those measurements tell us about the physical and biological properties of the cells. Our technology captures several mechanical and physical characteristics, and we use machine learning to identify patterns associated with higher breast cancer risk.

What we found is that certain combinations of these measurements can help distinguish cells from women who may be at higher risk from those who are not.

How could a test like this help personalize screening recommendations or reduce unnecessary procedures for some women?

ML: One example would involve women with dense breast tissue. Many women are told after a mammogram that dense breasts may increase breast cancer risk, but only a relatively small percentage of women with dense breasts actually develop breast cancer.

Our vision is that a woman with dense breasts could undergo a minimally invasive fine needle aspiration to collect breast cells. Those cells could then be analyzed using our device to determine whether there are biological signs of increased risk.

That could provide more individualized information instead of relying solely on broad risk categories from imaging alone.

Could this technology eventually support earlier prevention strategies or lifestyle interventions?

ML: That is one of the most exciting possibilities. Our research is based on the idea that higher-risk breast tissue may be biologically aging faster than expected.

Using related technologies, we have already observed that some women enrolled in breast cancer prevention trials show slower biological aging after interventions. In the future, this type of testing could potentially help monitor whether lifestyle changes — such as increased exercise or other preventive strategies — are slowing biological aging in breast tissue.

How is machine learning helping identify these cancer risk patterns, and what still needs to happen before this becomes routine care?

ML: Machine learning allows us to analyze large amounts of complex cellular data in ways humans cannot easily interpret on their own. Initially, we thought one measurement alone explained most of the risk signal, but machine learning showed us that multiple measurements together produced a much more accurate model.

So far, our work has been conducted using human samples, but on a relatively small scale. The next step is conducting larger and more diverse clinical studies to determine how well this technology performs in broader patient populations before it could potentially become part of routine care.

This transcript was edited for clarity.