Breast cancer risk is usually framed around familiar factors: genetics, age, family history, or what appears on a mammogram. But researchers at City of Hope, a cancer research and treatment organization, and the University of California, Berkeley are beginning to explore a different possibility – that risk may also be reflected in the physical behavior of cells themselves, long before cancer can be detected clinically.
Central to the study is the concept of “mechanical aging,” which suggests that cells can develop physical characteristics associated with older tissue regardless of chronological age. To explore this, the team uses tiny microfluidic devices that gently compress individual cells and observe how they recover. The goal is not simply to look for cancer, but to understand whether the physical state of healthy cells might carry information about future risk.
In this interview, Lydia L. Sohn, professor of mechanical engineering at the University of California, Berkeley and co‑senior author of the MechanoAge work, and Mark LaBarge, translational scientist at City of Hope focusing on breast aging and cancer risk, discuss a field that sits at the intersection of physics, engineering, aging biology, and cancer research – and why the mechanics of a single cell may reveal far more than previously imagined.
Can you explain “mechanical aging” in cells simply – like how a cell’s physical stiffness and bounce-back reveal cancer risk, separate from a woman’s actual age?
“Mechanical aging” is a way of describing how old a cell behaves physically, rather than how old the person is.
In our work, we compare the mechanical properties of breast cells from younger and older women. What we’ve seen is that some cells from younger women can actually look mechanically “older”: they are stiffer, less resilient, and don’t bounce back as easily after being deformed. In that sense, a cell’s physical behavior may reflect risk-related changes that are separate from a woman’s actual age.
Because each cell generates a large amount of mechanical data, we use machine learning to help identify patterns. The algorithm isn’t just looking at one simple feature like stiffness alone – it picks up subtle combinations of physical traits that distinguish mechanically young cells from mechanically old ones. So even if we can’t always point to a single measurement and say “this is the reason,” the overall mechanical profile can tell us whether a cell is behaving more like a younger or older breast cell.
Put simply, a cell’s stiffness and its ability to bounce back may act like a kind of physical aging signature – and that signature could help us understand cancer risk beyond chronological age alone.
For most women without genetic risks (over 90%), current tools like breast density are indirect – how could MechanoAge change screening or prevention talks with doctors?
MechanoAge could make breast cancer risk conversations more specific and actionable.
Right now, many women are told after a mammogram that they have dense or very dense breasts, and that this is a known risk factor for breast cancer. But that information is still fairly indirect. More than half of women undergoing mammography may receive some version of that message, yet most will never develop breast cancer. So the question becomes: what should an individual woman and her doctor actually do with that information?
MechanoAge, or a related MechanoRISQ score, could help refine that picture. Instead of simply saying, “You have dense breasts, so you may be at higher risk,” it could offer a more individualized readout of whether a woman’s breast cells show mechanical features linked to higher or lower risk.
For women with a higher MechanoRisk score, doctors might consider stepping up screening or prevention discussions, similar to how they already approach patients considered high risk under current guidelines. For women with a lower score, it could help avoid unnecessary escalation and provide reassurance that standard screening is appropriate.
In other words, MechanoAge could turn a vague density letter into a more personalized conversation about what level of screening or prevention actually makes sense for that individual woman.
Your study found that cells from young high-risk women act “mechanically older” than expected – what first clue did you spot in early tests, and how much older did they seem compared to actual age?
The first clue came from a simple mechanical comparison: how well breast cells recovered after being compressed, or “squished.”
Cells from older women didn’t bounce back as easily as cells from younger women. But then, when we looked at cells from young women with known high-risk factors, such as BRCA mutations, we saw something striking: their cells also failed to recover well after compression. Mechanically, they behaved much more like cells from older women than like cells from women their own age.
So even though these women were young chronologically, their breast cells appeared mechanically older. That was the key observation that first suggested cell mechanics might reflect risk-related aging that isn’t captured by chronological age alone.
You found mechanical changes even in healthy breasts of women with cancer elsewhere – what role did genetic factors play in speeding up this “mechanical aging”?
We actually don’t know what role genetics played in driving mechanical aging in that group.
What’s important is that these were women who had cancer in one breast, but the breast we measured was confirmed by pathologists to be cancer-free. We also selected them specifically because they had undergone genetic testing and did not carry known germline mutations – such as BRCA or other established inherited risk variants – that would explain their cancer risk.
That’s what makes this group particularly interesting. These are women we don’t currently know how to identify through standard genetic testing alone. Even if you sequenced their DNA, you might not find a known high-risk mutation. Yet MechanoAge appears to pick up something in their healthy breast tissue that suggests increased risk.
So the intriguing possibility is that MechanoAge could help identify some of the many women who develop breast cancer without a clear genetic explanation – potentially revealing aspects of risk biology that current tools simply don’t capture.
The study compared cells from healthy women, those with family history, and cancer patients – what was the biggest source of variation between cells from the same woman?
The biggest measurable source of variation appears to be recovery time — how long a cell takes to bounce back after being compressed.
In our current machine learning model, recovery time comes out as the most important parameter, which suggests that differences in how cells recover are driving much of what we see between cells from the same woman.
But the question also raises a broader biological issue. We still don’t know whether all breast cells within the same woman mechanically age at the same rate. There may be meaningful variation within a single breast – where some cells appear mechanically younger, while others look mechanically older.
So the concise answer is: recovery time is currently the strongest measured contributor, but the broader question of how much mechanical-age diversity exists among cells within the same woman remains open and scientifically important.
The platform is affordable and scalable – what’s the estimated cost per test today, and how could automation make it even cheaper?
Today, the test is very inexpensive at the device level. We currently make the cell-squeezing chips by hand, and the cost is roughly less than 50 cents per chip.
Automation should make that even cheaper. If the platform were scaled up, we could manufacture chips more efficiently and run multiple tests for a single patient – or for multiple patients – at the same time. That kind of batching would both lower costs and increase throughput.
Another important point is that the readout doesn’t require a complex or expensive imaging system. The instruments rely on relatively simple electronics, and in the lab version, we’ve shown that the setup can be built from readily available components for a few hundred dollars upfront.
So the idea is fairly straightforward: the chip itself is cheap, the instrumentation is relatively low-cost, and with automation, the system could become even more scalable and affordable over time.
What kinds of clinical endpoints and study design would likely be needed for regulators to evaluate MechanoAge as a breast cancer risk-assessment tool?
We do not yet know exactly what outcomes or group sizes regulators would require. This is a new device and a new type of risk-assessment technology, so the regulatory pathway is something we would need to define through discussions with regulators and clinicians.
Our vision, though, is that an early clinical trial would test whether a MechanoAge or MechanoRisk score can predict breast cancer diagnosis within a defined follow-up period – for example, two or five years after the score is measured.
So the key question would likely be: does the risk score meaningfully predict who goes on to develop cancer within that timeframe? But the exact trial size, endpoints, and approval criteria would ultimately need to be developed in collaboration with clinical and regulatory experts.
