MRI Sensor Links Rapid Tumor Pressure Rises to Traits That Favor Cancer Spread

A pressure surge may matter more than the peak

A new MRI sensor has allowed researchers to track local mechanical stress inside growing tumors, revealing that sudden increases, rather than simply high pressure, can push cancer cells toward a more mobile and invasive state. In mouse tumors, the team followed stress ranging from 0.5 to 12.4 kilopascals over 17 days, capturing differences between locations within the same tumor.

The research was led by Jinwoo Cheon, director of the Institute for Basic Science (IBS) Center for Nanomedicine and an Underwood Professor at Yonsei University in South Korea, and Young-wook Jun, a professor at the University of California, San Francisco. Their mechanoMR microparticle sensor, called M3 or M-cube, turns compression of a tiny gel particle into a measurable MRI signal.

The central finding concerns the pace of change. Gradually increasing stress allowed cells to activate protective responses. Abrupt increases were associated with oxidative stress and epithelial-mesenchymal transition (EMT), a process through which cells loosen their connections and acquire traits that can support cancer spread.

This is not a validated test for predicting metastasis in patients. The experiments establish a connection between changing mechanical stress and cell behavior in laboratory models and animals. They do not establish that a pressure reading can determine whether a person's cancer will spread to another organ.

What the publication record shows

The detailed academic account is a research preprint titled MechanoMR microparticle (M3) sensors reveal dynamic stress loading as a driver of epithelial-mesenchymal transition. Posted on December 13, 2025, its page explicitly states that it has not yet been reviewed by a journal.

Later reports describe publication in Nature Nanotechnology on September 28, 2026, followed by an IBS announcement on October 12. These are different stages of the publication record. The warning about peer review applies to the linked preprint version; it should not be confused with the later journal publication described in those reports.

The reported sequence is:

• December 13, 2025: Version 1 of the bioRxiv preprint was posted.
• September 28, 2026: The study was reported as published in Nature Nanotechnology.
• October 12, 2026: IBS announced the findings.

The preprint lists Minji An as first author. Later accounts identify the first author as Minji Ahn, an IBS Center for Nanomedicine postdoctoral researcher, using a different English spelling. The numerical details below are those documented in the preprint, while the later accounts describe the publication announcement and caution against interpreting the work as a clinical metastasis test.

How a tiny gel particle becomes a pressure sensor

Each M3 sensor consists of an alginate hydrogel particle approximately 70 micrometers across, containing magnetic nanoparticles measuring 13 nanometers. Alginate is a material obtained from seaweed that can form a soft gel. The embedded particles are zinc doped ferrite, a magnetic material that affects how nearby water produces an MRI signal.

When surrounding tissue compresses the gel, the particle deforms and expels some water. That changes both the movement of water near the magnetic particles and the spacing between those particles. Together, these changes alter the magnetic relaxation signal, which the researchers calibrate against known stresses.

The distinction between the particle's size and the imaging resolution matters. A sensor is about 70 micrometers wide, but individual particles were resolved by MRI at approximately 300 micrometers, or 0.3 millimeters. This supports the description of resolution below one millimeter, without implying that the scan resolves each cancer cell.

The device measures local solid stress within the tissue around it. It does not directly measure pressure inside an individual cancer cell, and it does not measure the pressure of fluid alone. Its purpose is to show how the mechanical environment surrounding cells changes as tissue grows or reorganizes.

Testing the signal before tracking tumors

The researchers first placed sensors inside agarose gel, a material used to mimic how tissue transmits force, and applied controlled compression. They compared MRI readings with direct optical measurements of particle deformation. That comparison tested whether the magnetic signal actually followed compression rather than some unrelated change.

The calibration experiments covered 0 to 15 kilopascals (kPa), while the later descriptions summarize the tumor pressure measurement range as 1 to 15 kPa. These figures refer to different descriptions of testing, not necessarily competing measurements. A kilopascal is a unit of stress or pressure, expressing force applied over an area.

As compression increased, the sensor signal changed in a reproducible way. By contrast, much stiffer particles made using a different polymer showed no detectable signal change within the same stress range. That control supports the proposed mechanism: deformation of the soft gel is essential to the measurement.

The signals also remained robust across changes in acidity, surrounding material and material stiffness, and for up to four weeks of monitoring. This persistence matters because a single scan cannot establish how quickly stress accumulated. Repeated readings are needed to distinguish a slow rise from a sudden surge.

Following stress through growing cancer models

The team tested M3 in three dimensional tumor spheroids, clusters of cells that reproduce selected features of a tumor's physical environment. The models used HCT116 colon cancer cells, A549 lung cancer cells and MDA-MB-231 breast cancer cells. Sensors were incorporated into the clusters so that local compression could be followed during growth.

All three cancer models accumulated stress, but at different rates. The researchers interpreted these differences as consistent with variations in cell stiffness and how each cluster grew within a surrounding matrix. The results show why one pressure trajectory cannot automatically describe every tumor type.

For the mouse experiments, sensors were mixed with HCT116 cells and injected beneath the skin to form xenografts, tumors grown from transplanted cells. Individual sensor locations were then followed over 17 days. Local stresses ranged from 0.5 to 12.4 kPa, with different rates of accumulation at different positions.

The maps did not reveal one uniform pressure level throughout each tumor. One example had a relatively scattered distribution of stress, while another contained two distinct stress regions. Averaging those locations into a single number would lose information about the conditions experienced by cells in particular areas.

Why gradual loading and sudden loading differ

The preprint identifies abrupt increases in stress, rather than accumulated stress or the highest stress reached, as key determinants of EMT induction in its experiments. This shifts the focus from how much compression a tumor contains at one moment to how its mechanical conditions change over time.

Under gradual stress escalation, gene activity analysis showed activation of protective antioxidant pathways, including FOXO and AMPK. These pathways help cells manage stress and maintain their existing state. The researchers describe this response as supporting epithelial stability, meaning that cells retain the connected organization typical of epithelial tissue.

Rapid stress increases were associated with stronger oxidative stress. This occurs when reactive molecules build up beyond the capacity of cellular defenses to control them. The later research descriptions also identify increased activity of genes related to YAP/TAZ, proteins involved in translating physical cues into changes in cell behavior.

In these conditions, EMT was promoted. Cells undergoing EMT loosen their attachments and acquire greater mobility and invasiveness. Those traits can help cancer cells leave their original tissue, but they are not equivalent to demonstrating successful metastasis, which also requires survival and growth at another location.

Why regions showing EMT had lower stress

One of the most revealing findings appears, at first, to run against the idea that rising pressure promotes invasive behavior. Regions showing EMT had lower local stress than regions without the same marker. The explanation proposed by the researchers is that a mechanical surge can help initiate a transition that subsequently releases compression.

In mouse tumors, sensors within regions expressing vimentin, a marker used to identify the mesenchymal state, measured 3.5 ± 2.8 kPa. Sensors in vimentin negative regions measured 7.4 ± 2.8 kPa. Comparing the reported means, stress in the regions expressing vimentin was about 53% lower, although the substantial variation means the averages do not define a diagnostic threshold.

Related spheroid experiments showed a similar sequence. After researchers induced EMT with 5-azacytidine, the sensor signals initially followed the same stress accumulation pattern as untreated controls. From around day five, the signals reversed, indicating release of compression as EMT markers developed through day nine.

The authors interpret this as a transition toward tissue behavior resembling a fluid, allowing cells to rearrange and release previously stored stress. The finding explains why a low reading at one moment does not necessarily mean a region never experienced a damaging surge. Its earlier trajectory may be more informative than its present value.

What the findings do not yet establish

The strongest supported conclusion is that stress dynamics can influence cell state in the tested models. The research does not provide a pressure threshold that predicts metastasis in patients, a validated screening test, or evidence that treatment decisions should be based on M3 readings.

Validation in human tumor tissue is still needed. Laboratory spheroids and mouse xenografts allow researchers to control and observe mechanical conditions, but they do not establish how reliably the same relationships operate across cancers in people.

There is also a practical distinction between MRI observation and sensor placement. MRI can read the particles without repeatedly opening the tissue, but the particles must first be introduced. In the mouse tumor experiments, they were mixed with cancer cells before the tumors developed. That setup does not demonstrate routine placement and monitoring in an existing human tumor.

Neither the publication descriptions nor the detailed preprint account give a scheduled human trial, a regulatory submission date or a clinical deployment deadline. The next scientific task is to test whether the measured stress histories and associated cellular responses hold in human tissue and can provide useful information beyond existing assessments.

Evidence beyond cancer spread

The research also included experiments relevant to fibrosis, a process involving excessive tissue scarring. Spheroids made from MRC5 cells, used as a pulmonary fibrosis model, contracted rather than simply expanding. Their sensors recorded a steeper accumulation of compression, consistent with active remodeling of the surrounding matrix.

Treatment with nintedanib slowed that stress buildup in the model. This experiment supports the use of M3 to study how cellular contraction and matrix remodeling alter local forces. It does not establish a new clinical use for the sensor or demonstrate a treatment benefit in patients.

The team also tested the platform during brain development in Xenopus laevis tadpoles. Their optical transparency allowed MRI readings to be compared with direct observation of particle deformation during rapid tissue changes. This provided another check that the sensor could function in living tissue undergoing substantial remodeling.

Cheon, the IBS Center for Nanomedicine director, described what the tumor experiments achieved:

This study tracked the 'invisible forces' within living tumors using MRI and demonstrated that changes in pressure stress acting on cancer cells are closely linked to their metastatic propensity

Cheon also identified tissue regeneration, inflammation and fibrosis as possible areas for future use. The present work gives that ambition a concrete starting point: a method for repeatedly measuring local mechanical stress and connecting its history with changes in cell behavior, rather than assuming that a tumor's highest pressure tells the whole story.

Key Points

• M3 combines a soft alginate particle with magnetic nanoparticles to translate local tissue compression into an MRI signal.
• Individual sensors were resolved at approximately 0.3 millimeters, and calibration covered 0 to 15 kPa.
• Mouse tumor measurements ranged from 0.5 to 12.4 kPa over 17 days.
• Abrupt stress increases promoted oxidative stress and EMT, while gradual increases activated protective antioxidant pathways.
• Regions showing EMT had lower average stress, consistent with compression being released after cells changed state.
• The linked December 2025 document is a preprint; later reports describe journal publication on September 28, 2026.
• The sensor is a research tool, not a validated test for predicting cancer spread in patients.