Tumour cell collectives break apart roughly more readily in disordered microenvironments than in regular ones — that is the central measured result of a study published in Science Advances by researchers at the Institute of Science and Technology Austria (ISTA) and the UK's Francis Crick Institute. The finding gives cancer researchers a reproducible biophysics platform for studying how metastasis begins: not only in genes, but in the geometry of the tissue a tumour is pushing through.

The team, led by Michael Sixt, head of ISTA's Cellular Morphodynamics Group, and Edouard Hannezo, leader of the Physical Principles in Biological Systems Group, built a microfluidic chip experiment plus an associated computer simulation to test a specific question: does heterogeneity around a tumour drive the detachment of individual cells, the first step of metastatic spread?

How did they measure it?

The researchers reproduced the pores between tissue fibres as forests of microscopic pillars inside microfluidic devices. In the ordered condition, the chip carried a square lattice of 9 µm-diameter pillars spaced with 9 µm gaps. In the disordered condition, each pillar was shifted from its lattice position by a random angle and distance.

Cancer cells were introduced at the centre of each forest and tracked with fluorescence imaging and light microscopy. Cell culture medium kept the cells alive for up to eight days, which set the practical observation window for the experiment.

The result was directional but clear: the tumour cell collective was more likely to break apart when navigating the heterogeneous geometry than the regular one.

"We found that the first cells detach, exactly as happens in a cancer metastasis," said Sixt, describing behaviour observed in the disordered condition. The team also recorded a second structural effect: pillars increasingly roughened the surface of the spreading cell interface, producing finger-like protrusions, and cells detached more often from the tips of those protrusions.

What did the tooling require?

The experimental design demanded new software, because no existing tools could specify mathematically defined disorder.

"There were no suitable tools for designing mathematically defined disorder. So I had to write completely new software that can define those patterns, and then translate that into designs that can be manufactured [via standard semiconductor lithography techniques] into microfluidic devices," said Saren Tasciyan, a biotech data science consultant who designed the devices during his PhD in the Sixt Group.

For labs considering replication, the manufacture route is standard semiconductor lithography, which keeps the barrier to adoption relatively low once the pattern-generation software is available.

Why add a simulation?

The eight-day survival limit constrained how long the researchers could watch the process. Sixt asked Hannezo's group to extend the timeline computationally and verify the experimental observations.

"Each cell was modelled as a bead that moves through an environment," said Zuzana Dunajová, a postdoc in the Hannezo Group and a computational scientist specialising in biological physics. The beads attract one another and move through geometries mimicking those on the chips.

"I kept seeing the same behaviour in my simulations – the beads detached from the collective more often in the disordered environment compared to the ordered," Dunajová said. She noted the result suggests a general principle of active or living systems rather than a property of one cell type: the roughness of cancer-cell invasion follows the same universal behaviour seen in systems such as burning paper, spreading fires or drying coffee drops.

That generality matters for anyone building metastasis models. If detachment driven by geometric disorder is a physics-level mechanism, it may operate across cell types and tissues, which would widen the platform's relevance beyond the specific cancer cells used here.

What are the limits of the data?

The study measured detachment in a controlled two-dimensional analogue of tissue — pillar forests on a chip — not in living tumours. The reported outcome is comparative (disordered versus ordered geometry) rather than a quantified detachment rate in vivo. Genetics, blood vessels, immune cells and signalling molecules all shape metastasis in the body, and the chip isolates only the geometric component. The eight-day cell survival ceiling is likewise an experimental constraint the simulation works around rather than eliminates.

The funding structure of the work is not detailed in the published summary, so readers should consult the paper itself for funder disclosures.

What comes next?

Sixt now plans to test whether cell detachment in a disordered microenvironment triggers epigenetic modifications — heritable changes that do not alter the DNA code — and, later, genetic modifications.

"I want to consider whether the detachment process can drive the single cell evolution into something malignant," he said. For therapy developers, the follow-up question is whether mechanically induced detachment feeds the evolution of malignant single cells, and whether intervening in that mechanical pathway offers a target upstream of genetics.