Cancer Research · eu
Brain Cancer Cells Can Switch Routes: GlioTrace Tracks the Invasion Paths of Thousands of Cells
Using tumor models derived from six patients, researchers observed glioblastoma cells changing shape and switching movement routes in living brain tissue slices. The method, which combines microscopy and machine learning, also offers a new tool for observing how treatment affects cancer cell invasion.
One reason glioblastoma is difficult to treat is that cancer cells cross the tumor boundary and infiltrate surrounding healthy brain tissue, leaving lesions that are difficult to remove surgically. Understanding this process requires more than seeing where cells end up; researchers also need to know how they move and how their behavior changes under treatment pressure.
A team at Uppsala University in Sweden presented GlioTrace in *Nature Communications*, combining automated microscopy, machine learning, and statistical models to track thousands of tumor cells. The study used glioblastoma cells from six patients, first allowing them to form tumors in mouse brains, then cutting the brain tissue into thin slices and keeping them alive in the laboratory. Images were taken at intervals of no more than two hours, with observation lasting up to five days.
These image sequences showed that tumors derived from different patients spread in different ways: some cells moved mainly along blood vessels, while others traveled more diffusely through surrounding tissue. A single cell could also change shape over several days, switching from movement along blood vessels to movement through tissue. This means that cells’ modes of invasion are not fixed; studying how to stop their spread also requires understanding how they switch routes.
The team further used GlioTrace to observe treatment responses. Two drug substances reduced cell movement and spread in the brain slices, but their effects varied according to the patient source and the cells’ movement behavior. This allowed researchers to break down the overall response and identify which cells’ behavior was affected. However, reduced movement in slices cannot yet be directly interpreted as a treatment benefit for patients.
Another finding requires even more cautious interpretation: after some tumor models received radiation and chemotherapy, the proportion of cells with a particular shape increased. These cells were among the fastest-moving types and also occupied a central position in the network of transitions between different cell states. The study raises the question of whether they may persist under treatment, but has not demonstrated that these cells are more resistant to treatment or that they cause tumor recurrence.
GlioTrace uses AI to identify and track individual cells in large volumes of microscopy images, then combines this with statistical analysis to reconstruct changes in behavior. The analysis software was developed by the team in collaboration with researchers at Chalmers University of Technology and the University of Gothenburg and has been made freely available to other researchers. Used with a standard microscope, it can reveal cell trajectories across an entire brain slice.
For now, evidence for this method still comes from experimental models consisting of patient-derived cells and mouse brain slices. It has not been evaluated as a diagnostic tool for patients, and it cannot yet be used to select therapies. Its immediate value is in turning cancer cell invasion from a series of static images into a process that can be analyzed step by step, helping researchers ask more precise questions: which cells has treatment changed, and which movement behaviors continue?