Scientists use AI-enhanced microscopy to understand glioblastoma brain tumours

New Delhi: Scientists from EMBL and Heidelberg University have made a breakthrough in glioblastoma brain tumours research using AI-supported microscopy. 

The new approach, detailed in a recent study, allows them to track glioblastoma cells in detail as they move along the brain’s nerve fibres, providing crucial insights into the early detection and spread of one of the deadliest brain cancers.

Glioblastomas travel primarily along the corpus callosum, a bundle of nerve fibres connecting the brain’s hemispheres. Previously, observing these tumour cells in deep within the brain’s white matter was challenging. With AI-enhanced microscopy, researchers can now monitor these cells in real time as they migrate along this “superhighway”, gaining a clearer understanding of their behaviour.

The study builds on a 2021 breakthrough in deep-tissue microscopy developed by EMBL Group Leader Robert Prevedel and his collaborators from around the globe. The team addressed longstanding challenges in visualising neurons and glial cells deep within brain tissue. Their work initially focused on studying communication between neurons in regions like the hippocampus. The recent enhancement of this technique with AI has allowed for even more precise imaging of brain tumour cells and their surrounding microenvironments.

“We have now gone from taking a snapshot of cells in a mouse brain to zooming in on specific cells and being able to follow them for many hours or even days,” Prevedel explained. “Also, incorporating custom AI approaches has allowed us to distinguish different parts of the microenvironment of the cells, which is also very important to understand their behaviour in context.”

The collaboration between Prevedel’s team and Dr. Varun Venkataramani of the University Hospital Heidelberg marks a significant step forward in glioblastoma research. Venkataramani, a neurooncologist, saw potential in the new microscopy method to extend the imaging capabilities into the white matter, where glioblastomas predominantly invade. His team was able to track tumour cells in the white matter, observing how they interact with the brain’s nerve fibres and spread through the corpus callosum.

“The 2021 paper by Robert’s group introduced a deep-tissue microscopy technique that I believed could extend our imaging capabilities to the white matter of the corpus callosum,” said Venkataramani. “This could potentially reveal novel biological processes and offer insights into the behaviour of these tumours in a critical, yet understudied niche.”

Marc Schubert, a lead author and medical student at Heidelberg University, added, “It has been fascinating to observe tumour cell invasion in the corpus callosum in real-time.” The ability to observe these cells in their natural microenvironment offers vital clues about the progression and adaptation of glioblastomas, which are notoriously difficult to treat and detect early with standard imaging techniques like MRIs.

“At this point, I think the most important aspect to this fundamental research is that it allows us to investigate these tumours in their most relevant microenvironmental niche for the first time,” Venkataramani said. “These findings also help explain the current challenges in detecting glioblastoma cells at the tumour’s infiltrative edges using conventional MRI techniques, which are the standard in clinical imaging.”

Artificial intelligence played a pivotal role in advancing the microscopy technique. “From a technical development point of view, the AI-based methods helped to ‘denoise’ our images, so the contrast now is much clearer,” Prevedel explained. “The AI can distinguish different structures inside the white matter, like myelinated fibres and blood vessels, which is important for a variety of reasons.”

Despite the promising results, Venkataramani emphasized that further development is needed before the technique can be applied in clinical settings. “It’s promising, but it’s much too soon to apply it clinically without further development,” he said, explaining that the next steps will integrate further advanced imaging modalities and refine the technique for practical use in diagnosing and treating glioblastomas.