Title: Three-Dimensional Molecular Cartography of the Glioblastoma–Brain Microenvironment Interface

Glioblastoma is the deadliest form of brain cancer, with most patients surviving less than two years after diagnosis. A major reason this disease is so hard to treat is that tumor cells invade deep into surrounding healthy brain tissue, making complete surgical removal impossible. Understanding how and why these cells spread is critical to developing better treatments.

Our project builds the first comprehensive 3D maps of glioblastoma tumors within the brain. We take surgically removed tumor tissue, cut it into hundreds of ultra-thin slices, stain each one, then digitally stack them using computer algorithms, essentially rebuilding the tumor in three dimensions on a screen.

We then apply artificial intelligence to go beyond what the eye can see. Our deep learning models predict which genes and proteins are active at every point in the tissue, based on how the stained tissue looks under a microscope. This lets us create a detailed molecular picture of the entire tumor without the enormous cost of running molecular tests on every slice.

With these 3D molecular maps, we can study the invasive front (the boundary where tumor meets healthy brain) and compare it to the tumor's interior. We aim to identify which interactions between cancer cells and normal brain cells, including neurons and immune cells, drive invasion. If we can pinpoint the molecular signals pushing tumor cells to spread, we may be able to identify new drug targets that slow or stop that process.

Ultimately, this work could change how we understand and treat the most aggressive form of brain cancer, giving patients more time and better outcomes.