Most medulloblastoma research has focused on the primary tumors growing in the cerebellum. However, most medulloblastoma patients do not die from the primary disease, but from leptomeningeal metastasis: the dissemination of tumor cells from the cerebellum into the brain and spinal cord. Metastatic lesions cannot be surgically removed, and there are no effective therapies to eliminate them or stop them from growing. To understand the molecular basis of metastasis, we are using animal models to study the differences between primary and metastatic tumors, and screening for genes that can promote or inhibit metastatic dissemination. We are also integrating our results with molecular data from medulloblastoma patients, to identify genes that are critical for metastasis. Understanding the function of these genes will allow us to design novel strategies for early detection, prevention, and treatment of metastasis in patients with medulloblastoma and other types of brain tumors.