Medulloblastomas and other embryonal brain tumors, the most common malignant brain tumors of childhood, have a 40-50% overall mortality. Survivors typically live with profound developmental and neurological disabilities, largely due to toxic therapies currently in use. In 1994, the Pomeroy lab was the first to show that a molecular marker, the neurotrophin-3 receptor TrkC, can predict overall survival of embryonal tumors with significantly greater accuracy than clinical or histological features. The group, then, was the first to apply genomic methods to show that favorable prognosis tumors have substantially different gene expression profiles than poor prognosis tumors, identifying a molecular signature that can distinguish tumors with different biological properties but identical histological features. Outcome prediction models developed by the Pomeroy lab based on gene expression profiles are by far the most accurate predictors of embryonal tumor outcome currently available. They are being tested at a national level in Children’s Oncology Group clinical trials focused on improving cognitive outcome by lowering radiation doses in good prognosis patients.