Children can be affected by a large variety of CNS tumor entities with very divergent outcomes. The etiology of childhood brain tumors is largely unknown. However, approximately 10% of cases are thought to arise from underlying pathogenic germline mutations, which after thorough human genetics assessment are undoubtedly considered to predispose to cancer. We have retrospectively and prospectively analyzed the frequency and spectrum of hereditary predisposition across 1022 patients with medulloblastoma, where we found 6 significant predisposition genes and up to 20% germline predisposition in the SHH (Sonic Hedgehog) subgroup in known predisposition genes. This does not yet account for novel discoveries, which will also be presented and increases the proportion of SHH medulloblastoma with underlying predisposition to more than one third of cases. Currently, we are prospectively analyzing the frequency of cancer predisposition in two population-based cohorts of childhood brain tumors at primary diagnosis and at relapse through our two molecular diagnostics pipelines, Molecular Neuropathology 2.0 (www.molecularneuropathology.org) and INFORM (http://www.dkfz.de/en/inform/index.html), respectively. The results of this interim analysis will be presented at the meeting including ethical and logistics issues, frequency & spectrum of pathogenic germline mutations per molecular entity, accompanying clinical symptoms, family history, and consequences for treatment stratification. Since Li-Fraumeni is the single most common cancer predisposition in the pediatric population, we have recently opened a national registry for these patients together with our colleagues from MHH Hannover (http://www.krebs-praedisposition.de), which at a later stage will also be opened for other predisposition syndromes. First experience with the registry will also be presented.