BACKGROUNDPediatric ependymoma is a heterogenous disease. Subgroup-specific clinical information on prospectively treated patients will help to improve treatment stratification.METHODSWithin the population based, prospective, multicenter E-HIT-trial (2001–2011) patients with localized ependymoma confirmed by neuropathological central-review, received hyperfractionated local radiotherapy (68Gy, 2x1Gy/day) followed by chemotherapy (stratum-A), or chemotherapy followed by local radiotherapy (54Gy, 1.8Gy/day) (children < 4years, stratum-B), or age-adapted radiotherapy with pre-/post-irradiation chemotherapy (residual tumor, diagnosis after 2005, stratum-C). Retrospective classification of DNA-methylation was available for n=164 E-HIT-trial participants, and n=80 patients with comparable treatment and prospective registration in the subsequent HIT-interim-registry (2012–2014). FINDINGS: For 291 E-HIT-trial patients, 5-year progression-free (PFS) and overall survival (OS) were 61±3%, and 81±2%. Five-year PFS/OS after complete resection were 71±4% and 87±3% in stratum-A (n=127), and 64±5% and 86±4% in stratum-B (n=86). Outcome was poor after incomplete resection, irrespective of treatment-stratum (n=78, 5-year PFS/OS: 43±6%, 68±5%). In the pooled trial- and registry-cohort, there were 152 patients with PF-EPN-A (5-year PFS/OS: 44±4%, 77±4%), 40 of them with 1q-gain (5-year PFS/OS: 28±7%, 66±8%), 21 with PF-EPN-B (5-year PFS/OS: 90±7%, 100%), 59 with ST-EPN-RELA (5-year PFS/OS: 63±7%, 87±5%), and 4 with ST-EPN-YAP1 (2 progression/relapse, no death).CONCLUSIONOutcome differed between molecular subgroups and insufficient survival rates were achieved for patients with PF-EPN-A with 1q-gain, despite combined radio- and chemotherapy treatment. Treatment reduction in the context of a clinical trial may be considered for PF-EPN-B.