PURPOSE: To evaluate the prognostic impact of large cell/anaplastic (LC/A) histology together with molecular and clinical risk factors in childhood medulloblastoma. METHODS: Three consecutive prospective medulloblastoma trials were screened for patients with the histological diagnosis of LC/A medulloblastoma. Tumors were considered as LC/A if they displayed areas of severe cytological anaplasia or a significant or predominant large cell component. Histology was centrally confirmed. Genomic DNA amplification of c-myc and n-myc, and mRNA expression of c-myc and trkC were analyzed. RESULTS: Twenty-eight patients with LC/A medulloblastoma with a median age of 6.1 years (1.4-16.5 years) and a median follow-up of 4.5 years were identified (5% of all medulloblastoma). Four-year event-free (EFS) and overall survival (OS) were 58% and 67%. Young age and metastases (n = 13, 4-year EFS 31% vs. 82% in 15 children >4 years and without metastases, P = 0.001), large cell histology (n = 9, 4-year EFS 22% vs. 75%, P = 0.005) and c-myc amplification (n = 9, 4-year EFS 22% vs. 89%, P < 0.0001) were negative prognostic factors. C-myc amplification was highly correlated with young age (P < 0.001), metastases (P = 0.002) and large cell histology (P = 0.007). Outcome of 12 patients with severely anaplastic tumors without these risk factors was not impaired (4-year EFS 86%). CONCLUSION: In a subgroup of patients without clinical and molecular risk factors outcome was favorable despite severely anaplastic histology. In contrast, c-myc amplification and large-cell histology were associated with an inferior outcome. Intensified treatment strategies should be considered for children with LC/A medulloblastoma and these characteristics.