Purpose: The HR-NBL1 trial of the European SIOP Neuroblastoma Group randomised 2 MAT regimens to demonstrate superiority based on event free survival (EFS).Method: Response eligibility criteria prior to randomisation after Rapid COJEC Induction (J Clin Oncol, 2010) 3 4 2 courses of TVD (Cancer, 2003) included complete bone marrow remission andA ^ 3, but improved, mIBG positive spots. The MAT regimens were BuMel (oral busulfan till 2006, 4_150 mg/m2 in 4 equal doses, or after 2006 intravenous use according to body weight and melphalan 140 mg/m__/day) and CEM (carboplatin ctn. infusion (4xAUC 4.1 mg/ml.min/day), etoposide ctn. infusion (4_338 mg/m__day or 4_200 mg/m__/ day*), melphalan (3_70 mg/m__/day or 3_60 mg/m__/day*. *reduced if GFR<100 ml/ min/1.73m__)). A minimum of 3_10E6 CD34/kgBW PBSC were requested. VOD prophylaxis included ursadiol, but not prophylactic defibrotide. Local control included surgery and radiotherapy of 21 Gy. A total of 598 high risk neuroblastoma patients were randomised (296 BuMel, 302 CEM). The median age at randomisation was 3 years (1-17.2).Results: A significant difference in EFS in favour of BuMel (3-years EFS 49% vs. 33%) was observed as well as for overall survival (3-years OS 60% vs. 48%, p¼0.004) with a median follow up of 3 years. This difference was mainly related to the relapse and progression incidence, which was significantly (p<0.001) lower with BuMel (48% vs. 60%). The severe toxicity rate up to day 100 (ICU and toxic deaths) was below 10%, but was significantly higher for CEM (p¼0.014). The acute toxic death rate was 3% for BuMel and 5% for CEM (NS). The acute MAT toxicity profile favours the BuMel regimen in spite of a total VOD incidence of 18% (grade 3:5%). Based on these results and following advice from the DMC, the randomisation was closed early.