Background: Evaluation of prognostic factors in high risk stage 4 neuroblastoma (HRNBL) patients (pts) treated with busulfan-melphalan (BUMEL) within HR-NBL1/SIOPEN. Methods: Between 2002-2009 (prior antiGD2 immunotherapy) BUMEL was given to 475 pts (275 males) in 149 centres / 20 countries. Pts < 1year (yr) had MYCN amplification (MNA). After induction (COJEC ± 2-4 TVD) BUMEL/SCR (stem cell reinfusion) was given once pts achieved at least PR. Local control aimed at gross surgical resection (achieved in 76%) and for radiotherapy (21Gy) to the primary tumour site. Maintenance was 13-cis retinoid acid only. The median age at diagnosis was 3 yrs (1 month-19 yrs). The median observation time is 7.4 yrs. Outcomes are reported as 5-yr EFS rates. Results: EFS was 0.64±0.12 for age < 1yr (17 pts), 0.62±0.08 for 1-1.5 yrs (42 pts), 0.40±0.03 for 1.5-5 yrs (317 pts) and 0.20±0.04 for pts > 5yrs (99 pts)(p = 0.0001). EFS was better with BUMEL/SCR given ≤ 240days (421pts) after diagnosis (0.41±0.02) than for 54 pts taking longer (0.16±0.05; p < 0.0001). Outcome was not different in 221 pts randomised for BUMEL (221/475). Pts with TVD have a significantly different EFS (0.41±0.03 (214pts) vs.106 pts +TVD: 0.30±0.05) (p = 0.026). CR pts prior to BUMEL (172 pts; 36 pts +TVD) had an EFS of 0.45±0.04, whilst EFS was 0.38±0.04 for VGPR pts (165 pts; 36pts +TVD) and 0.31 ±0.04 for PR pts (138 pts; 34pts+TVD) (p = 0.004). Earlier CR of pts without TVD appears superior (0.48±0.04 vs. 0.33±0.08 for CR after TVD: NS). EFS in PR pts prior BUMEL with ≤ 3 mIBG avid skeletal spots was better (0.38±0.05) than with more spots (0.13±0.06; p = 0.014). Also involvement of only one metastatic compartment results in significantly better EFS (0.50±0.07 vs. 0.37±0.02 for multiple compartments, p = 0.037). Severe toxicity rate was 7% (ICU, toxic deaths), the VOD rate 24% (grade 3:4%). Conclusions: Multivariable analysis (stage 4 patients > 1 yr and BUMEL ≤ 240days) reveals as independent unfavourable prognostic factors: age group 5-10yrs (Hazard-Ratio 2.77, p < 0.0001), age group 1.5-5yrs (Hazard-Ratio 1.73; p = 0.043) and PR prior BUMEL (HR = 1.46; p = 0.008). These findings enable guidance to future altered treatment approaches. Clinical trial information: NCT01704716