The inability to control the primary site, either as an isolated event or in concert with regional nodal or metastatic relapse, was the overwhelmingly greatest risk of treatment failure: nearly 60% of all events consisted of recurrence or progression of the primary tumour (about 30% of events in each group were due to isolated local relapse, local progression accounted for more than 16% of events, and an additional 13% of events were suboptimal response of the primary tumour at week 9); local recurrence, alone or in combination with regional nodal relapse or metastatic relapse accounted fully for 72% of all events; and an additional 5% of events occurred from recurrence of disease in nearby lymph nodes. Fewer than one-in-five events was an isolated metastatic recurrence (15 16% events in the IVA group vs 17 20% in the IVA plus doxorubicin group), the primary endpoint that adding doxorubicin would be most likely to affect favourably—but the proportion of patients at-risk for this type of event was too small to see any improvement in outcome even if the magnitude of benefit was large. The question of the EpSSG RMS 2005 trial was whether four cycles of doxorubicin given before definitive local control would improve outcomes in children with high-risk localised tumours. ...the factors associated with treatment failure—primarily local treatment failure, including primary radioresistance—are better understood, it is possible that future clinical trials in children with rhabdomyosarcoma focusing primarily or exclusively on chemotherapy questions are likely to also fail to improve outcome.