Improved survival from childhood cancer has been achieved at the cost of long-term comorbidities that reduce quality of life. The chemotherapy agent cisplatin is essential for treatment of many paediatric and adolescent malignancies but in a large proportion of patients results in cisplatin-induced hearing loss, a debilitating and permanent late effect, with consequences for neurocognition, academic achievement, employment, and social and psychological outcomes. 1, 2 In The Lancet Oncology, we previously reported the results of the Children’s Oncology Group study ACCL0431, an international, randomised, controlled trial of sodium thiosulfate for prevention of cisplatin-induced hearing loss in children and adolescents. 3 Participants were randomly assigned to sodium thiosulfate or observation in addition to their planned cisplatin-containing chemotherapy. Patients with any type or stage of cancer treated with cisplatin were eligible. In ACCL0431, we found compelling evidence for hearing protection from sodium thiosulfate, a finding replicated by the concurrent International Childhood Liver Tumours Strategy Group 6 trial (SIOPEL-6) in standard-risk (ie, localised) hepatoblastoma. 4 As described in the initial ACCL0431 report, survival was a prespecified secondary outcome to assess for potential interference with chemotherapy efficacy by sodium thiosulfate. Among all participants in aggregate, there was no significant difference in overall survival by randomised group at a relatively early median follow-up of 3· 5 years (IQR 3· 0–4· 5; relative hazard ratio 2· 03 [95% CI 0· 93–4· 44]). However, a nonsignificant trend towards lower overall survival among patients treated with sodium thiosulfate than in the control group prompted an unplanned survival analysis using post-hoc stratification of participants by extent of disease (localised or disseminated [ie, tumour identified in sites distant from the primary tumour at original diagnosis]). Among patients deemed to have localised disease, there was no difference in 3-year overall survival for observation (89%[95% CI 74–96]) versus treatment with sodium thiosulfate (83%[66–92]; log rank p= 0· 88), but among those deemed to have dissemi nated disease, 3-year overall survival was significantly lower for treatment with sodium thiosulfate (45%[95% CI 23–65]) than for observation (84%[62–94]; relative hazard ratio 4· 10 [95% CI 1· 30–12· 97], log rank p= 0· 0090; figure A). 3 Due to differing eligibility criteria between the two sodium thiosulfate trials (ACCL0431 and SIOPEL-6), survival data for disseminated disease were available only from ACCL0431. These findings resulted in a safety concern for children with disseminated disease, currently precluding routine use of sodium thiosulfate in this population. 5 realisation of this project; Francisco González-Romero for his contribution to the dissemination of the patient information sheets; and Venkata Pradeep Babu Koyyala for coordination of the Indian working group. RC received the COST Action 18122 virtual mobility grant to coordinate this project. CB received a Lord Kelvin Adam Readership from the University of Glasgow (Glasgow, UK). This article is based upon work from the COST Action European Cholangiocarcinoma Network, supported by COST, a funding agency for research and innovation networks.