Sodium thiosulfate as cisplatin otoprotectant in children: The challenge of when to use it.

Cisplatin is an essential chemotherapeutic agent for children treated with hepatic tumors, neuroblastoma, medulloblastoma, germ cell tumors, and osteosarcoma. Unfortunately, cisplatin-induced hearing loss (CIHL) occurs in about 50% of treated children,1-3 and is more frequent and more severe in those aged less than or equal to 5 years.4, 5 CIHL develops during the early stages of therapy,4 is irreversible,6 and in infants and young children causes speech and language deficits, impaired academic skills, and lower quality of life.7, 8 Prevention of CIHL has been variously approached by different disease clinical trial groups, some opting for individual cisplatin dose reduction or cisplatin replacement, but often without a solid evidence base. In other tumor groups, cisplatin dose reduction is being studied in randomized clinical trials. Either way, cisplatin remains an important treatment agent, particularly for children with high-risk and/or metastatic disease. Several otoprotective agents have been promising preclinically, but sodium thiosulfate (STS) has the best proven protection in clinical trials.9 STS induces antioxidant production to counter reactive oxygen species responsible for CIHL development.10 The licensing of STS (as PEDMARK) by the Food and Drug Administration (FDA), for use in children over 1 month as cisplatin otoprotectant, is a breakthrough, because the impact of permanent bilateral hearing loss is debilitating and costly. But the FDA has only approved STS for use in “localized, nonmetastatic” disease; so how has this current situation of limiting its application occurred? This commentary aims to help clinicians understand why STS has been restricted to localized cancer, how this situation has arisen, and guide the use of STS in the future. The efficacy of STS to reduce the incidence and severity of CIHL in children with cancer was studied in two randomized clinical trials: ACCL0431 (American Cancer Control Liver [study number] 0431) and SIOPEL 6 (International Society of Pediatric Oncology Epithelial Liver [group study number] 6).11, 12 Secondary end points were event-free survival (EFS) and overall survival (OS). In these two trials, STS was administered intravenously over 15 minutes, after an interval of 6 hours from the end of the cisplatin infusion. As demonstrated in pharmacokinetic studies, it is safe to administer STS at a time when active cisplatin is out of circulation.13 The trials were differently designed. ACCL0431 was a hearing trial, where children with any tumor were randomized to receive either cisplatin alone, or cisplatin and STS, and were stratified by age and cisplatin infusion time. They were not matched by tumor type, disease stage, clinical risk group or treatment, and disease prognostic factors were not collected. SIOPEL 6 was a randomized clinical trial for children with standard-risk hepatoblastoma stratified by clinical risk group and age and treated with six cycles of either single agent cisplatin alone, or cisplatin and STS, with delayed surgical resection. Patients were closely matched by tumor type, prognostic group, and treatment received. Both trials showed that STS significantly reduced hearing loss by ∼50% and as secondary end point; neither trial showed a statistically significant difference in EFS or OS between their treatment arms. However, in ACCL0431, a discernable, though statistically nonsignificant, trend toward lower OS in the STS-treated group prompted an unplanned, post hoc survival analysis using retrospective stratification of participants by initial extent of disease (localized or disseminated). No difference in EFS/OS was noted among patients retrospectively classified as having localized disease but among those retrospectively classified as having disseminated disease, a significantly lower OS was noted. But, in a subsequent comparison of expected outcomes, taken from historic published outcomes for a mixed disseminated tumor group, the control group with disseminated disease was found to have an unexpectedly favorable outcome, whereas the STS group with disseminated disease had the expected outcome.14, 15 Following the initial publication of ACCL0431, a misinterpretation about STS became widely spread; namely, that it was safe in localized disease but not safe in patients with metastatic disease. The more obvious conclusion is that the two arms of the ACCL0431 trial were not matched for tumor type or any prognostic factor. Crucially, the plausible reason for this observation was not tumor protection, but mismatched cohorts unmatched by tumor type, biology, pathology, stage, or treatment received. There are several hypotheses as to why the concern about STS, administered 6 hours after cisplatin, and thus not interfering with the antitumor activity of the chemotherapy, is unjustified. With regard to the ACCL0431 trial, these include: (a) overall tumor recurrence rate, which was equal in the STS group and the control group; (b) extent of disease was not a predetermined aim of the study; (c) the study design created large heterogeneity in tumor type and stage; (d) the number of each tumor type in the study was small; (e) the trial did not meet stopping rules that included tumor protection; and (f) most importantly, differences in treatment regimens and prognostic factors could have contributed, and likely did, to the retrospective deduction of tumor protection. Following publication of the ACCL0431 study, a widely held assumption has been made that STS administration can lead to reduced tumor control in children with disseminated disease, but without a biologically plausible rationale. It is clear from the mechanism of action of STS that if administered too early, while cisplatin is still active it is likely to have an adverse effect, but this should apply equally regardless of tumor stage. Thus a 6-hour delay after the cisplatin infusion before STS is administered is crucial. Fortunately, STS can still protect the inner ear from cisplatin's longer duration adverse effect on hearing, even when delayed by 6 hours. We need to design and support reliable, well-executed studies on appropriately delayed STS in children with metastatic disease for whom cisplatin is an essential therapy. In the meantime, we need to be brave enough to face the question together with each child and family: do we introduce appropriately delayed STS as a cisplatin otoprotectant in those most likely to benefit? And do we act in the best interest of the child and recognize the limitations of post hoc evidence and address the serious compromise of lifelong CIHL now that we have an effective antidote? Thanks to all participating families and patients in the above-mentioned clinical trials and to national charities for funding clinical trial research. Penelope R. Brock and James Geller are consultants for Fennec Pharmaceuticals. There are no other known conflicts of interest.