BANDICOOT: An Adaptive Platform Trial Designed to Improve the Complications, Cost-effectiveness and Health Outcomes for Children Receiving a Stem Cell Transplant

Background: In children with >120 rare diseases, haematopoietic stem cell transplant (HSCT) provides a medical "reset" for the body, replacing diseased or dysfunctional bone marrow with healthy donor-derived stem cells following high-dose chemotherapy and/or radiotherapy. However, severe and fatal complications are common with HSCT. There has been a lack of properly conducted clinical trials to decrease mortality and morbidity. Traditional randomised controlled trials (RCTs) have several critical limitations in children undergoing HSCT, including population heterogeneity, restrictive eligibility criteria and slow enrolment. Adaptive platform trials (APTs) may overcome these limitations through enhanced trial efficiency by sharing a control group, reducing sample size and allowing continuous learning from accumulating data. APTs also allow simultaneous evaluation of distinct interventions at different timepoints and in multiple subgroups of participants, facilitating tailored approaches across heterogeneous populations. When an intervention proves superior, it becomes the new standard of care, allowing additional interventions to be introduced.

To improve outcomes, we have developed an international APT - BANDICOOT. This trial will continuously enrol children and adolescents receiving HSCT and allow the assessment of multiple novel interventions simultaneously. The goal is to accelerate research findings, reduce duplication of efforts, and improve patient outcomes.

Objectives: The primary objective of BANDICOOT is to determine the effectiveness of a range of interventions to improve HSCT outcomes for children and adolescents.

The secondary objectives include:

• Assessing the cost-effectiveness of trial interventions
• Assessing the safety of a range of interventions to improve HCT outcomes
• Collection of a core data set for participants consenting to the platform regardless of domain eligibility.

Study design: BANDICOOT is a prospective, pragmatic, adaptive platform trial with interventions organised into domains. Domains may be open-label or blinded.

Study population: The trial population will be children aged 1-week old to 18 years old who are receiving an HSCT.

Trial outcomes: The primary outcome is an ordinal scale of HSCT outcomes based on organ support, viraemia, immune reconstitution and relapse status censored at Day 100 post HSCT. The selection and grading of components within this ordinal endpoint was informed by a formal endpoint development process, described in detail by Walker et al, 2025 (see References).

Interventions: Multiple interventions will be evaluated in BANDICOOT across multiple treatment modalities (domains). New interventions will be added over time, and interventions may be dropped for futility or included in standard care as the study progresses. The details of the interventions will be provided in separate clinicaltrials.gov Study Records, linked to this Master record.

Abbreviated methods: Inferences in this trial will be based on a Bayesian statistical model. The primary outcome will be analysed using a multinomial model with a cumulative logistic link, which is an extension of a binary logistic model to account for ordinal outcomes with more than two categories, and is commonly known as the 'proportional odds' model. Secondary outcomes will be analysed with parametric models specific to the type of outcome (e.g., the Bernoulli model with a logistic link for binary endpoints).