Comparative evidence on the independent contribution of granulocyte-macrophage colony-stimulating factor to anti-GD2 immunotherapy in neuroblastoma: a systematic review and gap analysis

Objective To assess comparative evidence on the independent contribution of GM-CSF to the efficacy and safety of anti-GD2 immunotherapy in neuroblastoma and to identify evidence gaps. Methods This systematic review was conducted in accordance with the PRISMA guidelines in February 2026. MEDLINE (via PubMed), EMBASE, and the Cochrane Library were searched to identify comparative studies evaluating anti-GD2 immunotherapy with and without GM-CSF in patients with neuroblastoma. In addition, the websites of key oncology societies were searched. Results No study directly compared otherwise equivalent anti-GD2 regimens that differed only in the inclusion of GM-CSF. Two studies met the broader comparative eligibility criterion and provided only indirect regimen-level information. Both evaluated older, non-licensed murine anti-GD2 antibodies (m3F8 and 14.G2a), and neither evaluated GM-CSF with currently licensed anti-GD2 antibodies. In the retrospective study by Cheung et al., survival outcomes differed across treatment eras; however, GM-CSF was co-administered with 13-cis-retinoic acid and treatment groups also differed with respect to prior stem cell transplantation, induction therapy, baseline characteristics, follow-up duration and supportive care. Therefore, the observed survival differences cannot be attributed to GM-CSF. Toxicity comparisons in this study were also limited by differences in toxicity ascertainment across treatment eras. The prospective phase I/Ib study by Frost et al. was primarily designed to evaluate dose finding and toxicity, included very small and heterogeneous treatment groups, and did not isolate GM-CSF as an independent treatment variable. No eligible comparative study was identified for dinutuximab, dinutuximab beta or naxitamab. Conclusion Available comparative evidence is insufficient to determine the independent contribution of GM-CSF to the efficacy or safety of anti-GD2 immunotherapy in neuroblastoma. These findings should not be used to justify modification of licensed product-specific anti-GD2 regimens. A randomized trial holding the anti-GD2 backbone and relevant co-interventions constant is needed to isolate the effect of GM-CSF. Systematic Review Registration https://www.crd.york.ac.uk/PROSPERO/ , identifier CRD420251116911.