Immunogenomic classification reveals prognostic immune signatures in pediatric solid and hematological tumors

The immune features in pediatric tumor are poorly explored. To characterize immune features of pediatric cancer, we performed an immunogenomic analysis of public database (TARGET) for pediatric solid tumor (PST) (n = 423) and pediatric hematological tumor (PHT) (n = 2302). We clustered PST and PHT samples into 5 subtypes (S1-S5) and 4 subtypes (H1-H4), respectively, based on immune features. In the PST cohort, cluster S1 with elevated expression of Wound_CSR (fibroblast core serum response in wound healing) and B cell signatures exhibited the worst overall survival. Conversely, cluster S4 (HR = 0.378, 95% CI: 0.24–0.59, P-value < 0.001) with down-regulated expression of these features was associated with prolonged survival. We also validated the prognostic significance of the S4 immune subtype in an independent neuroblastoma cohort from the ZJUCH (n = 127), which demonstrated favorable patient outcomes. In the PHT cohort, we observed that the relationships between immune clusters and prognosis differed between FLT3-ITD mutation-positive AML (AML-1) and FLT3-ITD mutation-negative AML (AML-2). In AML-1, cluster H2 featured upregulated infiltration of neutrophils, monocytes and antigen processing signatures, possibly leading to the worst overall survival. While in AML-2, cluster H2 exhibited a favorable outcome. The study highlights the potential of immune features as biomarkers for prognosis and treatment planning in pediatric cancers and provides novel insights into their immunological landscape.