Abstract B019: Mapping clonal architecture and evolution in pediatric brain cancers

Abstract Brain cancers are the most common solid tumors in children and adolescents and remain the leading cause of disease-related mortality in these populations. Despite their clinical burden, therapeutic progress has been limited, in part due to extensive intra- and inter-tumor heterogeneity. The clonal architecture and evolutionary dynamics that underlie pediatric brain tumors remain poorly characterized, yet are critical determinants of disease progression, treatment resistance, and clinical outcome. To address this knowledge gap, we investigated clonal evolution at the population level by analyzing longitudinal whole-genome sequencing data from matched primary and secondary tumors across 27 brain cancer types in 142 pediatric patients from the Children’s Brain Tumor Network. For each patient, we inferred tumor clonal composition and reconstructed a patient-specific phylogeny representing subclone relationships. We then quantified pairwise differences between phylogenies using a regularized Gromov-Hausdorff distance, embedded them into a shared metric space, and assessed their associations with molecular and clinical features. Tumor phylogenetic structure was significantly associated with mutational burden, mutational signatures, subtype classification, and patient's survival status. Notably, survival outcome, specific mutational signatures and microenvironment immune infiltration remained independently associated with tumor clonal architecture after adjusting for mutational burden, tumor subtype, and anatomical location, underscoring the clinical relevance of phylogenetic readouts. Temporal analysis of mutation signatures revealed that pediatric brain tumor genetic architecture is shaped by distinct mutational processes acting at different stages of tumor evolution. Early truncal mutations were enriched for age-related signature, whereas subclonal mutations in primary tumors showed increased contribution from homologous recombination repair deficiency- and hypermutation activity-related signatures. In contrast, mutations exclusive to secondary tumors were more dominated by late-stage replication error-associated signature. Finally, comparisons of matched primary and longitudinal phylogenetic trees demonstrated that patients with poor clinical outcomes, including deceased patients and those with aggressive disease, exhibited significantly greater evolutionary divergence, indicative of accelerated tumor evolution dynamics and reduced predictability of complete evolutionary trajectories from primary tumor sampling alone. Together, these findings demonstrate that pediatric brain tumor evolution is shaped by both cancer type and stage-specific mutational processes, and that patients with adverse clinical outcomes follow distinct evolutionary trajectories that are not fully explained by tumor subtype alone. By providing a population-level view of clonal evolution across pediatric brain cancers, our study highlights the value of tumor clonal phylogenetic dynamics for understanding tumor etiology, disease progression, and patient prognosis. Citation Format: Minh Anh Nguyen, Pablo G. Cámara. Mapping clonal architecture and evolution in pediatric brain cancers abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Brain Cancer; 2026 Mar 23-25; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86 (6Suppl): Abstract nr B019.