Pediatric Brain Tumors: Innovative Genomic Information Is Transforming the Diagnostic and Clinical Landscape

Abstract

Pediatric neuro-oncology has undergone an exciting and dramatic transformation during the past 5 years. This article summarizes data from collaborative group and institutional trials that have advanced the science of pediatric brain tumors and survival of patients with these tumors. Advanced genomic analysis of the entire spectrum of pediatric brain tumors has heralded an era in which stakeholders in the pediatric neuro-oncology community are being challenged to reconsider their current research and diagnostic and treatment strategies. The incorporation of this new information into the next-generation treatment protocols will unleash new challenges. This review succinctly summarizes the key advances in our understanding of the common pediatric brain tumors (ie, medulloblastoma, low- and high-grade gliomas, diffuse intrinsic pontine glioma, and ependymoma) and some selected rare tumors (ie, atypical teratoid/rhabdoid tumor and CNS primitive neuroectodermal tumor). The potential impact of this new information on future clinical protocols also is discussed. Cutting-edge genomics technologies and the information gained from such studies are facilitating the identification of molecularly defined subgroups within patients with particular pediatric brain tumors. The number of evaluable patients in each subgroup is small, particularly in the subgroups of rare diseases. Therefore, international collaboration will be crucial to draw meaningful conclusions about novel approaches to treating pediatric brain tumors.

Fig 1.

The genetic landscape of medulloblastoma. Recurrent genetic aberrations identified in medulloblastoma (derived from Northcott in 2012,^, Robinson et al, Pugh et al, Jones et al, and Northcott et al in 2014) averaged and displayed proportionally by height of terrain peaks. The figure reveals the unique subgroup-specific molecular aberration and highlights chromatin remodeling mutations as the unifying theme among all four medulloblastoma subgroups. Wingless (WNT) medulloblastoma (left; blue icy landscape), the most molecularly homogenous group, consists of CTNNB1 mutations in 85%, monosomy 6 in 80%, DDX3X mutation in 50%, TP53 mutation in 13%, and mutations in chromatin remodeling genes in 49.5% (composed of mutations in SMARCA4 [25%], MLL2 [12.5%], CREBBP [6%], TRAPP [3%], and MED13 [3%]). For the chromatin remodeling peaks (darker colored shading), only the most commonly mutated gene is labeled. Sonic hedgehog (SHH) medulloblastoma (bottom; red volcanic landscape) consists of PTCH1 mutation/deletion in 29%, TP53 mutation in 18%, DDX3X mutation in 11%, GLI2 amplification/mutation in 8%, MYCN amplification in 6%, SUFU mutation in 6%, SMO mutation in 3%, PTEN deletion in 2.5%, MYCL1 amplification in 2%, CDK6 amplification in 1%, MYCC amplification in 0.7%, and mutations in chromatin remodeling genes in 21% (composed of mutations in MLL2 [12%], BCOR [3%], LBD1 [3%], NCOR2 [1.5%], and SMARCA4 [1.5%]). Group 3 medulloblastoma (top; yellow desert rocky terrain) is characterized by GFI1/1B structural variants (eg, inversions, duplications) in 41%, isochromosome (iso) 17q in 26%, transforming growth factor (TGF) -β signaling in 20%, MYCC amplification in 17%, PVT1 alterations in 12%, OTX2 amplification in 8%, MYCN amplification in 4%, DDX3X mutation in 3%, CDK6 amplification in 1%, and mutations in chromatin remodelling genes in 28.5% (composed of mutations in SMARCA4 [10.5%], other KDM family members [5%], MLL2 [4%], KDMA6A [3%], GPS2 [3%], MLL3 [1%], CREBBP [1%], and CHD7 [1%]). Group 4 medulloblastoma (right; green forest mountain terrain) is characterized by iso 17q in 80%, GFI1/1B structural variants in 10%, SNCAIP tandem duplications in 10%, OTX2 amplification in 5.5%, MYCN amplification in 5%, CDK6 amplification in 5%, TP53 mutation in 1%, MYCC amplification in 1%, and mutations in chromatin remodeling genes in 30% (composed of mutations in KDMA6A [13%], other KDM family members [4%], MLL3 [3%], CHD7 [3%], ZMYM3 [3%], MLL2 [2%], GPS2 [1%], and BCOR [1%]).

Fig 2.

Subgroups of pediatric high-grade glioma that are based on German Cancer Research Center (DKFZ) methylation, age at onset, tumor location, oncogenic drives, gene expression, and median survival. IDH, isocitrate dehydrogenase; PXA, pleomorphic xanthoastrocytoma; RTK-I, receptor tyrosine kinase (subgroup 1).

Fig 3.

BRAF mutations and fusions by tumor histology and tumor location in pediatric low-grade gliomas.

Fig 4.

Several subtypes of ependymomas, including WHO grades 1 to 3 disease within all three compartments of the CNS—supratentorial (ST), posterior fossa (PF), and spinal (SP)—are illustrated. RELA-positive ependymomas, including YAP1 fusion–positive ependymomas and subependymomas, arise within the ST region of the brain. Both fusion-positive subtypes display histopathologic features of WHO grades 2 and 3 ependymomas. In the PF, the majority of ependymomas belong to subtype group A, and group B tumors are more infrequent. Both subtypes display the histologic pattern of anaplastic and WHO grade 2 ependymomas; in contrast, subependymomas can be classified as WHO grade 1. SP tumors are diagnosed as classic ependymomas that are WHO grade 2 or 3; myxopapillary ependymoma and spinal subependymomas are WHO grade 1. In children, group A and RELA-positive tumors are diagnosed most often and are associated with poor overall survival. SNV, single nucleotide variant.