Diffuse astrocytoma, AYA-type, frequently MAPK-altered: report of 45 patients

Abstract

A putative molecular subtype of IDH-wildtype diffuse glioma with recurrent MAPK pathway alterations has recently been reported. By dimensionality reduction analysis of genome-wide methylation profiling, these tumors form a distinct methylation cluster of gliomas. Characterization of 47 tumors from 45 patients reveals that these gliomas are predominantly supratentorial in young adults, are highly infiltrative, and harbor mitogen-activated protein kinase (MAPK) pathway alterations with high rates of CDKN2A/2B deletion, PDGFRA amplification, MYCN amplification, NF1 variants, and BRAF alterations. The tumors' epigenetics are distinct from other adult and pediatric gliomas in the 2021 World Health Organization (WHO) classification. The histology of the gliomas most often demonstrates high-grade astrocytic features, but can be variable from tumor to tumor, as well as fall into a spectrum of histologic grades. Outcomes show considerable variability based on histologic grade and molecular features, supporting grading within this group of tumors to ensure optimal care choices on an individual patient basis. These unifying epigenetic, sequencing, and infiltrative astrocytic features allow the tumors to be considered diffuse astrocytoma, adolescent, and young adult-type, with MAPK alterations (DAYA).

Keywords: AYA; Brain tumors; Epigenetics; Glioma; HGG_B; MAPK; Methylation profiling.

Fig. 1

The methylation cluster DAYA is epigenetically distinct from tumor types in the 2021 WHO classification. a Unsupervised clustering of DNA methylation data using UMAP embedding (n = 10,000, most variable probes). Samples from this proposed group (DAYA, n = 48) were embedded with relevant selected CNS tumors from the in-house cohort: GBM_MES_TYP (n = 103), GBM_RTK_I (n = 93), GBM_RTK_II (n = 94), HGAP (n = 83), DAYA (n = 47), pedHGG_A (n = 10), pedHGG_B (n = 8), pedHGG_MYCN (n = 26), pedHGG_RTK1A (n = 52), pedHGG_RTK1B (n = 24), pedHGG_RTK1C (n = 34), pedHGG_RTK2A (n = 25), pedHGG_RTK2B (n = 20), ANTCON/GTAKA (6), GG (94), HPAP (27), PA_CORT (81), PA_MID (100), PA_PF (108), PLNTY (8), and PXA (121). All samples were classified by the DKFZ classifier with scores > 0.9. b Unsupervised hierarchical clustering of DNA methylation data from the same set of tumors with differentially methylated 5000 probes

Fig. 2

a Oncoplot of DAYA. Tumor locations are indicated at the top, followed by whether each tumor is a primary or recurrent glioma. The genomic findings highlight involvement of the MAPK pathway. Alteration frequencies are shown to the left of the oncoplot and represent the proportion of alterations in overall cohort; counts for each alteration are given to the right. Alteration types are indicated by color as shown in the legend to the right. Patient age is displayed in a barplot at the bottom of oncoplot, as well as via violin plot to the right. b Cohort level copy-number plots for DAYA, GBM_RTKII, HGAP, and pedHGG_RTK1A

Fig. 3

Histologic examination of DAYA demonstrates a spectrum of histologic grades and morphologic findings. a, b Some gliomas demonstrate lower cellularity and low proliferative activity, but a clear infiltrative pattern. Midline involvement is noted in a subset of cases as in (b), where the pencil fiber indicates the basal ganglia location. c A few cases demonstrated perinuclear haloes and microcalcifications, reminiscent of oligodendroglioma. d Epithelioid morphology with multinucleated cells was appreciated. e Although most cases were supratentorial, rare cases involved the cerebellum. f Many tumors appeared high-grade due to dense cellularity with atypical cells, but lacked necrosis and microvascular proliferation. g, h Progression of a glioma over 12 years demonstrated higher histologic grade at recurrence (h) compared to histology at initial presentation (g). i, j Additional examples of high-grade histology, including a case with the morphologic features of a GBM (j). Scale bars: 100 microns

Fig. 4

The survival plots (Kaplan–Meier curves). a Survival curve of the overall DAYA cohort depicted in the context of TCGA GBM and IDH-mutant astrocytoma (histologic grades 2, 3) cohorts. DAYA show intermediate outcomes. b When divided into histologically high-grade and histologically low-grade gliomas, survival is significantly different between the two groups (p = 0.0014). c When divided into gliomas with and without gene amplifications, survival is significantly different between the two groups (p = 0.0014). d Overlaying TCGA GBM and IDH-mutant astrocytoma curves with DAYA divided by histologic grade demonstrates that patients with high-grade histologic DAYA have outcomes similar to GBM