Accurate pathological diagnosis is crucial for optimal management of patients with cancer. For the approximately 100 known tumour types of the central nervous system, standardization of the diagnostic process has been shown to be particularly challenging-with substantial inter-observer variability in the histopathological diagnosis of many tumour types. Here we present a comprehensive approach for the DNA methylation-based classification of central nervous system tumours across all entities and age groups, and demonstrate its application in a routine diagnostic setting. We show that the availability of this method may have a substantial impact on diagnostic precision compared to standard methods, resulting in a change of diagnosis in up to 12% of prospective cases. For broader accessibility, we have designed a free online classifier tool, the use of which does not require any additional onsite data processing. Our results provide a blueprint for the generation of machine-learning-based tumour classifiers across other cancer entities, with the potential to fundamentally transform tumour pathology.

Histological grading of choroid plexus tumors (CPTs) remains the best prognostic tool to distinguish between aggressive choroid plexus carcinoma (CPC) and the more benign choroid plexus papilloma (CPP) or atypical choroid plexus papilloma (aCPP); however, these distinctions can be challenging. Standard treatment of CPC is very aggressive and often leads to severe damage to the young child's brain. Therefore, it is crucial to distinguish between CPC and less aggressive entities (CPP or aCPP) to avoid unnecessary exposure of the young patient to neurotoxic therapy. To better stratify CPTs, we utilized DNA methylation (DNAm) to identify prognostic epigenetic biomarkers for CPCs.

Extra-cranial malignant rhabdoid tumors (MRTs) and cranial atypical teratoid RTs (ATRTs) are heterogeneous pediatric cancers driven primarily by SMARCB1 loss. To understand the genome-wide molecular relationships between MRTs and ATRTs, we analyze multi-omics data from 140 MRTs and 161 ATRTs. We detect similarities between the MYC subgroup of ATRTs (ATRT-MYC) and extra-cranial MRTs, including global DNA hypomethylation and overexpression of HOX genes and genes involved in mesenchymal development, distinguishing them from other ATRT subgroups that express neural-like features. We identify five DNA methylation subgroups associated with anatomical sites and SMARCB1 mutation patterns. Groups 1, 3, and 4 exhibit cytotoxic T cell infiltration and expression of immune checkpoint regulators, consistent with a potential role for immunotherapy in rhabdoid tumor patients.

Sarcomas are malignant soft tissue and bone tumours affecting adults, adolescents and children. They represent a morphologically heterogeneous class of tumours and some entities lack defining histopathological features. Therefore, the diagnosis of sarcomas is burdened with a high inter-observer variability and misclassification rate. Here, we demonstrate classification of soft tissue and bone tumours using a machine learning classifier algorithm based on array-generated DNA methylation data. This sarcoma classifier is trained using a dataset of 1077 methylation profiles from comprehensively pre-characterized cases comprising 62 tumour methylation classes constituting a broad range of soft tissue and bone sarcoma subtypes across the entire age spectrum. The performance is validated in a cohort of 428 sarcomatous tumours, of which 322 cases were classified by the sarcoma classifier. Our results demonstrate the potential of the DNA methylation-based sarcoma classification for research and future diagnostic applications.

Clear cell meningioma represents an uncommon variant of meningioma that typically affects children and young adults. Although an enrichment of loss-of-function mutations in the SMARCE1 gene has been reported for this subtype, comprehensive molecular investigations are lacking. Here we describe a molecularly distinct subset of tumors (n = 31), initially identified through genome-wide DNA methylation screening among a cohort of 3093 meningiomas, of which most were diagnosed histologically as clear cell meningioma. This cohort was further supplemented by an additional 11 histologically diagnosed clear cell meningiomas for analysis (n = 42). Targeted DNA sequencing revealed SMARCE1 mutations in 33/34 analyzed samples, accompanied by a nuclear loss of expression determined via immunohistochemistry and a decreased SMARCE1 transcript expression in the tumor cells. Analysis of time to progression or recurrence of patients within the clear cell meningioma group (n = 14) in comparison to those with meningioma WHO grade 2 (n = 220) revealed a similar outcome and support the assignment of WHO grade 2 to these tumors. Our findings indicate the existence of a highly distinct epigenetic signature of clear cell meningiomas, separate from all other variants of meningiomas, with recurrent mutations in the SMARCE1 gene. This suggests that these tumors may arise from a different precursor cell population than the broad spectrum of the other meningioma subtypes.

Glioblastoma IDH-wildtype presents with a wide histological spectrum. Some features are so distinctive that they are considered as separate histological variants or patterns for the purpose of classification. However, these usually lack defined (epi-)genetic alterations or profiles correlating with this histology. Here, we describe a molecular subtype with overlap to the unique histological pattern of glioblastoma with primitive neuronal component. Our cohort consists of 63 IDH-wildtype glioblastomas that harbor a characteristic DNA methylation profile. Median age at diagnosis was 59.5 years. Copy-number variations and genetic sequencing revealed frequent alterations in TP53, RB1 and PTEN, with fewer gains of chromosome 7 and homozygous CDKN2A/B deletions than usually described for IDH-wildtype glioblastoma. Gains of chromosome 1 were detected in more than half of the cases. A poorly differentiated phenotype with frequent absence of GFAP expression, high proliferation index and strong staining for p53 and TTF1 often caused misleading histological classification as carcinoma metastasis or primitive neuroectodermal tumor. Clinically, many patients presented with leptomeningeal dissemination and spinal metastasis. Outcome was poor with a median overall survival of only 12 months. Overall, we describe a new molecular subtype of IDH-wildtype glioblastoma with a distinct histological appearance and genetic signature.

Histone H3.3 lysine-to-methionine substitutions K27M and K36M impair the deposition of opposing chromatin marks, H3K27me3/me2 and H3K36me3/me2. We show that these mutations induce hypotrophic and disorganized eyes in Drosophila eye primordia. Restriction of H3K27me3 spread in H3.3K27M and its redistribution in H3.3K36M result in transcriptional deregulation of PRC2-targeted eye development and of piRNA biogenesis genes, including krimp. Notably, both mutants promote redistribution of H3K36me2 away from repetitive regions into active genes, which associate with retrotransposon de-repression in eye discs. Aberrant expression of krimp represses LINE retrotransposons but does not contribute to the eye phenotype. Depletion of H3K36me2 methyltransferase ash1 in H3.3K27M, and of PRC2 component E(z) in H3.3K36M, restores the expression of eye developmental genes and normal eye growth, showing that redistribution of antagonistic marks contributes to K-to-M pathogenesis. Our results implicate a novel function for H3K36me2 and showcase convergent downstream effects of oncohistones that target opposing epigenetic marks.

The large diversity of central nervous system (CNS) tumor types in children and adolescents results in disparate patient outcomes and renders accurate diagnosis challenging. In this study, we prospectively integrated DNA methylation profiling and targeted gene panel sequencing with blinded neuropathological reference diagnostics for a population-based cohort of more than 1,200 newly diagnosed pediatric patients with CNS tumors, to assess their utility in routine neuropathology. We show that the multi-omic integration increased diagnostic accuracy in a substantial proportion of patients through annotation to a refining DNA methylation class (50%), detection of diagnostic or therapeutically relevant genetic alterations (47%) or identification of cancer predisposition syndromes (10%). Discrepant results by neuropathological WHO-based and DNA methylation-based classification (30%) were enriched in histological high-grade gliomas, implicating relevance for current clinical patient management in 5% of all patients. Follow-up (median 2.5 years) suggests improved survival for patients with histological high-grade gliomas displaying lower-grade molecular profiles. These results provide preliminary evidence of the utility of integrating multi-omics in neuropathology for pediatric neuro-oncology.

Poorly differentiated chordoma (PDC) is a recently recognized subtype of chordoma characterized by expression of the embryonic transcription factor, brachyury, and loss of INI1. PDC primarily affects children and is associated with a poor prognosis and limited treatment options. Here we describe the molecular and immune tumour microenvironment profiles of two paediatric PDCs produced using whole-genome, transcriptome and whole-genome bisulfite sequencing (WGBS) and multiplex immunohistochemistry. Our analyses revealed the presence of tumour-associated immune cells, including CD8+ T cells, and expression of the immune checkpoint protein, PD-L1, in both patient samples. Molecular profiling provided the rationale for immune checkpoint inhibitor (ICI) therapy, which resulted in a clinical and radiographic response. A dominant T cell receptor (TCR) clone specific for a brachyury peptide-MHC complex was identified from bulk RNA sequencing, suggesting that targeting of the brachyury tumour antigen by tumour-associated T cells may underlie this clinical response to ICI. Correlative analysis with rhabdoid tumours, another INI1-deficient paediatric malignancy, suggests that a subset of tumours may share common immune phenotypes, indicating the potential for a therapeutically targetable subgroup of challenging paediatric cancers.

Myxopapillary ependymoma (MPE) is a heterogeneous disease regarding histopathology and outcome. The underlying molecular biology is poorly understood, and markers that reliably predict the patients' clinical course are unknown.

Atypical teratoid/rhabdoid tumor (ATRT) is an aggressive central nervous system tumor characterized by loss of SMARCB1/INI1 protein expression and comprises three distinct molecular groups, ATRT-TYR, ATRT-MYC and ATRT-SHH. ATRT-SHH represents the largest molecular group and is heterogeneous with regard to age, tumor location and epigenetic profile. We, therefore, aimed to investigate if heterogeneity within ATRT-SHH might also have biological and clinical importance. Consensus clustering of DNA methylation profiles and confirmatory t-SNE analysis of 65 ATRT-SHH yielded three robust molecular subgroups, i.e., SHH-1A, SHH-1B and SHH-2. These subgroups differed by median age of onset (SHH-1A: 18 months, SHH-1B: 107 months, SHH-2: 13 months) and tumor location (SHH-1A: 88% supratentorial; SHH-1B: 85% supratentorial; SHH-2: 93% infratentorial, often extending to the pineal region). Subgroups showed comparable SMARCB1 mutational profiles, but pathogenic/likely pathogenic SMARCB1 germline variants were over-represented in SHH-2 (63%) as compared to SHH-1A (20%) and SHH-1B (0%). Protein expression of proneural marker ASCL1 (enriched in SHH-1B) and glial markers OLIG2 and GFAP (absent in SHH-2) as well as global mRNA expression patterns differed, but all subgroups were characterized by overexpression of SHH as well as Notch pathway members. In a Drosophila model, knockdown of Snr1 (the fly homologue of SMARCB1) in hedgehog activated cells not only altered hedgehog signaling, but also caused aberrant Notch signaling and formation of tumor-like structures. Finally, on survival analysis, molecular subgroup and age of onset (but not ASCL1 staining status) were independently associated with overall survival, older patients (> 3 years) harboring SHH-1B experiencing relatively favorable outcome. In conclusion, ATRT-SHH comprises three subgroups characterized by SHH and Notch pathway activation, but divergent molecular and clinical features. Our data suggest that molecular subgrouping of ATRT-SHH has prognostic relevance and might aid to stratify patients within future clinical trials.

Atypical teratoid/rhabdoid tumors (AT/RT) are the most common malignant brain tumors manifesting in infancy. They split into four molecular types. The major three (AT/RT-SHH, AT/RT-TYR, and AT/RT-MYC) all carry mutations in SMARCB1, the fourth quantitatively smaller type is characterized by SMARCA4 mutations (AT/RT-SMARCA4). Molecular characteristics of disease recurrence or metastatic spread, which go along with a particularly dismal outcome, are currently unclear. Here, we investigated tumor tissue from 26 patients affected by AT/RT to identify signatures of recurrences in comparison with matched primary tumor samples. Microscopically, AT/RT recurrences demonstrated a loss of architecture and significantly enhanced mitotic activity as compared to their related primary tumors. Based on DNA methylation profiling, primary tumor and related recurrence were grossly similar, but three out of 26 tumors belonged to a different molecular type or subtype after second surgery compared to related primary lesions. Copy number variations (CNVs) differed in six cases, showing novel gains on chromosome 1q or losses of chromosome 10 in recurrences as the most frequent alterations. To consolidate these observations, our cohort was combined with a data set of unmatched primary and recurrent AT/RT, which demonstrated chromosome 1q gain and 10 loss in 18% (n = 7) and 11% (n = 4) of the recurrences (n = 38) as compared to 7% (n = 3) and 0% (n = 0) in the primary tumors (n = 44), respectively. Similar to the observations made by DNA methylation profiling, RNA sequencing of our cohort revealed AT/RT primary tumors and matched recurrences clustering closely together. However, a number of genes showed significantly altered expression in AT/RT-SHH recurrences. Many of them are known tumor driving growth factors, involved in embryonal development and tumorigenesis, or are cell-cycle-associated. Overall, our work identifies subtle molecular changes that occur in the course of the disease and that may help define novel therapeutic targets for AT/RT recurrences.

Pituicytoma (PITUI), granular cell tumor (GCT), and spindle cell oncocytoma (SCO) are rare tumors of the posterior pituitary. Histologically, they may be challenging to distinguish and have been proposed to represent a histological spectrum of a single entity. We performed targeted next-generation sequencing, DNA methylation profiling, and copy number analysis on 47 tumors (14 PITUI; 12 GCT; 21 SCO) to investigate molecular features and explore possibilities of clinically meaningful tumor subclassification. We detected two main epigenomic subgroups by unsupervised clustering of DNA methylation data, though the overall methylation differences were subtle. The largest group (n = 23) contained most PITUIs and a subset of SCOs and was enriched for pathogenic mutations within genes in the MAPK/PI3K pathways (12/17 [71%] of sequenced tumors: FGFR1 (3), HRAS (3), BRAF (2), NF1 (2), CBL (1), MAP2K2 (1), PTEN (1)) and two with accompanying TERT promoter mutation. The second group (n = 16) contained most GCTs and a subset of SCOs, all of which mostly lacked identifiable genetic drivers. Outcome analysis demonstrated that the presence of chromosomal imbalances was significantly associated with reduced progression-free survival especially within the combined PITUI and SCO group (p = 0.031). In summary, we observed only subtle DNA methylation differences between posterior pituitary tumors, indicating that these tumors may be best classified as subtypes of a single entity. Nevertheless, our data indicate differences in mutation patterns and clinical outcome. For a clinically meaningful subclassification, we propose a combined histo-molecular approach into three subtypes: one subtype is defined by granular cell histology, scarcity of identifiable oncogenic mutations, and favorable outcome. The other two subtypes have either SCO or PITUI histology but are segregated by chromosomal copy number profile into a favorable group (no copy number changes) and a less favorable group (copy number imbalances present). Both of the latter groups have recurrent MAPK/PI3K genetic alterations that represent potential therapeutic targets.

Atypical teratoid/rhabdoid tumor (AT/RT) is a highly malignant tumor that may not only contain rhabdoid tumor cells but also poorly differentiated small-round-blue cells as well as areas with mesenchymal or epithelial differentiation. Little is known on factors associated with histopathological diversity. Recent studies demonstrated three molecular subgroups of AT/RT, namely ATRT-TYR, ATRT-SHH, and ATRT-MYC. We thus aimed to investigate if morphological patterns might be related to molecular subgroup status. Hematoxylin-eosin stained sections of 114 AT/RT with known molecular subgroup status were digitalized and independently categorized by nine blinded observers into four morphological categories, that is, "rhabdoid," "small-round-blue," "epithelial," and "mesenchymal." The series comprised 48 ATRT-SHH, 40 ATRT-TYR, and 26 ATRT-MYC tumors. Inter-observer agreement was moderate but significant (Fleiss' kappa = 0.47; 95% C.I. 0.41-0.53; p < 0.001) and there was a highly significant overall association between morphological categories and molecular subgroups for each of the nine observers (p < 0.0001). Specifically, the category "epithelial" was found to be over-represented in ATRT-TYR (p < 0.000001) and the category "small-round-blue" to be over-represented in ATRT-SHH (p < 0.01). The majority of ATRT-MYC was categorized as "mesenchymal" or "rhabdoid," but this association was less compelling. The specificity of the category "epithelial" for ATRT-TYR was highest and accounted for 97% (range: 88-99%) whereas sensitivity was low [49% (range: 35%-63%)]. In line with these findings, cytokeratin-positivity was highly overrepresented in ATRT-TYR. In conclusion, morphological features of AT/RT might reflect molecular alterations and may also provide a first hint on molecular subgroup status, which will need to be confirmed by DNA methylation profiling.

Loss of nuclear SMARCB1 (INI1/hSNF5/BAF47) protein expression due to biallelic mutations of the SMARCB1 tumor suppressor gene is a hallmark of atypical teratoid/rhabdoid tumors (ATRT), but the presence of cytoplasmic SMARCB1 protein in these tumors has not yet been described. In a series of 102 primary ATRT, distinct cytoplasmic SMARCB1 staining on immunohistochemistry was encountered in 19 cases (19%) and was highly over-represented in cases showing pathogenic sequence variants leading to truncation or mutation of the C-terminal part of SMARCB1 (15/19 vs. 4/83; Chi-square: 56.04, p = 1.0E-10) and, related to this, in tumors of the molecular subgroup ATRT-TYR (16/36 vs. 3/66; Chi-square: 24.47, p = 7.6E-7). Previous reports have indicated that while SMARCB1 lacks a bona fide nuclear localization signal, it harbors a masked nuclear export signal (NES) and that truncation of the C-terminal region results in unmasking of this NES leading to cytoplasmic localization. To determine if cytoplasmic localization found in ATRT is due to unmasking of NES, we generated GFP fusions of one of the SMARCB1 truncating mutations (p.Q318X) found in the tumors along with a p.L266A mutation, which was shown to disrupt the interaction of SMARCB1-NES with exportin-1. We found that while the GFP-SMARCB1(Q318X) mutant localized to the cytoplasm, the double mutant GFP-SMARCB1(Q318X;L266A) localized to the nucleus, confirming NES requirement for cytoplasmic localization. Furthermore, cytoplasmic SMARCB1(Q318X) was unable to cause senescence as determined by morphological observations and by senescence-associated β-galactosidase assay, while nuclear SMARCB1(Q318X;L266A) mutant regained this function. Selinexor, a selective exportin-1 inhibitor, was effective in inhibiting the nuclear export of SMARCB1(Q318X) and caused rapid cell death in rhabdoid tumor cells. In conclusion, inhibition of nuclear export restores nuclear localization and residual tumor suppressor function of truncated SMARCB1. Therapies aimed at preventing nuclear export of mutant SMARCB1 protein may represent a promising targeted therapy in ATRT harboring truncating C-terminal SMARCB1 mutations.

Atypical teratoid/rhabdoid tumors (ATRTs) typically arise in the central nervous system (CNS) of children under 3 years of age. Despite intensive multimodal therapy (surgery, chemotherapy and, if age permits, radiotherapy), median survival is 17 months1,2. We show that ATRTs robustly express B7-H3/CD276 that does not result from the inactivating mutations in SMARCB1 (refs. 3,4), which drive oncogenesis in ATRT, but requires residual SWItch/Sucrose Non-Fermentable (SWI/SNF) activity mediated by BRG1/SMARCA4. Consistent with the embryonic origin of ATRT5,6, B7-H3 is highly expressed on the prenatal, but not postnatal, brain. B7-H3.BB.z-chimeric antigen receptor (CAR) T cells administered intracerebroventricularly or intratumorally mediate potent antitumor effects against cerebral ATRT xenografts in mice, with faster kinetics, greater potency and reduced systemic levels of inflammatory cytokines compared to CAR T cells administered intravenously. CAR T cells administered ICV also traffic from the CNS into the periphery; following clearance of ATRT xenografts, B7-H3.BB.z-CAR T cells administered intracerebroventricularly or intravenously mediate antigen-specific protection from tumor rechallenge, both in the brain and periphery. These results identify B7-H3 as a compelling therapeutic target for this largely incurable pediatric tumor and demonstrate important advantages of locoregional compared to systemic delivery of CAR T cells for the treatment of CNS malignancies.

Atypical teratoid/rhabdoid tumor (AT/RT) is a highly malignant brain tumor in infants that is characterized by loss of nuclear expression of SMARCB1 or SMARCA4 proteins. Recent studies show that AT/RTs comprise three molecular subgroups, namely AT/RT-TYR, AT/RT-MYC and AT/RT-SHH. The subgroups show distinct expression patterns of genes involved in ciliogenesis, however, little is known about the functional roles of primary cilia in the biology of AT/RT. Here, we show that primary cilia are present across all AT/RT subgroups with specific enrichment in AT/RT-TYR patient samples. Furthermore, we demonstrate that primary ciliogenesis contributes to AT/RT biology in vitro and in vivo. Specifically, we observed a significant decrease in proliferation and clonogenicity following disruption of primary ciliogenesis in AT/RT cell line models. Additionally, apoptosis was significantly increased via the induction of STAT1 and DR5 signaling, as detected by proteogenomic profiling. In a Drosophila model of SMARCB1 deficiency, concomitant knockdown of several cilia-associated genes resulted in a substantial shift of the lethal phenotype with more than 20% of flies reaching adulthood. We also found significantly extended survival in an orthotopic xenograft mouse model of AT/RT upon disruption of primary ciliogenesis. Taken together, our findings indicate that primary ciliogenesis or its downstream signaling contributes to the aggressiveness of AT/RT and, therefore, may constitute a novel therapeutic target.

Pediatric low-grade gliomas (pLGG) show heterogeneous responses to MAPK inhibitors (MAPKi) in clinical trials. Thus, more complex stratification biomarkers are needed to identify patients likely to benefit from MAPKi therapy. Here, we identify MAPK-related genes enriched in MAPKi-sensitive cell lines using the GDSC dataset and apply them to calculate class-specific MAPKi sensitivity scores (MSSs) via single-sample gene set enrichment analysis. The MSSs discriminate MAPKi-sensitive and non-sensitive cells in the GDSC dataset and significantly correlate with response to MAPKi in an independent PDX dataset. The MSSs discern gliomas with varying MAPK alterations and are higher in pLGG compared to other pediatric CNS tumors. Heterogenous MSSs within pLGGs with the same MAPK alteration identify proportions of potentially sensitive patients. The MEKi MSS predicts treatment response in a small set of pLGG patients treated with trametinib. High MSSs correlate with a higher immune cell infiltration, with high expression in the microglia compartment in single-cell RNA sequencing data, while low MSSs correlate with low immune infiltration and increased neuronal score. The MSSs represent predictive tools for the stratification of pLGG patients and should be prospectively validated in clinical trials. Our data supports a role for microglia in the response to MAPKi.

Ependymal tumors across age groups are currently classified and graded solely by histopathology. It is, however, commonly accepted that this classification scheme has limited clinical utility based on its lack of reproducibility in predicting patients' outcome. We aimed at establishing a uniform molecular classification using DNA methylation profiling. Nine molecular subgroups were identified in a large cohort of 500 tumors, 3 in each anatomical compartment of the CNS, spine, posterior fossa, supratentorial. Two supratentorial subgroups are characterized by prototypic fusion genes involving RELA and YAP1, respectively. Regarding clinical associations, the molecular classification proposed herein outperforms the current histopathological classification and thus might serve as a basis for the next World Health Organization classification of CNS tumors.

Rhabdoid tumors of early infancy are highly aggressive with consequent poor prognosis. Most cases show inactivation of the SMARCB1 (also known as INI1 and hSNF5) tumor suppressor, a core member of the ATP-dependent SWI/SNF chromatin-remodeling complex. Familial cases, described as rhabdoid tumor predisposition syndrome (RTPS), have been linked to heterozygous SMARCB1 germline mutations. We identified inactivation of another member of the SWI/SNF chromatin-remodeling complex, its ATPase subunit SMARCA4 (also known as BRG1), due to a SMARCA4/BRG1 germline mutation and loss of heterozygosity by uniparental disomy in the tumor cells of two sisters with rhabdoid tumors lacking SMARCB1 mutations. SMARCA4 is thus a second member of the SWI/SNF complex involved in cancer predisposition. Its general involvement in other tumor entities remains to be established.