Pilocytic astrocytomas (PAs) were recognized as a discrete clinical entity over 70 years ago. They are relatively benign (WHO grade I) and have, as a group, a 10-year survival of over 90 %. Many ...require merely surgical removal and only very infrequently do they progress to more malignant gliomas. While most show classical morphology, they may present a spectrum of morphological patterns, and there are difficult cases that show similarities to other gliomas, some of which are malignant and require aggressive treatment. Until recently, almost nothing was known about the molecular mechanisms involved in their development. The use of high-throughput sequencing techniques interrogating the whole genome has shown that single abnormalities of the mitogen-activating protein kinase (MAPK) pathway are exclusively found in almost all cases, indicating that PA represents a one-pathway disease. The most common mechanism is a tandem duplication of a ≈2 Mb-fragment of #7q, giving rise to a fusion between two genes, resulting in a transforming fusion protein, consisting of the N-terminus of KIAA1549 and the kinase domain of BRAF. Additional infrequent fusion partners have been identified, along with other abnormalities of the MAP-K pathway, affecting tyrosine kinase growth factor receptors at the cell surface (e.g., FGFR1) as well as BRAF V600E, KRAS, and NF1 mutations among others. However, while the KIAA1549-BRAF fusion occurs in all areas, the incidence of the various other mutations identified differs in PAs that develop in different regions of the brain. Unfortunately, from a diagnostic standpoint, almost all mutations found have been reported in other brain tumor types, although some retain considerable utility. These molecular abnormalities will be reviewed, and the difficulties in their potential use in supporting a diagnosis of PA, when the histopathological findings are equivocal or in the choice of individualized therapy, will be discussed.
With the number of prognostic and predictive genetic markers in neuro-oncology steadily growing, the need for comprehensive molecular analysis of neuropathology samples has vastly increased. We ...therefore developed a customized enrichment/hybrid-capture-based next-generation sequencing (NGS) gene panel comprising the entire coding and selected intronic and promoter regions of 130 genes recurrently altered in brain tumors, allowing for the detection of single nucleotide variations, fusions, and copy number aberrations. Optimization of probe design, library generation and sequencing conditions on 150 samples resulted in a 5-workday routine workflow from the formalin-fixed paraffin-embedded sample to neuropathological report. This protocol was applied to 79 retrospective cases with established molecular aberrations for validation and 71 prospective cases for discovery of potential therapeutic targets. Concordance of NGS compared to established, single biomarker methods was 98.0 %, with discrepancies resulting from one case where a
TERT
promoter mutation was not called by NGS and three ATRX mutations not being detected by Sanger sequencing. Importantly, in samples with low tumor cell content, NGS was able to identify mutant alleles that were not detectable by traditional methods. Information derived from NGS data identified potential targets for experimental therapy in 37/47 (79 %) glioblastomas, 9/10 (90 %) pilocytic astrocytomas, and 5/14 (36 %) medulloblastomas in the prospective target discovery cohort. In conclusion, we present the settings for high-throughput, adaptive next-generation sequencing in routine neuropathology diagnostics. Such an approach will likely become highly valuable in the near future for treatment decision making, as more therapeutic targets emerge and genetic information enters the classification of brain tumors.
Epileptogenic tumors affecting children and young adults are a morphologically diverse collection of neuroepithelial neoplasms that, as a group, exhibit varying levels of glial and/or neuronal ...differentiation. Recent advances in molecular profiling technology, including comprehensive DNA sequencing and methylation analysis, have enabled the application of more precise and biologically relevant classification schemes to these tumors. In this report, we describe a morphologically and molecularly distinct epileptogenic neoplasm, the polymorphous low-grade neuroepithelial tumor of the young (PLNTY), which likely accounts for a sizable portion of oligodendroglioma-like tumors affecting the pediatric population. Characteristic microscopic findings most notably include infiltrative growth, the invariable presence of oligodendroglioma-like cellular components, and intense immunolabeling for cluster of differentiation 34 (CD34). Moreover, integrative molecular profiling reveals a distinct DNA methylation signature for PLNTYs, along with frequent genetic abnormalities involving either B-Raf proto-oncogene (
BRAF
) or fibroblast growth factor receptors 2 and 3 (
FGFR2
,
FGFR3
). These findings suggest that PLNTY represents a distinct biological entity within the larger spectrum of pediatric, low-grade neuroepithelial tumors.
Pediatric glioblastoma (pedGBM) is amongst the most common malignant brain tumors of childhood and carries a dismal prognosis. In contrast to adult GBM, few molecular prognostic markers for the ...pediatric counterpart have been established. We, therefore, investigated the prognostic significance of genomic and epigenetic alterations through molecular analysis of 202 pedGBM (1–18 years) with comprehensive clinical annotation. Routinely prepared formalin-fixed paraffin-embedded tumor samples were assessed for genome-wide DNA methylation profiles, with known candidate genes screened for alterations via direct sequencing or FISH. Unexpectedly, a subset of histologically diagnosed GBM (
n
= 40, 20 %) displayed methylation profiles similar to those of either low-grade gliomas or pleomorphic xanthoastrocytomas (PXA). These tumors showed a markedly better prognosis, with molecularly PXA-like tumors frequently harboring BRAF V600E mutations and 9p21 (
CDKN2A
) homozygous deletion. The remaining 162 tumors with pedGBM molecular signatures comprised four subgroups: H3.3 G34-mutant (15 %), H3.3/H3.1 K27-mutant (43 %), IDH1-mutant (6 %), and H3/IDH wild-type (wt) GBM (36 %). These subgroups were associated with specific cytogenetic aberrations,
MGMT
methylation patterns and clinical outcomes. Analysis of follow-up data identified a set of biomarkers feasible for use in risk stratification: pedGBM with any oncogene amplification and/or K27M mutation (
n
= 124) represents a particularly unfavorable group, with 3-year overall survival (OS) of 5 %, whereas tumors without these markers (
n
= 38) define a more favorable group (3-year OS ~70 %).Combined with the lower grade-like lesions, almost 40 % of pedGBM cases had distinct molecular features associated with a more favorable outcome. This refined prognostication method for pedGBM using a molecular risk algorithm may allow for improved therapeutic choices and better planning of clinical trial stratification for this otherwise devastating disease.
In contrast to the relative morphological uniformity of histone H3 K27-mutant high-grade gliomas, H3 G34-mutant tumors present as a histopathologically heterogeneous group of neoplasms, with ...microscopic characteristics typical of either glioblastoma (GBM) or central nervous system primitive neuroectodermal tumors (CNS-PNET). In the current study, we performed an integrative clinical, histopathological and molecular analysis of 81 G34-mutant CNS tumors. Routinely prepared tumor tissues were investigated for genomic and epigenomic alterations. Despite their divergent histopathological appearance, CNS tumors with H3.3 G34 mutations displayed uniform epigenetic signatures, suggesting a single biological origin. Comparative cytogenetic analysis with other GBM subtypes disclosed a high frequency and high specificity of 3q and 4q loss across G34-mutant tumors.
PDGFRA
amplification was more common in cases with GBM than with PNET morphology (36 vs. 5 %, respectively), while
CCND2
amplifications showed the opposite trend (5 vs. 27 %). Survival analysis revealed the presence of amplified oncogene(s) and
MGMT
methylation as independent prognostic markers for poor and favorable outcomes, respectively. No difference in outcome was found between morphological variants (GBM vs. PNET). Thus, different histological variants of G34-mutant CNS tumors likely comprise a single biological entity (high-grade glioma with H3 G34 mutation, HGG_G34), which should be outlined in future diagnostic and therapeutic classifications. Screening for H3.3 G34 mutation should therefore be recommended as a routine diagnostic marker for supratentorial CNS tumors across a broad histological spectrum.
DNA methylation data-based precision cancer diagnostics is emerging as the state of the art for molecular tumor classification. Standards for choosing statistical methods with regard to ...well-calibrated probability estimates for these typically highly multiclass classification tasks are still lacking. To support this choice, we evaluated well-established machine learning (ML) classifiers including random forests (RFs), elastic net (ELNET), support vector machines (SVMs) and boosted trees in combination with post-processing algorithms and developed ML workflows that allow for unbiased class probability (CP) estimation. Calibrators included ridge-penalized multinomial logistic regression (MR) and Platt scaling by fitting logistic regression (LR) and Firth's penalized LR. We compared these workflows on a recently published brain tumor 450k DNA methylation cohort of 2,801 samples with 91 diagnostic categories using a 5 × 5-fold nested cross-validation scheme and demonstrated their generalizability on external data from The Cancer Genome Atlas. ELNET was the top stand-alone classifier with the best calibration profiles. The best overall two-stage workflow was MR-calibrated SVM with linear kernels closely followed by ridge-calibrated tuned RF. For calibration, MR was the most effective regardless of the primary classifier. The protocols developed as a result of these comparisons provide valuable guidance on choosing ML workflows and their tuning to generate well-calibrated CP estimates for precision diagnostics using DNA methylation data. Computation times vary depending on the ML algorithm from <15 min to 5 d using multi-core desktop PCs. Detailed scripts in the open-source R language are freely available on GitHub, targeting users with intermediate experience in bioinformatics and statistics and using R with Bioconductor extensions.
Diffuse gliomas are represented in the 2007 WHO classification as astrocytomas, oligoastrocytomas and oligodendrogliomas of grades II and III and glioblastomas WHO grade IV. Molecular data on these ...tumors have a major impact on prognosis and therapy of the patients. Consequently, the inclusion of molecular parameters in the WHO definition of brain tumors is being planned and has been forwarded as the “ISN-Haarlem” consensus. We, here, analyze markers of special interest including ATRX, IDH and 1p/19q codeletion in a series of 405 adult patients. Among the WHO 2007 classified tumors were 152 astrocytomas, 61 oligodendrogliomas, 63 oligoastrocytomas and 129 glioblastomas. Following the concepts of the “ISN-Haarlem”, we rediagnosed the series to obtain “integrated” diagnoses with 155 tumors being astrocytomas, 100 oligodendrogliomas and 150 glioblastomas. In a subset of 100 diffuse gliomas from the NOA-04 trial with long-term follow-up data available, the “integrated” diagnosis had a significantly greater prognostic power for overall and progression-free survival compared to WHO 2007. Based on the “integrated” diagnoses, loss of ATRX expression was close to being mutually exclusive to 1p/19q codeletion, with only 2 of 167 ATRX-negative tumors exhibiting 1p/19q codeletion. All but 4 of 141 patients with loss of ATRX expression and diffuse glioma carried either
IDH1
or
IDH2
mutations. Interestingly, the majority of glioblastoma patients with loss of ATRX expression but no IDH mutations exhibited an
H3F3A
mutation. Further, all patients with 1p/19 codeletion carried a mutation in
IDH1
or
IDH2
. We present an algorithm based on stepwise analysis with initial immunohistochemistry for ATRX and IDH1-R132H followed by 1p/19q analysis followed by
IDH
sequencing which reduces the number of molecular analyses and which has a far better association with patient outcome than WHO 2007.
Meningioma is the most common primary brain tumor and carries a substantial risk of local recurrence. Methylation profiles of meningioma and their clinical implications are not well understood. We ...hypothesized that aggressive meningiomas have unique DNA methylation patterns that could be used to better stratify patient management. Samples (
n
= 140) were profiled using the Illumina HumanMethylation450BeadChip. Unsupervised modeling on a training set (
n
= 89) identified 2 molecular methylation subgroups of meningioma (MM) with significantly different recurrence-free survival (RFS) times between the groups: a prognostically unfavorable subgroup (MM-UNFAV) and a prognostically favorable subgroup (MM-FAV). This finding was validated in the remaining 51 samples and led to a baseline meningioma methylation classifier (bMMC) defined by 283 CpG loci (283-bMMC). To further optimize a recurrence predictor, probes subsumed within the baseline classifier were subject to additional modeling using a similar training/validation approach, leading to a 64-CpG loci meningioma methylation predictor (64-MMP). After adjustment for relevant clinical variables WHO grade, mitotic index, Simpson grade, sex, location, and copy number aberrations (CNAs) multivariable analyses for RFS showed that the baseline methylation classifier was not significant (
p
= 0.0793). The methylation predictor, however, was significantly associated with tumor recurrence (
p
< 0.0001). CNAs were extracted from the 450k intensity profiles. Tumor samples in the MM-UNFAV subgroup showed an overall higher proportion of CNAs compared to the MM-FAV subgroup tumors and the CNAs were complex in nature. CNAs in the MM-UNFAV subgroup included recurrent losses of 1p, 6q, 14q and 18q, and gain of 1q, all of which were previously identified as indicators of poor outcome. In conclusion, our analyses demonstrate robust DNA methylation signatures in meningioma that correlate with CNAs and stratify patients by recurrence risk.
We have extended our understanding of the molecular biology that underlies adult glioblastoma over many years. By contrast, high-grade gliomas in children and adolescents have remained a relatively ...under-investigated disease. The latest large-scale genomic and epigenomic profiling studies have yielded an unprecedented abundance of novel data and provided deeper insights into gliomagenesis across all age groups, which has highlighted key distinctions but also some commonalities. As we are on the verge of dissecting glioblastomas into meaningful biological subgroups, this Review summarizes the hallmark genetic alterations that are associated with distinct epigenetic features and patient characteristics in both paediatric and adult disease, and examines the complex interplay between the glioblastoma genome and epigenome.
Astrocytoma and oligodendroglioma are histologically and genetically well-defined entities. The majority of astrocytomas harbor concurrent
TP53
and
ATRX
mutations, while most oligodendrogliomas carry ...the 1p/19q co-deletion. Both entities share high frequencies of
IDH
mutations. In contrast, oligoastrocytomas (OA) appear less clearly defined and, therefore, there is an ongoing debate whether these tumors indeed constitute an entity or whether they represent a mixed bag containing both astrocytomas and oligodendrogliomas. We investigated 43 OA diagnosed in different institutions employing histology, immunohistochemistry and in situ hybridization addressing surrogates for the molecular genetic markers
IDH1
R132H,
TP53
,
ATRX
and 1p/19q loss. In all but one OA the combination of nuclear p53 accumulation and ATRX loss was mutually exclusive with 1p/19q co-deletion. In 31/43 OA, only alterations typical for oligodendroglioma were observed, while in 11/43 OA, only indicators for mutations typical for astrocytomas were detected. A single case exhibited a distinct pattern, nuclear expression of p53, ATRX loss,
IDH1
mutation and partial 1p/19q loss. However, this was the only patient undergoing radiotherapy prior to surgery, possibly contributing to the acquisition of this uncommon combination. In OA with oligodendroglioma typical alterations, the portions corresponding to astrocytic part were determined as reactive, while in OA with astrocytoma typical alterations the portions corresponding to oligodendroglial differentiation were neoplastic. These data provide strong evidence against the existence of an independent OA entity.