RNA is a labile biomarker which has seen rather limited adoption in the clinical realm. RNA in situ hybridization is utilized in a research setting but no brain tumor biomarkers of clinical relevance have been discovered. While RNA expression profiling has been useful in the research setting, it has not seen rapid adoption with modern techniques commonly used in the clinic and most clinical assessments are performed to detect protein expression via IHC methods (Fig. 16.4a and b) [23, 24].

Cytogenetic techniques were some of the first methods employed to clinically distinguish gains and losses of DNA and specific regions of the genome or genes that were altered. Karyotypes from acutely cultured cells were used to determine associated gains or losses of chromosomes (e.g., gain of Chr 7 and loss of Chr 10 in GBM) in several brain tumors, however the application of this in practice was fairly limited compared to other cancers due to its fairly limited clinical value and is no longer commonly used in the diagnostic setting.

Fluorescence in situ hybridization (FISH), introduced in Chap. 2, is an extremely useful technique for hybridization and counting of copies of specific probes on tissue sections at targeted regions of the genome. This analysis has the advantage of being analyzed at a single cell level and, therefore, is extremely sensitive and specific. It is one of the most commonly used methods today to assess brain tumor copy aberrations, but given its targeted nature of analysis of only single regions of the genome, it is being replaced by more multiplexed or whole genome approaches.

Hybridization based copy number arrays (i.e., array CGH, SNP arrays) are highly sophisticated tools for multiplex assessment of the number of copies of genes or loci in the genome. These technologies now allow pathologists to simultaneously query up to one million different locations with good sensitivity and specificity in clinical FFPE samples (Fig. 16.4c) [25]. Next generation sequencing panels are also now used to assess copy number with good results for large (arm level) copy changes in cancers but has more limited ability to reliably detect smaller size and low level changes in copy number [26]. As example, there is often poor sensitivity for small single copy gains and deletions due to the cost of reaching broad coverage as is typically achieved in array based assays. New analytical methods and lowered sequencing costs are expected to improve this situation in the future.

Detection of single nucleotide variants (e.g., point mutations) was originally performed using Sanger sequencing and other targeted methods and has been most commonly used to identify mutations in IDH genes [27]. Targeted sequencing panels based on PCR amplification of multiple genes of interest are also commonly used in current practice [28]. However, the advent of next generation multiplex sequencing methods has rapidly advanced the ability to detect multiple mutations in cancer simultaneously and has been rapidly adopted for diagnostics and prognostic use in brain tumors [26, 29].

Most sequencing is currently focused on specific known oncogenes and tumor suppressors either at “hotspots” or at all exonic sequences of these genes, in order to efficiently detect mutations of interest in FFPE samples. The use of the targeted exome sequencing approach is particularly useful for brain tumors where tumor suppressor genes are frequently altered and lack consistent “hotspots” for effective evaluation. Current focus on next generation sequencing methods are on reduced turnaround time to allow more rapid use in clinical practice attempting to balance increased gene coverage with time required for analysis. The next generation sequencing techniques also allow for assessment of gene insertions or deletions (indels) and gene fusion events, however, analytical tools for such events are currently not reliable or widely available for clinical practice. These are expected to evolve rapidly to enable such events to be detected with clinical standards of sensitivity and specificity in the near future.

Brain cancers were one of the first indications where assessment of epigenetic modifications of the genome were of clinical relevance, specifically with respect to methylation status of the methylguanine methyltransferase (MGMT) gene in gliomas [30, 31]. Methylation events are characteristically assessed clinically using methylation specific PCR strategy with bisulfite conversion of methylated cytosine to uracil. The converted DNA is then sequenced using Sanger or pyrosequencing methods to infer which residues in the target regions of interest were methylated. As for copy number and single nucleotide variant detection methods, the use of multiplexed methylation arrays capable of whole genome methylation assessments has recently been entering into clinical use. Given the association of global methylation patterns with cell lineage, methylation arrays are most commonly used currently for recognition of diagnostic patterns of methylation in specific tumor types [32, 33]. Also such methods can reliably report out copy number information at the same time, in a similar manner to dedicated copy arrays. Interestingly, use of methylation arrays has not replaced targeted approaches for MGMT promoter assessment, although with further improvements and experience this is expected to be possible in the future.

The most common molecular events in meningiomas are loss of the tumor suppressor gene NF2 on chromosome 22q, which occurs in approximately 50 % of sporadic and nearly 100 % of syndromic tumors. Atypical WHO Grade II meningiomas are associated with losses of 1p and 14q most commonly. The overall burden of copy number changes appears to increase with histologic grade and recurrence risk in atypical meningioma and following gross total resection could be predicted using array CGH or other multiplex copy number assays to determine the overall genomic copy number changes within the tumor [34].

Recent sequencing studies of meningioma have identified strong clinico-pathologic correlation of mutational genotypes with histologic subtypes of meningiomas likely due to cell of origin differences. Oncogenic mutations in AKT1 and SMO were typically associated with an anterior/medial skull base location [35]. Mutations in TRAF7, a pro-apoptotic E3 ubiquitin ligase, have also been described in approximately 25 % of meningiomas as a common recurrent event and were present in association with KLF4 mutations or less commonly AKT1 mutations [36]. Drugs targeting the NF2, SMO, and AKT1 pathways are currently being evaluated in clinical trials of meningiomas and may help to address the lack of effective medical therapies for this disease.

These tumors were one of the first brain tumors to be recognized as associated with a specific diagnostic and prognostic molecular genetic signature. Essentially, all oligodendrogliomas exhibit codeletion of chromosomes 1p and 19q which are present in greater than 90 % of cases in carefully selected series [21]. The codeletion of these chromosomal arms was noted to result from a balanced translocation with subsequent loss of one rearranged allele leading to overall single copy loss of these regions [37]. More recently, these tumors were also noted to universally harbor mutations in either the IDH1 or IDH2 genes [21, 38]. These mutations alter the enzymatic activity of the IDH proteins leading to high levels of the onco-metabolite 2-hydroxyglutarate (2HG) which is normally expressed at only low levels in cells and may promote tumor growth. Furthermore, IDH mutations are strongly associated with a distinct methylation profile termed G-CIMP [39]. Such a profile may be induced by the IDH mutations or may represent the profile of the cell of origin for these tumors that has yet to be defined.

Other mutations in CIC, an HMG-box transcription factor with repressor activity, are identified in approximately 70 % of oligodendrogliomas and are unique to the disease. The presence of mutations in ATRX, TP53, and amplification of EGFR, hallmarks of other gliomas, are extremely rare in oligodendroglioma and typically 1p/19q codeletion are mutually exclusive with each of these events. While not yet formally utilized in the grading of oligodendrogliomas, the presence of CDKN2A loss is the most prominent change identified.

Astrocytomas are first defined by the presence of IDH1 mutations, most commonly the IDH1(R132H) variant (more than 80 % of patients). At a very high rate IDH1 mutations in this disease co-occur with loss of function mutations in the tumor suppressor genes ATRX and TP53 [21]. The most common mutations in ATRX lead to loss of expression of the protein in tumor cells as detected by IHC staining for the wild-type protein and is a useful and cost effective assay in clinical practice. Conversely, the most common mutations in TP53 lead to increased expression of P53 in tumor cells by IHC, with more than 10 % of cells generally having high-level nuclear expression. However, the use of these IHC markers as surrogates for detection of IDH1/2, ATRX, or TP53 mutations in oligodendrogliomas or astrocytomas can miss approximately 10–20 % of those cases with variant mutations whose nature does not alter protein expression in the predicted manner. Multiplex sequencing methods are, therefore, highly useful for reliable assignment of prognosis and diagnosis in all gliomas.

GBM is one of the best characterized cancers at the molecular level [40] and was one of the first cancers to be studied in large scale integrative genomic approaches of The Cancer Genome Atlas (TCGA) program established by the National Cancer Institute [41]. Collectively, these molecular studies have consistently identified three core signaling pathways which are disrupted in GBM: increased activation of receptor tyrosine kinase/RAS/PI3 K signaling, loss of function in P53 signaling, and reduced signaling of the RB pathway. Aberrant activation of RTK/RAS/PI3 K signaling is evident in 88 % of GBMs and most characteristically occurs due to amplification of the EGFR gene, along with rearrangements and overexpression of mutant EGFRvIII and extracellular domain mutants [42]. This pathway is also activated by amplification of other well known oncogenes including PDGRA, MET, AKT, or PIK3CA and/or aberrations that lead to loss of the PTEN tumor suppressor gene function. Alterations in the tumor suppressor TP53 are common in a subset of adult GBM including at least 42 % of adult tumors, either through direct mutations in TP53 or by amplification of MDM2 or MDM4 each seen in approximately 10 % of patients. Interestingly, with time studies have shown that TP53 mutations are more common in young adult astrocytomas and pediatric high-grade gliomas where they are seen in the majority of these tumors [43, 44]. The RB pathway is frequently targeted by different routes, including genomic losses combined with mutation of RB, as well as genomic losses of the CDKN2 family of tumor suppressor genes, or amplification of the negative regulators of the RB pathway, such as CDK4.