Diffuse gliomas with clear cell morphology are represented in the 2007 WHO classification of brain tumors as oligoastrocytomas and oligodendrogliomas of grades II and III. Although preoperative neuroimaging of these brain tumors are often successful, histopathologic evaluation of neurosurgically removed tumor specimens is still required for a definite diagnosis and subsequent molecular analysis. Such CNS tumors show an extensive variety of histological and cytological appearance making diagnosis in clear cell gliomas somewhat difficult, especially as oligoastrocytomas are often less clearly defined. There is an ongoing debate whether these tumors indeed constitute an entity or whether they represent a mixed bag containing both astrocytomas and oligodendroglioma. Recent advances with molecular data on these tumors have a major impact on neuopathological typing, prognosis and therapy of the patients. This neuro-oncologic review focuses on neuropathological and molecular features of the different types of diffusely infiltrating gliomas. Core features and distinct patterns and variants are also introduced and illustrated. Recent advances in immunohistochemistry and molecular biology have contributed to an improved classification and are discussed. Different methodologies for molecular analysis of 1p/19q codeletion, ATRX loss, MGMT promoter methylation and isocitrate dehydrogenase mutations are presented in detail and their prognostic and predictive implications for therapy are discussed.
Bailey was the first to recognize oligodendroglial tumor as a distinct brain tumor entity in 1926 . Later it became evident that these tumors were not biologically aggressive as astrocytic gliomas of identical histologic grade . Based on data from the Central Brain Tumor Registry of the United States registry (http://www.cbtrus. org), oligodendroglial tumors account for approximately 6% of all glial neoplasms and are the third most common glioma of the brain. A male gender predilection is noted in oligodendroglial neoplasms but there is no sufficient explanation for this fact . While almost all tumors occur sporadically, in individuals with Lynch syndrome, a constitutional mismatch repair defect, results in increase of brain tumors among families . The tumor peaks during the 5th and 6th decade and accounts for less than 1% of pediatric central nervous system neoplasms. Publications with higher numbers usually have applied a less stringent diagnostic approach. This probably results from recognition of the significant differences in prognosis between oligodendroglial and astrocytic tumors. It is believed that an increased awareness of favorable histopathologic features by neuropathologists have contributed to these tumors being diagnosed more eagerly. The current World Health Organization recognizes two tumor grades: Oligodendroglioma grade II (ICD-O: 9450/3) and Anaplastic Oligodendroglioma grade III (ICD-O: 9541/3). In addition, some closely related tumors show morphological features of both astrocytoma and oligodendroglioma. These mixed tumors are designed as Oligoastrocytomas (ICD-O: 9382/3) [5,6]. While many of the pathological criteria outlined below, are well defined, the diagnosis of oligodendroglioma remains challenging in atypical cases, especially in mixed gliomas and in tumors that are graded by mitotic activity alone. Furthermore, the initial determination of the tumor as pure oligodendroglioma or oligoastrocytoma may subsequently influence the decision whether the tumor is graded as grade II or grade III neoplasm. Not surprisingly, interobserver variability in such series is great, ranging from 9-80% as seen in a study on 155 tumors that were initially classified as oligoastrocytomas .
Oligodendrogliomas primarily affect the deep white matter of the cerebral hemisphere and are most common in the frontal lobes [8,9]. Spinal cord oligodendrogliomas are rare, but in children tumors with extensive leptomeningeal spread have been reported . As seen in most diffuse growing glial tumors, oligodendrogliomas have no clearcut borders. Rather the tumor cells intermingle with normal brain tissue and thus only main tumor components show macroscopically visible changes. The tumors are usually soft and have a grey to pink color and may show focally a mucoid, gelatinous matrix . They may appear hemorrhagic and / or calcified but this is not a specific feature for oligodendrogliomas. Superficial growth can expand the cortical grey matter and elicit eliptogenic potential . The anaplastic forms usually lack a central necrosis typically for glioblastoma but may show focally multiple smaller necroses. Rare disseminating cases may grow as superficial gelatinous masses . There are no macroscopically distinct differences between oligondendrogliomas and oligoastrocytomas. Many patients report a long history of seizures. Generalized seizures are associated with frontomesial locations, while tumors with partial seizures are more often located in orbitofrontal regions .
Oligodendroglial tumors are dominated by histologic monotony of the round to oval shaped tumor cells which are best seen in smears or at low magnification in paraffin-embedded specimens. Nuclei have bland chromatin and prominent nucleoli. Astrocytic tumors usually exhibit more nuclear pleomorphism at this stage . The very characteristic perinuclear halo (“fried egg”) – a fixation artifact resulting from autolytic water absorption – is absent in frozen sections or specimens that have been quickly processed resulting from a short fixation time. Delicate branching of capillaries and tumor calcifications that also may affect tumor vessels are more frequent in oligodendrogliomas than in other CNS tumors . Some tumors may exhibit nodules of higher cellularity and slightly increased mitotic activity. Such tumors are usually borderline cases between grade II and grade III. Prominent microcysts may be apparent in otherwise compact tumor architecture. Divergent growth pattern include columnar arrangement of tumor cells, the so-called spongioblastoma pattern , marked desmoplasia  signet-ring cell morphology  or perineuronal satellitosis of the tumor cells invading grey matter. These patterns have no impact on the clinical course or tumor grading. So called “secondary structures” of Schrerer, where tumor cells may concentrate along subpial structures are occasionally present. Rare cases of oligodendrogliomas may show focally neuropil islands, that have to be distinguished from neurocytomas. Another important differential diagnosis is the dysembryoplastic neuroepithelial tumor that are characterized by the presence of floating neurons. Oligodendrogliomas usually lack the multinodular appearance and the so called “glioneuronal element” of these tumors and are more deep-seated compared to the usually cortical located DNT. However, as some authors propose a non-specific form of DNT that is diagnosed in context of clinical and radiological presentation , molecular analysis might be required in some instances to discriminate these two entities safely. Cerebellar pilocytic astrocytomas may focally exhibit cluster of cells that resemble oligodendrocytes. Unless biopsies are very small, the other components of the tumor are usually visible and exclude oliogodendroglioma. Single scattered reactive astrocytes among otherwise oligodendroglial tumor cells should not prompt the diagnosis of an oligoastrocytoma. Some tumors show focal areas with eccentric nuclei and a fibrillar, almost glassy cytoplasm with short to none cell processes, the so called minigemistocytes or gliofibrillary oligodendrocytes. Again, these tumors should not classified as mixed gliomas. In the tumor edges severeal astrocytic cells might be present but unless clearly neoplastic in nature their presence does not warrant the diagnosis of mixed oligoastrocytoma. Mixed tumors may come in two flavors: The more readily to diagnose, but less often seen biphasic (“compact”) oligoastrocytoma with two distinct neoplastic tumor components and the more intermingled (“diffuse”) variant, in which sometimes the astrocytic cells are not always neoplastic in nature . There is no agreement to upon which minimum of astrocytic appearing tumor cells are required to diagnose a mixed tumor . Anaplastic oligodendrogliomas show increased nuclear pleomorphism and increased mitotic activity compared to grade II lesions while often retaining the monotonous architecture of more evenly distributed tumor cells. As the 2007 WHO classification only vaguely mentions “significant mitotic activity” as a delineator, many authors prefer a mitotic cutoff of 6 mitoses per 10 high power fields to discriminate between grade II and grade III lesions . Vascular proliferations or extended vascular hyperplasia are typical for anaplastic oligodendroglioma grade III. In addition smaller areas of necrosis may be present but not typically in the pseudopalisading forms of glioblastoma. The 2007 WHO classification however explicitly allows presence of pseudopalisading necroses in anaplastic oligendrogliomas and thus weakened a sufficient discrimination to the recently included glioblastomas with oligodendroglial component. Even more difficult is the discrimination of anaplastic oligoastrocytoma from glioblastoma. Recent data confirmed that presence of necrosis is an unfavorable predictor in mixed gliomas - but not in pure oligodendrogliomas . Based on this data, stratification of tumors as glioblastoma with oligodendroglial component is based mainly on presence of necrosis. As a rule of thumb, we expect in tumors that are not unequivocally classified as pure oligodendrogliomas, that the usually smaller necrosis should be limited to the oligodendroglial component before considering a mixed anaplastic oligoastrocytoma / small cell glioblastoma and strive to resolve the issue by additional immunohistochemistry and molecular analysis.
A specific marker for oligodendroglial tumors is still lacking. Therefore it is advisable to use a panel of different antibodies for which expression patterns in these neoplasms has been extensively studied and to combine the information gained from these stains with morphology of the tumor cells. In our institution we stain routinely gliomas for GFAP, MAP2, WT1, ATRX and IDH1(R132H). Expression of the glial fibrillary acidic protein GFAP is usually absent or only weakly seen in tumor cytoplasm of oligodendroglia, however in our daily practice sometime there is ample overlapping of the GFAPpositive fibrillary neuropil background with individual tumor cells . In addition minigemistocytes and interdispersed astrocytes are usually positive for GFAP. Olig-2 was initially considered specific for neural progenitors and developing oligodendrocytes and thus proposed as a oligodendroglioma-specific marker. While nuclear Olig2 immunoreactivity is slightly stronger in oligodendrogliomas, it is also constantly seen in other glial neoplasms such as astrocytomas and mixed glial-neuronal tumors . The marker however remains helpful as it is absent in clear cell ependymomas and neurocytomas . The microtubule-associated protein MAP2 is constantly expressed in oligodendrogliomas, but also found in up to 92% of astrocytomas and glioblastomas . A perinuclear “capped” expression pattern is more typical for oligodendroglial cells, while in astrocytomas the elongated cell processes are also immunoreactive for MAP2 . Entrapped neurons and primitive neuroepithelial tumor cells are however often strongly immunoreactive for MAP2 too. The Wilms-tumor protein WT1 is strongly expressed in the cytoplasm 83-92% of anaplastic astrocytomas and glioblastomas but in oligodendrogliomas, it is usually restricted to single WT1- positive tumour cells or completely absent . Similarly to GFAP, WT1 also strongly stains minigemistrocytes. Another recently introduced marker is the neurite outgrowth factor Nogo-A. In the prime publication, Nogo-A was found in 71% oligodendrogliomas and 24% glioblastomas but absent in astrocytomas . However subsequent studies confirming these results are lacking so far. Alpha internexin is found in approximately half of oligodendrogliomas and was introduced as associated marker for the molecular relevant 1p/19q codeletion . However because of higher specificity, it has been replaced recently by alpha thalassemia/mental retardation syndrome X-linked protein (ATRX) [27,28]. Nuclear ATRX immunoreactivity is retained in most oligodendroglioma cases and strongly associated with 1p/19q codeletion in oligoastrocytomas and oligodendrogliomas . Unlike diffuse astrocytomas, most oligodendrogliomas show only limited p53 immunreactivity . Recently a mutation specific IDH1(R132H) antibody has been developed. Positive IDH1(R132H) immunoreactivity is so frequent in oligodendroglial tumors (up to 91% in grade II and 94% in grade III lesions) that this marker is very useful to discriminate oligodendrogliomas from other brain tumors with clear cell morphology . The few remaining negative cases should be sequenced for rare IDH1/IDH2 mutations. IDH mutations are less common in pediatric oligodendroglial tumors . IDH1/2 mutation status however is not useful to separate oligodendrogliomas from mixed gliomas or pure astrocytomas. Synaptophysin and CD56 immunoreactivity may be observed in some oligondendrogliomas but in association with IDH1 immunoreactivity it is possible to exclude neurocytoma or neuroendocrine tumor metastatic to the brain . Electron microscopy is not used in regular routine practice as combined data from histology, immunistochemistry and molecular pathology is usually sufficient enough do diagnose an oligodendroglioma (Figure 1).
Figure 1: Example for integrated diagnosis in oligodendroglioma. Histopathology in H&E stain shows anaplastic glial tumor cells with perinuclear halo and mostly absent GFAP immunoreactivity in tumor cells. Together with strong MAP2 immunoreactivity in a perinuclear capped fashion and presence of IDH1(R132H) mutation in the tumor this indicates a canonical anaplastic oligodendroglioma grade III. The tumor however exhibited nuclear loss of ATRX and 1p/19q analysis showed no allelic losses in microsatellite- PCR (high-resolution gel bottom) as well in genome-wide copy number analysis (OncoScan FFPE top). The following clinical course of this case was compatible with an IDH1-mutated anaplastic astrocytoma.
Because of their prognostic and predictive potential, molecular data of brain tumors have become an integral part of modern neurooncology. Prognostic markers inform about likely disease outcome independent of the treatment received, while predictive markers provide information about expected outcomes with the application of specific interventions which is very important in oligodendrogliomas. Until twenty years ago diagnosis and grading of oligodendroglial tumor has been based on histopathological morphology. Since the discovery of recurrent codeletions of chromosome regions 1p36/19q13 in oligodendrogliomas in 1994  it became evident that certain molecular abberations carry important prognostic and predictive information on particular therapy response. Loss of heterozygosity in 1p and 19q is not restricted to oligodendroglial tumors but strongly associated with it and therefore useful as a diagnostic marker . The codeletion is found in 70-80% of oligodendrogliomas grade II, 40- 50% of oligoastrocytomas , in 10-20% of diffuse astrocytomas grade II WHO and in 2-8% of glioblastomas [34,35]. The number of codeletions may be even higher, when strict diagnostic criteria for the diagnosis of oligodendrogliomas are applied as outlined above. The 1p/19q loss results from a non-balanced translocation t(1:19)(q10:p10) that is also observed in tumor recurrences . This molecular signature is so robust that it is seen even in cases with increased intratumoral heterogeneity . The clinical course of 1p/19q codeleted tumors is significant better than for nondeleted tumors of similar grade [38,39]. This results from a so called “proneural” expression profile of the genomic landscape with less overall mutation rate . In addition in oligodendrogliomas, nearly all codeleted tumors also show either additional isocitrate dehydrogenase (IDH1 or IDH2) mutations . Not surprisingly journal reviewers often require 1p/19q deletions in oligodendrogliomas for sample homogeneity. There is an inverse correlation with ATRX loss / p53 mutation and codeletion of chromosomal arms 1p and 19q in oligodendrogliomas. This implicates that non-codeleted oligodendrogliomas and oligoastrocytomas have a clinical and genetic profile similar to astrocytoma and thus questions the existence of oligoastrocytoma as a distinct entity [41,42]. The 1p/19q codeletion is not only diagnostic and prognostic but also has therapeutic relevancy. Codeleted tumors are more responsive to a combined Procarbazine-CCNU-Vincristine (PCV) therapy regimen [43,44]. This benefit has been corrobated by independent long-term studies on IDH1/2 mutated and 1p/19q codeleted tumors [45,46]. IDH1/2 itself has a prognostic role, as oligodendroglial tumors with IDH1/2 mutation but without 1p/19q deletion still have a better clinical course than double-negative oligodendrogliomas without IDH1/2 and 1p/19q codeletion  (wich are most likely glioblastomas with an oligodendroglial component). A loss of heterozygosity is usually assessed through use of microsatellite marker PCR on high-resolution gels. This method requires corresponding blood samples to determine allele status . Therefore use of fluorescent in situ hybridisation is preferred by some laboratories but carries the risk of misdiagnosing cases with only partial loss. This risk can be covered by PCR that contains several loci along the chromosomal arms or by use of genomewide based array methods for copy-number alterations [15,16]. With the introduction of a mutation-specific antibody the determination of the IDH1/IDH2 mutation status has become easier. The IDH1 (R132H) mutation is detected by immunohistochemistry in approximately 75%- 90% of the cases . The remaining R132H-negative cases are then usually sequenced for non-canonical IDH1 and IDH2 mutations .
Other far less frequently employed biomarkers include the epidermal growth factor receptor (EGFR). Its amplification is mutually exclusive with the 1p/19q codeletion and in cases with amplification the diagnosis of (small cell) glioblastoma instead of anaplastic oligodendroglioma should be considered . The determination of telomerase reverse transcriptase (TERT) C228T and C250T mutation is useful in IDH1/2 wild-type tumors to detect potential high-grade lesions  but does not help in separating oligodendroglioma from glioblastoma as such mutations are present in both tumor entities . Additional mutations found in 1p/19q codeleted tumors are alterations in the putative tumor supressors far-upstream binding protein-1 (FUBP1, 15-30% of the tumors) and Drosophila capicua homolog (CIC, 50-70%) which are located on 19q and 1p respectively . The presence of mutation results in a loss of FUBP1 expression and can be determined by immunohistochemistry . CIC mutated, 1p/19q codeleted tumors seem to have a prognostic unfavorable course [14,48]. Genetic alterations are surprisingly similar between grade II and grade III oligodendrogliomas. One potential candidate driver for anaplastic progression is the oligodendrocyte-related transcription factor TCF12 .
Another marker employed routinely in gliomas is the determination of the O6-methylguanine-DNA-methyltransferase (MGMT) promotor status because of its prognostic and predictive role in glioblastoma . A methylated phenotype is associated with a single nucleotide polymoprhism within the MGMT gene and results in improved chemo-sensitivity of the tumors . In oligodendroglial tumors there is a strong association between MGMT promoter methylation and 1p/19q codeletion. However while MGMT alone is useful as a prognostic marker in oligodendrogliomas, it is not useful to predict outcome of adjuvant treatment in oligodendrogliomas . Almost all co-deleted oligodendrogliomas show a methylated MGMT promotor as part of a general hypermethylated genotype, the so called “G-CIMP” profile . While oligodendrogliomas in children are morphologically similar to their adult counterparts, they lack the same genetic background . One notable exception is the diffusely growing oligodendroglia-like neoplasms with markedly leptomeningeal spread. These tumors may carry the classical 1p/19q codeletion pattern seen in adult oligodendrogliomas in association to KIAA1549-BRAF fusion transcripts typically encountered in pilocytic astrocytoma [37,39].
Because of clinical relevance the molecular stratification of oligodendrogliomas is of utmost importance. The combination of IDH mutation and 1p/19q co-deletion is now considered the molecular signature of oligodendrogliomas [49,50] and assessment of the 1p/19q codeletion status guides the neuro-oncologist for chemotherapy responsiveness as seen in two large studies (RTOG 9402 and EORTC 26951) [51,52]. Not surprisingly the USA National Comprehensive Cancer Network treatment guidelines therefore recommends 1p/19q codeletion as the only molecular biomarker to be used for therapeutic stratification in anaplastic oligodendroglioma and oligoastrocytomas . As there is no sufficient delineator between grade II and grade III tumors, the histopathological analysis remains the first crucial step and identification of a clear-cell (“oligodendroglial”) phenotype by the neuropathologist is essential for subsequent targeted and costefficient molecular analysis. Based on the obtained data, prototypical oligodendrogliomas (defined as IDH1/2 mutated, 1p/19q codeleted, MGMT methylated (no ATRX loss and TERT mutated)) should be included in future studies when speaking strictly of oligodendrogliomas. Morphologically similiar tumors that do not meet these requirements may show a divergent, more unfavorable clinical outcome and thus may need to be treated accordingly. The proper designation of these tumors is currently debated and most likely included in the upcoming WHO classification of brain tumors. The integrated (layered) diagnostic by combining histopathology, immunohistochemistry and molecular data into a layered scheme will obviously become the future gold standard in neuropathology and neurooncology practice .
The author acknowledges support by Deutsche Forschungsgemeinschaft and Open Access Publishing Fund of University of Tübingen.