Oligodendrogliomas: a short history of clinical developments

Oligodendroglioma: the early days

Oligodendrogliomas have in particular attracted attention once it became clear that these are chemotherapy sensitive tumors. Although since the first description in 1926 and 1929 of oligodendroglial tumors it has been clear that these tumors are a special brand with a relatively good prognosis, they were taken for just another glioma subtype for many years thereafter [1]. Indeed, apart from a relatively more favorable prognosis of oligodendroglioma in comparison to their astrocytic counterparts and their often calcified lesions allowing a first diagnosis even at the time CT scans were not yet available, little pointed toward a specific significance for these tumors. This histological identification was somewhat complicated by the existence of another group with histological characteristics of both astrocytoma and oligodendroglioma, labeled mixed oligoastrocytoma. All these tumors were considered among the grade II and III glioma, and trials on grade II and high-grade glioma usually considered both astrocytoma, oligodendroglioma and mixed oligoastrocytoma. Furthermore, trials on high-grade glioma usually lumped both grade III and grade IV (glioblastoma) under the heading ‘anaplastic glioma’. As a consequence, grade III tumors were in general treated as grade IV tumors, albeit with the clear notion their outcome was better.

Oligodendroglioma: the chemotherapy sensitive glioma

This clinical significance of the diagnosis ‘oligodendroglioma’ radically changed when Cairncross et al. in 1988 and again in 1992 reported remarkably high response rates to procarbazine, CCNU (lomustine) and vincristine (PCV) chemotherapy in relapsing anaplastic oligodendroglioma [2]. This lead to confirmatory Phase II studies, and subsequently to studies in newly diagnosed anaplastic oligodendroglial tumors [3,4]. At that time, still no distinction was made between pure oligodendroglial tumors and mixed oligodendroglioma. At least in part driven by the wish not to miss oligodendroglial tumors the percentage of patients with oligodendroglioma showed a steep incline, from 5% of all glioma in historical series to no less than 20% of all glioma cases [5]. A clear change in diagnostic criteria was part of that increase, already pointing to the lack of objective criteria for the histological diagnosis of glioma in general but more in particular of oligodendroglial tumors.

Oligodendroglioma: the molecular tumor

From the early nineties on, the advance in molecular techniques began to make contribution to the understanding of oligodendroglioma. In 1994, Reifenberger et al. described combined loss of the entire short arm of chromosome 1 and the long arm of chromosome 19 (combined 1p/19q loss) as the most characteristic genetic lesion in oligodendroglioma [6]. In an important next paper, Mainz et al. showed that low-grade mixed oligoastrocytoma show either characteristic astrocytic lesions (TP53 mutations) or combined loss of 1p/19q, pointing out that at the molecular level mixed tumors do not exist [7]. Subsequent studies showed that anaplastic oligodendrogliomas showed in addition to 1p/19q loss other lesions (e.g., loss of 9p), but that typical genetic lesions found in glioblastoma were mutually exclusive with combined 1p/19q loss [8,9]. This pointed toward a glioblastoma-like genetic background of many anaplastic mixed oligoastrocytomas, further emphasizing that true mixed oligoastrocytic tumors do not exist.

Oligodendroglioma: 1p/19q loss & chemosensitivity

A second major breakthrough in this field was the observation by Louis and Cairncross that loss of 1p/19q was correlated to a 100% response rate to PCV chemotherapy, as opposed to 20–25% response rate in the absence of 1p/19q loss [10]. Further studies noted that typical oligodendroglial histological features were associated with a greater likelihood of 1p/19q loss, despite the occurrence of 1p/19q loss in tumors with atypical oligodendroglial or even pure astrocytic features [11]. Several series however showed that 1p/19q status was more predictive of outcome to chemotherapy than morphology. This resulted in tendency toward a more restrictive diagnosis of oligodendroglial features, and in many studies into the nature of the sensitivity of chemotherapy of 1p/19q co-deleted oligodendrogliomas [12]. These studies explored in particular the presence of other lesions on 1p and 19q, including epigenetic alterations on these chromosomes. A new findings coming out of these studies was the simultaneous discovery by Burger and Jenkins that the combined loss of 1p/19q in fact represents a balanced translocation, after which a copy of 1p and 19q is lost [13,14]. This translocation has so far not been observed in any other tumor type, and as opposed to other translocations the centromeric fusion of 1q and 19p does not result in an activated oncogene.

Oligodendrogliomas: initial results of the large prospective studies

These genetic studies allowed molecular analysis of the large prospective Phase III trials conducted by RTOG and EORTC that explored the role of (neo)adjuvant PCV chemotherapy in anaplastic oligodendroglial tumors. These trials, initiated in the mid-nineties were reported for the first time in 2004 and 2005, and although initially failed to show survival benefit from the addition of PCV to radiotherapy they already showed the major prognostic significance of combined 1p/19q loss [15,16]. Side studies showed that as anticipated ˜30% of included tumors did not have an oligodendroglial genetic signature, but evidence of alterations commonly observed in glioblastoma [17]. Moreover, a clear interobserver variation with regard to the histopathological diagnosis was present on both studies, even with respect for what is to be called a typical oligodendroglioma [18,19].

Oligodendroglioma: temozolomide

A major question that rose after the pivotal EORTC trial on temozolomide chemoirradiation in glioblastoma and the trials on PCV in newly anaplastic oligodendroglial tumors, was why temozolomide improved outcome in glioblastoma, a chemoresistant disease but PCV failed to improve outcome in a chemotherapy sensitive tumor [20]. One possibility was the different administration of chemotherapy in concurrent chemoirradiation with temozolomide that had been used in the glioblastoma trial. Here, chemotherapy was combined on a daily basis during the entire radiotherapy phase, which was not the case in the classical (neo)adjuvant PCV studies.

Oligodendroglioma: second-generation trials

In the aftermath of the initial results of EORTC 26951 and RTOG 9402, follow-up studies were designed that no longer embarked on histology per se, but on the presence or absence of combined 1p/19q loss in grade III tumors and mark the transition of trials based on histology to trials based on molecular signature. These trials, CATNON and CODEL aimed to define the role of temozolomide in grade III tumors, a question that rose from the outcome of the EORC trial on chemoirradiation with temozolomide in glioblastoma. Similar trials were initiated on low-grade tumors trying to define the role of chemotherapy in low-grade tumors.

Oligodendroglioma: the role of MGMT, IDH & methylation

The observations of the role of MGMT in sensitivity of glioblastoma to temozolomide resulted in studies of the significance of MGMT promoter methylation in the European study on anaplastic oligodendroglioma, but also in the German study (‘NOA4’) that compared radiotherapy to chemotherapy in anaplastic glioma [21,22]. Surprisingly, both studies found that MGMT promoter methylation was correlated with a improved outcome after radiotherapy alone, for which no clear biological explanation was present [22,23]. These and other studies also reported high MGMT promoter methylation rates (80% range) in grade II and grade III glioma [24]. This unexpected prognostic finding after radiotherapy alone was sort of resolved with the discovery of IDH mutations, which were found in 60–80% of all grade II and III glioma, and in up to 100% in 1p/19q co-deleted tumors in some studies. Not only were IDH mutations found to be prognostic, they also strongly correlated to the presence of MGMT promoter methylation. First of all, this revealed that the significance of MGMT promoter methylation in glioblastoma differed from that in grade III glioma. Genomic wide methylation studies subsequently showed that a CpG Island Methylated Phenotype also exists in glioma, similar to as observed in colon cancer. In vivo models have suggested that the metabolic alterations present in IDH-mutated cells may induce a hypermethylated state, thus leading to an explanation of some of the findings in the clinical studies. At present, this CpG Island Methylated Phenotype phenotype appears to be a very strong prognostic factor for outcome in grade III tumors, and perhaps predictive for benefit to adjuvant chemotherapy.

Oligodendroglioma: long-term follow-up of the clinical trials

The understanding of the impact of 1p/19q co-deletion was impacted again by the reports in 2012 of the long-term follow-up of the EORTC and RTOG studies on adjuvant PCV chemotherapy [25,26]. Surprisingly, after first negative overall survival results now both studies showed a clear survival benefit of the addition of PCV chemotherapy to radiotherapy in 1p/19q co-deleted tumors, with a separation of the survival curves 5–6 years after randomization. The similar observation in both studies give credit to this finding, even though it was made in subgroups both original studies. On top of this finding, the median survival of more than 12–14 years in 1p/19q co-deleted tumors following RT/PCV exceeds that of their anaplastic astrocytic counterparts significantly and points to a very significant long-term survivorship leading to new clinical questions. Moreover, an EORTC trial on low-grade glioma now gives evidence that in 1p deleted low-grade glioma, progression-free survival after temozolomide is equal to that after radiotherapy, whereas in 1p intact tumors radiotherapy is superior (Baumert et al., ASCO 2013, oral communication). This adds further to the notion molecular diagnostic are now part of clinical decision making in neuro-oncology.

Oligodendroglioma: next-generation sequencing

Very recently next-generation sequencing has identified novel mutations in 1p/19q co-deleted tumors: in the CIC and FUBP gene [27]. The clinical significance of these lesions is still unclear, and their functional impact remains speculative. Interest in telomerase activity has also been renewed in neuro-oncology, studies have shown that diffuse glioma are characterized by mutually exclusive TERT and ATRX mutations [28,29]. Surprisingly, 1p/19q co-deleted tumors have the same abnormality as glioblastoma (TERT mutations) whereas diffuse astrocytoma regardless of grade are characterized by ATRX mutations. Based on these findings a new classification of diffuse gliomas based on molecular alterations is being considered.

Oligodendroglioma: unanswered questions

Many questions though remain unanswered. It is clear PCV chemotherapy added to radiotherapy improves outcome of 1p/19q co-deleted patients. It is unclear whether that also holds also true for temozolomide. In the past decade temozolomide has replaced PCV because of its better tolerability, but trials on the combination of RT and temozolomide in oligodendroglial tumors are simply unavailable. Several groups have reported on upfront treatment with chemotherapy alone (temozolomide, or combinational regimens with autologous bone marrow transplant) [30,31]. Lack of long-term follow-up and possible bias by patient selection make the interpretation of these series difficult. Although the ongoing CODEL study will give an answer to this question, this answer will not come in the near future. An important question is whether radiotherapy can be safely postponed, whether chemotherapy alone and radiotherapy at progression will be equally effective. If combinational treatment is indeed making a survival difference this may adversely affect outcome, but delaying radiotherapy in favorable prognosis patients and thus avoiding radiotherapy induced sequelae on cognition may improve quality of survival. Again, no answer from currently available data exists.

On a more basic level, it is still unclear what induces oligodendroglial tumors, and what drives 1p/19q loss. We now know that IDH mutations are pivotal in diffuse glioma, but why some develop TP53 mutations and others 1p/19q co-deletion remains a mystery. The answer to these key questions is hampered by the paucity of 1p/19q loss in vitro and in vivo models. Although these models do exist, in general these tumors are almost impossible to culture (which also holds for IDH mutated tumors). Absence of good models hinder the development of more effective treatments and a better understanding of oncogenesis of 1p/19q co-deleted tumors. We still do not know what other factors (inhibited tumor suppressor genes, oncogenes) drive these tumors. And, perhaps most important, we lack innovative agents. It is good we have a better understanding of the responsiveness to classical radiotherapy and chemotherapy of oligodendroglial tumors. But, all these treatments were available in the mid-nineties, and since then no other treatment has increased our therapeutic arsenal.

This special focus issue will give a detailed overview of the current understanding, diagnosis and treatment of oligodendroglial tumors. The ins and outs of classical histology are not reviewed here, for that topic we refer our readers to the 2007 edition of the WHO text book on neuro-oncology [32]. Each article in this special issue is written by a specialist in the field, with a track record in that area. We are very happy that these colleagues have been willing to contribute to this special issue, which gives a overview of all aspects of oligodendroglioma and will present the standard of both understanding and clinical care.