Oligodendroglioma (OD) is a rare tumor of the central nervous system. Its current World Health Organization (WHO) definition is “a well-differentiated, diffusely infiltrating tumor of adults, typically located in the cerebral hemispheres and composed predominantly of cells morphologically resembling oligodendroglia” [1]. By definition, mixed oligoastrocytic (OA) tumors have morphologic characteristics of both astrocytic tumors and pure ODs. Recent European Organization for Research and Treatment of Cancer (EORTC) and Radiation Therapy Oncology Group (RTOG) trials used the presence of 25% oligodendroglial elements as the arbitrary cutoff point for the diagnosis of a mixed tumor. The recent 2007 WHO classification of brain tumors no longer accepts the diagnosis of anaplastic oligodendroglioma (AOD) if necrosis is present [2]. These tumors are now considered amongst the glioblastomas, perhaps with the addition of “with oligodendroglial features.” All oligodendrogliomas constitute 4-15% of glial tumors. An interesting feature of oligodendrogliomas is the common possibility to state the timing and clinical onset and evolution of the disease, compared to other tumors. This characterisitic allowed us a survival analysis not only from the operation time, but also from the beginnning of the clinical symptoms. We present a large clinical series of oligodendrogliomas with long follow-up and we present and discuss our results.

We describe 90 patients treated in Nottingham University Hospital between May 18, 1962 and October 31, 1989. We have studied their clinical files during hospitalization and at follow-up until October 31, 1989. All pathological diagnoses were reevaluated according to the modern classification. The histological results were all revised and fulfilled the WHO diagnostic criteria for oligodendroglioma. The outcome is classified according to the Glascow outcome scale. Outcome is described at 3 months post-op. Post-op mortality is defined as mortality within one month after the operation.

We examine their exact pathological diagnosis, age, sex, incidence, symptoms and signs, hydrocephalus, location, surgical approach, extent of removal, reoperations, radiotherapy, postoperative outcome and prognostic factors of the tumors. We have studied in particular the survival of the patients after initial symptoms and after operation, by using life tables and logrank tests. The survival was studied in relation to histological grading, sex, age, epilepsy as initial symptom, infiltration of eloquent area, extent of removal, lobectomy, reoperation and when the patient was operated. 88 of them were operated and 2 of them died in the hospital before any surgical intervention. For another 11 patients with this diagnosis there was no available clinical follow up and were excluded from the study.

See Table 1. Two-tiered WHO grading system: 71=79% of the tumors were grade 2 and 19=21% were grade 3.

58% were men and 42% women. Mean age was 40 years (see Table 2).

101 patients in 26½ years in 3 million people, that is 1.26 cases:1000000/year.

See Tables 3, 4, 5.

Clinical and neuroradiological signs 11. Among these, 4 were operated with shunt. In the other cases tumor removal relieved the hydrocephalus. Another 4 with probable hydrocephalus.

37 gross total removal; 47 partial removal; 4 biopsy.

Twenty-five patients were operated more than once. Mean time from first operation 3 years and 6 months. In 4 of them it was not real recurrence but rather completion of the operation in the first year. In 4 out of 21 real reoperations there was increased grade of the tumor. Second reoperation was done in 3 patients. Mean time from first reoperation 2 years and 1 month.

2 patients entered the operating theatre in coma and died post-op without recovering. In these patients as well as in 2 patients who died before any surgical intervention, a large intracerebral hemorrhage was found. 2 patients died with severe neurological deficits after the operation. In one of them ligation of the carotid artery was done for intractable hemorrhage. Another three were in worse condition after the operation, 2 of them with medium and 1 with bad outcome. 3 post-op hematomas; one died. 1 severe infection; removal of bone flap. 1 pulmonary embolism.

The outcome according to the Glascow Outcome Scale was: Good outcome for 51 patients, medium for 17, bad for 8, while 5 died. Perioperative mortality: 5,7% for the whole series; Decreased to 1,2% after 1974 (post CT-scan era); Morbidity 9%; Returned to previous occupation 58%.

Figure 1. Survival of grade 2 vs. grade 3 tumors. Black line: grade 2. Green line: grade 3.

62% had radiotherapy. There were no strict criteria for applying post-op radiotherapy. Usually whole-brain 5000-5500 cGy were given in total divided in daily doses of 180-200 cGy. Usually after operation, but in 2 cases after recurrence. One death from post radiation pneumonia.

Life table analysis and Logrank test for histological grade, sex, age, epilepsy, eloquent area, total or partial removal, lobectomy, reoperation, radiotherapy, time of operation (before or during CT-scan era).

5-years survival 72%; 10-years survival 45%; 20-years survival 9%; half of the patients die within 9 years. Grade 2 vs. grade 3 tumors: see Figure 1. Total vs. partial removal: see Figure 7.

Figure 2. Life table of tumors with gross total (black line) vs. partial (green line) removal.

5-years survival 75%; 10-years survival 46%; 20-years survival 9%; half of the patients die within 9 years. The results are very similar to those of the whole series.

5-years survival 55%.

5-years survival 79%; 10-years survival 62%; 20-years survival 14%; Half of the patients die within 10 years.

5-years survival 66%; 10-years survival 38%; 20-years survival 8%; half of the patients die within 8 years.

Figure 3. Life table of tumors with (black line) vs. without (green line) radiotherapy.

See Figure 3. 5-years survival after RXT 68,2%, 10 years 48,5%, 20 years 6,4%, median survival 9 years. 5-years survival without RXT 71,5%, 10years 33,5%, median survival 8 years.

See Table 6.

Even if there is a tendency for better survival after operation with younger age (p=0.1147) and radiotherapy (p=0.1399), only three factors reached statistical significance:

The first is gross total removal (p=0.0250). The second is epileptic seizures as first symptom (p=0.0216). The third is operation after 1974 (p=0.0352).

These significances were lost if we measure survival after the patients first symptoms and not after operation. The stronger aggravating factor predicting survival after initial symptoms was infiltration of eloquent areas, which anyhow did not reach statistical significance (p=0.9360).

Oligodendrogliomas constitute 4-8% of all gliomas. The mean age of patients at diagnosis is 40 years. Ratio of males to females is around 3:2. They usually arise in the cerebral hemispheres, especially the frontal lobes. They are rare in the posterior fossa. Cerebrospinal fluid metastases occur in 1%. Spinal ODs comprise only about 2.6% of intramedullary tumors of the cord and filum [3]. Our results are in accordance with the published bibliography.

Our series is a clinical series. It spans a long time period of 27 years. During this time period major innovations took place. The computed assisted tomography CT scan) has significantly improved our results. Magnetic resonance imaging (MRI scan) has also been used in few of the most recent cases. The operating microscope has begun to be used. Many different surgeons (including 6 consultant neurosurgeons) have operated these patients. There was no uniform protocol related to the use of radiotherapy or chemotherapy. Despite these facts the results were very good and comparable to the best results in recent series. The mean post-op survival was 9 years.

ODs typically present with a long history of seizures. Seizures are the presenting symptoms in 50-80% of cases. In our series seizures were the presenting symptom in 59% of cases. Our results show that seizures as presenting symptom led to worse prognosis.

Before the CT-scan era patients were sometimes operated after an apoplectic event due to peritumoral intracerebral hemorrhage. Two of the deaths in our series were in patients in coma from intracerebral hemorrhage before the CT-scan period. These patients were operated successfully but they did not recover from coma. Also, the only two patients who did not have any surgical intervention were admitted in coma and after death they were found to have intracerebral hemorrhage.

The remainder of the presenting symptoms are not specific to ODs, but are more often related to local mass effect or increased intracranial pressure.

Three prognostic factors were statistically significant for survival after operation: Complete vs. partial removal, epileptic seizures as first symptom (aggravating factor) and operation before vs. after 1974, when the CT-scan has started to be used.

We have studied the survival of our patients not only after time of operation but also from first symptoms. The collection of these data is more difficult and imprecise. Anyhow, we believe that this is the only way of taking into consideration delays in treatment and if and how they can affect prognosis.

None of the prognostic factors has reached statistical significance when time measurement was after first symptoms. Infiltration of eloquent area was the most important aggravating factor, but it did not reach statistical significance. Obviously this infiltration did not allow total surgical removal.

History of allergies and/or asthma is associated with a decreased risk of anaplastic oligodendroglioma and history of asthma only is associated with a decreased risk of oligodendroglioma. A family history of brain tumors is associated with an increased risk of anaplastic oligodendroglioma. Having had chicken pox is associated with a decreased risk of oligodendroglioma and anaplastic oligodendroglioma [4].

All histological diagnoses of oligodendroglioma were revised before this study, so that they are in accordance with the recent WHO classification. Intraventricular tumors diagnosed in the past as such, can be neurocytomas (2 cases). A case of anaplastic oligodendroglioma with necrosis was reclassified as glioblastoma. The presence of 25% oligodendroglial elements was used for the diagnosis of OA.

Isolated tumor cells of ODs consistently penetrate largely intact brain parenchyma; an associated solid tumor component may or may not be present. Classical pathology of oligodendrogliomas includes: “Honeycomb” appearance. “Fried-egg” appearance of delayed fixation. “Chicken-wire” vasculature. Microcalcifications 60-90%. Perineuronal satellitosis. Subpial infiltration and perivascular satellitosis (secondary structures of Scherer). 16% of hemispheric ODs are cystic (cysts form from coalescence of microcysts from micro-hemorrhages, unlike astrocytomas which can actively secrete fluid). Typically GFAP(-) but can exhibit cells (minigemistocytes and gliofibrillar oligodendrocytes) that exhibit small amounts of GFAP(+) cytoplasm. Microgemistocytic oligodendroglioma is strongly GFAP(+).

A number of differentiation antigens that are specifically expressed by normal oligodendrocytes are not diagnostically significant markers for oligodendrogliomas: MBP is no longer expressed. Others are expressed in a minority of cases, ex. PLP, MAG.

As oligoastrocytoma is diagnosed the oligodendroglioma which has a component of more than 25% neoplastic oligodendrocytes. It can present in compact or diffuse variants. Immunoreactivity for the neurotrophin receptors TrkA, B and C is positive only in normal and neoplastic astrocytes, but not in normal or neoplastic oligodendrocytes.

Oligodendroglioma grading systems are less predictive than those for astrocytomas, particularly with regard to intermediate-grade tumors.

The anaplastic oligodendroglioma displays increased cellularity, marked atypia, microvascular proliferation and frequent mitoses. Necrosis classifies the tumor as glioblastoma [2].

Grade 2 tumors have 1-3 mitoses per 10 high power fields (HPFs), compared to 2-5 mitoses per 10 HPFs for anaplastic tumors [24].

The anaplastic oligodendroglioma has a better prognosis than the anaplastic oligoastrocytoma.

Glioblastomas with oligodendroglial component have probably better prognosis than classical glioblastomas.

Previous grading systems that have been used:

• 1. Ringertz (1950) grading system [5] : Well differentiated, anaplastic/malignant and glioblastoma multiforme. Criteria: nuclear pleomorphism, mitotic activity, vascular proliferation, necrosis. Unlike astrocytomas, the presence of necrosis does not automatically upgrade a tumor from anaplastic oligodendroglioma to glioblastoma.
• 2. St. Anne/ Mayo grading system [6] : Four-tiered system. Nuclear abnormalities, mitoses, endothelial proliferation, necroses. Grade 1 lesion has none of the characteristics; grade 2 has one; grade 3 has two; grade 4 has 3 or 4.
• 3. Daumas-Duport criteria [7] : They suggest that oligodendrogliomas are composed of isolated neoplastic oligodendrocytes that are entrapped in a fibrillary background of axons and gliosis. They stress that they are underdiagnosed and often interpreted as fibrillary astrocytomas. Using proposed criteria ODs constitute 33% of all adult supratentorial glial tumors. These views are controversial. They have also shown that oligodendrogliomas with even a single mitosis have a worse outcome than those without.

Patients with 1p19q deletion have a better prognosis and respond better to treatment, especially to chemotherapy. It was recently demonstrated that the combined loss of 1p and 19q is mediated by an unbalanced translocation of 19p to 1q [8]. Loss of heterozygosity (LOH) on 19q13.3 in 50-80% of cases.

Compared with low-grade ODs, AODs usually have additional chromosomal deletions, in particular loss of heterozygosity for 9p and/or deletion of the CDKN2A gene (p16) [9]. These occur in 33%–50% of AODs and are associated with tumor progression. As a rule, other chromosomal abnormalities are found in low-grade and in high-grade ODs with atypical morphological features. Deletions on chromosome 10 have been reported in 19%–25% of anaplastic oligodendroglial tumors (AODs). Amplification of the EGFR gene on chromosome 7p occurs in 20%–30% of AODs, but reportedly never in those with deletion of 1p and 19q, and correlates with a poor prognosis.

Proliferation data from 20 series with 108 cases note a tendency toward correlation between labeling index (LI) and histological malignancy. Mean LI is 1,2-6 for low-grade and 5,3-11,1 for high-grade oligodendrogliomas. MIB-1 LI data of 8 series with 513 patients show that there exists a significant association between labeling index and grade or survival. They also show that a threshold of around 5 can reliably identify a subset of more aggressive low grade oligodendrogliomas [10].

Flow cytometry of 60 oligodendrogliomas has revealed S-phase factor to be independent from histologic grade predictor of survival. The SPF identified both more aggressive grade 2 tumors and less aggressive grade 3 tumors [11].

CT typically demonstrates a well demarcated hypodense mass, often near the cortical surface. Frequently, intratumoral calcification is seen and occasionally subclinical or symptomatic hemorrhage is detected. Contrast enhancement, when seen, is typical of aggressive or anaplastic tumors.

CT-scan improved post-op survival to a significant degree in our series. It did not only improve surgical planning and post-op follow-up and treatment, but also allowed an earlier diagnosis, and consequently improved significantly post-op survival. Before the CT-scan period some patients arrived to the Hospital in a preterminal state.

The T1 images of MRI-scan usually reveal a hypointense mass, and the T2 images disclose a hyperintense lesion that extends beyond what is appreciated on most CT scans. FLAIR imaging may prove to be even more effective in showing extent of disease. FLAIR images usually reveal a hyperintense mass that is larger than is seen on T2 images. As with CT, anaplastic or aggressive oligodendrogliomas usually enhance following intravenous gadolinium infusion [12]. The number of patients with an MRI-scan in the current series was low and we cannot draw any conclusions concerning its use.

Neuroradiology cannot reliably differentiate ODs from astrocytomas, especially taking into account that many of the tumors are oligoastrocytomas. Particularities: 30-40% have visible calcification on plain radiographs and up to 90% on CT. 50% of all grades, especially the anaplastic ones present moderate patchy enhancement.

MR spectroscopy shows, as expected, depressed NAA levels and absent lactate. MR spectroscopic imaging can help to better define the grading of the tumor. Functional MRI can help in pre-op planning, especially in cases where the tumor infiltrates eloquent areas. Diffusion tension imaging can be useful for demonstrating adjacent white matter tracts.

Timing of operation: These patients are usually operated at the time of diagnosis. Deferring aggressive treatment (surgery and/or radiation) until symptoms clearly indicate tumor progression does not compromise outcome for a select group of patients who are diagnosed at a young age with a nonenhancing mass on MR and seizures as the lone presenting symptom [13].

Gross total removal (90-95%) if possible. Gross total removal led to better post-op outcome in our series also. This is not always possible, because the tumor is infiltrating and tumor within eloquent areas cannot be removed. The proof of better outcome with gross total removal is less convincing for anaplastic tumors.

Can malignant transformation be prevented by operation? The answer is equivocal. 19% of our patients presented with tumors of increased grade at reoperation at a mean time of 3 and a half years after first operation.

Serial sampling can be used intraoperatively. Stereotactic methods or neuronavigation and intraoperative imaging ultrasonography or MRI have been used in recent years, but not in this series.

The high rate of epileptic seizures (60-90%) obliges us to address specifically this problem during the operation, ex. intraoperative electrocorticography for excision of epileptic foci is helpful.

In the usual frontal or temporal location there is early infiltration of the basal ganglia and the thalamus, making their total removal difficult.

Their distribution to eloquent areas is, together with malignant transformation that occurs in 20% of cases, the main factor prohibiting reoperation. Otherwise, a patient has been operated 8 times in 41 years and is still alive.

Gross total removal: The existing bibliography supports the conclusion that gross total removal increases post-op survival (see Table 7).

Antiepileptic treatment has practically never been stopped in our patients with ODs, even after a successful operation. One of the main differences between our series and those which are more up to date is that today we use chemotherapy in almost all these patients. Chemotherapy is, generally speaking, more efficacious than radiotherapy [18]. We do not know why patients with anaplastic oligoastrocytomas have better response to chemotherapy than patients with astrocytomas. Possibly, methyl guanine methyltransferase (MGMT), which mediates at least part of the cell resistance to alkylating and methylating agents, is less expressed in ODs and perhaps even more so in 1p/19q codeleted tumors [19]. We also do not know the exact reason, why patients with 1p19q deletion respond better to treatment. Their gene expression profile probably resembles that of normal brain. Radiation therapy (RXT) and chemotherapy must be given at diminished tumor burden but in proliferating cells. The radical operation should consequently be offered at this critical period, but our current grading system is not particularly useful in this regard.

The timing of radiotherapy is controversial for low-grade gliomas. Α prospective, randomized study providing class 1 data [20] has not shown significant survival advantage after its use. Radiotherapy increases the time period until recurrence at the expense of intellectual decline. No radiotherapy is needed after gross total removal. Radiotherapy improves survival in anaplastic oligodendrogliomas. 5400 cGy over 5 weeks. It is obligatory in cases of AODs and AOAs and in cases of malignant transformation.

Grade 2 oligodendrogliomas, especially those with loss of 1p and 19q, are sensitive to combination chemotherapy with procarbazine, lomustine and vincristine (PCV), with 60-65% of patients responding and median response duration of 1-1.5 years. The response rate in mixed oligoastrocytomas may be slightly less but is still better than that of pure astrocytic tumors [21]. Patients with recurrence of anaplastic oligodendroglioma and grade 3 oligoastrocytoma do better with PCV. PCV: Procarbazine 60mg/m² orally for 14 days (days 8 to 21) every 6 weeks; CCNU 110mg/m² orally once (day 1) every 6 weeks; and vincristine 1,4mg/m² intravenously twice (days 8 and 29) every 6 weeks. Temozolomide has also been tried in recurrent oligodendroglial tumors. It has less toxicity. The standard dosing schedule of temozolomide is 5 consecutive days of daily dosing during 28 day cycles. However, different dosing schedules may produce better results. In one study, researchers compared patients who received temozolomide for at least 12 months on the 5/28 day cycle, dividing such patients into two groups: "short term" patients receiving temozolomide for 12-18 cycles and those "long term" patients receiving 19 or more cycles (range was 19 to 32 cycles). Researchers found that there was a statistically significant advantage for "long term" treatment (median progression free survival for "short term" patients was 95 weeks (follow up of 73 weeks), but for "long term" patients the median progression free survival was not yet reached (follow up of 134 weeks) [22]. The results of chemotherapy are better for younger patients and 1p/19q(-) tumors.

• Anaplastic oligodendrogliomas: A recent long-term study does affirm that radiation combined with adjuvant chemotherapy is significantly more efficatious for anaplastic oligodendroglioma patients with 1p 19q co-deleted tumors and has become the new standard of care [23]. Combined treatment does not increease the risk of intellectual decline compared to RXT alone [24].

  According to a recent study for AODs: In 1p19q codeleted cases, chemotherapy alone did not appear to shorten overall survival in comparison with chemotherapy plus RXT, and PCV appeared to offer longer disease control than temozolomide but without a clear survival advantage. Combined chemotherapy plus RXT led to longer disease control and survival than did chemotherapy or RXT alone in cases with no 1p19q deletion [25].

• Oligodendrogliomas in children: They constitute 4-8% of supratentorial tumors in children, 7,7% in our series. 80% have seizures (55% complex and 25% simple partial). MRI shows a poorly delineated tumor with that may have heterogeneous enhancement and mild surrounding vasogenic edema. Cystic degeneration can also be seen. Calcification is better appreciated on CT. There was a tendency towards better survival of younger patients (<40 years old) in our series.

Oligodendroglioma, with is relentless course but long survival, is a very interesting tumor. Surgery remains its main treatment. We have found better post-op survival with gross total removal. Epileptic seizures are very common, aggravate survival when they are the presenting symptom and require specific treatment. Radiotherapy tends to improve survival, but the question when it should be given remains unanswered. Chemotherapy is given today to almost all these patients. It is especially important in cases of 19q, 1p (-) tumors.

We wish to thank the Consultant Neurosurgeons Mr. John Firth, Terrence Hope, Jonathan Punt and the Consultant Neuropathologist Dr. James Lowe of Nottingham University Hospital for allowing us to use the data of the patients of the Department.