Abstract
Background: Low-grade oligodendrogliomas and mixed gliomas can be indolent and remain unchanged for years. Optimal timing and effectiveness of initial treatment is uncertain and therapy can be associated with toxicity.
Methods: Retrospective review of patients diagnosed between 1979 and 1997 with low-grade oligodendroglioma or mixed glioma. Time to progression, survival, prognostic factors, and treatment toxicities were evaluated.
Results: A total of 106 patients (77 oligodendroglioma, 29 mixed glioma) were identified; median age was 36.7 years. Initial presenting symptoms were seizures in 76 (72%) and headache in 11 (10%); tumor was diagnosed as an incidental finding in five patients. Tumor progression was diagnosed in 72 patients (68%). Overall median time to progression (MTTP) was 5.0 years (range 0.5 to 14.2). Median overall survival (OS) was 16.7 years. No prognostic factors reached statistical significance. MTTP and OS were not significantly affected by treatment. Of 62 patients who received radiation therapy, 9 (15%) developed radiation necrosis and 13 developed radiation therapy–related cognitive changes, requiring ventriculoperitoneal shunting in six. Significant myelosuppression was seen in 35 of 76 (46%) patients treated with chemotherapy.
Conclusions: Low-grade oligodendroglioma and mixed glioma have a long median overall survival. There were no apparent differences in either immediate versus deferred treatment or choice of initial therapy on disease-free or overall survival. Chemotherapy was associated with significant acute toxicity in almost one half of patients; radiation therapy produced late neurotoxicity in one third, justifying deferred treatment until clinically necessary.
Oligodendroglioma and mixed glioma (oligoastrocytoma) comprise 5 to 10% of all gliomas.1-4 Low-grade tumors tend to occur in younger patients, where they frequently present with seizures and can be indolent for years.5 Comparable to astrocytic tumors, favorable prognostic factors include age younger than 40 years, tumor grade, and extent of resection6; other possible favorable prognostic factors include seizure presentation and normal neurologic examination.7,8 Lack of contrast enhancement on neuroimaging studies suggests a low-grade malignancy but is not sufficiently reliable.9,10
Low-grade oligodendrogliomas have historically been treated with surgery followed by radiation therapy (RT) or observation.6 Most studies are retrospective and suggest no benefit or slight benefit with postoperative RT.4-6,11-20 Indeed, the major controversy surrounding low-grade gliomas in general is the timing and effectiveness of RT. Two recent prospective trials have helped clarify the role of postoperative RT in newly diagnosed patients with low-grade gliomas.21,22 Most patients enrolled in these trials (60 to 65%) had astrocytomas; 22 to 25% had oligodendrogliomas and 9 to 10% had mixed gliomas. The recent European Organization for Research and Treatment of Cancer (EORTC) trial22 showed no overall difference in 5-year survival comparing postoperative RT (54 Gy) with observation; however, postoperative RT did significantly prolong the estimated 5-year progression-free survival compared with observation (44% versus 37%, p = 0.02). Randomized prospective trials by the EORTC21 and Mayo/North Central Cancer Treatment Group (NCCTG)23 demonstrated that lower dose immediate RT (45 versus 59.4 and 50.4 versus 64.8 Gy) was as effective as higher doses with a lower rate of radiation necrosis. Quality of life analysis revealed significantly lower levels of functioning and more symptom burden with higher RT doses.24 Together, these data demonstrate that 45 to 50 Gy are sufficient for treatment when RT is used.
Radiation-induced toxicities, including dementia and radiation necrosis, are of particular concern in patients with low-grade tumors given their longer survival. The potential for radiation neurotoxicity and the recognition of the chemosensitivity of anaplastic oligodendrogliomas by Cairncross and MacDonald in 198825 has led to the use of chemotherapy in low-grade and “aggressive” oligodendrogliomas and mixed gliomas with positive results.26-35 Nonetheless, there are limited data to assess the efficacy and toxicity of this approach. We review the experience at Memorial Sloan-Kettering Cancer Center with low-grade oligodendrogliomas and mixed gliomas diagnosed during the imaging era between 1979 and 1997.
Methods.
A retrospective chart review was conducted from hospital, neurology, and pathology databases for patients diagnosed between 1979 and 1997 with low-grade oligodendroglioma or mixed glioma. All initial histologic specimens were reviewed at our institution by a single neuropathologist; most subsequent samples were also reviewed. A minimum of 1 year of clinical follow-up information was required. A total of 143 patients were identified; 10 patients were excluded for lack of initial pathology review, 2 for pathology reclassification, and 25 for lack of adequate follow-up information. Therefore, 106 patients were identified who met inclusion criteria. Nine of these patients have been reported previously.31
Age, gender, original pathologic diagnosis, initial symptom, tumor location, imaging findings, extent of resection, and initial treatment were recorded. Recurrence information included pathology, time to first recurrence, and subsequent treatment. Tumor progression was defined as a change in radiographic characteristics such as increased tumor size or new enhancement with or without clinical worsening. Only one patient had clear clinical deterioration in the absence of any radiographic changes. Chemotherapy selection, radiation dose, treatment toxicities, and duration of follow-up and survival were recorded.
Date of diagnosis was the date of initial surgery in 61 patients. However, in 45 patients abnormal imaging consistent with a low-grade brain tumor was documented more than 1 month before obtaining pathologic material. In these patients, the initial imaging date was used as the date of diagnosis; all of these patients had subsequent pathologic confirmation of an oligodendroglioma. Eighteen of the 45 patients were followed clinically after initial imaging and had histology obtained for the first time at tumor progression. In these patients, the initial diagnosis was presumed to be low-grade if at least 12 months had passed from initial imaging to progression.
Statistical analysis.
Progression-free and overall survival distributions were estimated using Kaplan-Meier methodology.36 Variables considered as potential prognostic factors for progression-free and overall survival were age (≤40 years, >40 years), pathology (oligodendroglioma, mixed), and type of resection (gross total resection, other); whether the tumor was initially enhancing was considered for progression-free survival only. Univariate analyses were performed using the log-rank test and multivariate analyses using the Cox proportional hazards model.37,38 For the prognostic factor analysis, time to progression and overall survival were calculated from the date of pathologic confirmation rather than initial diagnosis, as tumor histology and type of resection were determined at the time of surgery. Patients initially observed without surgery (n = 18) were excluded from prognostic factor analysis.
Results.
Patient characteristics.
A total of 106 patients were identified with a median follow-up of 6.0 years. Seventy-seven had oligodendrogliomas; 29 had mixed gliomas. Median age was 36.7 years (range 18 to 64). There were 54 men and 52 women. Tumors were frontal in 59, multilobed (usually frontotemporal or frontoparietal) in 25, temporal in 17, and parietal in 5. Patient characteristics are summarized in table 1.
Patient characteristics
The initial presenting symptom was seizure in 76 (72%), headache in 11 (10%), and cognitive change in 6 (6%) (table 2). During this study, seizures developed in a total of 95 patients (90%). Five tumors were incidentally diagnosed after imaging was performed for other reasons; two following trauma and one each during endocrine evaluation, myelopathy/possible MS evaluation, and breast cancer restaging.
Initial presenting symptom
Diagnosis.
The median time from symptom onset to diagnosis was 18 days (range 0 to 137 months). MRI and/or CT were obtained in all patients. Contrast enhancement was noted in 13 patients (15% of the 86 patients where documented). Calcifications were noted in 21 patients. Diagnosis was established pathologically after initial presentation in 61 patients. Eighteen had no histologic diagnosis until tumor progression (range 19 to 170 months). Patients diagnosed by imaging (n = 27) had a median delay of 2.4 months before histologic confirmation.
Six patients were not diagnosed until 48 months or more after the onset of seizures. Two had initial noncontrast CT scans that failed to identify their tumor. Idiopathic and posttraumatic epilepsy were diagnosed in one each, and imaging was not performed. One patient had a single seizure and did not seek medical attention. The longest delay occurred in a man who had completely recovered from measles encephalitis (age 3) and was diagnosed at age 39 with an oligodendroglioma 137 months after the onset of seizures. Other causes of delayed diagnosis included misdiagnoses of stroke, postpartum depression, benign headache, and history of cocaine abuse.
Treatment.
Surgery at initial diagnosis (table 3) consisted of biopsy in 28, subtotal resection in 41, and gross total resection in 19 (22%). Initial treatment selection was observation in 68 (including the 18 without initial biopsy), RT in 20, chemotherapy in 12, and radiochemotherapy in 6 (see table 3). Median dose of radiation was 5940 cGy (range 4800 to 6500 cGy). RT was eventually administered to 62 patients. Initial chemotherapy at diagnosis was procarbazine, lomustine, and vincristine (PCV) in 14; carmustine in one; and cisplatin in one. Chemotherapy was eventually administered to 76 patients during this study. High-dose chemotherapy with stem cell rescue was used or is underway in eight patients, including six following recurrence with documented malignant transformation and no prior therapy.
Initial treatment and surgical resection
Recurrence.
Recurrence was diagnosed in 72 patients (68%) after a median follow-up of 6.0 years. The median time to progression (MTTP) from initial diagnosis for all 106 patients was 5.0 years (figure 1). Multiple recurrences were diagnosed in 31 patients. New enhancement was observed in 77% at recurrence. Tumor recurrence was histologically confirmed as anaplastic or high grade in 35 and low grade in 20; no pathologic confirmation at recurrence was obtained in 17 patients.
Figure 1. Kaplan-Meier curve shows time to progression and overall survival of all 106 patients. Tick mark indicates last follow-up.
The prognostic factor analysis was based on the 88 patients whose pathology was confirmed at presentation. The MTTP for oligodendroglioma was 4.9 years versus 4.3 years for mixed glioma but this was not significant. MTTP was 8.4 years following gross total resection compared with 4.6 years following biopsy or subtotal resection but this was not significant. Other factors such as age, pathologic diagnosis, or enhancement on initial scan did not significantly affect time to progression or survival.
The MTTP based on initial treatment is summarized in figure 2. Patients observed after diagnosis had MTTP of 4.2 years. Excluding the 18 patients observed without initial tissue diagnosis, MTTP was 3.9 years. MTTP was 5.7 years after postoperative RT and not reached (median follow-up of 5.5 years) after initial treatment with chemotherapy. Patients treated with radiochemotherapy had MTTP of 8.6 years.
Figure 2. Kaplan-Meier curve demonstrates time to progression after initial treatment. RT = radiation therapy. Tick mark indicates last follow-up.
Treatment at first recurrence in the 72 patients was chemotherapy in 32, radiochemotherapy in 16, radiation in 13, surgery alone in 9, and no treatment or pending in 1 each. Chemotherapy selection was PCV in most patients unless previously administered.
Survival.
Only 19 patients died during this study. Median overall survival was 16.7 years. After excluding the 18 patients observed without initial tissue diagnosis, median overall survival was 13.5 years. Survival for oligodendroglioma was 12.1 years and was not reached (median follow-up of 7.1 years) for mixed glioma. No prognostic factor had a significant impact on survival.
Survival based on initial treatment is summarized in figure 3. Patients observed after diagnosis had a median survival of 16.7 years. Excluding the 18 patients observed without initial tissue diagnosis, survival after observation was 10.3 years. Survival was 13.5 years following postoperative RT and not reached (median follow-up of 5.5 years) after initial treatment with chemotherapy. Median survival was not reached in the six patients initially treated with both radiation and chemotherapy (median follow-up of 9.8 years).
Figure 3. Kaplan-Meier curve demonstrates survival after initial treatment. RT = radiation therapy. Tick mark indicates last follow-up.
Toxicity.
Treatment toxicities are summarized in table 4. Radiotherapy was used in 62 patients (58%) over the course of their illness. Radiation necrosis was diagnosed in 9 patients (15%), occurring a median of 52 months following RT (range 3 to 108 months). The median dose (5940 cGy) in these patients was identical to the group as a whole. One patient died from treatment-related brain necrosis documented at autopsy 3 months after finishing intra-arterial carmustine followed by RT (6020 cGy). Delayed cognitive impairment due to radiotherapy was noted in 13 patients and required ventriculoperitoneal shunting in six.
Toxicities of treatment
Chemotherapy was used in a total of 76 (72%) patients; PCV was the initial regimen in 57. Significant myelosuppression (requiring dose reduction or delay, colony stimulating factor, transfusion, or treatment of neutropenic fever) was recorded in 35 (46%) patients treated with chemotherapy. Procarbazine rash was noted in 15 patients. Shingles or other herpetiform infections occurred in 10 patients including one with fatal disseminated zoster. Carmustine pulmonary toxicity was diagnosed in four patients. Intra-arterial carmustine caused visual loss in one patient. One patient had renal failure attributed to cisplatin that required renal transplantation. One patient had fatal chemotherapy complications following transplant at another institution. Allergic reactions to anticonvulsants were documented in 18 patients.
Surgical complications included CNS infections, postoperative pulmonary embolism, and subdural hematoma (one after shunt placement) in two patients each. Other surgical complications included postbiopsy hematoma, ventriculitis, and intraventricular hemorrhage in one patient each.
Other malignancies were encountered 10 times in seven patients unrelated to their CNS tumor. Four patients were diagnosed before their brain tumor with breast cancer; one also had melanoma and another had probable meningioma. Three patients were diagnosed with second malignancies after their brain tumor. Melanoma and squamous cell and basal cell carcinoma (not in radiated ports) were completely resected without recurrence in two patients. The third developed testicular lymphoma 5 years after diagnosis; he had been treated with surgery alone and went on to receive chemotherapy for his lymphoma without recurrence of either tumor.
Of note, one patient had pre-existing MS; he was treated with PCV chemotherapy without recurrence for more than 6 years. Radiation was specifically avoided due to underlying demyelination and potentially synergistic toxicity.39
Discussion.
We reviewed our experience with a large group of patients with low-grade oligodendroglioma and mixed glioma (oligoastrocytoma) diagnosed during the imaging era. Our results are similar to others, revealing a comparable median age at onset (37 years), tumor location (predominantly frontal and temporal), and initial presenting symptom (seizure in 78%),5 as well as rate of gross total resection at initial diagnosis (22%).6 However, the median time from symptom onset to diagnosis in our series was only 18 days, significantly shorter than the 8 months18 or 2.9 years6 reported in studies preceding modern neuroimaging. Earlier diagnosis has created a lead time bias that is reflected in the long median survival of 16.7 years in our patients. CT, and particularly MRI, can also identify asymptomatic or incidental low-grade tumors, as seen in five of our patients, which will influence both recurrence and survival times. However, diagnosis can still be delayed or difficult when the history contains potentially confounding features such as prior trauma, substance abuse, or pre- or coexisting neurologic disease.
Patients were selected for initial therapy by individual treating physicians. Our bias has been to observe patients initially diagnosed with low-grade oligodendroglioma or mixed glioma to avoid potential treatment toxicities unless they have progressive or disabling symptoms or uncontrolled seizures. Delaying surgery with a presumptive diagnosis of a low-grade glioma based on favorable clinical and imaging findings is controversial.8,40-42 However, most agree that this does not compromise patient outcome and may reduce the duration of treatment-related morbidity. Our study included 18 such patients who had MTTP and overall survival better than the group as a whole. This suggests that clinical features can identify a subpopulation of patients likely to do well and in whom treatment can be safely deferred. Even in such patients, if a gross total resection is feasible, it should be performed, as it can occasionally be curative.
The greatest selection bias exists between the observed group and all patients who received immediate treatment regardless of type. Deferral of treatment until progressive symptoms dictate therapeutic intervention did not appear to compromise survival or outcome. Our data are similar to those seen in the prospective EORTC study comparing survival of low-grade gliomas (mostly astrocytomas) following observation versus immediate radiation; unlike the EORTC experience, we did not detect a benefit of immediate treatment on MTTP. Moreover, in our data, it would appear that the choice of treatment has little impact on overall outcome.
Our study did not reveal statistically significant differences for age, histologic diagnosis, extent of resection, or enhancement on time to progression or overall survival. Others have reported these features to be important prognostic factors.11 Our study was limited by the relatively few patients who died or had gross total resection or enhancement on neuroimaging, and therefore, our data may not have had sufficient power to detect the impact of these factors on outcome.
Treatment toxicity is a particular concern given the long survival of patients with low-grade oligodendroglioma and mixed glioma. Indeed, this is the basis for much of the controversy surrounding immediate treatment of low-grade gliomas in general.40,42 We found a high rate (15%) of radiation necrosis and RT-related cognitive impairment, even without prospective neuropsychologic testing. Recent prospective trials examining radiation dose in low-grade gliomas21,23 conclude no survival difference between higher and lower doses and a poorer quality of life in the higher dose group.24 The median dose in our patients was 5960 cGy, which is relatively high and likely accounts for the high incidence of delayed toxicity we observed. However, one third received RT after malignant transformation of their tumor, which usually requires treatment with higher doses. Nevertheless, these data support the findings that substantial RT-induced neurotoxicity accrues in patients with long survival and argues that RT doses should not exceed 4500 to 5000 cGy in patients with low-grade glioma.
Chemotherapy toxicity in our patients is similar to other reports regarding frequency of significant myelosuppression and other side effects. These toxicities are primarily acute and reversible, unlike RT toxicities, which are delayed and irreversible. However, follow-up of our chemotherapy patients may be insufficient to detect delayed consequences of chemotherapy, such as second malignancies. Considering all these factors, when treatment is indicated, our preference is to start with chemotherapy and reserve radiotherapy for malignant transformation or chemotherapy failures. Identification of specific chromosomal changes may assist this decision as described for anaplastic oligodendrogliomas,43 but this remains controversial and unproven in low-grade oligodendrogliomas.44
Our study confirms that low-grade oligodendroglioma and mixed glioma can be indolent for years and have long survival. Treatment timing (immediate versus deferred) and treatment modality (observation, RT, chemotherapy, or radiochemotherapy) did not appear to affect progression or survival. The acute toxicities associated with chemotherapy in one half and delayed neurotoxicities associated with RT in one third of our patients argue for a judicious approach to patients and justify withholding treatment until necessary. Predictable clinical or laboratory features that accurately define the optimal timing and choice of treatment have yet to be fully defined and will require randomized studies in these relatively uncommon brain tumors. Nevertheless, our data suggest that an overall therapeutic approach based on clinical characteristics alone will provide many patients with long survival and minimal treatment-related toxicity.
Footnotes
-
See also page 1402
-
J.D.O. is currently affiliated with The Neuromedical Center, Baton Rouge, LA.
-
Presented in part at the 51st annual meeting of the American Academy of Neurology; Toronto; April 17–24, 1999.
- Received November 2, 1999.
- Accepted December 20, 1999.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.