Fax: (011) 31-10-4391031
Successful treatment of low-grade oligodendroglial tumors with a chemotherapy regimen of procarbazine, lomustine, and vincristine
Version of Record online: 6 JAN 2005
Copyright © 2005 American Cancer Society
Volume 103, Issue 4, pages 802–809, 15 February 2005
How to Cite
Stege, E. M. B.-t., Kros, J. M., de Bruin, H. G., Enting, R. H., van Heuvel, I., Looijenga, L. H. J., van der Rijt, C. D. D., Smitt, P. A. E. S. and van den Bent, M. J. (2005), Successful treatment of low-grade oligodendroglial tumors with a chemotherapy regimen of procarbazine, lomustine, and vincristine. Cancer, 103: 802–809. doi: 10.1002/cncr.20828
- Issue online: 3 FEB 2005
- Version of Record online: 6 JAN 2005
- Manuscript Accepted: 18 OCT 2004
- Manuscript Revised: 7 OCT 2004
- Manuscript Received: 17 JUN 2004
- European Organization for Research and Treatment of Cancer Grant, Astra Zeneca. Grant Number: AZ/01/02
- low grade;
- fluorescence in situ hybridization
Anaplastic oligodendroglioma (OD) tumors, especially those with the combined loss of the short arm of chromosome 1 (1p) and the long arm of chromosome 19 (19q), are sensitive to chemotherapy. Only limited data are available on the role of chemotherapy in low-grade OD. The authors retrospectively studied the outcome of the procarbazine, lomustine, and vincristine (PCV) chemotherapy regimen in a group of 16 patients with newly diagnosed OD and 5 patients with recurrent low-grade OD.
Two groups of patients were studied: newly diagnosed patients with large OD and mixed oligoastrocytomas (OA) and patients with recurrent OD and OA after radiotherapy who still showed nonenhancing tumors. Treatment consisted of standard PCV chemotherapy. In the newly diagnosed and responding patients, radiotherapy was withheld until the time of disease recurrence. Responses were assessed by T2-weighted magnetic resonance image (MRI) scans. Loss of chromosome 1p and 19q was assessed using fluorescent in situ hybridization with locus-specific probes.
Three of five patients with recurrent tumors responded. Thirteen of the 16 newly diagnosed patients showed evidence of response. The median time to disease progression in this group was > 24 months. Only one of these patients experienced disease progression while receiving chemotherapy. Several patients showed a signficant clinical improvement despite only a modest improvement of the tumor on the MRI scans. Even patients without loss of 1p or 19q showed satisfactory responses. No TP53 mutations were found.
Newly diagnosed patients with OD tumors, with or without loss of 1p/19q, responded to PCV chemotherapy. Up-front chemotherapy may be indicated especially for patients with large tumors. MRI scans were of limited value for the assessment of response. A Phase III trial should be initiated to compare radiotherapy with chemotherapy. Cancer 2005. © 2005 American Cancer Society.
The remarkable chemosensitivity of anaplastic oligodendrogliomas (OD) and oligoastrocytomas (OA) is now well established.1, 2 Most trials studying the role of chemotherapy in ODs focused on anaplastic or enhancing tumors, and investigated either chemotherapy with procarbazine, lomustine, and vincristine or with temozolomide.3 Clinical data regarding the chemotherapy regimen received by newly diagnosed patients with low-grade and nonenhancing ODs are scarce. One report described a small series of nine patients treated with PCV, with evidence of response in six patients.4 In another recent series, PCV chemotherapy was followed immediately by radiotherapy, for which reason time to disease progression (TTP) could not be assessed.5 A study of the effect of temozolomide on low-grade glioma included 10 newly diagnosed ODs.6 In that study, two partial (PR) and three minor responses (MR) were observed. Other series on chemotherapy in so-called progressive low-grade gliomas, including oligodendroglial tumors, investigated recurrent and often enhancing tumors, suggestive of anaplastic transformation.7, 8
Recent articles show that approximately two-thirds of ODs and approximately one-half of mixed gliomas show a combined loss of 1p and 19q. The subset of patients with OD and OA with a loss of 1p/19q have a better prognosis and a marked sensitivity to PCV chemotherapy compared with patients with OA.9–12 It is unclear though whether this relation remains true for low-grade ODs treated with up-front chemotherapy.5
Since the initial reports in 1998 on chemotherapy in anaplastic OD we started treating patients with newly diagnosed and large low-grade oligodendroglial tumors (“gliomatosis cerebri”) with up-front PCV chemotherapy, but withheld radiotherapy until the time of disease progression.13 All these patients had large tumors and radiotherapy would have required very large irradiation fields. All required treatment for progressive deficits, headache, or intractable seizures. We now report on the outcome of these patients, including a subset of patients with recurrent and nonenhancing, low-grade ODs. We investigated the presence of tumor loss of 1p and/or 19q and TP53 gene mutations in patients who had available tumor samples.
MATERIALS AND METHODS
Since 1998, all patients with newly diagnosed OD or mixed OA were offered up-front PCV chemotherapy if they had large and multilobe tumors, for which radiotherapy would suggest very large treatment volumes (> 50% of the hemispheres). If they had a favorable response to chemotherapy, radiotherapy was deferred until disease progression occurred. The clinical data of all these patients were retrieved. We included patients with disease recurrence of a histologically proven low-grade oligodendroglial tumor after previous radiotherapy if the magnetic resonance image (MRI) scan showed no or almost no enhancement at the time of disease progression. Gliomatosis cerebri was defined as a diffuse glial tumor extensively infiltrating the brain and involving more than two lobes according to the World Health Organization (WHO) classification.13 Adequate hematologic, hepatic, and renal function and a WHO performance status score of ≥ 2 were required for chemotherapeutic treatment.
All patients received the standard PCV schedule, consisting of lomustine 110 mg/m2 on Day 1, procarbazine 60 mg/m2 on Days 8–21, and vincristine 1.4 mg/m2 (maximum of 2 mg) on Days 8 and 29 in cycles of 6 weeks for a maximum of 6 cycles. Dose reductions were made as described previously.2
Follow-up data were obtained with MRI scans after every second cycle during chemotherapy and thereafter every 3 months. The response to treatment was measured using the modified criteria of Macdonald et al.14 All MRI scans (baseline and follow-up) were accomplished with and without contrast. Tumor size was defined as the largest cross-sectional diameter on any slice of a transversal sequence multiplied by the largest diameter perpendicular to it on the same slice. Response was assessed on T2-weighted images. A 50% decrease of the abnormal area with steroids stable or decreased was considered a partial response (PR). A 25% increase was considered evidence of progressive disease (PD). Neurologic deterioration and/or increased need of steroids were also indicative of PD. Patients had an minor response (MR) if they experienced radiologic improvement but did not have evidence of a PR, and if they had a clear decrease in mass effect of the tumor and had steroids stable or decreased. All other responses were considered stable disease (SD). All MRI scans performed before and during chemotherapy were reviewed by an independent observer who was unaware of the clinical outcome (HGdB). The location of the tumor was noted, as were the boundaries of the tumor (which were either sharp and circumscript or diffuse and vague). The TTP was measured from the start of chemotherapy to the first day of clinical or radiologic disease progression (months). TTP and survival time were measured in months, from the first day of chemotherapy to the date of the event. Toxicity was assessed with the National Cancer Institute Common Toxicity Criteria Version 2.0.
Tumor Material and Histopathology
All tumor specimens were reviewed histologically by JMK, who was kept unaware of the clinical data. For patients from whom tumor samples were available for genotyping, the best tumor area was selected. This selection was based on tumor content, taking the predominant tumor morphology of each individual into account. For patients diagnosed as having OD, the pathologist stated whether he considered the tumor a “classical” OD. Tumors were considered classical when they comprised neoplastic cells with round nuclei and perinuclear halos arranged in a honeycomb fashion. Tumors with these cells, but without an orderly arrangement or with less round nuclei and/or without perinuclear halos and without astrocytic features or with astrocytic cells, were considered as atypical OD. By definition, OAs were defined as atypical ODs.
From each selected tumor block, multiple parallel sections of 3-μm thickness were prepared for the different studies described.
Immunohistochemistry for TP53
Paraffin sections of 3-μm thickness were treated for P53 immunohistochemistry as previously described.12 As the P53-specific antibody we used the mouse monoclonal anti-P53, Do-7, 1:50 (Dako, Glostrup, Denmark). Only nuclear staining was considered positive. In each case, the number of clearly positive cells among 100 cells in a representative area was scored.
P53 Mutation Analysis
Exons 5–8 of the P53 gene were investigated by polymerase chain reaction (PCR)–single-strand conformation polymorphism (SSCP) as described previously.12 In short, each exon was amplified in two overlapping fragments. PCR was performed with 100 ng DNA isolated from a representative tumor area in a final reaction volume of 15 μL containing 1.5 mM MgCl2, 0.02 mM dATP, 0.2 mM dGTP, dTTP, and dCTP each, 0.8 μCi α-32PdATP (Amersham, Buckinghamshire, UK), 20 pmol of each primer, and 0.2 U Taq polymerase (Promega, Madison, WI). PCR analysis was performed for 35 cycles (denaturing at 95 °C for 30 seconds, annealing at 55 °C for 45 seconds, and extension at 72 °C for 1 minute) in a Biometra thermocycler (Biometra, Göttingen, Germany). A final extension was carried out at 72 °C for 10 minutes. PCR products were diluted with loading buffer (95% formamide, 10 mM ethylenediaminetetraacetic acid [pH 8.0], 0.025% bromophenol blue, and 0.025% xylene cyanol), denatured at 95 °C for 5 minutes, and snap-cooled on ice. For SSCP analysis, the samples were run overnight at 7 W on a nondenaturing 6% polyacrylamide gel containing 10% glycerol in 1 × Tris-borate-EDTA (TBE) running buffer. After electrophoresis, gels were fixed in 10% acetic acid, dried on blotting paper on a vacuum gel dryer, and exposed to X-ray film overnight at −70 °C, using intensifying screens. Films were evaluated by visual inspection.
Fluorescence In Situ Hybridization
Locus-specific probes for 19q13 and 19q13.4 (BAC127F23 and RPCI11-426G3, respectively; Research Genetics, Huntsville, AL), 19p13 (BAC2310A1), 1p36 (D1S32), and centromere 1 (pUC1.77) were fluorescently labeled to generate specific fluorescence in situ hybridization (FISH) markers for these loci as has been described previously.10, 12, 15 Briefly, tumor sections were deparaffinized, dehydrated, microwave treated in citrate buffer (pH 6.0) for 5 minutes, digested in pepsin solution (0.005% in 0.1 M NaCl, pH 1.5–2.0) for 15 minutes at 37 °C, dehydrated through a series of 70%, 90%, and 100% ethanol, and air-dried. Dual-probe hybridization was performed using differentially labeled probes. For the indirect labeling, we used digoxigenin-16-dUTP and biotin-16-dUTP (both manufactured by Roche Diagnostics, Mannheim, Germany). Probes and target DNA were denatured simultaneously on a hotplate at 80 °C for 3 minutes, followed by overnight incubation at 37 °C. Slides were then washed in 1.5 M urea/0.1 × standard saline citrate (SSC) at 45 °C for 15 minutes, rinsed in 2 × SSC at room temperature, and dehydrated. Slides were air-dried after dehydrating. The digoxigenin and biotin-labeled probes were detected using sheep-anti-digoxigenin labeled with fluorescein isothiocyanate (FITC; Roche Diagnostics) and avidin labeled with CY3 (Brunschwig Chemie, Amsterdam, The Netherlands), respectively. Nuclei were counterstained with 4',6-diamidino-2-phenylindole (DAPI) in antifade solution. A Zeiss Axioplan microscope (Zeiss, Thornwood, NY) equipped with single, dual, and triple-pass filters (DAPI/FITC/tetramethylrhodamine isothiocyanate [TRITC]) was used to count the number of FISH signals for each locus-specific FISH probe. Sixty nonoverlapping nuclei were enumerated per hybridization. Ratios were calculated for 1p versus centromere 1 and 19q versus 19p by dividing the number of signals of the marker by the number of signals of the reference. A ratio < 0.80 was considered as allelic loss.
Between July 1998 and September 2003, 5 patients with recurrent and 16 patients with newly diagnosed low-grade OD or mixed OA were treated with PCV chemotherapy. Eleven patients (9 newly diagnosed) had gliomatosis cerebri. The median patient age was 43.5 years (range, 25–63 years). Patient characteristics are listed in Table 1. All but one patient had epileptic seizures as the initial symptom. Three of these patients had additional symptoms at the time of presentation, such as headache, paresthesias, or sensory aphasia. At the time of our report, the median follow-up time is > 25 months. At central pathology review, 7 patients were diagnosed with classical OD and 14 were diagnosed with atypical OD or OA.
|UPN||Group||Gender||Age||Histology||Predominant localization||1p||19q||Response||TTP||OS||Survival status|
|66||Rec||M||40||Atypical||GC||No loss||No loss||PD||0||20||DOD|
|144||Rec||F||56||Classic, Grade II||GC||Loss||UK||MR||31||38||DOD|
|145||Rec||M||31||Atypical, Grade II||Frontotemporal||No loss||Loss||MR||55+||55+||Alive|
|152||Rec||M||25||Atypical, Grade II||Temporal||No loss||No loss||SD||4||4||DOD|
|153||Rec||F||54||Atypical, Grade II||Temporal||UK||UK||PR||11||12||DOD|
|67||New||M||53||Atypical, Grade II||GC||Loss||Loss||MR||47||47||DU|
|69||New||M||53||Atypical, Grade II||GC||UK||UK||MR||36||58+||Prog, alive|
|146||New||M||41||Classic, Grade II||GC||Loss||Loss||PR||24||52+||Prog, alive|
|149||New||F||40||Atypical, Grade II||Parietotemperal||Loss||Loss||PR||31+||31+||Alive|
|154||New||M||52||Atypical, Grade II||Temporal||Loss||Loss||MR||37+||37+||Alive|
|155||New||F||37||Atypical, Grade III||Temporal||No loss||Loss||MR||20||24||DOD|
|182||New||M||61||Atypical, Grade II||GC||No loss||No loss||MR||20||20||DU|
|190||New||M||39||Classic, Grade II||GC||Loss||Loss||SD||68||68+||Alive|
|191||New||F||46||Classic, Grade II||Bifrontal||Loss||Loss||MR||25+||25+||Alive|
|194||New||M||30||Classic, Grade II||Temporal||No loss||No loss||PR||16||23+||Prog, alive|
|196||New||M||46||Classic, Grade II||GC||Loss||No loss||MR||30+||30+||Alive|
|245||New||M||56||Classic, Grade II||Temporal||Loss||Loss||MR||20+||20+||Alive|
|279||New||M||63||Atypical||Frontal||No loss||No loss||MR||14+||14+||Alive|
A total of 98 cycles of PCV were administered (median, 5; range, 1–6 cycles). Twelve cycles were dose delayed, and another 11 cycles were dose reduced. Most of the dose reductions or delays were due to hematologic toxicity. Three patients experienced Grade 3 bone marrow toxicity, one patient experienced Grade 3 diarrhea, one patient experienced Grade 3 fatigue, and three patients experienced hepatotoxicity. Seven patients received the intended six cycles. Eleven patients had to discontinue treatment due to reversible and asymptomatic myelosuppression (n = 9) or hepatotoxicity (n = 2). One patient with Grade 3 hepatotoxicity continued treatment with temozolomide after recovery. In another patient, treatment was discontinued at the time of diagnosis of a secondary, unrelated malignancy.
Table 1 summarizes the individual responses in both groups. Three of five patients with recurrent tumors showed some response. One patient in this group died during treatment, probably due to a seizure. Because seizures were the main reason to initiate treatment in this patient and the seizure frequency did not decrease, he is considered to have PD. At the time of the current report, 4 of the 5 patients treated for recurrent tumors had PD and died (a median survival of 20 months). In three of the newly diagnosed patients, a PR was observed, with another 10 patients showing neuroradiologic evidence of a MR. With a median follow-up of 25 months, the median TTP in this group has not been achieved yet and is > 24 months. In one patient with a mixed OA showing loss of 1p, treatment had to be discontinued after the first cycle because of severe behavioral abnormalities. He died 6 months later. Three other patients in the newly diagnosed group have died, two of these due to unrelated secondary malignancies. One patient in the newly diagnosed group with a MR to chemotherapy was irradiated immediately after the end of chemotherapy. In this patient, chemotherapy was started on the intensive care unit because of impending cerebral herniation. After completion of the chemotherapy regimen, it was decided that further delay of radiotherapy would bring the risk of a rapid deterioration at the time of disease recurrence, and she received radiotherapy to the tumor area. Despite this, the patient experienced disease recurrence 1 year after the end of radiotherapy and died shortly thereafter. Apart from the patient with early discontinuation of treatment, 6 other patients have PD after a median interval of 22 months.
No relation was observed between any pretreatment imaging characteristics and response to treatment. A striking observation in this group of patients was the often marginal improvement of MRI abnormalities on T2-weighted images, often despite impressive clinical improvement (Fig. 1). Four patients improved two points on the WHO performance scale despite only a MR after neuroimaging. Three patients with a MR improved one point on the WHO scale. Most of the other responding patients showed more subtle improvements, like an improvement in seizure control. Furthermore, in some patients, the MRI abnormalities continued to improve after the discontinuation of PCV chemotherapy (Fig. 2). Recently, we started to use C11-methionine positron emission tomography (PET) scans to follow these patients. PET data are available for one of the patients from this series (Fig. 3). Despite a MR on an MRI scan, a C-11 methionine PET scan showed an almost complete disappearance of the area with hyperactivity.
For all but three patients, material was available for genotyping. In two, reliable assessment of 19q status was not possible. Seven patients were diagnosed with combined 1p/19q loss, 4 of whom had a PR, 2 a MR, and 1 SD for 68 months. In only two other patients a PR was observed (one without either 1p/19q loss; in the other, no material for genotyping was made available). However, good responses (two MRs, one PR) were observed in three newly diagnosed patients without either loss of 1p or19q. Two of these patients had disease recurrence after 16 and 20 months, respectively, which is shorter than the time to disease progression in any of the patients with 1p/19q loss. Although in 3 tumor specimens, > 20% of nuclei showed positive staining with the P53 antigen, no mutations were identified after SSCP mutation analysis. One of these specimens had a single loss of 19q and another specimen had a single loss of 1p. All three patients responded to chemotherapy.
Despite the known chemosensitivity of anaplastic OD to chemotherapy, much less is known of the results of chemotherapy of low-grade OD. The current study shows that up-front PCV chemotherapy is a promising approach in newly diagnosed low-grade oligodendroglial tumors, with lasting responses in most of the newly diagnosed patients. So far, only a few patients in this group have either failed or had disease recurrence, with a median disease progression-free survival that currently is > 24 months. The shorter response duration in the patients with recurrent low-grade tumors resembles the response duration of patients with recurrent anaplastic ODs.2
We used up-front chemotherapy instead of radiotherapy to avoid irradiating large gliomas with oligodendroglial elements. In these patients, we expected a relatively favorable prognosis and survival of > 2–5 years. Radiotherapy would have required large radiotherapy portals, which would have put these patients at risk to develop delayed leukoencephalopathy. Although a recent study on cognitive deficits in patients with low-grade gliomas detected only an increased risk on radiotherapy-induced cognitive deficits if the radiotherapy fraction size was > 2 gray (Gy), this is likely to be different for patients who receive whole-brain irradiation.16, 17
Our data show that we have succeeded in postponing radiotherapy. So far, five newly diagnosed patients have experienced disease recurrence, one of who had been irradiated because of her previous disease history. In the other four patients, routine follow-up detected disease progression at an asymptomatic stage. Despite the frequent but asymptomatic hematologic toxicity, PCV was generally well tolerated although most patients did not receive the full six cycles because of clinically asymptomatic and reversible myelosuppression.
In recurrent anaplastic OD, a strong relation exists between the combined loss of 1p and 19q and the response to chemotherapy.12, 18 In the current series, newly diagnosed patients with tumors without loss of either 1p or 19q were also responding. This is in line with observations of others who noted responders in patients with OPD regardless of chromosomal loss, or a similar response rate in patients with low-grade astrocytomas compared to patients with low-grade OD.5, 6 Although the response duration seems shorter than in the tumors with loss of 1p and or 19q, and relatively more “true” PR were observed in patients with loss of 1p/19q, the small numbers prevent further conclusions. More data with longer follow-up are clearly needed. With the currently available data, in individual patients, the presence or absence of specific genetic lesions in low-grade oligodendroglial tumors should not be used for treatment decisions.
An important observation in the current study is the poor relation between conventional MRI scan response criteria and clinical outcome. In several patients, only some decrease in mass effect on the T2-weighted MRI scan abnormalities was observed, despite impressive and lasting clinical improvement. The T2-weighted images do not allow a differentiation between gliosis in responding patients and active tumor, and hence may show only a decrease in mass effect and a more clear distinction between white and grey matter in case of a response. Furthermore, in one case, a continuous improvement was noted for until one year after the end of chemotherapy, which has also been observed by others.4
To overcome the difficulties of response assessment in the treatment of low-grade glioma, other imaging modalities have been considered. Thallium single-photon emission computed tomography scanning was found useful for the evaluation of response to chemotherapy of anaplastic glioma, but is less useful for low-grade gliomas, as it often fails to show accumulation.19 18-Fluorodeoxyglucose PET scanning is also less useful for brain tumors, because of the high glucose metabolism of the normal brain. PET scanning with radioactive-labeled amino acids like C11-methionine appears to be more useful to evaluate low-grade gliomas.20–22 Because of the short half-life time of C11-methionine, a cyclotron for on-site synthesis is required, which limits the availability of this technique to only a very few institutions. In another study of temozolomide chemotherapy in low-grade glioma, magnetic resonance spectroscopy did not yield additional information over standard response assessment using MRI fluid-attenuated inversion recovery (FLAIR) sequences.23
Our results and those of others show the limited value of smaller Phase II studies comprising patients with low-grade gliomas with response as the primary end point. Because TTP may be a much more reliable and relevant end point than responses observed on MRI scan images, further information on the role of chemotherapy in low-grade (oligodendro)glioma will only be obtained from larger Phase III studies comparing chemotherapy with radiotherapy. A further rationale for such a study design is the moderate efficacy of radiotherapy in low-grade glioma.24 The European Organization of Research and Treatment of Cancer (EORTC) is embarking on such a study. As the currently available studies suggest that neither oligodendroglial morphology nor combined loss of 1p/19q is a prerequisite for response to chemotherapy, the EORTC study will include both low-grade astrocytoma and OD. The study will randomize newly diagnosed patients with low-grade glioma to receive radiotherapy (50.4 Gy in fractions of 1.8 Gy) and daily continuous temozolomide (75 mg/m2 on Days 1–21 every 4 weeks for 12 cycles). The patients will be stratified for 1p loss.
Outside trials, the decision whether to use either chemotherapy or radiotherapy as the primary treatment, should be based on the expected side effects of treatment. The best available data do not support a major effect of radiotherapy on the cognitive function of patients with low-grade glioma of limited size, provided fractions < 2 Gy are used to a total dose of 50–55 Gy.16 For patients with small tumors, radiotherapy offers a shorter treatment without systemic side effects.
In conclusion, we show that PCV chemotherapy is effective in patients with large low-grade oligodendroglial tumors. Patients with oligodendroglial tumors or mixed OAs without loss of either 1p or 19q may also respond to this treatment. MRI scanning is of modest value to assess response to chemotherapy in patients with low-grade glioma. A Phase III study is required to assess the role of chemotherapy in newly diagnosed patients with low-grade (oligodendro)glioma. The role of metabolic imaging, in particular C11-methionine PET scanning and magnetic resonance spectroscopy, needs further exploration.
- 13Gliomatosis cerebri. In: KleihuesP, CaveneeWK, editors. Pathology and genetics. Tumours of the nervous system. Lyon: IARC, 2000: 92–93., .
- 24Randomized trial on the efficacy of radiotherapy for cerebral low-grade glioma in the adult: European Organization for Research and Treatment of Cancer study 22845 with the Medical Research Council study BRO4: an interim analysis. Int J Radiat Oncol Biol Phys. 2002; 52: 316–324., , , et al.