|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Neuro-Oncology Working Group 01 Trial of Nimustine Plus Teniposide Versus Nimustine Plus Cytarabine Chemotherapy in Addition to Involved-Field Radiotherapy in the First-Line Treatment of Malignant GliomaFrom the Neuro-Oncology Working Group of the German Cancer Society, Germany. Address reprint requests to Michael Weller, MD, Department of Neurology, University of Tübingen, Medical School, Hoppe-Seyler-Strasse 3, 72076 Tübingen, Germany; e-mail: michael.weller{at}uni-tuebingen.de.
Purpose: The role of chemotherapy in the primary treatment of malignant glioma remains controversial. The results from the German-Austrian Glioma trial (GAG, 1983 to 1988) demonstrated a survival benefit for chemotherapy using carmustine (BCNU) plus teniposide (VM26) over BCNU alone in addition to radiotherapy in patients with a Karnofsky performance score (KPS) more than 60. The Neuro-Oncology Working Group (NOA) of the German Cancer Society therefore compared the efficacy of nimustine (ACNU) plus VM26 and ACNU plus cytarabine (Ara-C) chemotherapy in addition to standard radiotherapy in patients with newly diagnosed malignant glioma. Patients and Methods: From 1994 to 2000, 375 patients were randomly assigned to receive radiotherapy and cycles of ACNU 90 mg/m2 intravenously (IV) on day 1 and VM26 60 mg/m2 IV on days 1 to 3 (n = 183), or ACNU 90 mg/m2 IV on day 1 and Ara-C 120 mg/m2 IV on days 1 to 3 (n = 179), in 6-week intervals. Thirteen patients were not eligible after central neuropathology review. The remaining 362 patients had glioblastoma (n = 301) or anaplastic glioma (n = 61). Results: Median survival and 2-year survival rates were 17.3 months and 25% for ACNU plus VM26, and 15.7 months and 29% for ACNU plus Ara-C in glioblastoma, and 60 months and 88% for ACNU plus VM26 and 62.5 months and 72% for ACNU plus Ara-C in anaplastic glioma. Multivariate analysis revealed no survival advantage for either arm or for subpopulations defined by histology, age, or KPS. Hematologic toxicity was more prominent in the ACNU plus Ara-C arm. Conclusion: The median survival times and 2-year survival rates for patients with anaplastic glioma and glioblastoma achieved in the NOA-01 trial compare favorably with historical trials and with the Radiation Therapy Oncology Group database. The toxicity profile favors ACNU plus VM26 for further evaluation.
THE STANDARD treatment for malignant gliomas includes surgical resection and postoperative involved-field radiotherapy. These therapeutic measures result in median survival times of approximately 1 year for patients with glioblastoma (World Health Organization [WHO] grade 4) and of 2 or more years for patients with anaplastic astrocytoma (WHO grade 3).1–3 Several chemotherapy regimens have shown moderate efficacy in malignant gliomas that recur or progress after surgery and radiotherapy in phase II trials.4,5 The role of chemotherapy in addition to radiotherapy in the first-line treatment of malignant gliomas is less well defined.3 No phase III trial of the last two decades has further substantiated the notion that carmustine (BCNU) or any other chemotherapy regimen plus radiotherapy is more effective than radiotherapy alone.6 However, a reanalysis of two former phase III trials6,7 confirmed a small but significant advantage of radiotherapy plus BCNU chemotherapy compared with radiotherapy alone, independent of the major prognostic factors.2 In addition, a meta-analysis of 12 trials on the basis of individual patient data concluded that there is a significant prolongation of median survival of 2 months when chemotherapy is part of the first-line treatment of malignant glioma.3
The design of the Neuro-Oncology Working Group (NOA)-01 trial reported in this article was strongly influenced by the results of the earlier German-Austrian Glioma (GAG) trial that recruited 501 malignant glioma patients from 1983 to 1988. The GAG trial had compared whole-brain radiotherapy plus combination chemotherapy using BCNU and teniposide (VM26) with BCNU alone. The results from the GAG trial indicated an advantage of the combination chemotherapy that was significant for progression-free survival, but lacked significance for overall survival. A strong interaction between Karnofsky performance score (KPS) and type of chemotherapy became apparent on subgroup analysis and was confirmed in the multivariate analysis: patients with a KPS of 70 or more benefited from combination chemotherapy, whereas combination chemotherapy reduced survival in patients with a KPS below 70 (GAG Study Group, manuscript submitted for publication; Table 1
The high rate of pulmonary toxicity of BCNU in the GAG trial provided the rationale for the NOA executive committee to substitute nimustine (ACNU) for BCNU in the subsequent NOA-01 trial. Unpublished pilot studies of ACNU alone had shown no relevant pulmonary toxicity in 51 patients with recurrent and 49 patients with newly diagnosed malignant glioma who had their pulmonary function closely monitored. Furthermore, the superior activity of BCNU plus VM26 over BCNU alone for patients with a KPS of 70 or more in the GAG trial suggested that ACNU plus VM26 be evaluated as one arm of the ensuing NOA-01 trial. Cytarabine (Ara-C) was chosen as the experimental agent to be compared with VM26 as an adjunct to ACNU because of favorable single-center experience with Ara-C in 38 patients with recurrent and 50 patients with newly diagnosed malignant glioma. Finally, when the trial was initiated, whole-brain radiotherapy was no longer the standard type of radiotherapy for malignant glioma. Therefore, the NOA-01 trial compared ACNU plus VM26 and ACNU plus Ara-C in addition to involved-field radiotherapy in the first-line treatment of malignant glioma.
Study Design NOA-01 opened as a prospective trial in 1994. The study intended to randomly assign patients according to KPS to receive ACNU, no further chemotherapy after radiotherapy (KPS 50 to 60), ACNU plus VM26, or ACNU plus Ara-C (KPS 70 to 100). Eligible patients had WHO grade 3 or 4 glioma; were older than 15 but younger than 71 years of age; gave informed consent; and had WBC counts  4,000/µL, platelets 100,000/µL, creatinine  2 mg/dL, and bilirubin 3 mg/dL (inclusion criteria). Patients with a history of restrictive pulmonary disease, chronic heart failure, multiple sclerosis, or stroke were excluded. Patients were stratified according to biopsy or resection, as indicated by the neurosurgeons, and among the resected patients, according to study center. The planned sample size was based on the results of the GAG trial. The study required 90 patients per low KPS arm to demonstrate a survival advantage of 1.55, and 210 patients per high KPS arm to demonstrate a survival advantage of 1.3 (alpha = 5%; beta = 20%).8 Survival time was the primary end point. The low KPS arms accrued 18 patients and were closed in 1997 because of poor patient accrual. An interim analysis for the high KPS arms revealed that a difference of 1.3 could not be expected even when recruitment was completed to enroll 210 patients per arm. Conversely, it was clear at the same interim analysis that the toxicity profile and the survival data were favorable in the NOA-01 trial compared with the GAG trial. These were secondary end points. The trial was therefore prematurely closed in 2000. The high KPS arms recruited 375 patients from 1994 to 2000. Of these 375 patients, 13 were later excluded because central neuropathology review performed at the German Brain Tumor Reference Center (Institute of Neuropathology, University Bonn, Bonn, Germany) did not confirm the diagnosis of anaplastic glioma (WHO grade 3) or glioblastoma (WHO grade 4). The remaining 362 patients represent the valid study group.
Of these patients, central neuropathology review was available for 331 patients. All tumors were classified according to the WHO classification and grading guidelines.9 Patients with only a local neuropathologic diagnosis had glioblastoma (n = 25), anaplastic astrocytoma (n = 5), and anaplastic oligodendroglioma (n = 1). The data summarized in Table 2
Surgery The neurosurgeons were requested to indicate whether they had performed a biopsy or a resection. They were also asked to estimate whether they had removed 20% to 50%, 50% to 90%, or more than 90% of the tumor, or had performed a macroscopically complete resection. Early postoperative cranial computed tomography or magnetic resonance imaging was not mandatory.
Radiotherapy
Chemotherapy
Statistical Analysis
The influence of prognostic variables was explored by a multivariate analysis with the Cox proportional hazard model with stepwise selection of the potential prognostic factors. The resulting model with six variables should assume rough proportionality, although inhomogeneities were indicated by the residuals. Nevertheless, differences between both treatment arms were not apparent either in the univariate log-rank test or in the multivariate analysis. The procedures for the statistical analysis of the NOA-01 trial as delineated in the study protocol included a comparison of the NOA-01 data with the prior data of the GAG trial. This comparison was done once for all GAG patients with a KPS
Patient Characteristics, Therapy, and Follow-Up The patient characteristics for the valid study group are summarized in Table 2  . No relevant differences between the two populations emerged. Glioblastoma was the most common histologic diagnosis. The majority of patients underwent open resection; however, as indicated, the extent of resection was estimated by the neurosurgeons, but was not assessed by mandatory early postoperative neuroimaging. The mean total doses of irradiation were 57.7 ± 6.9 Gy (range, 13 to 65 Gy) and 57.2 ± 8.8 Gy (range, 4 to 74 Gy), and the interval from surgery to first dose 30 ± 8.4 days (range, 12 to 73 days) and 32 ± 13.2 days (range, 18–145 days) in the ACNU plus VM26 (data from 154 patients) and ACNU plus Ara-C (data from 137 patients) arms, respectively. No relevant difference in the type of radiotherapy between the two treatment arms became apparent. Six patients in the ACNU plus VM26 arm (13, 16, 18, 29, 44, and 47 Gy) and eight patients in the ACNU plus Ara-C arm (4, 9, 16, 30, 35, 38, 42, and 47 Gy) received less than 50 Gy total dose, mostly because of early neurologic deterioration and tumor progression. The number of chemotherapy cycles and the total dose of ACNU administered in both arms were comparable (Table 3 ). Follow-up data as available are summarized in Table 4. Chemotherapy was halted in almost half of the patients because of tumor progression. Additional tumor-specific treatments at first recurrence or progression were documented in 114 patients (32%).
Toxicity Hematologic toxicity was moderate, in general (Table 3 ). There were more relevant nadirs attributable to ACNU than to VM26 or Ara-C. In the ACNU plus Ara-C arm, there was a higher incidence of early and late grade 3 to 4 toxicity than in the ACNU plus VM26 arm. In the course of treatment, ACNU was reduced in 6% of the patients in each arm. VM26 was also reduced in 6% of the patients, whereas almost half of the patients (47%) required a dose reduction of Ara-C, indicating that the differential reduction strategy prevented an underdosage of ACNU in the Ara-C group. Conversely, the proposed individual dose escalation could be performed in 40% of the patients for ACNU and 5% of the patients for VM26 in the ACNU plus VM26 arm, and in 20% of the patients for ACNU and 4% of the patients for Ara-C in the ACNU plus Ara-C arm. Nonhematologic acute toxicity was comparable in both arms (Table 5). Despite the relative frequency of WHO grade 3 and 4 hematologic toxicity, hemorrhages or infections were uncommon, presumably because of the short duration of myelosuppression. As indicated in Table 4, three patients in the ACNU plus VM26 arm and one patient in the ACNU plus Ara-C arm died from septicemia during treatment-related myelosuppression.
Survival Survival data broken down by treatment arm, histology, age, and KPS are listed in Table 6 and Figure 1. The median survival times were 60 months for anaplastic gliomas and 16.5 months for glioblastomas. At a median follow-up of 30 to 40 months, only one of 10 patients with oligodendroglial or mixed anaplastic gliomas has died. Multivariate analysis revealed no difference in survival between the ACNU plus VM26 and ACNU plus Ara-C arms determined for the whole group, or separately for anaplastic glioma and glioblastoma, or any subgroup defined by age or KPS. Histology, age, and KPS were the most powerful predictors of survival (Table 7). A comparison of the NOA-01 and GAG trials showed a significant improvement in survival in the NOA-01 trial irrespective of whether the total NOA-01 population was compared with all GAG patients with a KPS 70 (data not shown) or only with those GAG patients with a KPS 70 who received the combination chemotherapy of BCNU and VM26 (Table 8).
Major recent meta-analyses concluded that nitrosourea-based chemotherapy in addition to standard radiotherapy in the first-line treatment of malignant gliomas provides a small but significant and reproducible gain in median survival.1–3 The NOA-01 trial reported here was initiated in the early 1990s when it was believed to be unethical (in Germany) to include a radiotherapy-only arm in a randomized phase III trial for malignant glioma patients with favorable prognostic factors (KPS 70). This trial, therefore, does not provide additional data for the ongoing debate about whether chemotherapy should be considered standard in the first-line treatment of malignant glioma. However, it would appear unlikely that the results of the NOA-01 trial could have been achieved with radiotherapy alone.
This prospective multicenter phase III trial reports better median survival times and 2-year survival rates for glioblastoma and anaplastic glioma (Table 6
To exclude that the patients enrolled onto the NOA-01 trial represent a selected patient sample with favorable prognostic factors, we analyzed the NOA-01 study population with the partition algorithm developed by the RTOG that has been validated with more than 7,000 patients enrolled onto RTOG trials.11,12 The RTOG RPA identifies six prognostic classes of malignant glioma patients on the basis of the following criteria: age, KPS, histology, mental status, duration of clinical symptoms preceding treatment, resection or biopsy, and radiation dose. Each individual NOA-01 patient was assigned the appropriate RTOG-RPA class. The estimates for the median survival times and the 2-year survival rates were significantly improved in the NOA-01 trial in most RTOG-RPA classes (Table 9
The substitution of ACNU for BCNU in the NOA-01 trial compared with the GAG trial was introduced because of the severe pulmonary toxicity encountered in the GAG trial, with an incidence of symptomatic restrictive lung disease of 12% per annum and 4% related deaths (GAG Study Group, manuscript submitted for publication). The comparison of the survival data in the NOA-01 and the GAG trials (Table 8 ; Fig 1E) illustrates that the change in nitrosourea from BCNU to ACNU did not compromise the overall outcome for the study population. Moreover, the pulmonary toxicity observed in the NOA-01 trial (Table 5) played almost no role compared with the incidence of pulmonary toxicity in the GAG trial mentioned above, although the dose of ACNU was escalated in 30% of the patients (Table 3). These data indicate that ACNU may be equally effective but better tolerated than BCNU. Of note, given the lack of a difference between the VM26 and Ara-C arms, it is impossible to derive any conclusion regarding the contribution of VM26 or Ara-C, or their synergy with ACNU, to the encouraging results of the trial. This raises the important question of differences in efficacy and toxicity of ACNU plus VM26 versus ACNU alone. The median survival for glioblastoma patients as observed in the NOA-01 trial is similar to the 16 months of the phase II trial of concomitant and adjuvant temozolomide in addition to radiotherapy for newly diagnosed glioblastoma.15 This is remarkable in that phase II trials commonly achieve better results than phase III trials that evaluate the same regimen. Results from the European Organization for Research and Treatment of Cancer (EORTC) 26981/22981 phase III trial on the basis of this phase II temozolomide protocol will be available soon. If temozolomide plus radiotherapy is determined to be superior to radiotherapy alone in the EORTC 26981 trial, one may well consider a comparison of the chemotherapy arm of EORTC trial 26981/22981 with the ACNU plus VM26 arm of the NOA-01 trial in a future phase III trial. Such a trial would not only have to compare efficacy and side effects, but also (among other factors) quality of life and cost issues for radiochemotherapy regimens that may provide a modest advantage over radiotherapy alone.
Administration responsibility: Neuro-Oncology Working Group (NOA) of the German Cancer Society. Writing committee: Michael Weller, Tübingen; Bettina Müller, Kreischa; Rainer Koch, Dresden; Michael Bamberg, Tübingen; Peter Krauseneck, Bamberg. Heads of clinical trial: Michael Bamberg, Tübingen; J. Bock, Düsseldorf. Clinical coordinators: Peter Krauseneck, Bamberg; Bettina Müller, Kreischa. Central pathology review: Otmar Wiestler, Bonn. Statistical analysis: Rainer Koch, Dresden; Bettina Müller, Kreischa. The participating centers, in order of the number of patients entered (n), were as follows: University of Tübingen, Departments of Neurology and Radiation Oncology, n = 105; Bamberg Clinic, Department of Neurology, n = 40; Bad Berka Clinic, n = 34; Ulm, Günzburg, and Augsburg Clinics, n = 30; University of Berlin Charité, Department of Radiation Oncology, n = 29; Ingolstadt Clinic, n = 25; Cologne-Merheim Clinic, n = 23; Trier Clinic, n = 17; University of Lübeck, n = 16; Erfurt Clinic, n = 13; Sande Clinic, n = 11; University of Giessen, n = 8; Bielefeld Clinic and University of Magdeburg, n = 6 each; Braunschweig Clinic and University of Essen, n = 4 each; Technical University Munich, n = 3; and University of Bochum, n = 1.
1. Fine HA, Dear KB, Loeffler JS, et al: Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 71:2585–2597, 1993[CrossRef][Medline] 2. DeAngelis LM, Burger PC, Green SB, et al: Malignant glioma: Who benefits from adjuvant chemotherapy? Ann Neurol 44:691–695, 1998[CrossRef][Medline] 3. Glioma Meta-Analysis Trialists Group: Chemotherapy in adult high-grade glioma: A systematic review and meta-analysis of individual patient data from 12 randomised trials. Lancet 359:1011–1018, 2002[CrossRef][Medline]
4. Wong ET, Hess KR, Gleason MJ, et al: Outcomes and prognostic factors in recurrent glioma patients enrolled onto phase II clinical trials. J Clin Oncol 17:2572–2578, 1999
5. Yung WKA, Prados MD, Yaga-Tur R, et al: Multicenter phase II trial of temozolomide in patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. J Clin Oncol 17:2762–2771, 1999 6. Walker MD, Green SB, Byar DP, et al: Randomized comparisons of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. N Engl J Med 303:1323–1329, 1980[Abstract] 7. Green SB, Byar DP, Walker MD, et al: Comparisons of carmustine, procarbazine, and high-dose methylprednisolone as additions to surgery and radiotherapy for the treatment of malignant glioma. Cancer Treat Rep 67:121–132, 1983[Medline] 8. George SL, Desu MM: Planning the size and duration of a clinical trial studying the time to some critical event. J Chron Dis 27:15–24, 1974[CrossRef][Medline] 9. Kleihues P, Cavenee WK: World Health Organization Classification of Tumours: Pathology & Genetics—Tumours of the Nervous System. Lyon, France, IARC Press, 2000 10. Marubini E, Valsecci MG: Analysing Survival Data from Clinical Trials and Observational Studies. Chichester, United Kingdom, Wiley, 1995, p. 58
11. Curran WJ, Scott CB, Horton J, et al: Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials. J Natl Cancer Inst 85:704–710, 1993 12. Scott CB, Scarantino C, Urtasun R, et al: Validation and predictive power of Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis classes for malignant glioma patients: A report using RTOG 90-06. Int J Radiat Oncol Biol Phys 40:51–55, 1998[CrossRef][Medline]
13. The Medical Research Council Brain Tumor Working Party: Randomized trial of procarbazine, lomustine, and vincristine in the adjuvant treatment of high-grade astrocytoma: A Medical Research Council trial. J Clin Oncol 19:509–518, 2001 14. Rainov NG, on behalf of the GL1328 International Study Group: A phase III clinical evaluation of herpes simplex virus type 1 thymidine kinase and ganciclovir gene therapy as an adjuvant to surgical resection and radiation in adults with previously untreated glioblastoma multiforme. Hum Gene Ther 11:2389–2401, 2000[CrossRef][Medline]
15. Stupp R, Dietrich PY, Ostermann Kraljevic S, et al: Promising survival for patients with newly diagnosed glioblastoma multiforme treated with concomitant radiation plus temozolomide followed by adjuvant temozolomide. J Clin Oncol 20:1375–1382, 2002 Submitted March 3, 2003; accepted June 20, 2003.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
