Fax: (011) 39 0516225057
Long-term results of a prospective study on the treatment of medulloblastoma in adults
Article first published online: 6 SEP 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 9, pages 2035–2041, 1 November 2007
How to Cite
Brandes, A. A., Franceschi, E., Tosoni, A., Blatt, V. and Ermani, M. (2007), Long-term results of a prospective study on the treatment of medulloblastoma in adults. Cancer, 110: 2035–2041. doi: 10.1002/cncr.23003
- Issue published online: 18 OCT 2007
- Article first published online: 6 SEP 2007
- Manuscript Accepted: 3 JUL 2007
- Manuscript Revised: 1 JUL 2007
- Manuscript Received: 5 MAR 2007
Because medulloblastoma (MB) is rare in adults, the few studies on this condition have been retrospective, and the follow-up has tended to be short. Furthermore, the different therapeutic strategies used in these patients has made it difficult to assess survival rates and prognostic factors.
In 1989, a prospective Phase II trial was initiated to evaluate the efficacy of treatment for adults with MB. Patients were staged completely with a neuroradiologic examination of the brain and neuroaxis and by cerebrospinal fluid cytology, according to Chang's staging system. Low-risk patients received radiotherapy alone, whereas high-risk patients received 2 cycles of upfront chemotherapy followed by radiotherapy and adjuvant chemotherapy. The current article reports on the long-term results from that trial.
After a median follow up of 7.6 years, among a total of 36 adults with MB, the overall progression-free survival (PFS) and overall survival (OS) rates at 5 years were 72% and 75%, respectively. In low-risk patients, the 5-year PFS rate was 80%, and the 5-year OS rate was 80%; in high-risk patients, the 5-year PFS rate was 69%, and the 5-year OS rate was 73%.
In adult patients with MB, long-term follow-up was essential for evaluating the real impact of treatments. Low-risk and high-risk patients did not differ significantly in terms of PFS or OS. Cancer 2007. © 2007 American Cancer Society.
Medulloblastoma (MB) is rare in adults and is diagnosed in approximately 0.5 of 100,000 patients per year. Because of its rarity, most available studies on MB in adults have been retrospective and have reported on limited patient populations. Moreover, the radiotherapy and chemotherapy strategies and schedules used vary, and the median follow-up ranges from 51 months to 104 months.1, 2 Clearly, such methodological and clinical differences have an important influence on the reliability of any attempt to identify prognostic factors and to predict the outcomes of the different treatment strategies in these studies.
MATERIALS AND METHODS
In 1989 a prospective trial was started, and 36 adult patients with MB were accrued for the study.3 All patients provided their fully informed consent to take part in the study by signing a form, which was approved by the Institutional Review Board of Padova University Hospital (Italy). The study was conducted according to the principles of the Declaration of Helsinki and the rules of good clinical practice. The patients were staged using the classic Chang system definitions for tumor (T) and metastasis (M) parameters4 (Table 1). Residual disease after surgery (defined as disease that measured ≥1.5 cm2) also was taken into account in the classification into low-risk and high-risk groups. The low-risk group comprised patients who had T1, T2, T3a, and M0 disease and no residual disease after surgery. The high-risk group comprised patients who had T3b through T4 and M1, M2, M3 disease and patients who had postoperative residual tumor. Low-risk patients received radiation therapy alone, which started within 28 days of surgery. High-risk patients were given 2 cycles of upfront chemotherapy before radiotherapy. After receiving this treatment, patients with M1, M2, or M3 disease were given maintenance chemotherapy, which consisted of 4 cycles; otherwise, 2 cycles more than the number required for complete remission were planned. Radiotherapy consisted of craniospinal radiation at a dose of 36 grays (Gy) in 20 fractions of 1.8 Gy/5 fractions per week, followed by a boost of 18.8 Gy in 10 fractions to the posterior fossa (up to a total of 54.8 Gy). In patients with M3 spinal disease, the dose was 39.6 Gy in 22 fractions delivered only to the spinal cord. Between 1989 and 1995, in accordance with the findings of a multicenter trial of the International Society of Pediatric Oncology,5 patients received a mechlorethamine, vincristine, procarbazine, and prednisone (MOPP)-like regimen that consisted of intravenous nitrogen mustard (3 mg/m2 on Days 1 and 8), intravenous vincristine (1.4 mg/m2 [maximum 2 mg] on Days 1 and 8), oral prednisone (40 mg/m2 on Days 1 to 10), and procarbazine (50 mg on Day 1, 100 mg on Day 2, and 100 mg/m2 on Days 3-10). The protocol was amended in 1995 in view of reports6 indicating that cisplatin-based regimens were more effective in adults with advanced embryonic central nervous system tumors, and MOPP chemotherapy was replaced with the following regimen: cisplatin 25 mg/m2 daily for 4 days, etoposide 40 mg/m2 daily for 4 days, and cyclophosphamide 1000 mg/m2 on Day 4: This cycle was repeated every 4 weeks. Twenty-two of 26 high-risk patients agreed to be treated with upfront chemotherapy, which consisted of an MOPP regimen in 6 patients and a cisplatin-based combination in 16 patients. Drug doses were modified in patients with hematologic toxicity, which was diagnosed on the basis of neutrophil and platelet counts at nadir and on the day of treatment, as described elsewhere.3
|Variable||Greatest tumor dimension and disease spread|
|T3a||>3 cm with spread into the aqueduct of Sylvius and/or foramen of Luschka, cerebral subarachnoid space, third or lateral ventricles|
|T3b||>3 cm with unequivocal spread into the brainstem; for T3b, surgical staging may be used in the absence of involvement at imaging|
|T4||>3 cm with spread beyond the aqueduct of Sylvius and/or the foramen magnum|
|M0||No evidence of gross subarachnoid or hematogenous metastasis|
|M1||Microscopic tumor cells in cerebrospinal fluid|
|M2||Gross nodular seeding in cerebellum|
|M3||Gross nodular seeding in spinal subarachnoid space|
|M4||Metastasis beyond cerebrospinal axis|
In low-risk patients, contrast-enhanced brain magnetic resonance imaging (MRI) studies were obtained every 3 months. Spinal MRI studies were obtained every 6 months for the first 2 years, they were repeated every 6 months thereafter for up to 5 years, and they were obtained yearly thereafter. Patients with metastatic disease at the time of diagnosis underwent cerebrospinal fluid (CSF) examination and spinal MRI studies 6 weeks after the end of radiotherapy and after every 2 cycles of postradiation chemotherapy until the complete disappearance of disseminated disease. Follow-up controls were made after metastatic disease disappearance every 6 months with brain MRI studies, and spinal MRI studies also were obtained every 12 months.
In patients without metastasis at diagnosis, CSF was examined only if there was a clinical suspicion of leptomeningeal tumor recurrence. After the identification of disease recurrence, patients underwent MRI studies of the entire neuroaxis and CSF cytology.
Progression-free survival (PFS) and overall survival (OS) were measured from the time of surgery to the date of disease progression or death, respectively, or to the date of last follow-up and were analyzed using the Kaplan-Meier method. The 95% confidence intervals were calculated using associated estimated standard errors. The log-rank test was used to compare low-risk and high-risk groups and to test the significance of the following prognostic variables: sex, the presence of a shunt, residual disease, T stage, M stage, histologic subtype, location of lesions, postoperative Karnofsky performance score, duration of radiotherapy, and the time between surgery and the start of radiotherapy.
Between January 1989 and February 2001, 36 patients (27 men and 9 women) who had a histologic diagnosis of MB were treated according to the above-described protocol. Patients' characteristics and study treatments are reported in Table 2. Thirty-six patients were treated according to the protocol: Twenty-six patients were classified as high risk, and 10 patients were classified as low risk.
|Characteristic||No. of patients|
In the current study, after a median follow-up of 7.6 years, the overall median PFS in low-risk patients was 9.3 years (range, from 5.9 years to not achieved), whereas no median PFS was achieved in high-risk patients. The PFS rate at 5 years was 80% (95% CI, 59-100%) and 69% (95% CI, 54-89%) in low-risk patients and high-risk patients, respectively (P = not significant [NS]) (Fig. 1).
In subgroup analyses, neither metastatic status at the time of diagnosis (5-year PFS rate: patients with metastasis-negative [M0] disease, 78%; 95% CI, 63-97%; patients with metastasis-positive [M1, M2, or M3] disease, 61%; 95% CI, 40%-95%; P = NS), nor the presence of residual disease after surgery (5-year PFS rate: patients without residual disease, 82%; 95% CI, 62%-100%; patients with residual disease, 68%; 95% CI, 52-89%) had a significant effect on the 5-year PFS rate. However, the 5-year PFS rate was 82% (95% CI, 68-98%) in patients with T1 through T3a disease and 44% (95% CI, 21-92%) in patients with T3b through T4 disease (borderline significance; P = .06).
The 5-year PFS rate was not significant for high-risk patients who received upfront chemotherapy with MOPP or diethylcarbamazine (P = .9). At the time of this protocol, patients with T3b M0 disease were considered high risk; thus, data were reanalyzed on the basis of the current risk classification (high risk [M1, M2, M3] and residual disease >1.5 cm2). However, PFS at 5 years remained nonsignificant (P = .82).
At a median follow-up of 7.6 years, the median OS (10.7 years in low-risk patients) was not achieved in high-risk patients. The 5-year OS rate was 80% (95% CI, 58-100%) and 73% (95% CI, 58-92%) in low-risk and high-risk patients, respectively (P = NS) (Fig. 2). Survival at 5 years was not influenced by metastatic status at the time of diagnosis (5-year OS rate: M-negative patients, 83%; 95% CI, 69-100%; M-positive patients, 62%; 95% CI, 40-95%; P = NS) or by the presence of residual disease after surgery (5-year OS rate: patients without residual disease, 82%; 95% CI62-100%; patients with residual disease, 72%; 95% CI, 56-92%). However, the 5-year OS rate was 82% (95% CI, 68-98%) in patients with T1 through T3a disease and 56% (95% CI, 31-100%) in patients with T3b and T4 disease (P = .04).
The median duration of radiotherapy was 46 days (range, 35-70 days) overall, and it was 46.5 days (range, 39-79 days) and 45 days (range, 35-59 days) for patients who did and did not receive preradiation chemotherapy, respectively. Toxicity during radiotherapy was thrombocytopenia (grade 2, 20%; grade 3, 20%) and leukopenia (grade 2, 20%) in low-risk patients who did not receive preradiation chemotherapy and thrombocytopenia (grades 2, 11%; grade 3, 15%; grade 4, 3%) and leukopenia (grade 2, 20%; grade 3, 7%; grade 4, 3%) in high-risk patients. In 2 high-risk patients (7%), treatment was interrupted because of radiation-related toxicity (thrombocytopenia) for 22 days and 7 days, respectively.
At a median follow-up of 7.6 years, in total, 17 patients developed recurrent disease (47%); the median time to disease progression was 36 months (range, 4-129 months), and 15 patients (41.7%) had died. Eleven patients (42.3%) in the high-risk group and 6 patients (60%) in the low-risk group developed recurrent disease. Figure 3 illustrates that the risk of recurrence appeared to increase markedly after 7 years of follow-up in low-risk patients and after 10 years of follow-up in high-risk patients.
The therapeutic approaches used for recurrence were stereotactic radiotherapy in 1 patient, chemotherapy in 14 patients, and no treatment in 2 patients. OS after disease recurrence was 15 months (range, 1-44 months).
The widespread tendency almost always to treat adult patients with MB according to pediatric protocols has been based on the assumption that, in adults, this tumor has the same properties that it has in children. However, few prospective clinical trials are available in adults, and those studies are retrospective. In addition, the different treatments given spanned decades, during which diagnostic procedures, neurosurgical skills, and radiotherapy techniques have changed.
Surgery and postoperative radiotherapy long have been considered the cornerstones for the treatment of MB in adult and pediatric patients. Conventional radiotherapy doses of approximately 36 Gy in adults are delivered to the craniospinal axis together with a boost of 18 to 20 Gy to the posterior fossa.7–9 The authors of retrospective studies have reported that from 40% to 70% of adult patients are progression-free 5 years after diagnosis. Because risk assessment and late toxicity have been taken into account in randomized clinical trials in pediatric patients, patients with average-risk MB currently receive dose-reduced radiotherapy (craniospinal irradiation, 23.4 Gy with a boost up to 55.8 Gy to the posterior fossa) followed by adjuvant chemotherapy, and no difference has been reported between survival after this regimen and survival after 36 Gy to the craniospinal axis without chemotherapy.10 However, because of the differences in terms of long-term toxicities between adult and pediatric patients, this approach has not been proposed for adult patients, and the role of adjuvant chemotherapy in this setting is not clear. Some studies failed to demonstrate that it significantly improved PFS, others have suggested that it does (Table 3). In 5 studies,1, 11–14 adjuvant chemotherapy (using different regimens) was given to adults with MB, and the findings reported for outcome were inconclusive, mainly because the studies were not prospective and because chemotherapy was delivered to both high-risk and low-risk patients, and only a trend toward improvement in survival was observed in the overall series without a distinction being made between low-risk and high-risk adult patients.7
|Carrie et al., 199411||156||75 Patients, 8 in 1 CT: SIOP protocol; IFO/CDDP/VCR||5-Y PFS rate: 58% in M0 group vs 51% in M+ group; 59% in R0 group vs 64% in R+ group|
|Prados et al., 199512||47||32 Patients||5-Y PFS 5 rate: 54% in LR group vs 38% in HR group|
|Chan et al., 200013||32||24 Patients: 3-5 cycles of CT||5-Y PFS rate: 59% in M0 group vs 47% in M+ group; 86% in R0 group vs 27% in R+ group|
|Louis et al., 2002 (unpublished results)||51||24 SR patients, no CT; 27 HR patients, CT with CARBO/VP-16||5-Y PFS rate: 60% in SR group vs 63% in HR group|
|Abacioglu et al., 20021||30||10 Patients||5-Y PFS rate: 69% in CT-treated patients; 60% in untreated patients|
|Greenberg et al., 200114||17||17 Patients: Packer regimen or POG protocol||PFS, 48 mo|
The current article reports the long-term results from a prospective pilot trial in which only radiotherapy was delivered to low-risk patients, whereas radiotherapy plus chemotherapy was reserved for high-risk adult MB patients. A debate is ongoing about whether it is more useful to administer chemotherapy before or after radiotherapy. The rationale for up-front chemotherapy is: 1) the destruction of the CSF-tumor barrier caused by the operation may allow drugs to penetrate more easily into the remaining tumor tissue; 2) aggressive chemotherapy is tolerated better, because bone marrow reserves are not compromised by craniospinal radiotherapy; 3) toxicity profiles potentially are favorable, allowing the delivery of cisplatin with a decreased likelihood of severe ototoxicity; and 4) this setting is ideal for testing the efficacy of chemotherapy in patients with measurable postoperative disease. However, those against this approach postulate that chemotherapy may delay the initiation of radiotherapy, thus increasing the risk of tumor progression15–17 and potentially hindering the completion of craniospinal radiotherapy. Two randomized studies have been conducted to compare the efficacy of preoperative and postoperative chemotherapy in high-risk pediatric patients. In the study by Zeltzer et al.,16 a reduction in the 5 year PFS was reported when 8 in 1 chemotherapy preceded radiotherapy. In the other study, which was conducted by Kortmann et al.,18upfront chemotherapy was compared with maintenance chemotherapy, and the results indicated that the latter treatment was more beneficial. More recently, a large randomized trial on the treatment of nonmetastatic MB19 compared chemotherapy plus radiation with radiotherapy alone and reported improved survival in patients who received preirradiation chemotherapy. In our high-risk patients, the receipt of radiotherapy chemotherapy before radiotherapy neither affected the final outcome nor resulted in tumor progression during its delivery, as postulated in children, nor did it affect the dosage and timing of subsequent radiotherapy. These favorable results may have depended on the fact that we employed a brief, intense, and effective chemotherapy regimen. The same regimen, when administered after radiotherapy, cannot be delivered at full dosage, and its toxicity is more sustained.
The results of our study, after a relatively long follow-up, indicate that the initial differences in both PFS and OS at 5 years were lost at 10 years, with late recurrences occurring in 6 of 10 low-risk patients, including 3 patients who developed recurrences after 100 months. However, statistical comparisons on only 10 low-risk patients are limited. The prolonged follow-up allowed us to evaluate long-term outcomes in high-risk patients who received chemotherapy: Fifteen of 18 patients (83%) without disease progression at the first analysis3 were alive and progression-free more than 3 years later.
Overall, the data discussed above suggest that recurrences tend to occur early in low-risk patients and late in high-risk patients. In our study, the median OS after disease recurrence was 15 months. Although 3 of 6 low-risk patients with recurrent disease were alive at the time of data analysis, and all 11 patients who had a high-risk of recurrence had died, no significant difference was observed between low-risk patients and high-risk patients in the median OS after recurrence (10 months vs 19 months, respectively; P = .19).
To our knowledge, few authors have evaluated survival analyses stratified on the basis of risk evaluation in adults with MB who received radiotherapy and chemotherapy. In their study on 47 adults with MB using risk assessment based on the University of California–San Francisco risk criteria (high-risk, >25% residual disease after surgery, brainstem invasion, and/or tumor cells in CSF or metastatic disease), Prados et al.12 observed a significant difference between high-risk and low-risk patients for 5-year PFS and OS, and a significant association between the administration of adjuvant chemotherapy and OS. Louis et al. used the same risk assessment that they used in their series of 72 adult MB patients; 28 high-risk patients received a preirradiation carboplatin-etoposide regimen, whereas low-risk patients received a regimen of radiotherapy alone (unpublished results). After a median follow-up of 24 months, no significant difference was observed between the 2 groups for PFS or OS. Despite the differences between the characteristics of the populations studied and the duration of follow-up, those investigators, like us, observed no significant difference between low-risk patients and high-risk patients in terms of PFS, OS, or the OS rate.
In the current study, chemotherapy administered to high-risk patients reduced the risk of recurrence and death and, even more intriguingly, appeared to delay the risk of recurrence by approximately 3 years. The risk of late recurrence in low-risk patients is of prime importance, because it was reported in a series of patients who were treated in the modern era in which the chemotherapy and radiotherapy treatment protocol remains largely unchanged. The lack of any significant difference between the PFS and OS for low-risk and high-risk patients may suggest that, like pediatric patients, adult patients with low-risk MB also may be given chemotherapy. However, although chemotherapy can be given to low-risk patients, it is too early to consider reducing the radiotherapy dose like what has been done for low-risk pediatric patients.
In conclusion, our data show that long-term follow-up is of crucial importance in patients with MB, because it allows a more accurate definition of the impact of treatment. In view of the rarity of this disease, referral to specialized centers and a plea to launch a multicenter randomized clinical trial could be considered for low-risk adult MB patients to verify whether adjuvant chemotherapy can increase the survival of this category of patients.
- 17Prospective randomised trial of chemotherapy given before radiotherapy in childhood medulloblastoma. International Society of Paediatric Oncology (SIOP) and the (German) Society of Paediatric Oncology (GPO): SIOP II. Med Pediatr Oncol. 1995; 25: 166–178., , , et al.
- 18Postoperative neoadjuvant chemotherapy before radiotherapy as compared to immediate radiotherapy followed by maintenance chemotherapy in the treatment of medulloblastoma in childhood: results of the German prospective randomized trial HIT '91. Int J Radiat Oncol Biol Phys. 2000; 46: 269–279., , , et al.
- 19Results of a randomized study of preradiation chemotherapy versus radiotherapy alone for nonmetastatic medulloblastoma: the International Society of Paediatric Oncology/United Kingdom Children's Cancer Study Group PNET-3 Study. J Clin Oncol. 2003; 21: 1581–1591., , , et al.