✓ To determine the value of radiographically assessed response to radiation therapy as a predictor of survival in patients with glioblastoma multiforme (GBM), the authors studied a cohort of 301 patients who were initially treated according to uniform clinical protocols. All patients had newly diagnosed supratentorial GBM and underwent the maximum safe resection followed by external-beam radiation treatment (60 Gy in standard daily fractions or 70.4 Gy in twice-daily fractions of 160 cGy). The radiation response and survival rates were assessable in 222 patients. The extent of resection and the immediate response to radiation therapy were highly correlated with survival, both in a univariate analysis and after correction for age and Karnofsky performance scale (KPS) score in a multivariate Cox model (p < 0.001 for radiation response and p = 0.04 for extent of resection). A subgroup analysis suggested that neuroimaging obtained within 3 days after surgery served as a better baseline for assessment of radiation response than images obtained later. Imaging obtained within 3 days after completion of a course of radiation therapy also provided valid radiation response scores. The impact of the radiographically assessed radiation response on survival time was comparable to that of age or KPS score. This information is easily obtained early in the course of the disease, may be of value for individual patients, and may also have implications for the design and analysis of trials of adjuvant therapy for GBM, including volume-dependent therapies such as radiosurgery or brachytherapy.
Fred G. Barker II, Michael D. Prados, Susan M. Chang, Philip H. Gutin, Kathleen R. Lamborn, David A. Larson, Mary K. Malec, Michael W. McDermott, Penny K. Sneed, William M. Wara, and Charles B. Wilson
A retrospective analysis of 140 patients treated from 1967 to 1990
Brian J. Goldsmith, William M. Wara, Charles B. Wilson, and David A. Larson
✓ The authors retrospectively analyzed 140 patients treated at the University of California, San Francisco, from 1967 to 1990 to evaluate the results of radiation therapy (median 5400 cGy) given as an adjuvant to subtotal resection of intracranial meningioma. Of the 140 meningiomas, 117 were benign and 23 were malignant. The median follow-up period was 40 months. The overall survival rate at 5 years was 85% for the benign and 58% for the malignant tumor groups (p = 0.02); the 5-year progression-free survival rates were 89% and 48%, respectively (p = 0.001). For patients with benign meningioma, the 10-year overall and progression-free survival rates were 77%. An improved progression-free survival rate in that group was not related to tumor size but was associated with a younger age (p = 0.01) and treatment after 1980 with innovative technologies (p = 0.002); none of those variables affected the progression-free survival rate in the patients with malignant meningioma. Increased progression-free survival in the benign tumor group was also significantly associated with increasing the minimum radiation dose (p = 0.04). The 5-year progression-free survival rate for patients with benign meningioma treated after 1980 (when computerized tomography or magnetic resonance imaging was used for planning therapy) was 98%, as compared with 77% for patients treated before 1980 (p = 0.002). There were no second central nervous system tumors. Morbidity (3.6%) included sudden blindness or cerebral necrosis and death. When total resection of benign meningioma is not feasible, subtotal resection combined with precise treatment planning techniques and adjuvant radiation therapy can achieve results comparable to those of total resection.
Michael S. B. Edwards, William M. Wara, Raul C. Urtasun, Michael Prados, Victor A. Levin, Dorcas Fulton, Charles B. Wilson, John Hannigan, and Pamela Silver
✓ Fifty-three patients (19 adults and 34 children) harboring brain-stem glioma were treated with hyperfractionated radiation therapy (100 cGy twice a day, 5 days/wk, to a total dose of 7200 cGy). For the entire group, the median time to tumor progression (TTP) was 59 weeks (adults 66 weeks, children 44 weeks) and the median survival time was 74 weeks (adults 92 weeks, children 64 weeks). Statistically significant prognostic factors associated with a decrease in TTP and median survival times (adults < children) were: patient's age, a clinical history of less than 2 months, widespread brain-stem dysfunction, and a diffuse tumor as seen on magnetic resonance imaging. A finding of glioblastoma multiforme at histological analysis was associated with a statistical trend toward poorer survival, but in general tumor histology was not predictive of outcome. No evidence of an increase in acute or delayed radiation toxicity was seen with this fractionation schedule and total dose. This study suggests that hyperfractionation prolongs the TTP and survival time for many patients with brain-stem glioma. However, there remains a group of patients who are only moderately helped by this technique and for whom more aggressive treatment is warranted.
Michael S. B. Edwards, Roger J. Hudgins, Charles B. Wilson, Victor A. Levin, and William M. Wara
✓ The authors believe that the preferred treatment for pineal region tumors in children requires definitive surgery with a histological diagnosis and that a conservative approach consisting of shunting and radiation therapy no longer seems to be appropriate. The results are reported of a retrospective review of the presentation, treatment, and outcome of 36 children under the age of 18 years treated between 1974 and 1986. Eleven children had germinomas (two-cell type), seven had astrocytomas, and the remaining 18 had 15 histologically different tumor types. Surgery was performed on 30 patients; there were no deaths, but a 10% rate of persistent morbidity was found. The median follow-up period was 4 years. Nine (82%) of 11 patients with germinomas are alive without evidence of recurrence; one child died from recurrent tumor in the pineal region and another is presently being treated for recurrent tumor of the spinal cord. Six (86%) of the seven patients with astrocytoma are well after biopsy and radiation therapy. Of the remaining 18 children, five (28%) died from tumor progression.
The cerebrospinal fluid (CSF) tumor markers α-fetoprotein and β-human chorionic gonadotropin were helpful in determining the presence of malignant germ-cell tumors, particularly those with a poor prognosis. Magnetic resonance imaging was useful for diagnosis and for planning the operative approach. Magnetic resonance images showed the presence of pineal region tumors in four children with hydrocephalus who had no evidence of tumor on computerized tomography scans.
Because the great variety of tumor types found in the pineal region must be treated in different ways and because improved microsurgical and stereotaxic surgical techniques have made mortality and morbidity rates acceptably low, a biopsy diagnosis should be obtained in all patients. Preoperative assessment of CSF tumor markers and cytology is useful for the identification of patients who have a poor prognosis.
Victor A. Levin, William M. Wara, Richard L. Davis, Pamela Vestnys, Kenneth J. Resser, Kathleen Yatsko, Stephen Nutik, Philip H. Gutin, and Charles B. Wilson
✓ The authors report the results of a randomized study conducted to evaluate the relative benefit of treatment with 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) or the combination of procarbazine, 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea, and vincristine (PCV) administered after radiation therapy with hydroxyurea to 76 evaluable patients with glioblastoma multiforme and 72 patients with other anaplastic gliomas. The primary end-point of the study was time to tumor progression. For better-risk patients with Karnofsky performance scores of 70 to 100, results suggest that PCV was of greater benefit than BCNU (p = 0.15 for glioblastoma multiforme; p = 0.13 for other anaplastic gliomas). Median times to tumor progression were 31 and 32 weeks for patients with glioblastoma multiforme; 25th percentile times to progression were 70 and 40 weeks for patients treated with PCV and BCNU, respectively. For patients with other anaplastic gliomas treated with PCV and BCNU, median times to progression were 123 and 77 weeks, respectively. Multivariate analysis showed that the prognostic variables of age and Karnofsky scores were important for patients with glioblastoma multiforme and other anaplastic gliomas, and that the extent of surgical resection was important for those with other anaplastic gliomas.
Results of treatment with radiation therapy and chemotherapy
Surasak Phuphanich, Michael S. B. Edwards, Victor A. Levin, Pamela S. Vestnys, William M. Wara, Richard L. Davis, and Charles B. Wilson
✓ Twenty-seven patients aged 1 to 18 years harboring supratentorial (20 in the cerebrum and seven in the thalamus) malignant gliomas were treated between 1975 and 1982. There were four glioblastomas multiforme, 14 anaplastic astrocytomas, and nine malignant gliomas. All patients had a subtotal resection or biopsy as the initial procedure and received postoperative radiation therapy (RT). Fifteen of 27 patients were treated by RT alone; 14 had tumor progression with a median time to tumor progression (MTP) of 65 weeks. Twelve patients were treated with chemotherapy as an adjuvant to RT; only seven had tumor recurrence, with an MTP of 130 weeks. Of the 21 patients with recurrent tumors in both groups, 18 were treated with chemotherapy alone, or chemotherapy with a second surgical procedure or second course of RT.
For all histological grades of tumor, the MTP for first recurrence was 75 weeks and the median survival time was 180 weeks. Age at initial diagnosis was found to be a statistically significant prognostic factor, with patients younger than 10 years of age surviving longer than patients aged over 10 years (p = 0.02).
Long-term follow-up review and therapeutic guidelines
Mitchel S. Berger, Michael S. B. Edwards, William M. Wara, Victor A. Levin, and Charles B. Wilson
✓ Primary cerebral neuroblastoma is a distinct pathological and clinical entity that differs from other primitive neuroectodermal tumors. To characterize the clinical course of this lesion, the authors performed a retrospective analysis in 11 patients who ranged in age from 17 months to 26 years. The tumor had no predilection for either sex. Signs and symptoms were mostly those associated with increased intracranial pressure. The lesions commonly involved the parietal and occipital lobes. Computerized tomography scans of nine patients showed five solid and four cystic lesions; calcifications were found more commonly in the solid lesions. Contrast enhancement was seen in all tumors, yet angiograms typically showed an avascular mass. Total removal of tumor was possible in only two patients, both with cystic tumors. The remaining nine underwent subtotal resection of a solid lesion (in five) or a cystic lesion (in four). All 11 patients underwent postoperative irradiation that included the spinal axis in two cases; only one received adjuvant chemotherapy (solid tumor). Four patients, all with solid tumors that initially were subtotally resected, had evidence of tumor recurrence. The only patient with a subtotally resected solid lesion who did not have recurrence received adjuvant chemotherapy. The six patients who had cystic lesions are free of recurrent tumor at 26 to 109 months after surgery. Based on follow-up analysis of the 11 patients, recommendations are proposed for the treatment of primary cerebral neuroblastomas.
Dorcas S. Fulton, Victor A. Levin, William M. Wara, Michael S. Edwards, and Charles B. Wilson
✓ Forty-five children harboring brain-stem tumors were treated at the University of California, San Francisco, between 1969 and 1979. Pathological diagnoses were made in 19 patients. All patients received radiation therapy (RT). Thirteen patients received chemotherapy before, during, or immediately after RT. Twenty-four patients were treated with chemotherapy at the time of tumor progression, after initial treatment with RT alone. No statistically significant difference in time to tumor progression or survival was found for treatment with chemotherapy as an adjuvant to RT compared to treatment with RT alone followed by chemotherapy administered at the time of tumor progression. There were, however, more long-term survivors in the group that was first treated with chemotherapy at the time of tumor progression. There was no statistically significant correlation between survival and tumor pathology or location, although there were more long-term survivors among patients harboring low-grade gliomas and among patients with tumors confined to the midbrain. The authors documented the response of some brain-stem tumors to chemotherapy; however, cooperative controlled studies will be required to determine the optimum treatment for this disease.
Victor A. Levin, Charles B. Wilson, Richard Davis, William M. Wara, Tana L. Pischer, and Lowell Irwin
✓ This Phase III clinical trial compared the effectiveness of the combination of 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), radiation therapy, and hydroxyurea (BHR group) to the combination of BCNU and radiation therapy (BR group) for the treatment of malignant gliomas. In both arms of the study, BCNU was administered intravenously for 3 consecutive days before the initiation of radiation therapy, and at 8-week intervals thereafter until unequivocal tumor progression. In the BHR arm of the study, hydroxyurea was administered orally on alternate days during radiation therapy. Patients in each arm were stratified almost equally by tumor type (glioblastoma multiforme (GM) or other nonglioblastoma multiforme malignant gliomas (NGM)) and extent of surgical resection of tumor. Patients were also evaluated with the Karnofsky Performance Status (KPS) scale. Time to progression was determined by comparing the results of sequential neurological examinations and radionuclide and computerized tomographic scans.
Of the 130 patients entered into the study, 99 constitute the valid study group. Data were analyzed with Kaplan-Meier representations and the statistical methods of Gehan and Cox. The NGM patients with KPS ratings of ≥60 did better on both arms of the study, with median times to tumor progression (MTP's) of 50 and 72 weeks for BHR and BR, respectively. However, GM patients showed statistically significant differences (p = 0.03) between the two arms of the study, with MTP's of 41 and 31 weeks for BHR and BR, respectively. The GM patients with subtotal tumor resection did slightly better on BHR than on BR, with MTP's of 49 weeks (p = 0.03) and 31 weeks for the respective groups.