Response criteria for phase II studies of supratentorial malignant glioma.

We suggest "new" response criteria for phase II studies of supratentorial malignant glioma and favor rigorous criteria similar to those in medical oncology, with important modifications. Four response categories are proposed: complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Response in this scheme is based on major changes in tumor size on the enhanced computed tomographic (CT) or magnetic resonance imaging (MRI) scan. Scan changes are interpreted in light of steroid use and neurologic findings. We advocate careful patient selection, emphasize pitfalls in the assessment of response, and suggest guidelines to minimize misinterpretations of response.

Effects of radiation and chemotherapy on cognitive function in patients with high-grade glioma.

PURPOSE: The effect of radiotherapy on the long-term cognitive performance of patients treated for intracranial neoplasm is a major concern to clinicians and patients, particularly as long-term survival or cure is possible for a small minority of patients. To assess the effects of cranial radiotherapy and chemotherapy on the cognitive performance of high-grade glioma patients, we analyzed cognitive performance data collected in a series of prospective clinical trials. METHODS: We studied 701 high-grade brain tumor patients entered onto two consecutive North Central Cancer Treatment Group (NCCTG) randomized treatment trials designed to compare radiotherapy and carmustine (BCNU) versus radiotherapy and 1-(2-chloroethyl)-3(2,6 dioxo-l-piperidyl)-1-nitrosource a (PCNU) (first trial) and radiotherapy and BCNU and interferon alfa (IFN) versus radiotherapy and BCNU (second trial). Folstein Mini-Mental Status Exam (MMSE) score and Eastern Cooperative Oncology Group (ECOG) performance score (PS) recorded at baseline and 6, 12, 18, and 24 months were analyzed to assess cognitive and physical function over time. Patients who did not demonstrate tumor progression within 60 days of the assessment time were considered nonprogressors at that evaluation. A loss of greater than 3 points on the MMSE was considered significant deterioration. RESULTS: The number of patients who experienced a greater than 3-point decrease in MMSE from baseline was 13 of 119 nonprogressors (10.9%; 95% confidence interval [CI], 6.3% to 18.9%) at 6 months, three of 54 nonprogressors (5.5%; 95% CI, 0.5% to 12.8%) at 12 months, three of 30 nonprogressors (10%; 95% CI, 2.1% to 26.5%) at 18 months, and four of 22 nonprogressors (18.2%; 95% CI, 5.2% to 40.3%) at 24 months. The CIs at all times overlapped, which indicates no statistically significant increase in the percentage of patients who experienced a significant decrease in their MMSE score. Patients who demonstrated a significant decrease in their MMSE score were significantly older than those who did not (P = .0017) at 6 months and remained so throughout follow-up; moreover, they had a significantly shorter time to progression and death. ECOG PS was strongly negatively correlated with MMSE score throughout the study, and MMSE score at all time intervals was correlated with baseline PS. CONCLUSION: In this population of glioma patients who received radiotherapy, there is no clear trend to cognitive worsening. Factors such as older age, poorer PS, and subclinical tumor progression may be more significant factors in those patients who did demonstrate a significant cognitive decline.

Phase III placebo-controlled trial of capsaicin cream in the management of surgical neuropathic pain in cancer patients.

PURPOSE: A minority of cancer survivors develops long-term postsurgical neuropathic pain. Based on evidence that capsaicin, the pungent ingredient in hot chili peppers, might be useful for treating neuropathic pain, we developed the present clinical trial. PATIENTS AND METHODS: Ninety-nine assessable patients with postsurgical neuropathic pain were entered onto this study. After stratification, patients were to receive 8 weeks of a 0.075% capsaicin cream followed by 8 weeks of an identical-appearing placebo cream, or vice versa. A capsaicin/placebo cream was to be applied to the painful site four times daily. Treatment evaluation was performed by patient-completed weekly questionnaires. RESULTS: During the first 8-week study period, the capsaicin-cream arm was associated with substantially more skin burning, skin redness, and coughing (P < .0001 for each). Nonetheless, treatment was stopped for patient refusal or toxicity just as often while patients were receiving the placebo as compared with the capsaicin. The capsaicin cream arm had substantially more pain relief (P = .01) after the first 8 weeks, with an average pain reduction of 53% versus 17%. On completion of the 16-week study period, patients were asked which treatment period was most beneficial. Of the responding patients, 60% chose the capsaicin arm, 18% chose the placebo arm, and 22% chose neither (P = .001). CONCLUSION: A topical capsaicin cream decreases postsurgical neuropathic pain and, despite some toxicities, is preferred by patients over a placebo by a three-to-one margin among those expressing a preference.

Comprehensive study of diagnosis and treatment of trigeminal neuralgia secondary to tumors.

Among 5,058 patients seen at the Mayo Clinic from 1976 through 1990 for face pain, we diagnosed trigeminal neuralgia in 2,972. Tumors were causing the face pain in 296 patients. Sex and pain distributions paralleled those in idiopathic trigeminal neuralgia; however, patients with tumors causing trigeminal neuralgia were younger than those with idiopathic pain. Meningiomas and posterior fossa tumors were the most common. Neurologic deficits developed on follow-up evaluation in 47% of the patients, often precipitating further study and eventual diagnosis of the tumor. Delay in tumor diagnosis averaged 6.3 years. CT with contrast was the most frequently used initial diagnostic radiographic technique, detecting a tumor in 40 of 43 examinations. MRI was subsequently used to confirm and better delineate the tumor in five of five cases. Carbamazepine was the most effective drug for relieving trigeminal neuralgia, but relief was usually temporary. Of the surgical treatment options, total removal of the tumor was the most effective in completely relieving tic pain. In patients at high surgical risk, however, temporarily or permanently blocking afferent impulses with radiofrequency ablation, glycerol rhizotomy, or alcohol blocks was a good alternative to craniotomy.

CT of the brachial plexus in patients with cancer.

To assess the usefulness of CT, we reviewed 51 patients with clinically diagnosed brachial plexopathy who were seen between 1977 and 1981. The established etiology was metastatic tumor in 46 and radiation fibrosis in 5. CT was abnormal in 89% of tumor patients. Myelography, bone scan, and plain cervical spine radiographs were less useful. In four of five patients with radiation fibrosis, CT showed distortion of normal tissue planes without a discrete mass, but was not always distinguishable from tumor infiltration. CT of the brachial plexus provides the best two-dimensional view of tumor infiltration and detects bony changes earlier than standard radiographs. CT is a useful guide for surgical exploration of the brachial plexus, but does not differentiate tumor infiltration from radiation fibrosis.

Differential diagnosis between radiation and tumor plexopathy of the pelvis.

We studied 20 patients with lumbosacral radiculoplexopathy from radiation treatment and 30 patients with plexus damage from pelvic malignancy. Indolent leg weakness occurred early in radiation disease, whereas pain marked the onset of tumor plexopathy. Eventually, all radiation cases had weakness, which was bilateral in most of them and painless in one-half of them. Tumor patients typically had unilateral weakness, which was painful in all of them. Radiation disease often resulted in serious neurologic disability. Of the tumor patients, 86% were dead within 3 1/2 years after onset of neurologic symptoms.

Hemimasticatory spasm: clinical and electrophysiologic observations.

Hemimasticatory spasm is a rare disorder of the trigeminal nerve that produces involuntary jaw closure due to paroxysmal unilateral contraction of jaw-closing muscles. We report three patients with this disorder. Electrophysiologic studies demonstrated normal blink and masseter reflexes. The masseter inhibitory reflex was absent during periods of spasm. Needle electromyography demonstrated irregular bursts of motor unit potentials that were identical to the pattern observed in hemifacial spasm. The electrophysiologic findings suggest ectopic excitation of the trigeminal motor root or its nucleus, an abnormality that is analogous to ectopic excitation of the facial nerve in hemifacial spasm. One patient improved temporarily with surgery, one improved while on treatment with carbamazepine, and another responded favorably to botulinum toxin injection.

Distinction between neoplastic and radiation-induced brachial plexopathy, with emphasis on the role of EMG.

The results of clinical, radiologic, and electrophysiologic studies are retrospectively reviewed for 55 patients with neoplastic and 35 patients with radiation-induced brachial plexopathy. The presence or absence of pain as the presenting symptom, temporal profile of the illness, presence of a discrete mass on CT of the plexus, and presence of myokymic discharges on EMG contributed significantly to the prediction of the underlying cause of the brachial plexopathy. The distribution of weakness and the results of nerve conduction studies were of no help in distinguishing neoplastic from radiation-induced brachial plexopathy.

Phase I and pharmacokinetic study of preirradiation chemotherapy with BCNU, cisplatin, etoposide, and accelerated radiation therapy in patients with high-grade glioma.

PURPOSE: We conducted a Phase I study of bischloroethylnitrosourea (BCNU), cisplatin, and oral etoposide administered prior to and during accelerated hyperfractionated radiation therapy in newly diagnosed high-grade glioma. Pharmacokinetic studies of oral etoposide were also done. METHODS AND MATERIALS: Patients started chemotherapy after surgery but prior to definitive radiation therapy (160 cGy twice daily x 15 days; 4800 cGy total). Initial chemotherapy consisted of BCNU 40 mg/m2 days 1-3, cisplatin 30 mg/m2 days 1-3 and 29-31, and etoposide 50 mg orally days 1-14 and 29-42, repeated in 8 weeks concurrent with radiation therapy. BCNU 200 mg/m2 every 8 weeks x 4 cycles was given after radiation therapy. RESULTS: Sixteen patients, 5 with grade 3 anaplastic astrocytoma and 11 with glioblastoma were studied. Grade 3-4 leukopenia (38%) and thrombocytopenia (31%) were dose-limiting. Other toxicities were anorexia (81%), nausea (94%), emesis (56%), alopecia (88%), and ototoxicity (38%). The maximum tolerated dose was BCNU 40 mg/m2 days 1-3, cisplatin 20 mg/m2 days 1-3 and 29-31, and oral etoposide 50 mg days 1-21 and 29-49 prior to radiation therapy and repeated in 8 weeks with the start of radiation therapy followed by BCNU 200 mg/m2 every 8 weeks for 4 cycles. Median time to progression and survival were 13 and 14 months respectively. Responses occurred in 2 of 9 (22%) patients with evaluable disease. In pharmacokinetic studies, all patients achieved plasma concentrations of >0.1 microg/ml etoposide (the in vitro radiosensitizing threshold), following a 50 mg oral dose. The mean +/- SD 2 hr and 6 hr plasma concentrations were 0.92 +/- 0.43 microg/ml and 0.36 +/- 0.12 microg/ml, respectively. Estimated duration of exposure to >0.1 microg/ml etoposide was 10-17 hr. CONCLUSIONS: Preirradiation chemotherapy with BCNU, cisplatin, and oral etoposide with accelerated hyperfractionated radiation therapy in high-grade gliomas is feasible and merits further investigation. Sustained radiosensitizing concentrations can be achieved with low oral doses of etoposide.

Phase I evaluation of radiation combined with recombinant interferon alpha-2a and BCNU for patients with high-grade glioma.

PURPOSE: A Phase I study to determine the safety, toxicity, and maximum tolerated dose (MTD) of carmustine (BCNU) and interferon alpha-2a (IFN-a) when combined with radiation as initial therapy in high-grade glioma. METHODS AND MATERIALS: Patients with newly diagnosed Grade 3 or 4 astrocytoma, oligoastrocytoma, or gliosarcoma were enrolled after surgery. All received radiation therapy to the brain (64.8 Gy/36 fractions), combined with a single dose of BCNU (200 mg/m2) at the start of radiation. Chemotherapy after completing radiation consisted of BCNU 150 mg/m2 once every 7 weeks, and IFN-a 12 x 10(6) units/m2 subcutaneously Days 1-3 each week of a 7-week cycle. Subsequent dose modification was based on constitutional symptoms for IFN-a and on myelosuppression for BCNU. RESULTS: Fifteen patients were entered on the study. Four were excluded because they did not receive IFN-a (3 refused treatment and 1 patient left the study due to multiple medical problems). Eleven were evaluable for toxicity and efficacy. Nonhematological toxicity, mainly lethargy and flu-like symptoms, were dose-limiting for IFN-a. After the first 6 patients were treated per the initial protocol, the frequency of IFN-a administration was decreased to Days 1-3 on weeks 1, 3, and 5 of the 7-week cycle for 5 additional patients. Lethargy, fever, chills, myalgias, alopecia, and anorexia occurred in all patients. Other toxicities included nausea and vomiting (91%), central-nervous-system depression or mood changes (64%), headaches (55%), and elevation of liver enzymes (36%). Grade 3-4 leukopenia occurred in 4 (45%) of 11 patients, and Grade 3-4 thrombocytopenia in 3 (27%) of 11 patients. Due to myelosuppressive effects, BCNU dose was not escalated. Median survival of the cohort was 44 months. Objective responses occurred in 5 (56%) of 9 patients and median duration of response was 33 months. The MTD of this combination after radiation therapy is IFN-a 12 x 10(6) units/m2 Days 1-3, on Weeks 1, 3, and 5 of a 7-week cycle and BCNU 150 mg/m2 Day 1, every 7 weeks. CONCLUSIONS: Treatment with radiation, IFN-a, and BCNU is feasible and effective in patients with high-grade gliomas, although constitutional symptoms from IFN-a are substantial.

A phase 3 randomized study of radiotherapy plus procarbazine, CCNU, and vincristine (PCV) with or without BUdR for the treatment of anaplastic astrocytoma: a preliminary report of RTOG 9404.

PURPOSE: This study was an open label, randomized Phase 3 trial in newly diagnosed patients with anaplastic glioma comparing radiotherapy plus adjuvant procarbazine, CCNU, and vincristine (PCV) chemotherapy with or without bromodeoxyuridine (BUdR) given as a 96-hour infusion each week of radiotherapy. METHODS AND MATERIALS: Only patients 18 years or older with newly diagnosed anaplastic glioma were eligible; central pathology review was accomplished, but was not mandated prior to registration. The study had initially opened as a Northern California Oncology Group (NCOG) trial in 1991, becoming an Intergroup RTOG, SWOG, and NCCTG study in July 1994. Total accrual of 293 patients was planned as the sample size, using survival and time to tumor progression as the primary endpoints. The experiment arm (RT/BUdR plus PCV) was to be compared to the control arm (RT plus PCV) using an alpha = 0.05, one-tailed, with a power of 85% for detecting an increase in median survival from 160 to 240 weeks, assuming a 3-year follow-up after completion of enrollment. RESULTS: As of July 1996, 281 patients had been randomized; 53 (20%) were ineligible, primarily based upon central pathology review, and another 39 cases were canceled. In total, 30% of cases were excluded from analysis. The treatment arms were well balanced despite this rate of exclusion. The RTOG Data Monitoring Committee recommended suspension of enrollment in July 1996 based upon a stochastic curtailment analysis which strongly suggested that the addition of BUdR would not be associated with increased survival. In February 1997, the study was closed prior to full enrollment. At that time, the 1-year survival estimates were 82% versus 68% for RT plus PCV and RT/BUdR plus PCV respectively (one-sided, p = 0.96). The conditional power analysis indicated that even with an additional 12 months of additional accrual and follow-up the probability of detecting the prespecified difference was less than 0.01%. The differences in the two arms seem to be due to early deaths in the BUdR arm, not related to toxicity of the treatment. CONCLUSIONS: Despite encouraging Phase 2 results with BUdR, it is unlikely that a survival benefit will be seen. A final study analysis will not be done for at least 3 more years.

Nonconvulsive status epilepticus in a patient with leptomeningeal cancer.

A 56-year-old man sought medical assistance because of recurrent nonconvulsive status epilepticus without a history of prior seizure activity. Examination of the cerebrospinal fluid disclosed leptomeningeal cancer. To our knowledge, the association of partial complex status epilepticus and leptomeningeal cancer has not been reported previously. If the results of computed tomographic and magnetic resonance imaging studies are normal, examination of cerebrospinal fluid should be considered in patients with nonconvulsive status epilepticus.

Results of computer-assisted stereotactic laser resection of deep-seated intracranial lesions.

A computer-assisted stereotactic system has been developed for the precise resection of deep-seated intracranial neoplasms. After the tumor volume is reconstructed from computed tomographic and magnetic resonance imaging data, a computer-monitored, stereotactically directed carbon dioxide laser is used to vaporize the intracranial tumor. A computer graphics terminal is used to monitor the position of the laser in relationship to the planar slices through the tumor, which are reformatted orthogonally to the surgical plane of view. This procedure produced satisfactory postoperative neurologic results in 36 of the 41 patients who underwent treatment. The system provides precise surgical control in three-dimensional space for the safe resection of substantial amounts (as assessed by postoperative computed tomography) of intra-axial neoplasms.

High-grade astrocytomas: resource use, clinical outcomes, and cost of care.

OBJECTIVE: To describe the clinical course, survival, resource use, and direct medical costs of care for patients with high-grade astrocytomas. MATERIAL AND METHODS: All patients with grade 3 or 4 astrocytoma who resided in Olmsted County, Minnesota, or one of the six adjacent counties and had a tissue diagnosis first made between 1987 and 1992 were studied. Clinical characteristics, initial management, use of resources, clinical course, survival, and medical charges were analyzed. RESULTS: Sixty-four patients, with a mean age of 62 years, were identified; 81% had glioblastoma multiforme. Approximately 60% underwent surgical resection, 80% had radiotherapy, and 50% had chemotherapy for initial management. After initial treatment (median duration, 116 days), approximately 75% of patients had a course with stable disease (median duration, 198 days). The overall median duration of survival was 323 days; lower grade and younger age were significantly associated with longer median survival-for example, 1,493 days for patients younger than 65 years with grade 3 astrocytomas and 205 days for patients 65 years old or older with grade 4 astrocytomas. The mean total direct medical charges were $67,887. CONCLUSION: In most patients with high-grade astrocytomas, a substantial period elapsed before disease progressed. Although the overall median duration of survival was less than 1 year, younger patients, especially those with grade 3 astrocytomas, had a longer survival. The management of patients with high-grade astrocytomas uses substantial health-care resources.

Magnetic resonance imaging in cancer-related lumbosacral plexopathy.

OBJECTIVE: To study the relative utility of computed tomography (CT) and magnetic resonance imaging (MRI) of the lumbosacral plexus in patients with systemic cancer and plexopathy. DESIGN: In a retrospective study, we identified all patients encountered at Mayo Clinic Rochester between 1987 and 1993 with a diagnosis of lumbosacral plexopathy, and we selected for analysis those with MRI scans of the plexus (an abnormal finding was not necessary for inclusion) and a clinical and electrophysiologic appearance consistent with a diagnosis of metastatic lumbosacral plexopathy. MATERIAL AND METHODS: The study group consisted of 31 patients (20 men and 11 women). The types of tumor were as follows: prostatic, 10 patients; colorectal, 7; bladder, 3; cervical, 3; and other, 8. Eighteen patients had received pelvic radiotherapy before diagnosis of lumbosacral plexopathy. All available MRI scans (in 27 patients) were reviewed blinded; the initial imaging report was used if the actual scans were unavailable (in 4). CT had been done in 22 patients, and results for 16 were available for blinded review. Original reports were available for the other six. RESULTS: Direct involvement of the lumbosacral plexus by tumor was evident on 23 MRI studies, and 6 others showed widespread metastatic disease in the region of the plexus. On 13 CT examinations, direct involvement of the lumbosacral plexus by tumor was noted. In four patients, MRI findings were abnormal and CT findings were normal. No patient had abnormal CT findings and normal MRI findings. CONCLUSION: In this retrospective review, MRI was more sensitive than CT for diagnosing cancer-induced lumbosacral plexopathy. Thus, use of MRI should be considered in the diagnostic work-up of patients with clinical and electrophysiologic evidence of plexopathy and suspected systemic cancer.

Paraneoplastic brachial plexopathy in a patient with Hodgkin's disease.

We describe a case of inflammatory brachial plexopathy that occurred in the context of a mild, diffuse sensorimotor peripheral neuropathy associated with Hodgkin's disease. Clinical, electrophysiologic, and pathologic studies helped distinguish this disorder from other causes of brachial plexopathy in patients with cancer. Treatment with corticosteroids seemed beneficial in this patient. We suggest that this may be another type of paraneoplastic condition associated with Hodgkin's disease.

Phase II study of antineoplastons A10 (NSC 648539) and AS2-1 (NSC 620261) in patients with recurrent glioma.

OBJECTIVE: To assess the pharmacokinetics, toxicity, and efficacy of antineoplastons A10 (NSC 648539) and AS2-1 (NSC 620261). DESIGN: We initiated a phase II trial in order to determine whether evidence of antitumor activity of A10 and AS2-1 could be documented. MATERIAL AND METHODS: Patients with anaplastic astrocytoma or glioblastoma multiforme recurring after radiation therapy were eligible for enrollment in the trial. Patients received escalating doses of A10 and AS2-1 by multiple intermittent intravenous injections with use of a portable programmable pump to the target daily dose of 1.0 g/kg for A10 and of 0.4 g/kg for AS2-1. RESULTS: Nine patients were treated, in six of whom the treatment response was assessable in accordance with protocol stipulations. No patient demonstrated tumor regression. Reversible grade 2 or 3 neurocortical toxicity, consisting of transient somnolence, confusion, and exacerbation of an underlying seizure disorder, was noted in five patients. Mean steady-state plasma concentrations of phenylacetate and phenylacetylglutamine after escalation to the target doses of A10 and AS2-1 were 177+/-101 microg/mL and 302+/-102 microg/mL, respectively. Patients who exhibited confusion tended to have higher phenylacetate levels. CONCLUSION: Although we could not confirm any tumor regression in patients in this study, the small sample size precludes definitive conclusions about treatment efficacy. Antineoplaston-related toxicity was acceptable in most patients with appropriate dose modification, although severe neurocortical toxicity may occur. Steady-state plasma concentrations of phenylacetate with use of A10 and AS2-1 were similar to those reported with use of similar doses of phenylacetate alone.

Neurologic complications of systemic cancer.

Neurologic complications of systemic cancer are common. Metastatic complications include intracranial metastasis, spinal metastasis, leptomeningeal metastasis, and metastasis to peripheral nerves. Treatment is often effective in preventing further neurologic disability and pain. Early diagnosis is important. Nonmetastatic complications of systemic cancer include a wide variety of neurologic illnesses that can also occur in the noncancer population. The most common is toxic metabolic encephalopathy. Other nonmetastatic complications include cerebrovascular complications, complications of treatment, immunocompromised infections, and paraneoplastic syndromes.

Radiation-associated atheromatous disease of the cervical carotid artery: report of seven cases and review of the literature.

The natural history of postirradiation extracranial cerebrovascular disease is uncertain. Previous reported cases spanning 20 years of carotid surgery are difficult to evaluate, because patients may sometimes have unspecified symptoms, physical examinations, postoperative results, and follow-up. Also, the evolution of carotid surgery over the past two decades makes it impossible to compare earlier operative technique with the state-of-the-art technique of today. Our series of 7 patients underwent 9 carotid endarterectomies with an average follow-up period of 46 months. The number of patients is small, and although technically this is a more difficult operation, we feel the results are favorable and may be comparable with endarteerctomy procedures in nonirradiated patients. These patients should be approached as if radiation changes are not a major factor when they are considered for reconstructive arterial surgery.

Phase I evaluation of preirradiation chemotherapy with carmustine and cisplatin and accelerated radiation therapy in patients with high-grade gliomas.

OBJECTIVE: A Phase I study was conducted to determine the safety, toxicity, and maximum tolerated dose of preirradiation chemotherapy using carmustine (BCNU) and cisplatin in the treatment of high-grade gliomas. METHODS: Patients with newly diagnosed high-grade gliomas received BCNU and cisplatin after surgery, both before and during definitive radiation therapy. Preirradiation chemotherapy consisted of an administration of 40 mg/m2 BCNU on Days 1 through 3 and 30 mg/m2 cisplatin on Days 1 through 3 and 29 through 31 and repeated at 8 weeks to coincide with the start of radiation therapy. Postradiation chemotherapy consisted of an administration of 200 mg/m2 BCNU once every 8 weeks for four cycles. Radiation therapy consisted of 160-cGy fractions administered twice daily for 15 days, yielding a total dose of 4800 cGy. Dose escalation of BCNU was planned. If hematological toxicity was mild, the dose of cisplatin was to be held constant and BCNU dose escalated to 50 mg/m2 on Days 1 through 3. RESULTS: Eighteen patients were studied. The hematological toxicity was dose-limiting. Grade 3 or 4 leukopenia occurred in each of 10 patients (56%), and Grade 3 or 4 thrombocytopenia occurred in each of 9 patients (50%). Other toxicities included anorexia (94%), nausea (83%), emesis (33%), alopecia (94%), mild ototoxicity (50%), and, in one patient, death as a result of BCNU pulmonary toxicity. The median survival time was 14 months. Objective responses occurred in 45% of the patients evaluable for response. The maximum tolerated dose of this combination was 50 mg/m2 BCNU on Days 1 through 3 and 30 mg/m2 cisplatin on Days 1 through 3 and 29 through 31 before radiation and repeated in 8 weeks to coincide with the start of radiation. CONCLUSION: This schedule of the preirradiation administration of BCNU and cisplatin with accelerated hyper-fractionated radiation therapy for the treatment of high-grade gliomas provides a less toxic alternative to that of previous studies of preirradiation chemotherapy with these agents and merits further investigation.