Adult patients (age 18 to 70 years) were eligible if they had histologically proven inoperable GBM as determined by gadolinium-enhanced magnetic resonance imaging (Gd-MRI), no prior history of RT or chemotherapy, a KPS of ≥ 60, and adequate laboratory values (absolute neutrophil count ≥ 1,500 cells/μL, platelet count ≥ 100,000/μL, hemoglobin ≥ 8 g/dL, serum creatinine and bilirubin levels < 1.5× the upper limit of normal, and AST and ALT levels ≤ 3× the upper limit of normal). The study protocol was approved by the local medical ethics committee, and all patients provided written informed consent.

This prospective phase II study was conducted at a single center in Marseille, France. The primary objective was to evaluate objective response rate. Secondary objectives included assessment of progression-free survival (PFS), overall survival (OS), and safety. Patients received oral temozolomide at 150 mg/m²/d on days 1 to 7 and on days 15 to 21 of every 28-day cycle (7 days on/7 days off). Study treatment was administered for up to four cycles before RT (neoadjuvant) and for up to four additional cycles starting 1 month after RT (maintenance) until unacceptable toxicity or disease progression. RT was delivered after four cycles of temozolomide on response or stable disease or earlier if tumor progression occurred during neoadjuvant chemotherapy. Focal RT was delivered with a linear accelerator once daily at 2 Gy per fraction for 5 days a week for a total of 60 Gy. Treatment volumes were determined based on initial Gd-MRI scan, including contrast-enhanced lesions plus a 2- to 3-cm margin, depending on tumor volume and location. Repeat cycles of temozolomide were administered based on strict hematologic criteria that required an absolute neutrophil count of ≥ 1,500 cells/μL and a platelet count of ≥ 100,000/μL; otherwise, the subsequent treatment cycle was delayed until recovery, and blood cell counts were reassessed weekly. A treatment delay of 2 or more weeks required a decrease of temozolomide in increments of 25 mg/m²/d. Prophylactic antiemetics, such as 5-hydroxytryptamine-3, were administered each day of temozolomide treatment. Anticonvulsants were administered as needed, and corticosteroid dose was adjusted based on the neurologic condition of the patient.

Complete physical and comprehensive neurologic examinations, KPS, Mini-Mental State Examination, laboratory tests, and Gd-MRI were performed at baseline within 14 days (28 days for imaging) of study entry. All tests, except for hematology and chemistry assessments, were performed after each 28-day cycle for the first four cycles and then before RT and 1 month after completion of RT. Overall tumor response was assessed according to the criteria of Macdonald et al¹⁵ and was based on consecutive Gd-MRI scans of patients who were under stable corticosteroid use for 7 days before imaging. All patient scans in which a response (complete or partial response) was observed were centrally reviewed by an independent radiologist.

Weekly blood tests were performed during the first two treatment cycles and every 2 weeks thereafter. During maintenance therapy, clinical evaluations were performed after each cycle, and Gd-MRI was performed at least every two cycles. Adverse events occurring during treatment or up to 30 days after treatment were recorded and graded according to National Cancer Institute Common Toxicity Criteria (version 3.0).

Twenty-five of 30 tumor biopsy samples were available for MGMT analysis. The three remaining assessable patients had insufficient tumor biopsy material for analysis. Immunohistochemistry was performed according to the method of Mollemann et al.¹⁶ Briefly, tissue sections were incubated with a mouse monoclonal antibody (MT-1; Chemicon, Temecula, CA) at a dilution of 1:100 for 1 hour at room temperature. Cytoplasmic-only and granular nuclear reactivity were considered negative. High expression was defined as ≥ 35% of tumor cell nuclei expressing detectable MGMT protein; low expression was defined as less than 35% of nuclei expressing detectable MGMT.

PFS was measured from date of study entry to first sign of radiologic or clinical disease progression, whichever came first. OS was measured from date of study entry to date of death or last follow-up. PFS and OS were estimated using the Kaplan-Meier method. The results of the log-rank test, which was used to compare PFS and OS between groups, are reported separately for the intent-to-treat (ITT) and assessable populations. Correlation between MGMT and response rate was calculated using the χ² test. P ≤ .05 was considered significant.

A total of 30 patients (20 men and 10 women) with a median age of 60 years were enrolled between June 2002 and July 2004, and 29 patients were eligible to receive treatment (ITT population; Table 1). One patient was ineligible after giving consent because of unacceptable hepatic function. Another patient was lost to follow-up without evidence of disease progression after initiation of treatment cycle 2 and was not assessable for response. Twenty-eight patients were assessable for response, and all patients were assessable for survival and safety. Tumor materials from 30 patients were available for central histopathologic review. All patients had confirmed GBM. At the time of initial diagnosis, two patients had undergone partial surgical resection, and 28 patients were diagnosed on the basis of stereotactic biopsy. By recursive partitioning analysis classification, 17 patients were class V, and three patients were class VI. Median time from diagnosis to study entry was 12 days (range, 6 to 12 days). No patient received prior RT or chemotherapy.

A total of 97 treatment cycles of temozolomide were administered (median, three cycles per patient; range, one to eight cycles; Table 2). During treatment cycle 1, temozolomide was administered at 150 mg/m²/d in all patients (n = 29). Among 27 patients who received more than one treatment cycle, the dose was reduced in six patients (22%), including two patients who discontinued treatment after cycle 2 because of prolonged thrombocytopenia. Of the six patients who received dose reductions, two patients received a dose reduction to 125 mg/m²/d, and four patients received a dose reduction to 100 mg/m²/d. Forty-five percent of patients completed four cycles of neoadjuvant temozolomide, and 45% of patients had progressive disease after one (n = 2), two (n = 7), or three (n = 4) cycles and proceeded immediately to RT. Twenty-two patients (76%) started RT, and 16 patients (55%) completed RT. Twelve patients (41%) were unable to complete RT because of clinical and/or radiologic disease progression, and one patient (3%) was lost to follow-up. Of the 13 patients with progressive disease who proceeded to RT, eight started RT, and two completed RT. Three patients (10%) started and completed maintenance chemotherapy. Ten patients (35%) with controlled disease before RT experienced progression after RT and discontinued study participation before maintenance therapy.

In the ITT population, seven patients (24%; 95% CI, 8.6% to 39.7%) had a partial response, nine patients (31%) had stable disease for 3 months, and 12 patients (41%) had progressive disease. In the assessable population (n = 28), the objective response rate was 25% (95% CI, 8.63% to 41.37%). In the ITT group, study treatment resulted in an overall disease control rate (partial response + stable disease) of 55%. All responses were confirmed at least 4 weeks apart and were maintained at least until the initiation of RT. All responses occurred during the first three cycles; objective responses were observed after two cycles in five patients and after three cycles in two patients. An objective response was observed in two of eight patients with tumors of the corpus callosum, in two of seven patients with multifocal tumors, and in three of 14 patients with lobe tumors.

At the time of analysis, the median follow-up time was 6 months (range, 0.9 to 19 months) for the entire population. In the ITT population, median PFS time was 3.8 months (95% CI, 3.4 to 4.2 months), and median OS time was 6.1 months (95% CI, 0.9 to 11.3 months). Twelve- and 18-month survival rates were 28% (95% CI, 14.2% to 47.9%) and 14% (95% CI, 11.3% to 43.9%), respectively. The 12-month PFS rate was 7%, and all patients had disease progression at 18 months. Results for the ITT and assessable patient population were similar. At the time of analysis, a total of three patients (10%) were alive. The median follow-up time for these patients was 11.7 months (range, 4.2 to 11.9 months).

MGMT expression was low in normal brain and restricted to endothelial cells and a few oligodendrocytes (< 10%). Expression of MGMT in tumor cells and positive control cells was mainly visible in nuclei, although some cytoplasmic expression might have occurred. Of 25 tumor biopsy samples available for MGMT analysis, low MGMT expression (< 35% of tumor cells) was detected in 11 patients, and high MGMT expression was detected in 14 patients (Table 3). RT was started in 10 patients (91%) with low MGMT expression, and all 10 patients completed RT. In contrast, among patients with high MGMT expression, nine patients (64%) started RT, and six patients (43%) completed RT. The objective response rate was 55% in the 11 patients with low MGMT expression compared with 7% in the 14 patients with high MGMT expression (P = .004). Analysis of MGMT expression as a continuous variable (percentage of tumor cells) showed a trend toward lower MGMT expression in patients with stable disease or objective response compared with patients with progressive disease (Fig 1). Patients with low MGMT expression, compared with patients with high MGMT expression, also had significantly longer PFS (5.5 v 1.9 months, respectively; P = .009; Fig 2A) and OS (16 v 5 months, respectively; P = .003; Fig 2B).

Overall, temozolomide was well tolerated and associated primarily with hematologic toxicity. A total of 13 grade 4 hematologic adverse events occurred in seven patients (thrombocytopenia, n = 5; neutropenia, n = 4; leukopenia, n = 3; and lymphopenia, n = 1; Table 4). In addition, a total of five grade 3 hematologic adverse events (17%; lymphopenia) occurred. Of the four patients who received RBC transfusions, three patients also received 13 platelet transfusions. Two patients discontinued chemotherapy after cycle 2 because of severe and prolonged thrombocytopenia. Temozolomide was tolerated in 22 patients (81%) during cycle 1, and these patients followed the planned course of treatment. Febrile interstitial pneumonitis occurred in five patients (18%) and was associated with either profound lymphopenia (n = 3) or profound neutropenia (n = 2); however, all occurrences of pneumonitis resolved without sequelae. Pneumocystis carinii pneumonia was not confirmed in any patient but was suspected in one patient who, therefore, received trimethoprim. Emesis, constipation, and asthenia were moderate with the exception of one patient with grade 3 constipation. A low incidence of other treatment-related nonhematologic grade 3 or 4 adverse events was observed.