Multi-institutional phase II study of temozolomide administered twice daily in the treatment of recurrent high-grade gliomas†
We dedicate this article to the memory and courage of Jenna Kamil, who died from a brain tumor at age 7.
The prognosis for patients with recurrent high-grade gliomas is poor and treatment options are limited. Current chemotherapeutic regimens can improve clinical outcomes, but extend survival by only a few months. Temozolomide is a methylating agent that is typically administered once daily. Because preclinical studies suggested that a twice-daily dosing schedule might be more effective, the safety and efficacy of twice-daily dosing of temozolomide were studied in patients with recurrent gliomas at their first, second, or third recurrence.
This multi-institutional trial enrolled 120 patients with recurrent glioblastoma multiforme (GBM), anaplastic astrocytoma (AA), or anaplastic oligodendroglioma (AO). An initial oral dose of 200 mg/m2 of temozolomide was followed by 9 consecutive doses of 90-mg/m2 every 12 hours. Treatment cycles were repeated every 28 days. Doses were escalated to 100 mg/m2 twice daily in the absence of unacceptable toxicity or were reduced if unacceptable toxicity occurred.
For GBM, AA, and AO patients, respectively, the median progression-free survival (PFS) was 4.2 months, 5.8 months, and 7.7 months, whereas the median overall survival (OS) was 8.8 months, 14.6 months, and 18 months. The overall response rate (partial and complete) for the GBM, AA, and AO patients was 31%, 46%, and 46%, respectively. Grade 3/4 toxicities included neutropenia (1.1%), thrombocytopenia (3.6%), and anemia (0.3%) (graded according to the World Health Organization grading system).
Twice-daily dosing may enhance the efficacy of temozolomide in the treatment of recurrent gliomas without increasing toxicity. This regimen compares favorably with other dosing schedules of temozolomide reported in the literature. Cancer 2008. © 2008 American Cancer Society.
High-grade gliomas are highly aggressive tumors that are rarely cured by conventional treatments such as surgery and radiotherapy. The prognosis for patients with recurrent gliomas is poor because treatment options are limited and few patients survive more than 2 years.1 In these patients repeat surgery may not be feasible because of tumor infiltration of critical structures. Additional radiation treatment is of limited utility in controlling tumor growth and may potentiate neurologic toxicity. Although chemotherapy improves outcome in the adjuvant setting, it extends survival only by a few months.2 Moreover, high-grade gliomas are heterogeneous in nature and the response to chemotherapy has been shown to differ considerably according to tumor type. The anaplastic oligodendroglioma (AO) subgroup of gliomas has demonstrated chemosensitivity, but responses are less frequent for patients with anaplastic astrocytoma (AA) or glioblastoma multiforme (GBM).2–4
Temozolomide (Temodar, Schering, Kenilworth, NJ), a second-generation DNA methylating agent, is rapidly absorbed orally.5 It readily penetrates the blood-brain barrier, with a reported cerebrospinal fluid (CSF)-plasma concentration ratio of up to 40%.6 Temozolomide has demonstrated antitumor activity in several human brain tumor xenograft models, including carmustine-resistant tumors.7, 8 In addition, clinical studies have shown single-agent activity in patients with recurrent/refractory anaplastic gliomas and GBMs.9–11 In clinical trials, temozolomide was generally well tolerated, with minimal toxicity and predictable myelosuppression; grade 3/4 thrombocytopenia was noted in only 7% of patients (graded according to the World Health Organization grading system). The standard dosage is 150 to 200 mg/m2 once daily for 5 consecutive days in a 28-day treatment cycle. However, both experimentally and clinically, the response to temozolomide has been found to be dose-dependent and schedule-dependent; more dose-intense regimens may be more effective than single-dose therapy or the standard 5-day dosing regimen.5, 12–14
Temozolomide undergoes nonenzymatic chemical conversion to its active metabolite, monomethyl triazenoimidazole carboxamide (MTIC).5 The cytotoxic effects of temozolomide are by methylation of O6 guanine residues in DNA, which forms the cytotoxic adduct O6-methylguanine (O6-mG).12 Repair of this adduct is by O6-alkylguanine DNA alkyltransferase (AGAT) and mediates resistance to temozolomide and other alkylating drugs.15 During DNA adduct repair, AGAT is irreversibly inactivated and new synthesis is required to restore repair activity. It has been hypothesized that increasing the frequency of administration of temozolomide might deplete more AGAT.16 Gerson et al.17 initially reported that in cancer patients a temozolomide dosing schedule with a bolus dose of 200 mg/m2 followed by 9 doses (50–100 mg/m2) given twice daily for 5 days resulted in rapid depletion of AGAT activity (>90%) in peripheral blood mononuclear cells (PBMCs).
MATERIALS AND METHODS
This multicenter, single-arm phase 2 study was designed to determine the efficacy and safety of a twice-daily temozolomide regimen in patients with recurrent high-grade gliomas. As in many previous studies in glioma the primary objective was to determine progression-free survival (PFS). Secondary objectives were to assess tumor response rates and toxicity. All patients underwent a baseline neurologic and general medical examination before starting chemotherapy. Complete blood counts were performed weekly and liver function tests and electrolyte measurements were performed monthly. Contrast-enhanced magnetic resonance imaging (MRI) was performed every 2 treatment cycles within 7 days of the next cycle.
In the study protocol, the required number of events (ie, progression or death) was determined such that there was approximately 80% power to detect a 50% improvement in median PFS compared with historical controls (3 months for GBM and 6 months for AA and AO)12, 18–21 using a 1-sided Wald test under exponential survival. This required enrollment of approximately 30 patients with GBM, 30 patients with AA, and 40 patients with AO to achieve a power of 86%, 83%, and 92%, respectively. The study enrolled 68 patients with GBM, 28 with AA, and 24 with AO. With this sample size at the observed median survival, the empiric powers were 71%, 32%, and 28%, respectively.
This protocol was approved by the Institutional Review Board (IRB) of each participating institution and was conducted in accordance with the principles of the Helsinki Declaration and HIPAA guidelines. All patients provided written informed consent. Eligible patients were those with recurrent/progressive (up to and including 3 recurrences) malignant gliomas (GBM, AA, or AO) who had MRI-measurable disease within 3 weeks of the initiation of treatment. Patients upgraded from low-grade glioma were not eligible. Last prior radiation and chemotherapy had to be 12 and 6 weeks, respectively, before study entry. Patients were excluded if they had received prior treatment with temozolomide for glioma. Other entry criteria included a Karnofsky performance status >60, life expectancy ≥12 weeks, age ≥18 years, and absence of pregnancy or lactation. Required baseline laboratory measurements were as follows: an absolute neutrophil count >1500/mm3, a platelet count >100,000/mm3, hemoglobin >10 g/dL, blood urea nitrogen and creatinine <1.5× the upper limit of normal (ULN), and aspartate aminotransferase and alanine aminotransferase levels <3× ULN. Patients were excluded if they had not recovered from the toxicities of previous treatments or had tested positive for the human immunodeficiency virus.
Temozolomide was administered orally in a fasting state on a twice-daily schedule for 5 days. An oral dose of 200 mg/m2 on the morning of the first day was followed 12 hours later by 9 doses of 90 mg/m2 every 12 hours. Treatment cycles were repeated every 28 days. Responses to chemotherapy were evaluated every 2 cycles by MRI and were tabulated according to the criteria of Macdonald et al.22 Patients without disease progression while receiving temozolomide therapy received a maximum of 12 treatment cycles. In some instances, IRB approval was requested to allow for treatment beyond 12 months, up to 20 months. In the absence of unacceptable toxicity after the first cycle, dosing for subsequent cycles was escalated to 100 mg/m2 twice daily after the initial dose of 200 mg/m2. For patients with hematologic or any other therapy-related toxicity, prespecified dose reductions were made, as shown in Table 1. Patients who required a dose reduction in the previous cycle were required to recover to nadir toxicity level 0 with a hematocrit ≥30 before the next treatment cycle to continue therapy.
Table 1. Dose Modification Criteria for Twice-Daily Temozolomide
|0||≥2000||≥100,000||Dose unchanged from previous|
|1||1500–1999||75–99,999||Dose unchanged from previous|
|2||1000–1499||50–74,999||Dose unchanged from previous|
|3||500–999||25–49,999||Decrease to next lower dose level|
|4||<500||<25,000||Decrease to next lower dose level|
|Dose Level (mg/mm2) for Twice-Daily Temozolomide|
|Dose Level||Oral Bolus Dose*||Dose Every 12 Hours†||Total Dose|
All patients had a baseline bidimensional measurement of their tumor. If multiple lesions were present, the largest lesion was used as the index lesion. This lesion was monitored on subsequent contrast MRI scans for response to chemotherapy. Patients with stable disease (SD), partial response (PR), or complete response (CR) received 2 more treatment cycles with temozolomide followed by another assessment of response. CR was defined as complete disappearance of all radiographically measurable lesions that was sustained for at least 4 weeks. PR was defined as a ≥50% but <100% reduction in the enhancing component of all brain lesions without the appearance of any new lesions. Progressive disease (PD) was defined as a ≥25% increase in the enhancing tumor or the appearance of new lesions. SD described all other situations (ie, <25% volume reduction of all measurable lesions and no new lesions for >8 weeks). For CR, PR, and SD, the patient had to be on a stable or decreased corticosteroid dose and demonstrate stable or improved neurologic function. All data were analyzed on an intention to treat (ITT) basis.
Patients were separated into 3 groups according to tumor histology: GBM, AA, and AO. Baseline characteristics of all patients were summarized using descriptive statistics. The response rate and median PFS were estimated with a 95% confidence interval (95% CI). The Cox proportional hazards regression (HR) model was used to correlate each potential prognostic factor with survival in univariate analyses within each histologic stratum. A variable selection procedure was conducted based on Collett,23 and a logistic regression model was used to identify important factors for objective clinical responses (CR or PR). Data were collated based on ITT analysis.
The actual accrual was different from the planned accrual in the original protocol; therefore, the post hoc power for each histological group was recomputed based on the observed median PFS and a 1-sided Wald test at a .05 level under exponential survival. The probability of death in each group was computed using formula (9.12) published in Collett.23
A total of 120 patients were enrolled by October 2002 at 7 institutions. Patient characteristics at baseline are summarized in Table 2. The median age at time of treatment was 50 years (range, 24–77 years). Histologic diagnosis was GBM in 68 patients (57%), AA in 28 patients (23%), and AO in 24 patients (20%). All but 2 patients (98%) had received previous radiation treatment, and 69 patients (58%) had progressed on or after previous chemotherapy, which did not include temozolomide. The majority of patients (n = 79; 66%) were treated at the time of first recurrence; 26 patients (22%) were treated at second recurrence, and 15 patients (13%) were treated at third recurrence. No patients were naive to radiation or chemotherapy.
Table 2. Demographic and Baseline Characteristics
|No. of patients (%)||68 (57)||28 (23)||24 (20)||120|
|Median KPS, %|| || || ||80|
|Gender, no. (%)|
|Female||28 (41)||11 (39)||9 (38)||48 (40)|
|Male||40 (59)||17 (61)||15 (63)||72 (60)|
|Recurrence, no. (%)|
|1||48 (71)||19 (68)||12 (50)||79 (66)|
|2+||20 (29)||9 (32)||12 (50)||41 (34)|
|No. of treatment cycles|
|Previous chemotherapy, no. (%)|| |
|None||28 (41)||10 (36)||13 (54)||51 (43)|
|Non-nitrosourea||7 (10)||1 (4)||0 (0)||8 (7)|
|Nitrosourea||33 (49)||17 (61)||11 (46)||61 (51)|
A median of 4 temozolomide treatment cycles (range, 1–20 cycles) per patient was administered. All patients discontinued study treatment because of disease progression; there were no discontinuations because of adverse events or hematologic toxicity.
Response outcomes by tumor histology (ITT population) are summarized in Table 3. The overall response rate (CR + PR) was 31% (95% CI, 20–43%) for GBM, 46% (95% CI, 26–67%) for AO, and 46% (95% CI, 28–66%) for AA. The overall response rate for all tumors combined was 38%. The majority of responders (85%) achieved their best response after 2 cycles of chemotherapy. The median duration of response was 6 months for GBM, 17 months for AO, and 10 months for AA. Eleven percent of responders achieved best response after 4 cycles, 3% after 6 cycles, and 1% after 8 cycles. By the end of the study, 31 responders experienced disease progression; however, at the time of last follow-up, 24 patients were alive, 21 of whom remained free of disease progression. The median follow-up for all patients was 21months.
Table 3. Response by Histology by Intention-to-Treat Analysis
|No. of patients||68||28||24||120|
|CR, no. (%)||3 (4)||5 (18)||4 (17)||12 (10)|
|PR, no. (%)||18 (26)||8 (29)||7 (29)||33 (28)|
|SD, no. (%)||22 (32)||8 (29)||5 (21)||35 (29)|
|PD, no. (%)||19 (28)||4 (14)||7 (29)||30 (25)|
|Overall response (CR + PR), no. (%)||21 (31)||13 (46)||11 (46)||45 (38)|
The 6-month PFS rate was 43% for all patients. Based on tumor histology, the 6-month PFS was 35% for GBM patients, 50% for AA patients, and 58% for AO patients (Table 4). The median PFS was 5 months and the median OS was 11 months for all patients combined. The median PFS was 4 months for GBM patients, 6 months for AA patients, and 8 months for AO patients, which was comparable to or slightly improved compared with historical controls; however, the median PFS was not improved by 50%, which was the primary objective of this study. Among the 45 responders (CR + PR), the median time to progression was 8 months for GBM patients, 12 months for AA patients, and 19 months for AO patients. For all patients, the 6-month survival rate was 71% for GBM patients, 86% for AA patients, and 75% for AO patients. The 12-month survival rates were 35%, 59%, and 71% for GBM patients, AA patients, and AO patients, respectively.
Table 4. Overall and Progression-Free Survival by Intention-to-Treat Analysis
|Median PFS, mo (95% CI)||4 (3–6)||6 (4-NA)||8 (3–21)||5 (4–7)|
|Median OS, mo (95% CI)||9 (7–11)||15 (11-NA)||18 (13-NA)||11 (9–13)|
|6-mo PFS probability, % (95% CI)||35 (26–49)||50 (34–72)||58 (42–82)||43 (35–53)|
|6-mo OS probability, % (95% CI)||71 (61–82)||86 (73–100)||75 (60–95)||75 (68–83)|
|12-mo PFS probability, % (95% CI)||13 (7–24)||33 (18–60)||42 (26–70)||23 (16–32)|
|12-mo OS probability, % (95% CI)||35 (25–48)||59 (41–85)||71 (55–92)||46 (38–57)|
Association Between Prognostic Factors and Clinical Outcome
The following prognostic factors were analyzed to determine correlation with clinical outcome: age, previous treatments, histology, and number of prior recurrences. The association between response and histology was not statistically significant (chi-square test: P = .42). The number of prior recurrences was the only significant factor for PFS. Surprisingly, a high number of prior recurrences was found to be correlated with a reduced risk of disease progression (HR, 0.75; P = .03). The impact of this factor on OS was similar but did not reach statistical significance (HR, 0.83; P = .16). Age was the only factor significantly affecting OS, with an HR of 1.0 (P = .04).
Safety of the Twice-Daily Regimen
Of the 624 total treatment cycles, grade 3/4 thrombocytopenia, neutropenia, and anemia were observed in 3.6%, 1.1%, and 0.3% of cycles, respectively. Overall, 18% of patients experienced a grade 3/4 toxicity. Grade 1/2 lymphopenia was observed in approximately 35% of the patients; no grade 3/4 lymphopenia or opportunistic infections occurred.
There were 2 treatment-related deaths. One patient died of septicemia while neutropenic; the other died with severe thrombocytopenia. One patient experienced a transient myelodysplastic syndrome after 1 cycle of chemotherapy. Doses were escalated for 20 patients with no grade 3/4 toxicities. Eleven patients (9%) required dose reduction for transient hematologic toxicity. No dose delays were required because of hematologic toxicity. It is interesting to note that no correlation was found between the severity of hematologic toxicities and previous chemotherapy.
The optimal dosing regimen for temozolomide remains unknown. On the basis of studies suggesting increased AGAT depletion and increased clinical efficacy with more frequent or prolonged exposure to temozolomide, different treatment regimens have been investigated. These include temozolomide daily for 7 days on and 7 days off,24, 25 daily for 3 weeks on and 1 week off,24 and daily for 6 or 7 weeks followed by 4 weeks off.14 When compared with the standard once-daily dosing regimen, these extended temozolomide dosing schedules resulted in a 2.1-fold to 2.8-fold14, 24 greater drug exposure (ie, total dose per cycle). They were also associated with significant decreased AGAT activity in PMBCs (ie, 72% decrease on Day 7 and recovery to only 55% of baseline before dosing on Day 15 in a trial of a 7 days-on/off schedule24). Moreover, the safety profile of temozolomide using extended dosing schedules was similar to the standard once-daily, 5-day dosing schedule.
The twice-daily dosing regimen in this phase 2 study was based on a phase 1 study in which a bolus dose of 200 mg/m2 of temozolomide was followed by 9 consecutive doses given every 12 hours to patients with metastatic tumors.17 The rationale for twice-daily dosing after an initial loading dose was that AGAT activity may partially recover by 24 hours. Gerson et al.17 initially reported that a twice-daily regimen resulted in a >90% reduction in AGAT activity in PBMCs at 24 hours, compared with only 50% in those receiving once-daily dosing. As a caveat, this and a second study from Gerson's group, which measured AGAT activity from pretreatment and posttreatment tumor biopsies, showed that AGAT depletion in tumor tissue was variable and did not correlate with depletion of AGAT activity in PBMCs.17, 26 Therefore, although it is possible that AGAT depletion in brain tumors may be enhanced by twice-daily dosing, as noted in PBMCs, there is no clear evidence that this is the case in tumors.
The results from the current study indicate that the temozolomide twice-daily regimen has clinical efficacy in the treatment of recurrent high-grade gliomas, achieving response rates that compare favorably with those reported by others in similar patient populations, with a safety profile that is comparable to the standard once-daily regimen.10, 11, 19–21 A summary of phase 2 studies that used the standard once-daily temozolomide dosing regimen18 for recurrent gliomas is shown in Table 5. It is important to note that clinical studies that make a direct comparison between different temozolomide dosing regimens have not been published to date; therefore, caution must be exercised in comparing data from these studies. However, some of the efficacy endpoints evaluated in this phase 2 study suggest that twice-daily temozolomide dosing might provide some advantage over the standard dosing regimen. For example, the6-month PFS and OS rates for GBM patients in the current study were 35% and 71%, respectively, compared with 21% and 60% reported by Yung et al.11 with the once-daily 5-day regimen (note: PFS was 1 of the primary objectives of that study). In addition, the overall response rate of 31% observed in our GBM patients was greater than that reported by Yung et al. (5%),11 Harris et al. (20%),10 and Brada et al. (8%).20 Three CRs (5%) were achieved in our GBM patients on twice-daily temozolomide, whereas no CRs were observed in 2 of the 3 studies where GBM patients received the once-daily temozolomide dosing. The results of the current study are particularly striking given that the pivotal trial of temozolomide in patients with recurrent GBM reported by Yung et al.11 restricted eligible patients to those in first recurrence, whereas the present study enrolled patients at first (66%), second (22%), and third (13%) recurrence.
Table 5. Results of Published Studies Using Temozolomide for the Treatment of Recurrent Malignant Gliomas
|AA||97 (EH)||6 (6)||27 (28)||31 (32)||33 (34)||5.5||49||14.2||78|
|AOA||14 (EH)||2 (14)||4 (29)||1 (7)||6 (43)||5.8||46||14.8||79|
|AA/AOA||162 (ITT)||13 (8)||44 (27)||44 (27)||57 (35)||5.4||46||13.6||75|
|AO/AOA||48 (ITT)||8 (17)||13 (27)||19 (40)||21 (44)||6.7||50||10.0||77|
|van den Bent 200128*|
|AO||30 (ITT)||3 (10)||6 (20)||8 (27)||9 (30)||—||43||7||—|
|AA and AA/AO||31 (ITT)||5 (16)||5 (16)||7 (23)||10 (32)||7||—||13||—|
|GBM||25 (ITT)||0 (0)||5 (20)||5 (20)||5 (20)||1||—||9||—|
|All||56 (ITT)||5 (9)||10 (18)||12 (21)||15 (27)||6||—||11||—|
|GBM||112 (ITT)||0 (0)||6 (5)||45 (40)||6 (5)||Approx. 4||21||Approx. 7.5||60|
|GBM||138 (ITT)||2 (2)||9 (6)||60 (43)||11 (8)||2.1||19||5.4||46|
|GBM||28 (ITT)||0 (0)||0 (0)||—||0 (0)||2.3||19||7.7||60|
|GBM/AA/AO/AOA||35 (ITT)||0 (0)||2 (6)||—||2 (6)||2.5||27||8.7||67|
|GBM||21 (ITT)||0 (0)||2 (10)||17 (81)||2 (10)||4.9||48||—||—|
The survival rates of the GBM subgroup in the current study also compare favorably with those observed with extended-dosing regimens of temozolomide, which, as discussed earlier, demonstrated promise in phase 1 studies based on significant AGAT depletion in PMBCs. The Khan et al.21 trial of 6 weeks on/4 weeks off and the Wick et al.27 study of a 7 days on/7 days off regimen reported 6-month PFS for GBM patients of 19% and 48%, respectively, as compared with 35% in this study. A randomized intergroup study (RTOG 0525) is comparing 2 schedules of temozolomide for patients with newly diagnosed GBMs. Patients will receive either standard daily for 5 days or an extended schedule in which patients receive drug for 21 days of a 28-day cycle.
A direct comparison of the results for recurrent AA and AO from the current study with reported data are also possible. However, to our knowledge only 2 published clinical studies provided data for pure AA or AO histologies, whereas most other studies combined AA or AOs with tumors of mixed histology. In the study reported by Yung et al.,18 the overall response rate for AA patients was 34% compared with 46% in the current study. The median PFS (6 months) and OS (15 months) observed in the present study for AA were comparable to those reported for AA patients in the Yung study18 (5.5 and 14.2 months, respectively).
In the study reported by van den Bent et al.,28 which used the standard once-daily 5-day schedule, the overall response rate in patients with recurrent AO was 30% compared with 46% in our study. The median OS was 7 months for patients with AO in that study,28 whereas in the present study, median OS for patients with AO was 18 months. These results suggest that the twice-daily regimen is at least as efficacious, if not better than, once-daily dosing in patients with recurrent AA and AO, but this is based on very limited data.
A parallel comparison of the demographics and baseline patient characteristics of the current study with those from the 2 studies discussed above was performed to identify any potential confounding factors that may account for the difference in response rates. The median age and percentage of chemotherapy-naive patients in the study of AA and AO reported by Yung et al.18 (42 years and 40%, respectively) were comparable to that for the current AA and AO populations. The median age in the GBM study by Yung et al.11 (52 years) is comparable to our GBM population. An interesting difference between the studies by Yung et al.11, 18 and the present study is that Yung's studies were restricted to patients at first recurrence of disease. Because the current study did not have this restriction, 34% of the total study population was in their second or third recurrence.
In summary, the results of the current study with twice-daily dosing of temozolomide compare favorably with published reports of patients with recurrent gliomas, particularly with respect to response and PFS in patients with GBM, and suggest that a twice-daily dosing regimen is an effective treatment schedule in the recurrent setting and is associated with less hematologic toxicity than the dose-dense regimens that deliver a 2-fold higher dose per cycle. Based on these findings, phase 3 testing of twice-daily versus once-daily dosing is warranted. Pending confirmation of the benefits of twice-daily dosing, this regimen could potentially be used as an alternative to the standard once-daily dosing schedule during the adjuvant phase of chemoradiotherapy for newly diagnosed malignant glioma or as second-line therapy either in patients not previously treated with temozolomide or possibly in those achieving a good initial response to first-line temozolomide. The twice-daily dosing regimen many have several advantages over once-daily dosing, including the potential to deplete MGMT more effectively and increased response rates, particularly in patients with GBM. The responses observed in the present study were long-lasting for patients with GBM, with some patients maintaining a PR or SD for 20 months. Up to 4 to 6 cycles may be necessary before a response is achieved, and this should be considered when determining whether or not to continue temozolomide. Further studies of alternative temozolomide dosing regimens to optimize treatment efficacy are warranted.
We thank the following investigators for their invaluable contributions to the study: K.S. Nisbett, M. Odaimi, and C. De Rosa (Experimental Therapeutics Program, Division of Medical Oncology, New York Presbyterian, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY); K. Peterson (Stanford University Medical Center, Stanford, Calif); H. Krouwer (Froedert Memorial Lutheran Hospital, Milwaukee, Wis); S. Weaver (Albany Medical College, Albany, NY); L. McAllister (The Neurological Clinic, Portland, OR); L. Heller (Integrated Therapeutics Inc., Flushing, NY); and G. McKhann (Department of Neurosurgery, New York Presbyterian, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY). We also thank the following data managers who worked with us to prepare the article: Jennifer Vanderwerker, Carmela Ramnes, Lynn Adler, Jose DeCorte, Rebecca Moore, Margaret Coursen, Harriet Patrick, and Nancy O'Connor. We thank our families for silently supporting our endeavors and, last, but no less important, the 120 patients in this protocol for believing in us and allowing us the privilege to treat them.