Phase II trial of intracerebrospinal fluid etoposide in the treatment of neoplastic meningitis


  • Marc C. Chamberlain M.D.,

    Corresponding author
    1. Department of Interdisciplinary Oncology, Moffitt Cancer Center and Research Institute, Tampa, Florida
    • Department of Interdisciplinary Oncology, Moffitt Cancer Center and Research Institute, 12902 Magnolia Ave., Tampa, FL 33612-0804
    Search for more papers by this author
    • Fax: (813) 745-3731

  • Denice D. Tsao-Wei M.S.,

    1. Department of Preventive Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author
  • Susan Groshen Ph.D.

    1. Department of Preventive Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author



The purpose of the current study was to determine the toxicity and response of a fixed dose intracerebrospinal fluid (CSF) etoposide in the treatment of patients with newly diagnosed neoplastic meningitis (NM). NM reportedly occurs in 1% to 5% of patients with known cancer. Currently available treatment options are limited and provide only modest benefit.


Twenty-seven patients (median age, 55 yrs) with clinically and cytologically documented NM received intra-CSF etoposide. Tumor histologies included lung (8 patients), breast (5 patients), primary brain tumor (4 patients), non-Hodgkin lymphoma (4 patients), melanoma (4 patients), colon (1 patient), and prostate (1 patient). Concurrent involved-field radiotherapy (19 of 27 patients) or systemic chemotherapy (17 of 27 patients) was administered based on clinical indications. Etoposide was administered at a fixed dose (0.5 mg every day given 5 days per week every other week for 8 weeks [induction]). Patients were evaluated by CSF cytology and neurologic examination at the conclusion of induction therapy. Responding patients continued to receive etoposide (5 consecutive days every 4 weeks) with monthly evaluations.


Seven of 27 patients (26%) treated with etoposide had a cytologic response and either stable or improved neurologic status at the conclusion of induction. Eight patients (30%) developed disease progression during induction therapy and did not complete the 8-week induction course of therapy. At the conclusion of induction therapy, 12 patients (44%) had persistently positive CSF cytology, although they were clinically stable. In responding patients, time to neurologic disease progression ranged from 8 weeks to 40 weeks (median, 20 wks). Toxicity manifested as transient chemical arachnoiditis (5 of 27 patients; 13% of all treatment cycles). The 6-month neurologic disease progression-free survival was 11%.


Etoposide appears to have modest activity against NM and easily managed toxicity. Cancer 2006. © 2006 American Cancer Society.

Neoplastic meningitis (NM) is a devastating complication of systemic or primary brain cancers. The overall frequency of NM varies as a function of tumor type and in general occurs in 5% of all patients with cancer.1–18 Treatment of NM is almost never curative and treatment goals therefore are palliative, with attempted preservation of neurologic function.

There is a paucity of randomized clinical trials with respect to NM and, as a consequence, there is no standard treatment.4–7 Four chemotherapy agents (liposomal cytarabine, methotrexate, cytarabine, and thio-TEPA) are used in the treatment of NM, with only modest differences noted among the agents with respect to toxicity or efficacy.4–7, 10–12

Etoposide is an attractive agent for use in NM for several reasons.9–23 First, etoposide has efficacy against a variety of cancers, including both hematologic (i.e., lymphomas) and solid tumors (i.e., small cell lung cancer, ovarian cancer, germ cell tumors, and brain tumors). Second, intracerebrospinal fluid (CSF) etoposide has been used to treat NM in small studies and has demonstrated modest efficacy and acceptable toxicity.19–23 Lastly, etoposide appears to have a unique mechanism of action (i.e., functions as a topoisomerase 1 inhibitor) compared with presently available chemotherapy agents in use for the treatment of NM.

This Phase II trial had 2 principal objectives: to document the toxicity of a fixed dose of intra-CSF etoposide and to determine the 6-month neurologic disease progression-free survival (PFS) rate of etoposide in the treatment of NM.


Study Design

This was an open-label, nonrandomized, single institution trial in which patients with cytologically documented NM irrespective of tumor histology and no prior intra-CSF therapy were treated with a fixed dose of etoposide.


The inclusion criteria required that patients have histologically proven cancer with a positive (recorded as positive, not suspicious or atypical) CSF cytology within 14 days before study entry; no prior intra-CSF therapy (i.e., newly diagnosed NM); a Karnofsky performance status (KPS) >50; no uncontrolled infection, including human immunodeficiency virus (HIV); and that they understood and signed an informed consent document regarding the experimental nature of the study. Required laboratory values were liberal and included a platelet count >80,000 and a white blood cell count (WBC) >3000 (or an absolute neutrophil count [ANC] >1500), creatinine <3 times the upper limit of normal, and all other blood chemistries <3 times normal. Prior central nervous system (CNS) radiation of any type was allowed but prior treatment of NM with intra-CSF chemotherapy was not. However, patients could have received prior intra-CSF chemotherapy for CNS prophylaxis. CSF compartmentalization identified on a radioisotope flow study or the need for a ventriculoperitoneal shunt excluded patients from participation.6–18 Concurrent systemic chemotherapy was permitted for treatment of disease outside of the meninges, with the exception of high-dose methotrexate (at a dose of >500 mg/m2), high-dose cytosine arabinoside (>2.0 gm/m2), high-dose thio-TEPA (>300 mg/m2), or investigational agents.19–21 Patients entering the study with symptomatic or radiographically visible CNS disease were required to receive local radiotherapy before the induction period but concurrent whole brain or craniospinal radiation was not permitted.6–8 This study was opened in 2000 with the approval of the University of Southern California Institutional Review Board and closed in 2004 after the accrual of 27 patients (Table 1). Patient entry onto study was discontinued due failure to achieve greater than 5 patients alive and neurologic disease PFS at 6 months.

Table 1. Characteristics of Study Subjects
NumberAge, yGenderTumor HistoryNeurologic PresentationConcurrent RadiotherapyConcurrent Systemic ChemotherapyDuration of Cytologic Response in wkSurvival, wk
  1. M: male; F: female; +: administered; CR: complete response (cytology converted positive to negative); PD: progressive disease.

152FBreastSpinal++ [Vinorelbine]CR/2429
255MNonsmall cell lungCerebral and cranial nerve+ PD/46
355MColonSpinal++ [Oxaliplatin]PD/810
460FBreastCranial nerve + [Exemestane]CR/1216
564FBreastSpinal++ [Gemcitabine]CR/4052
655MNon-Hodgkin lymphomaSpinal++ [Ibritumomab tiuxetan]CR/816
747MNonsmall cell lungSpinal++ [gefitinib]PD/89
826MMedulloblastomaSpinal + [temozolomide]CR/2024
949MNonsmall cell lungSpinal+ PD/45
1055MMelanomaCerebral+ PD/45
1161FNonsmall cell lungCranial nerve + [Gemcitabine and carboplatin]PD/46
1276MNonsmall cell lungSpinal and cranial nerve+ PD/811
1327MEpendymomaSpinal+ CR/2430
1456FNon-Hodgkin lymphomaCerebral and cranial nerve++ [Ibritumomab tiuxetan]PD/46
1534MMelanomaCranial nerve + [Temodar and thalidomide]PD/810
1677MProstateSpinal+ CR/1216
1751MMelanomaSpinal+ PD/45
1827FIntracranial germ cell tumorSpinal++ [Ifosphamide, methotrexate and vincristine]PD/811
1956MNonsmall cell lungSpinal and cranial nerve++ [gefitinib]PD/810
2063FBreastCranial nerve + [Trastuzumab]PD/812
2156FNon-Hodgkin lymphomaCranial nerve + [Ibritumomab Tiuxetan]PD/813
2229MMedulloblastomaSpinal++ [Carboplatin]PD/810
2361MNonsmall cell lungSpinal+ PD/49
2454FBreastCranial nerve + [vinorelbine]PD/811
2534MMelanomaCerebral and cranial nerve++ [temozolomide]PD/24
2664MSmall cell lungSpinal+ PD/810
2761FNon-Hodgkin lymphomaCranial nerve  PD/812

Intra-CSF Chemotherapy

Patients received 0.5 mg of etoposide (Schering-Plough, Kenilworth, NJ) by intraventricular injection every day for 5 consecutive days per week every other week for a total of 8 weeks (20 total doses). This period (2 months) constituted induction and patients who were clinically stable or improved and who had converted from positive to negative CSF cytology at all previously positive sites continued to maintenance therapy. During maintenance therapy, patients received 0.5 mg of etoposide every day for a total of 5 consecutive days every 4 weeks (5 total doses). In patients who were clinically stable or improved and who maintained negative CSF cytology, maintenance therapy continued until disease progression. A cycle of etoposide was defined as 1 week of therapy during which patients received 5 doses of etoposide. All patients received the study drug as prescribed by the protocol. Steroids were not prescribed for prophylaxis but were permitted for treatment of etoposide -induced arachnoiditis. Patients who developed disease progression during induction therapy or had persistently positive CSF cytology at the conclusion of induction therapy were offered alternative therapy as clinically appropriate and according to patient preference.

Clinical and Laboratory Monitoring

Before each cycle of therapy (defined as 1 week of 5 doses of etoposide), patients underwent a complete neurologic history and examination, measurement of hematologic and serum chemistry parameters, and a ventricular CSF cytology. In responding patients, lumbar CSF cytology was assessed monthly. CSF cytology results were reported by the cytopathologist as unsatisfactory, negative, abnormal, suspicious, or malignant. For purposes of analysis, cytology examination interpreted as suspicious were scored as positive and those interpreted as abnormal were scored as negative.6–8

Patients were considered responders if their CSF cytology converted from positive to negative at all sites previously shown to be positive and they remained neurologically stable or improved at the conclusion of induction.6–8 Patients were considered nonresponders if they had positive or suspicious CSF cytology at the end of the induction period (2 months) or if they demonstrated neurologic progression regardless of CSF cytology. A complete response was defined as conversion of positive to negative CSF cytology and resolution of all pretreatment neurologic deficits.4–8 A partial response was defined as conversion from positive to negative CSF cytology and either stable or improved pretreatment neurologic deficits. All other outcomes were considered progressive disease. Patients were considered evaluable for toxicity if they received any study drug as prescribed by the protocol and had at least 1 CSF examinations while receiving treatment.

Identification of episodes of drug-related arachnoiditis was based on an operational definition of whether within 3 days of drug injection patients developed neck rigidity, neck pain, meningismus, nausea, emesis, headache, fever, confusion, or a sterile CSF pleocytosis.6, 7 Arachnoiditis was graded on the basis of the highest grade of any of the constellation of adverse events characterized as mild (Grade 1), moderate (Grade 2), severe (Grade 3), or life-threatening (Grade 4). Fatigue was defined by the National Cancer Institute Common Toxicity criteria (version 2).

Statistical Analysis

The current study was a Phase II study to investigate the efficacy of intra-CSF etoposide in the treatment of NM. The primary objective of the statistical analysis was to estimate the 6-month neurologic disease PFS after study entry.24, 25 In that response to intra-CSF chemotherapy does not predict the duration of response, the response rate was not thought to be an acceptable endpoint for assessment of treatment efficacy. Therefore, the primary endpoint for statistical analysis is the 6-month neurologic disease PFS status such that a patient who was alive and free of neurologic disease progression at 6 months was considered to have a favorable outcome. Secondary objectives included response rate, median survival, median time to disease progression, and the rate of toxicity, especially those that result in a significant modification or cessation of intra-CSF etoposide. There was no stratification in this study. A Simon minimax 2-stage design was used: if the chance of remaining alive at 6 months was 15% or less (6-mo survival rates noted in previous studies of NM), then there would be little interest in further studying intra-CSF etoposide on this schedule. However, if there was a 40% chance or greater that a patient would remain alive at 6 months, then consideration would be given to studying intra-CSF etoposide further, provided safety and other endpoints were also favorable. With this design, in the first stage 16 patients were entered. If 2 or fewer patients survived to 6 months or more, the study would be stopped. If 3 or more patients demonstrated a survival of 6 months or more, then an additional 11 patients would be enrolled and treated. If 6 or fewer of the 27 patients (≤22%) were alive at 6 months, there would be little interest in studying this schedule of intra-CSF etoposide further. With this design, there was a 95% chance that we would recommend intra-CSF etoposide for further study if the true 6-month survival were 40% or greater (power) and there was a 10% chance of deciding to study intra-CSF etoposide further if the true chance of 6-month of survival was 15% or less (type 1 error). In patients who discontinued intra-CSF etoposide therapy because of toxicity, response was the best-recorded response through the cessation of intra-CSF etoposide. For all patients, overall survival was measured from the date of the initiation of etoposide therapy to death. Time to disease progression was measured from the date of initiation of intra-CSF etoposide therapy to the date of neurologic disease progression. The Kaplan-Meier method was used to estimate the distribution of overall survival and time to tumor progression. A log-rank test was used to assess the influence of baseline characteristics on the effect of treatment for overall survival and time to tumor progression. Relative risk and its 95% confidence interval (95% CI) were calculated. Kaplan-Meier plots were constructed to display the estimated probabilities of overall survival and time to progression. The Fisher exact test was used in comparing cause of death in patients responding to treatment (responders) and those not responding to treatment (nonresponders). Determination of cause of death was subjective and therefore subject to bias. Cause of death was defined clinically and utilized all available clinical data to estimate cause of death. One author (M.C.C.) made the determination with regard to cause of death. No prospective criteria were utilized in determining cause of death; rather, this was performed based on clinical analysis and judgment.


Patient Characteristics

Twenty-seven consecutive eligible patients (17 men and 12 women) were entered onto the study (Table 1). Age ranged from 26 years to 77 years with a median age of 55 years. All patients received drug via the intraventricular route through an Ommaya reservoir. Primary tumor histology was as follows: lung (8 patients [7 nonsmall cell and 1 small cell]), breast (5 patients), primary brain tumor (4 patients [2 medulloblastoma, 1 intracranial germ cell tumor, and 1 ependymoma]), non-Hodgkin lymphoma (4 patients), melanoma (4 patients), colon (1 patient), and prostate (1 patient). Twelve patients had bulky neuroradiographic disease and 19 patients received limited-field radiotherapy (before the initiation of intra-CSF etoposide) for either bulky or symptomatic disease. At the conclusion of limited-field radiotherapy, all 19 patients who underwent radiation had positive CSF cytology. Seventeen patients received concurrent tumor-specific systemic chemotherapy (Table 1). Before enrollment in the trial, no patient had been treated with either brain or spine radiotherapy. Patients received a total of 112 cycles of etoposide with a median of 4 cycles per patient (range, 1-12 cycles).


All patients were evaluable for toxicity; however, only 19 patients (70%) completed etoposide induction therapy (Fig. 1). Eight of the 27 patients (30%) clinically progressed during induction therapy and were not evaluable for response (progressive nonresponding patients). Seven of the 27 patients (26%) treated with etoposide attained a response (6 partial responses and 1 complete response) at the conclusion of the induction period and continued onto maintenance therapy. Twelve patients (44%) completed induction, but did not respond to therapy (i.e., persistently positive CSF cytology) and were offered alternative intra-CSF therapy (clinically nonprogressive nonresponding patients) as clinically indicated and according to patient preference. In responding patients, time to neurologic disease progression ranged from 8 weeks to 40 weeks, with a median of 20 weeks. The difference in response in lymphoma (1 of 4 patients; 25%) compared with solid tumor (6 of 23 patients; 26%) was not found to be statistically significant. Aside from histology, there were no apparent differences (age, pretreatment KPS, application of radiotherapy, systemic chemotherapy, or salvage intra-CSF chemotherapy) in patients who responded versus those who did not responde to etoposide. The overall median survival after the diagnosis of NM was 10 weeks (range, 4-52 wks), with 3-month, 6-month, and 12-month survival probabilities of 30% (95% CI, 12-47%), 11% (95% CI, 1-21%), and 4% (95% CI, 0-11%), respectively. Cause of death differed between nonresponders and responders. All 8 progressive nonresponders died of NM (100%), whereas of the 12 nonprogressive nonresponders 8 died of NM and 4 died of NM in association with progressive systemic disease. Among the 7 responders, 3 (43%) died of NM and 4 (57%) died of progressive systemic disease.

Figure 1.

Overall survival and disease progression-free survival (PFS) in 27 patients with neoplastic meningitis who were treated with etoposide.


Toxicity manifested as a transient arachnoiditis appearing within 1-5 days of treatment with etoposide. Arachnoiditis of any grade occurred in 4 patients (15%) and in 13.4% of all treatment cycles (15 of 112 cycles). However, only 2 patients (7%) had >Grade 2 arachnoiditis, noted in 9 of 112 (8%) of all treatment cycles. Arachnoiditis was easily managed with oral steroids, analgesics, antiemetics, and antipyretics. Arachnoiditis appeared after 2-3 days of treatment and resolved within 1-3 days after the last administration of etoposide in all patients. No patient required hospitalization or delay in treatment, nor was a treatment-related death recorded. Furthermore, no evidence of hematologic toxicity related to etoposide was observed. Chronic fatigue presumed to be treatment-related, although recognized as a symptom of NM, was common and was noted in 10 patients (37%). In no patient was fatigue greater than Grade 2 by NCI Common Toxicity criteria. Psychostimulant medication (amphetamines or modafinil) was inconsistently utilized for treatment-related chronic fatigue. Chronic fatigue appeared to be a cumulative toxicity of intra-CSF etoposide.


To our knowledge, only a limited selection of chemotherapeutic agents exists for the treatment of NM.6, 7, 9–12 Three agents are frequently administered (cytosine arabinoside, methotrexate, and thio-TEPA) based on a limited number of Phase III trials conducted in the treatment of patients with NM.4–7 Recently, a slow-release liposomal formulation of cytosine arabinoside, DepoCyt, has been released and is indicated for the treatment of NM.6, 7 Nonetheless, new agents for the treatment of NM would be welcome and are much needed.

Intra-CSF etoposide has been used for the treatment of NM in several prior studies.19–23 In 1992, a report of 2 patients treated with intra-CSF etoposide (0.5 mg per day for 5 consecutive days given every 3 weeks) first appeared and suggested both antitumoral activity and limited toxicity.20 In 2001, a pharmacokinetic study was undertaken in 14 patients (10 children and 4 adults) with intra-CSF etoposide administered at a dose of 0.5 mg per day for 5 consecutive days every 2-3 weeks.21 Pharmacokinetic data analysis in the CSF (11 courses, 4 patients) revealed a terminal half-life of 7.4 hours and an area under the curve (AUC) of 25.0 μg-h/mL. Comparison with 2 patients treated with systemic etoposide demonstrated CSF peak values more than 100-fold those achieved by intravenous infusion. Furthermore, the etoposide concentrations achieved by systemic infusion (standard dose regimen) were below 0.1 μg/mL, a concentration shown in vitro to be cytotoxic (etoposide concentrations of 0.1-10 μg/mL are cytotoxic in vitro and dependent on cell line and drug exposure time). In addition, etoposide concentrations achieved by intraventricular administration exceed by 2–10-fold those resulting from high-dose intravenous administration. In patients treated with intra-CSF etoposide, cytotoxic concentrations were observed at 24 hours after a single dose of 0.5 mg of etoposide. Four of 14 patients were evaluable for response, among whom all converted from positive to negative CSF cytology. The duration of response ranged from 3 weeks to 15 weeks. This study confirmed that the 5-day regimen was well tolerated, produced sustained cytotoxic CSF etoposide concentrations, and had limited and acceptable toxicity. The toxicity seen was confined to transient treatment-related arachnoiditis, a common side effect of intraventricular drug administration. Another small study in 11 children with NM, reported in 2003, demonstrated acceptable tolerability when etoposide was administered at a dose of 0.5 mg per day for 5 consecutive days every 2-4 weeks for a median of 8.5 months and a median cumulative dose 27 mg.23

In the current study, adult patients with newly diagnosed NM were treated with a fixed dose and schedule of intra-CSF etoposide based on the above-mentioned studies. Admittedly, no dose-response relations with intra-CSF etoposide have been established to date, and perhaps warrant further study, especially in view of the limited toxicity observed in both the current and past studies.

The results of the current study did demonstrate evidence of antitumoral activity both in solid tumors and lymphoma, with a 26% response to treatment noted after 8 weeks (4 cycles) of induction treatment. The 6-month neurologic disease PFS of 11%, the primary study endpoint, is comparable although slightly less robust to that reported in prior studies and suggests that intra-CSF etoposide may be another agent to consider in the treatment of patients with NM. The response to intra-CSF etoposide reflects both the use of involved-field radiotherapy to sites of symptomatic or bulky disease and the use of systemic chemotherapy in appropriate patients. However, the majority of prior studies of patients with NM have used a similar multimodal approach to the treatment of NM and the current study does not differ in that regard. Nonetheless, several studies suggest that NM may be managed as well by site-selective radiotherapy and concurrent tumor-specific chemotherapy without the use of intra-CSF chemotherapy. This issue regarding the role of intra-CSF chemotherapy in patients with NM warrants further study.

The results of the current study suggest that intra-CSF etoposide has minimal toxicity that, when observed, manifests as transient sterile chemical meningitis easily managed with oral steroids. The response to intraventricular etoposide suggests this agent may have a role in the treatment of NM; however, a confirmatory trial would be of value to validate the findings of this small study.