Chemotherapy dose-intensity and survival for childhood medulloblastoma.

AIM: To determine the relationship between prescribed dose-intensity of chemotherapy and survival in childhood medulloblastoma. MATERIALS AND METHODS: A total of 55 trials from 1970-2009 were identified, 30 were eligible for analysis, with individual treatment regimes with 5-year (or more) outcome figures. Relationships of outcome to dose-intensity were analysed using weighted regression. RESULTS: Overall, 2,434 patients were identified, 1,010 were classified as 'standard'- and 671 as 'high'-risk patients, with 5-year overall survivals (OS) of 67.2% (95% Confidence Interval=60.5%-73.6%) and 47.6% (95% Confidence Interval=39.5%-55.7%), respectively. A protective effect for chemotherapy versus craniospinal radiotherapy alone (5-year OS of 58.2% versus 51.6%) was found. Individually, vincristine, cisplatin, lomustine (CCNU) and cyclophosphamide appear to confer the most beneficial effect, particularly for high-risk patients. Positive relationships between OS and dose-intensity were found, except for lomustine, with cyclophosphamide offering the greatest protection. CONCLUSION: Consideration of chemotherapy dose-intensity may further optimise treatment, particularly in the context of risk stratification.

Biochemical changes and cytotoxicity associated with methionine depletion in paediatric central nervous system tumour cell lines.

BACKGROUND: The aim of this study was to investigate the importance of the extent and duration of methionine depletion as a cause of cytotoxicity for CNS tumour cell lines, and also to investigate the associated in vitro cellular biochemical responses. MATERIALS AND METHODS: Cell growth inhibition was assayed by the SRB assay. Intracellular methionine levels were measured by GC/MS following dervatization with MTBSTFA. After methionine depletion, methionine synthase and MGMT activities were also determined. Glutathione levels were assayed by HPLC after derivatization with OPA. RESULTS: Medulloblastoma (Daoy) and glioma (D54) cells were found to be methionine dependent and effects on proliferation, apoptosis and clonogenic survival were dependent on time and degree of methionine depletion. Methionine depletion also caused a demonstrable decrease in L-methionine levels and an increase in glutathione levels for both cell lines, with a decrease in MGMT activity for Daoy cells. CONCLUSION: Daoy and D54 cells are methionine dependent; the degree and duration of methionine depletion is related to cell death. The associated biochemical changes in MGMT and glutathione may be expected to modulate chemosensitivity and this will be investigated in future studies.

Methionine restriction reduces the chemosensitivity of central nervous system tumour cell lines.

BACKGROUND: The aim of this study was to investigate the effects of methionine depletion with cytotoxic agents that are potentially influenced by depletion of methionine, and are known to have a role in CNS tumour treatments for children. MATERIALS AND METHODS: Cytotoxicity studies and synergistic interactions were assayed by SRB assay. Glutathione levels were assayed by HPLC after derivatization with OPA. MGMT activity was determined by a restriction endonuclease inhibition assay. RESULTS: Methionine depletion causes a demonstrable increase in glutathione levels for medulloblastoma (Daoy) and glioma (D54) cells, with a decrease in MGMT activity for Daoy cells. For both cell lines, methionine depletion reduces their sensitivity to a range of chemotherapy agents that interface at the level of methionine metabolism, namely temozolomide, cisplatin and methotrexate. CONCLUSION: The results show that methionine depletion increases the resistance of tumour cells to the chemotherapeutic agents tested. However, in methionine-replete conditions, we have demonstrated synergistic activity for various combinations of chemotherapeutic agents that are hitherto unreported and may have clinical utility for the treatment of children with CNS tumours.

Preclinical efficacy of the bioreductive alkylating agent RH1 against paediatric tumours.

BACKGROUND: Despite substantial improvements in childhood cancer survival, drug resistance remains problematic for several paediatric tumour types. The urgent need to access novel agents to treat drug-resistant disease should be expedited by pre-clinical evaluation of paediatric tumour models during the early stages of drug development in adult cancer patients. METHODS/RESULTS: The novel cytotoxic RH1 (2,5-diaziridinyl-3-[hydroxymethyl]-6-methyl-1,4-benzoquinone) is activated by the obligate two-electron reductase DT-diaphorase (DTD, widely expressed in adult tumour cells) to a potent DNA interstrand cross-linker. In acute viability assays against neuroblastoma, osteosarcoma, and Ewing's sarcoma cell lines RH1 IC(50) values ranged from 1-200 nM and drug potency correlated both with DTD levels and drug-induced apoptosis. However, synergy between RH1 and cisplatin or doxorubicin was only seen in low DTD expressing cell lines. In clonogenic assays RH1 IC(50) values ranged from 1.5-7.5 nM and drug potency did not correlate with DTD level. In A673 Ewing's sarcoma and 791T osteosarcoma tumour xenografts in mice RH1 induced apoptosis 24 h after a single bolus injection (0.4 mg/kg) and daily dosing for 5 days delayed tumour growth relative to control. CONCLUSION: The demonstration of RH1 efficacy against paediatric tumour cell lines, which was performed concurrently with the adult Phase 1 Trial, suggests that this agent may have clinical usefulness in childhood cancer.

Diffuse brainstem gliomas in children: should we or shouldn't we biopsy?

The decision to biopsy diffuse pontine gliomas in children remains controversial. There have been many publications over the last 30 years aiming to address this issue. The prognosis for these patients remains extremely poor regardless of treatment and many authors advocate that biopsy carries significant risk for little or no clinical benefit. However, with an increasing knowledge of tumour biology and genetics there is the potential for specific treatments tailored for individual tumours based on their biological or genetic characteristics. The progress of such science in the first instance requires histological diagnosis as part of well conducted clinical trials, then, when treatments have been developed, biopsy samples will be needed to identify the tumours that may respond to such treatments. The authors believe that there is an increasing need for performing a biopsy of these lesions.

The cellular adaptations to hypoxia as novel therapeutic targets in childhood cancer.

Exposure of tumour cells to reduced levels of oxygen (hypoxia) is a common finding in adult tumours. Hypoxia induces a myriad of adaptive changes within tumour cells which result in increased anaerobic glycolysis, new blood vessel formation, genetic instability and a decreased responsiveness to both radio and chemotherapy. Hypoxia correlates with disease stage and outcome in adult epithelial tumours and increasingly it is becoming apparent that hypoxia is also important in paediatric tumours. Despite its adverse effects upon tumour response to treatment hypoxia offers several avenues for new drug development. Bioreductive agents already exist, which are preferentially activated in areas of hypoxia, and thus have less toxicity for normal tissue. Additionally the adaptive cellular response to hypoxia offers several novel targets, including vascular endothelial growth factor (VEGF), carbonic anhydrase, and the central regulator of the cellular response to hypoxia, hypoxia inducible factor-1 (HIF-1). Novel agents have emerged against all of these targets and are at various stages of clinical and pre-clinical development. Hypoxia offers an exciting opportunity for new drug development that can include paediatric tumours at an early stage.

Clinical and cellular pharmacology in relation to solid tumours of childhood.

Our understanding of the clinical and cellular pharmacology of drugs commonly used in the treatment of childhood cancer have increased greatly over the past two decades. However, with the exception of childhood acute lymphoblastic leukaemia (ALL), our current knowledge of factors such as inter-patient pharmacokinetic variability and cellular determinants of chemosensitivity has not been utilized in the design of integrated clinical studies. Recent pre-clinical and clinical evaluation of the topoisomerase I inhibitors topotecan and irinotecan has highlighted the potential importance of pharmacological factors in their effectiveness as cytotoxics. In this review, the clinical and cellular pharmacology of vincristine, actinomycin D, doxorubicin, cyclophosphamide, ifosfamide, cisplatin, carboplatin and etoposide will be discussed in relation to the major paediatric solid tumours. For each disease type, knowledge of the clinical and cellular pharmacology of a candidate drug will be related to pharmacodynamic responses such as response, toxicity and prognosis. For diseases such as Wilms' tumour, osteogenic sarcoma and Ewing's tumour, histological response to initial induction chemotherapy is of prognostic significance, and the depth of response is increasingly recognised as an important determinant of prognosis for high-risk neuroblastoma. For several of these tumour types, the dose-intensity of chemotherapy may be an important variable in determining prognosis. However the relationship between pharmacokinetic variability, cellular pharmacology and the major determinants of chemosensitivity to those drugs employed in first line therapy is unknown. The study of these relationships, by means of up front window studies in children who present with high-risk disease, may be as important as the introduction of new agents. Indeed, the optimisation of current therapies may be required to allow a fully informed selection of those children for whom novel therapies are truly needed. Funding and international collaboration at the clinical and scientific level would be required to achieve these aims.

Novel targets for therapy in paediatric oncology.

Although the majority of children with cancer are now cured of their disease, a significant number either have disease resistant to current therapy, or are unable to tolerate the short and long term complications of their treatment. Therefore new therapeutic strategies which optimise existing agents by use of their clinical and molecular pharmacology are needed, along with the development of new agents. Accordingly, the agents chosen for the study need to be prioritized, and are thus selected on the basis of categories such as encouraging pre-clinical data from xenografts of paediatric tumours, novel mechanisms of action, drugs that modify or overcome cellular resistance and drugs that are active in adult studies. In this review, novel targets for chemotherapy such as topoisomerase I, angiogenesis and signal transduction will be discussed. In addition, the circumvention of methotrexate resistance by novel antifolate thymidylate synthase inhibition, and the modulation of alkylating agents by inhibition of 0(6)-alkylDNA-alkyltransferase will be discussed as strategies to overcome potentially important resistance mechanisms in paediatric oncology. Finally, recent advances in biological therapies, tumour vaccination and gene therapy will be highlighted. In the future, investigation of cancer biology, selection of new drugs, and securing of funds to support the conduct of integrated early clinical studies that maximise the pharmacological, cellular biological and molecular pathological information gained, will be the major challenges to be faced by paediatric oncologists.

A phase I study in paediatric patients to evaluate the safety and pharmacokinetics of SPI-77, a liposome encapsulated formulation of cisplatin.

Pre-clinical studies indicate that cisplatin encapsulated in STEALTH((R))liposomes (SPI-77) retains anti-tumour activity, but has a much reduced toxicity, compared to native cisplatin. A phase I study was conducted to determine the toxicity and pharmacokinetics of SPI-77 administered to children with advanced cancer not amenable to other treatment. Paediatric patients were treated at doses ranging from 40 to 320 mg m(-2)by intravenous infusion every 4 weeks. Blood samples taken during, and up to 3 weeks after, administration and plasma and ultrafiltrate were prepared immediately. Urine was collected, when possible, for 3 days after administration. SPI-77 administration was well tolerated with the major toxicity being an infusion reaction which responded to modification of the initial infusion rate of SPI-77. Limited haematological toxicity and no nephrotoxicity were observed. No responses to treatment were seen during the course of this phase I study. Measurement of total plasma platinum showed that cisplatin was retained in the circulation with a half life of up to 134 h, with maximum plasma concentrations approximately 100-fold higher than those reported following comparable doses of cisplatin. Comparison of plasma and whole blood indicated that cisplatin was retained in the liposomes and there was no free platinum measurable in the ultrafiltrate. Urine recovery was less than 4% of the dose administered over 72 h. Results from this phase I study indicate that high doses of liposomal cisplatin can safely be given to patients, but further studies are required to address the issue of reformulation of liposomally bound cisplatin.

A phase I study of nolatrexed dihydrochloride in children with advanced cancer. A United Kingdom Children's Cancer Study Group Investigation.

A phase I study of nolatrexed, administered as a continuous 5 day intravenous infusion every 28 days, has been undertaken for children with advanced malignancy. 16 patients were treated at 3 dose levels; 420, 640 and 768 mg/m(2)24 h(-1). 8 patients were evaluable for toxicity. In the 6 patients treated at 768 mg/m(2)24 h(-1), dose-limiting oral mucositis and myelosuppression were observed. Plasma nolatrexed concentrations and systemic exposure, measured in 14 patients, were dose related, with mean AUC values of 36 mg(-1)ml(-1)min(-1), 50 mg ml(-1)min(-1)and 80 mg ml(-1)min(-1)at the 3 dose levels studied. Whereas no toxicity was encountered if the nolatrexed AUC was <45 mg ml(-1)min(-1), Grade 3 or 4 toxicity was observed with AUC values of >60 mg ml(-1)min(-1). Elevated plasma deoxyuridine levels, measured as a surrogate marker of thymidylate synthase inhibition, were seen at all of the dose levels studied. One patient with a spinal primitive neuroectodermal tumour had stable disease for 11 cycles of therapy, and in two patients with acute lymphoblastic leukaemia a short-lived 50% reduction in peripheral lymphoblast counts was observed. Nolatrexed can be safely administered to children with cancer, and there is evidence of therapeutic activity as well as antiproliferative toxicity. Phase II studies of nolatrexed in children at the maximum tolerated dose of 640 mg/m(2)24 h(-1)are warranted.

Remission induction therapy for childhood acute lymphoblastic leukaemia: clinical and cellular pharmacology of vincristine, corticosteroids, L-asparaginase and anthracyclines.

Remission induction therapy with vincristine, a corticosteroid, L-asparaginase and an anthracycline has been the mainstay of the initial phase of treatment for childhood acute lymphoblastic leukaemia (ALL) for the past 25 years. The speed and depth of the early response to remission induction therapy has become an important determinant of the intensity of subsequent therapy in many protocols worldwide. Moreover, the detection of significant levels of minimal residual disease at the end of remission induction may have an important bearing on subsequent outcome. Although these clinical observations may reflect, in part, the inherent sensitivity of lymphoblasts to remission induction therapy, the pharmacology of these agents in relation to childhood ALL may also play an important part in early response to therapy. In-vitro studies of human leukaemia cell lines indicate that both the extracellular fluid concentration and duration of exposure to vincristine and anthracyclines are important determinants of cytotoxicity. For L-asparaginase and corticosteroids, the cellular and molecular pharmacological determinants of chemosensitivity have been partially characterized, but further work is needed in this area. The clinical pharmacology of vincristine and L-asparaginase have been well characterized in relation to childhood ALL, and considerable interpatient pharmacokinetic variability exists for these drugs. For corticosteroids and anthracyclines, pharmacology studies are needed in order to fully characterize and understand the factors influencing interpatient pharmacokinetic variability for these agents in relation to childhood ALL. Whereas the relationship between the clinical pharmacology, and potentially important pharmacodynamic effects such as asparagine depletion, has been well characterized for therapy with L-asparaginase, similar studies have yet to be performed for the other drugs that form the mainstay of remission induction therapy for childhood ALL. Therefore, further studies are required to investigate the relative importance of the clinical and cellular pharmacology of vincristine, corticosteroids, L-asparaginase and anthracyclines in the speed and depth of response to remission induction therapy for childhood ALL. Where these have been studied, interindividual differences in the clinical and cellular pharmacology of anticancer agents have been shown to be important determinants of the long-term disease-free survival for children with ALL.

Consolidation therapy for childhood acute lymphoblastic leukaemia: clinical and cellular pharmacology of cytosine arabinoside, epipodophyllotoxins and cyclophosphamide.

The intensification of post-remission induction therapy has been shown to improve the relapse-free survival for childhood acute lymphoblastic leukaemia (ALL), and is now a standard component of the treatment of childhood acute lymphoblastic leukaemia. For cytosine arabinoside (ara-C), methotrexate, vincristine and corticosteroids, in-vitro studies indicate that the extracellular drug concentration and exposure time are important determinants of cytotoxicity for human leukaemia cell lines. For L-asparaginase, epipodopyllotoxins and cyclophosphamide, there have been few studies of the relationship between cellular pharmacology and cytotoxicity in relation to ALL. The clinical and cellular pharmacology of methotrexate and cytosine arabinoside have been studied in relation to childhood ALL in vivo. For these drugs, there is evidence to suggest that maintenance of plasma concentrations that are biochemically optimal is necessary to maximize anti-leukaemic effects. For cytosine arabinoside in particular, optimal extracellular fluid concentrations are not likely to be achieved or maintained by bolus or short-duration i.v. infusions. A potentially important example of this may be served by the success of antimetabolite-based intrathecal chemotherapy for CNS-directed treatment of childhood ALL. Intrathecal administration of both methotrexate and cytosine arabinoside results in prolonged leukaemic cell exposure to cytotoxic concentrations of the drug. For vincristine, anthracyclines and asparaginase, the actual dose intensity received by children during consolidation therapy may be important, and there is considerable interpatient variation in the pharmacokinetics of cyclophosphamide and teniposide in the therapy of childhood cancers. The importance of this relationship to childhood ALL is not known. The pharmacological and cellular pharmacological studies performed at St Jude Children's Research Hospital (Memphis, TN, USA) have allowed investigation of the relationships between the clinical and cellular pharmacology of methotrexate and prognosis, and have supported the individualization of consolidation therapy with this drug. Cytosine arabinoside has been less well studied in relation to childhood ALL, although evidence exists to suggest that the administration of conventional-dose bolus or infusion schedules may not be optimal in terms of the antileukaemic efficacy of this antimetabolite. For L-asparaginase, ongoing studies may allow the relationship between dose and schedule of administration to be related to pharmacodynamic measures such as asparagine depletion and prognosis. Therefore, through knowledge of clinical and cellular pharmacological properties, it may be possible to optimize the consolidation phase of therapy for childhood ALL, without disrupting the fundamental principles by which the overall treatment is administered. This may be particularly important for children with disease that has inherent or acquired resistance to therapy.

Continuing therapy for childhood acute lymphoblastic leukaemia: clinical and cellular pharmacology of methotrexate, 6-mercaptopurine and 6-thioguanine.

Across the world, therapy with 6-mercaptopurine (6-MP) and methotrexate (MTX) forms the basis of the continuing therapy of childhood acute lymphoblastic leukaemia (ALL). In this review, the pharmacological determinants of the sensitivity of human leukaemia cell lines and lymphoblasts derived from children with ALL will be discussed. In addition, clinical pharmacological studies of 6-MP and MTX in relation to the continuing therapy with childhood ALL will be reviewed. For 6-MP in vitro, prolonged exposure times to relatively high extracellular drug concentrations are necessary for cytotoxicity, and these concentrations are much higher than those achieved during continuing therapy for childhood ALL. For MTX, plasma concentrations are achieved during continuing therapy that would be cytotoxic to human leukaemia cells during prolonged exposures in vitro. For both MTX and 6-MP, wide inter- and intrapatient variation in plasma pharmacokinetic parameters has been described. For 6-MP and MTX, cellular pharmacological studies have been largely restricted to erythrocytes as a surrogate of the possible effects in leukaemic blasts. Although measures of the pharmacology of 6-MP and MTX in erythrocytes has been related to prognosis in many studies, 6-MP systemic exposure and the dose intensity of 6-MP and MTX actually received by children during this phase of therapy seems to be the most important determinant of efficacy. Further studies will be needed to determine the importance of pharmacokinetic variability during continuing therapy as a determinant of outcome for children with ALL. In this respect, minimal residual disease status during this phase of treatment may prove to be a useful pharmacodynamic endpoint.

Phase I trials in pediatric oncology: perceptions of pediatricians from the United Kingdom Children's Cancer Study Group and the Pediatric Oncology Group.

PURPOSE: To identify areas of concern regarding the conduct of phase I trials, the perceived expectations and motivations of the parents of children entered, the expectations of toxicity and benefit, and the ethical concerns of pediatric hematologists and oncologists in the United Kingdom and North America. METHODS: A survey instrument consisting of 19 open- and closed-ended questions was sent to United Kingdom Children's Cancer Study Group (UKCCSG)- and Pediatric Oncology Group (POG)-affiliated pediatricians. RESULTS: Fifty-three UKCCSG- and 78 POG-affiliated pediatricians responded. Thirty-two UKCCSG and 51 POG respondents had previously entered at least one child into a phase I study. Overall, respondents believed that parents entered their children for medical benefit, altruism, and hope of cure. Although many respondents believed that children could benefit from medical improvement, feelings of altruism, and maintenance of hope, the chance of cure or complete remission was thought to be small. Similarly, parents were thought to potentially benefit through altruism and maintenance of hope. Whereas 83% of UKCCSG respondents indicated that phase I trials were associated with ethical difficulties, this was a concern for 48% of POG respondents. The main ethical concerns of respondents were risk of toxicity, consent of the child, unrealistic hope, and coercion. CONCLUSION: The respondents in this survey expressed mainly ethical concerns regarding the conduct of phase I trials and had realistic expectations of the potential for toxicity and benefit for those children who participate in these studies.

Impact of polyglutamation on sensitivity to raltitrexed and methotrexate in relation to drug-induced inhibition of de novo thymidylate and purine biosynthesis in CCRF-CEM cell lines.

The aim of this study was to investigate the influence of folylpolyglutamyl synthetase (FPGS) activity on the cellular pharmacology of the classical antifolates raltitrexed and methotrexate (MTX) using two human leukemia cell lines, CCRF-CEM and CCRF-CEM:RC2Tomudex. Cell growth inhibition and drug-induced inhibition of de novo thymidylate and purine biosynthesis were used as measures of the cellular effects of the drugs. CCRF-CEM:RC2Tomudex cells had <11% of the FPGS activity of CCRF-CEM cells, whereas MTX uptake and TS activity were equivalent. In CCRF-CEM:RC2Tomudex cells, MTX polyglutamate formation was undetectable after exposure to 1 microM [3H]MTX for 24 h. After exposure to 0.1 microM raltitrexed, levels of total intracellular raltitrexed-derived material in CCRF-CEM:RC2Tomudex cells were 30- to 50-fold lower than in the CCRF-CEM cell line. CCRF-CEM: RC2Tomudex cells were >1000-fold resistant to raltitrexed and 6-fold resistant to lometrexol but sensitive to MTX and nolatrexed when exposed to these antifolates for 96 h. After 6 h of exposure, CCRF-CEM cells retained sensitivity to MTX and raltitrexed but were less sensitive to lometrexol-mediated growth inhibition. In contrast, CCRF-CEM: RC2Tomudex cells were markedly insensitive to raltitrexed, lometrexol, and to a lesser degree, MTX. Simultaneous measurement of de novo thymidylate and purine biosynthesis revealed 90% inhibition of TS activity by 100 nM MTX in both cell lines, whereas inhibition of de novo purine synthesis was only observed in CCRF-CEM cells, and only after exposure to 1000 nM MTX. Ten nM raltitrexed induced >90% inhibition of TS activity in CCRF-CEM cells, whereas in CCRF-CEM:RC2Tomudex cells, there was no evidence of inhibition after exposure to 1000 nM raltitrexed. These studies demonstrate that polyglutamation is a critical determinant of the cellular pharmacology of both raltitrexed and MTX, markedly influencing potency in the case of raltitrexed and locus of action in the case of MTX.

The role of pharmacokinetic and pharmacodynamic studies in the planning of protocols for the treatment of childhood cancer.

The chemosensitive nature of many childhood cancers means that chemotherapy has a greater role in therapy than in adult practice. However, the present methods, schedules of administration and combinations have often been derived form historical precedent rather than from pharmacological knowledge. For many drugs, paediatric phase I and II studies have never been performed and reliance on adult studies will be inadequate as children may show differences in drug disposition or susceptibility to toxicity. In this review, we examine pharmacokinetic and pharmacodynamic studies as they relate to the treatment of a 'model' childhood cancer in the UK: acute lymphoblastic leukaemia (ALL). Each of the drugs used is examined in the light of pharmacological evidence. For the drugs L-asparaginase, methotrexate, cytarabine and the thiopurines, this evidence suggests that the current use of these drugs is not optimal and that significant improvements in cure for ALL might be achieved by pharmacologically guiding their use. We highlight an important recent study demonstrating a 10% increase in long-term survival in childhood ALL by the use of pharmacologically guided dosing compared to standard (by body surface area) dosing. Since significant improvements in survival may depend upon such effective use, we suggest that pharmacological studies become an integral part of phase II and phase III trials of treatments for childhood cancer.

Phase I study of temozolomide in paediatric patients with advanced cancer. United Kingdom Children's Cancer Study Group.

A phase I study of temozolomide administered orally once a day, on 5 consecutive days, between 500 and 1200 mg m(-2) per 28-day cycle was performed. Children were stratified according to prior craniospinal irradiation or nitrosourea therapy. Sixteen of 20 patients who had not received prior craniospinal irradiation or nitrosourea therapy were evaluable. Myelosuppression was dose limiting, with Common Toxicity Criteria (CTC) grade 4 thrombocytopenia occurring in one of six patients receiving 1000 mg m(-2) per cycle, and two of four patients treated at 1200 mg m(-2) per cycle. Therefore, the maximum-tolerated dose (MTD) was 1000 mg m(-2) per cycle. The MTD was not defined for children with prior craniospinal irradiation because of poor recruitment. Plasma pharmacokinetic analyses showed temozolomide to be rapidly absorbed and eliminated, with linear increases in peak plasma concentrations and systemic exposure with increasing dose. Responses (CR and PR) were seen in two out of five patients with high-grade astrocytomas, and one patient had stable disease. One of ten patients with diffuse intrinsic brain stem glioma achieved a long-term partial response, and a further two patients had stable disease. Therefore, the dose recommended for phase II studies in patients who have not received prior craniospinal irradiation or nitrosoureas is 1000 mg m(-2) per cycle. Further evaluation in diffuse intrinsic brain stem gliomas and other high-grade astrocytomas is warranted.