Eight drugs in one day chemotherapy in children with brain tumors: a critical toxicity appraisal.

The toxicity of a new chemotherapeutic regimen for CNS tumors using eight anticancer agents administered in a 12-hour period was evaluated in 34 children with newly diagnosed and recurrent tumors treated at the Children's Hospital and Medical Center in Seattle. The chemotherapy included methylprednisolone, vincristine, lomustine (CCNU), procarbazine, hydroxyurea, cisplatin (CDDP), and either cyclophosphamide or imidazole carboximide (DTIC). The first five patients additionally received high-dose methotrexate (HDMTX), but because of excessive toxicity this was replaced by cytarabine. Of 125 courses administered without HDMTX, red cell transfusions were required in 15% of the courses and platelet transfusions were required in 8%. Hospitalization for empiric treatment of fever associated with neutropenia occurred in 6% of courses. Three episodes of documented sepsis occurred. Virtually all hematologic toxicity occurred in patients with recurrent disease. Renal toxicity occurred in 14% of patient courses. One patient died of renal failure and sepsis following therapy. Neurologic and hepatic complications were infrequent. High-frequency hearing loss above the voice range was common, but clinically significant hearing loss occurred in only three patients. Severe nausea and vomiting were infrequent. Overall, the toxicity of "eight in one" chemotherapy in newly diagnosed patients was minimal; toxicity was greater in patients with recurrent disease.

Asymptomatic recurrence detection with surveillance scanning in children with medulloblastoma.

PURPOSE: To assess the utility of surveillance neuroimaging in detecting recurrent disease in patients treated for medulloblastoma. PATIENTS AND METHODS: Records and scans of 59 consecutive patients treated for medulloblastoma between 1984 and 1993 in one institution were retrospectively reviewed. RESULTS: Nineteen of 59 patients had recurrence of tumor, of which 17 were available for this study. Eleven of the 17 recurrent patients were asymptomatic at the time of detection. The median time to recurrence was 13 months (range, 3 to 90). CONCLUSION: Surveillance scanning detected a majority of recurrences before onset of symptoms. Although the outcome of those with recurrent disease remains poor, early detection with minimum disease provides the best setting in which to test newer therapies. Patients and their parents also were more likely to elect pursuing further treatment when relapse was detected asymptomatically.

Phase I trial of intrathecal liposomal cytarabine in children with neoplastic meningitis.

PURPOSE: We performed a phase I trial of intrathecal (IT) liposomal cytarabine (DepoCyt; Enzon Pharmaceuticals, Piscataway, NJ and SkyePharma Inc, San Diego, CA) to determine the maximum-tolerated dose, the dose-limiting toxicities, and the plasma and CSF pharmacokinetics of IT lipsomal cytarabine in children >/= 3 years of age with advanced meningeal malignancies. PATIENTS AND METHODS: Eighteen assessable patients received IT liposomal cytarabine through either an indwelling ventricular access device or via lumbar puncture. Liposomal cytarabine was given once every 2 weeks during induction, once every 4 weeks during consolidation, and once every 8 weeks during the maintenance phase of treatment. The initial dose was 25 mg, with subsequent escalations to 35 and 50 mg. CSF pharmacokinetic samples were obtained in a subset of patients. RESULTS: Arachnoiditis, characterized by fever, headache, nausea, vomiting, and back pain was noted in the first two patients at the 25 mg dose level. Therefore, subsequent patients were treated with dexamethasone, beginning the day of liposomal cytarabine administration and continuing for 5 days. Headache (grade 3) was dose limiting in two of eight patients enrolled at the 50 mg dose level. Eight of the 14 patients assessable for response demonstrated evidence of benefit manifest as prolonged disease stabilization or response. CONCLUSION: The maximum-tolerated dose and recommended phase II dose of liposomal cytarabine in patients between the ages of 3 and 21 years is 35 mg, administered with dexamethasone (0.15 mg/kg/dose, twice a day for 5 days). A phase II trial of IT liposomal cytarabine in children with CNS leukemia in second or higher relapse is in development.

Survival and prognostic factors following radiation and/or chemotherapy for primitive neuroectodermal tumors of the pineal region in infants and children: a report of the Childrens Cancer Group.

PURPOSE: To describe the biologic and clinical features of children with primitive neuroectodermal tumors (PNETs) arising in the pineal region (pineoblastomas) and evaluate prospectively the efficacy of radiation therapy (RT) and/or chemotherapy. PATIENTS AND METHODS: Between 1986 and 1992, 25 children with PNETs of the pineal region were treated as part of a Childrens Cancer Group study. Eight infants less than 18 months of age were nonrandomly treated with eight-drugs-in-1-day chemotherapy without RT. The remaining 17 patients were treated with craniospinal RT and randomized to receive either vincristine, lomustine (CCNU), and prednisone or the eight-drugs-in-1-day regimen. RESULTS: Of 24 completely staged patients, 20 (83%) had localized disease at diagnosis. All infants developed progressive disease a median of 4 months from the start of treatment. Of the 17 older patients treated with RT and chemotherapy, the Kaplan-Meier estimate of progression-free survival (PFS) at 3 years is 61% +/- 13%. This is superior to the PFS of children with other supratentorial PNETs (P = .026). Following RT, 12 of 17 patients (70.6%) had a residual pineal region mass, which persisted for as long as 5 years before resolving; only four subsequently developed progressive disease. CONCLUSION: (1) Eight-in-1 chemotherapy without RT appears to be ineffective therapy for young children with PNETs of the pineal region. (2) For children more than 18 months of age at diagnosis treated with craniospinal RT and chemotherapy, the PFS is superior to that of children with other supratentorial PNETs. (3) A residual enhancing mass following RT is not predictive of treatment failure.

Randomized phase III trial in childhood high-grade astrocytoma comparing vincristine, lomustine, and prednisone with the eight-drugs-in-1-day regimen. Childrens Cancer Group.

PURPOSE: In a previous randomized trial, the addition of adjuvant chemotherapy to postoperative radiotherapy proved beneficial in the treatment of childhood high-grade astrocytomas. The present study tests the hypothesis that an eight-drug adjuvant chemotherapy regimen would improve survival in such children compared with the three-drug regimen of the prior study. PATIENTS AND METHODS: Between April 1985 and May 1990, patients between the ages of 18 months and 21 years with newly diagnosed high-grade astrocytomas were eligible for this study, as determined by the treating institution's histopathologic diagnosis. Treatment consisted of postoperative local-field radiotherapy and adjuvant chemotherapy, either lomustine (CCNU), vincristine, and prednisone (control regimen) or eight-drugs-in-1-day chemotherapy (experimental regimen). Two cycles of postoperative preirradiation chemotherapy were administered in the experimental regimen. Patients were evaluated radiographically every 3 months after irradiation. RESULTS: Eighty-five eligible patients were randomized to the control regimen and 87 to the experimental regimen. The progression-free survival (PFS) and overall survival (OS) at 5 years were 33% (SE = 5%) and 36% (SE = 6%), respectively. There was no statistical difference in outcome between the two chemotherapy regimens. In patients with confirmed diagnoses of anaplastic astrocytoma (AA) or glioblastoma multiforme (GBM), anaplastic astrocytoma, greater than 90% resection, and nonmidline tumor location were characteristics predictive of an improved PFS. There was a difference in toxicity between the two chemotherapeutic regimens, with greater myelosuppression and hearing loss in the experimental regimen. Tumor recurrence occurred primarily within the primary tumor site. CONCLUSIONS: There is no benefit to the treatment of high-grade astrocytomas in children with eight-drugs-in-1-day chemotherapy compared with CCNU, vincristine, and prednisone. Extent of tumor resection and histopathologic diagnosis are significant prognostic variables. The overall outcome for children with high-grade astrocytomas remains poor.

Prognostic factors and treatment results for supratentorial primitive neuroectodermal tumors in children using radiation and chemotherapy: a Childrens Cancer Group randomized trial.

PURPOSE: To determine clinical characteristics and response to treatment for children with supratentorial primitive neuroectodermal tumors (S-PNETs). PATIENTS AND METHODS: After surgery and staging, 55 patients aged 1.5 to 19.3 years with S-PNETs were randomized to receive craniospinal radiotherapy (RT) followed by eight cycles of 1-(2-chloro-ethyl)-3-cyclohexylnitrosourea (CCNU), vincristine (VCR), and prednisone (standard treatment) or two cycles of 8-in-1 chemotherapy followed by RT and then eight additional cycles of 8-in-1. RESULTS: Three-year Kaplan-Meier estimates (estimate +/- SE) of survival and progression-free survival (PFS) rates for patients with confirmed diagnoses of S-PNET were 57% +/- 8% and 45% +/- 8%, respectively; survival and PFS rates for children with PNETs located in the pineal region were 73% +/- 12% and 61% +/- 13%, respectively, and were significantly different from the other S-PNETs (P < .03). The 8-in-1 arm had greater toxicity than the standard-treatment arm. Distributions of PFS between the two treatment groups were not significantly different (P > .5). Other univariate prognostic factors that influenced PFS included metastasis (M) stage (P < .03: M0 50% +/- 9% v M1-4 0%) and age (P < .02: 1.5 to 2 years 25% +/- 13% v > or = 3 years 53% +/- 9%). CONCLUSION: In this first randomized treatment trial for S-PNETs in children, no significant differences were detected between the two treatment groups. M0 and pineal site of involvement were independent predictors of a better outcome. However, survival was better than previously reported.

Survival of infants with primitive neuroectodermal tumors or malignant ependymomas of the CNS treated with eight drugs in 1 day: a report from the Childrens Cancer Group.

PURPOSE: Very young children with CNS primitive neuroectodermal tumors (PNETs) and ependymomas have a poor prognosis and commonly have impairment of growth and cognitive abilities, in part resulting from radiotherapy. Thus, an intensive chemotherapeutic regimen was used to treat children less than 18 months of age at diagnosis. PATIENTS AND METHODS: Children were treated on a Childrens Cancer Group (CCG) protocol with an eight-drug chemotherapeutic regimen (vincristine, carmustine, procarbazine, hydroxyurea, cisplatin, cytarabine, prednisone, and cyclophosphamide) following surgery and postoperative staging. Delayed or reduced-volume radiotherapy was to be administered to all patients, but, in fact, was omitted in most cases. RESULTS: On central review of pathology, 82 children had diagnosis concordant with study entry criteria. Of these, 46 (56%) had posterior fossa (PF) PNET, eight (10%) had pineal PNET, 11 (12%) had nonpineal supratentorial PNET, 15 (18%) had ependymoma, and two had rhabdoid tumors. Fifty percent of tumor resections were complete, as verified by postoperative computed tomographic (CT) scan, and 23% of patients had metastatic disease at the time of diagnosis. Objective tumor response was documented following two cycles of chemotherapy in 28% of assessable patients. Toxicity of chemotherapy was primarily hematopoietic. Five children died of chemotherapy-related complications. Radiotherapy was administered to only nine patients before tumor progression. The 3-year progression-free survival (PFS) rates for PF PNET, pineal PNET, supratentorial nonpineal PNET, and ependymoma are 22% (SE = 6%), 0%, 55% (16%), and 26% (11%), respectively. The 3-year PFS rate for those children without metastatic disease was 29% (6%), as compared with 11% (6%) for those with metastatic disease. The only independent predictors of PFS were metastasis stage and location of the tumor within the pineal region. The median time to progression was 6 months. Twenty-four children completed the chemotherapeutic regimen without tumor progression; 19 are event-free survivors more than 2 years from diagnosis, only three of whom received radiation therapy. CONCLUSION: While overall survival in this group of very young patients is poor, a subset of children who have received only chemotherapy as adjuvant treatment remain free from tumor recurrence.

Eight drugs in one day chemotherapy for brain tumors: experience in 107 children and rationale for preradiation chemotherapy.

The development of a new multidrug chemotherapy regimen for primary brain tumors was based upon the cellular heterogeneity within individual tumors, the Goldie-Coldman and Price-Goldie-Hill hypotheses, and known agonistic effects of certain drug combinations and sequences. Eight drugs (vincristine [VCR], hydroxyurea, procarbazine, CCNU, cisplatin, cytosine arabinoside [Ara-C] high-dose methylprednisolone, and either cyclophosphamide or dacarbazine) were administered within 12 hours in an attempt to minimize myelosuppression. Courses were repeated at 2- to 4-week intervals. The regimen was devised to include lipid and water soluble drugs, polar and nonpolar agents, phase-specific and cell-cycle independent agents, and antineoplastics with different mechanisms of action. More than 330 courses of the regimen were administered to 107 children with brain tumors whose tumor had recurred or had been incompletely resected at diagnosis. Tumor response according to protocol-specified criteria and independent review was evaluable in 78% of the patients. After just two cycles of chemotherapy and within a 4- to 6-week interval, 50% had an objective tumor response including 15.5% who had a complete response (CR). Nephrotoxicity and high-frequency hearing losses were noted after three to five courses of therapy in approximately half of the patients. Transfusions with red cells or platelets and use of antibiotics for fever and neutropenia were required in 10% to 25% of patients. The regimen appears satisfactory for preradiation chemotherapy in newly diagnosed patients with residual tumor after surgery, but it must be compared with standard therapeutic approaches in prospective controlled trials before its relative value can be established.

Metastasis stage, adjuvant treatment, and residual tumor are prognostic factors for medulloblastoma in children: conclusions from the Children's Cancer Group 921 randomized phase III study.

PURPOSE: From 1986 to 1992, "eight-drugs-in-one-day" (8-in-1) chemotherapy both before and after radiation therapy (XRT) (54 Gy tumor/36 Gy neuraxis) was compared with vincristine, lomustine (CCNU), and prednisone (VCP) after XRT in children with untreated, high-stage medulloblastoma (MB). PATIENTS AND METHODS: Two hundred three eligible patients with an institutional diagnosis of MB were stratified by local invasion and metastatic stage (Chang T/M) and randomized to therapy. Median time at risk from study entry was 7.0 years. RESULTS: Survival and progression-free survival (PFS) +/- SE at 7 years were 55%+/-5% and 54%+/-5%, respectively. VCP was superior to 8-in-1 chemotherapy, with 5-year PFS rates of 63%+/-5% versus 45%+/-5%, respectively (P = .006). Upon central neuropathology review, 188 patients were confirmed as having MB and were the subjects for analyses of prognostic factors. Children aged 1.5 to younger than 3 years had inferior 5-year estimates of PFS, compared with children 3 years old or older (P = .0014; 32%+/-10% v 58%+/-4%, respectively). For MB patients 3 years of age or older, the prognostic effect of tumor spread (MO v M1 v M2+) on PFS was powerful (P = .0006); 5-year PFS rates were 70%+/-5%, 57%+/-10%, and 40%+/-8%, respectively. PFS distributions at 5 years for patients with M0 tumors with less than 1.5 cm2 of residual tumor, versus > or = 1.5 cm2 of residual tumor by scan, were significantly different (P = .023; 78%+/-6% v 54%+/-11%, respectively). CONCLUSION: VCP plus XRT is a superior adjuvant combination compared with 8-in-1 chemotherapy plus XRT. For patients with M0 tumors, residual tumor bulk (not extent of resection) is a predictor for PFS. Patients with M0 tumors, > or = 3 years with < or = 1.5 cm2 residual tumor, had a 78%+/-6% 5-year PFS rate. Children younger than 3 years old who received a reduced XRT dosage had the lowest survival rate.

Pilot study of high-dose thiotepa and etoposide with autologous bone marrow rescue in children and young adults with recurrent CNS tumors. The Children's Cancer Group.

PURPOSE: This study was designed to determine the toxicity, radiographic response rate, and outcome following high-dose thiotepa, etoposide, and autologous bone marrow rescue (ABMR) for young patients with recurrent malignant brain tumors. METHODS: Eligibility criteria required adequate renal, hepatic, and pulmonary function, and no bone marrow infiltration. Thiotepa 300 mg/m2 and etoposide 500 mg/ m2 were infused on 3 consecutive days, and autologous bone marrow was infused 72 hours following chemotherapy. RESULTS: Forty-five patients with recurrent high-grade brain tumors, aged 8 months to 36 years (median, 8 years), were treated. Seven patients (16%) died of treatment-related toxicities within 56 days of marrow reinfusion. Delayed platelet engraftment occurred in 44% of patients who survived beyond day 56. Of 35 patients with radiographically measurable disease who survived at least 28 days following ABMR, there were two complete responses (CRs) and six partial responses (PRs), for an overall response (CRs plus PRs) rate of 23% (SE = 7%). Objective responses were observed in four of 14 assessable patients with high-grade glioma and in two of six with primitive neuroectodermal tumors (PNETs)/ medulloblastoma. Survival was significantly improved in patients treated with minimal residual disease (P < .0005). Five of 18 patients (28%) with high-grade gliomas remain free of disease at 39+, 44+, 49+, 52+, and 59+ months post-ABMR. CONCLUSION: The combination of high-dose thiotepa and etoposide has activity against a variety of recurrent childhood brain tumors. These results merit further evaluation in children and young adults with both recurrent and newly diagnosed high-grade brain tumors.

O6-Methylguanine-DNA methyltransferase in pediatric primary brain tumors: relation to patient and tumor characteristics.

The DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) confers resistance to methylating and chloroethylating agents in pediatric medulloblastoma- and glioma-derived cell lines and xenografts. Here, we assayed MGMT activity in 110 pediatric brain tumors to establish correlates with patient and tumor characteristics. We also assayed MGMT in histologically normal brain adjacent to 22 tumors to characterize changes in activity accompanying neurocarcinogenesis. MGMT activity was detected in 94% of tumors, ranging ca. 1,500-fold from 0.34 to 498 fmol/10(6) cells (approximately 205-300,000 molecules/cell). Mean activity was 25 +/- 66 fmol/10(6) cells, including six specimens with undetectable activity (Mer- phenotype; <0.25 fmol/10(6) cells or 151 molecules/cell). MGMT content varied 10-fold among diagnostic groups and was associated with degree of malignancy, as evidenced by a 4-fold difference in activity between high- and low-grade tumors (P = 0.03). Tumor MGMT content was age dependent, being 5-fold higher in children 3-12 years old than in infants (P = 0.015) and adolescents (P = 0.015). Mean activity in tumors was 9-fold higher than in adjacent histologically normal brain (21 +/- 44 versus 2.4 +/- 4.0 fmol/10(6) cells; P = 0.05). By comparing tumor and adjacent normal tissue from the same patient, we found that 68% of cases exhibited an elevation of tumor activity that ranged from 2- to >590-fold. Moreover, 67% of Mer- normal tissue was accompanied by Mer+ tumor. These observations indicate that MGMT activity is frequently elevated during pediatric neurocarcinogenesis. Significantly, enhanced MGMT activity may heighten resistance to alkylating agents, suggesting a potential role for MGMT inhibitors in therapy.

Radiation, methotrexate, and white matter necrosis: laboratory evidence for neural radioprotection with preirradiation methotrexate.

To investigate the interactions between methotrexate (MTX) and irradiation of the central nervous system, adult rats were infused with MTX via the lateral cerebral ventricle either before, during, or following single fraction irradiation to the cervical spine. Single doses ranging from 1600 cGy to 3200 cGy were administered and the dose-response curve for forelimb paralysis was compared with that seen in irradiated animals which did not receive MTX. There was no effect on the dose-response curve when MTX was administered simultaneously with or following irradiation compared to radiation alone. When MTX was given prior to irradiation, however, the entire dose-response curve shifted in the direction of radioprotection by approximately 225 cGy. Histopathologic examinations were consistent with this observation, with animals pretreated with methotrexate demonstrating significantly less white matter necrosis than observed in untreated controls. Protection of normal CNS tissue from radionecrosis, and from the associated paralysis, may be achieved with preradiation methotrexate.

Computed tomography in the evaluation of febrile neutropenic pediatric oncology patients.

AIM: To evaluate the diagnostic utility of computed tomography (CT) obtained during prolonged febrile neutropenia in pediatric oncology patients. METHODS: We evaluated the medical records of all patients with a malignant disease who had a CT examination during an episode of febrile neutropenia lasting for 4 days or more at Children's Hospital and Regional Medical Center in Seattle, WA, between January 1, 1997, and June 1, 1999. RESULTS: CT was performed on 83 patients to evaluate 109 episodes of prolonged febrile neutropenia. Sixty-eight (62%) of the initial CT scans demonstrated abnormalities, leading to changes in therapy in 42 (39%). The diagnostic and therapeutic utility of CT varied by anatomic site. Abdominal and head/neck CT detected abnormalities in only 19 and 8% of studies, respectively, resulting in therapy changes in 9 and 4%, respectively. Sinus CT demonstrated abnormalities in 41% of cases and altered therapy in 24%. Chest CT had the highest diagnostic utility, with 49% of cases demonstrating abnormalities, leading to therapy alteration in 30%. CT was rarely abnormal in the absence of localizing signs or symptoms. In 55 instances 1 or more follow-up scans were done. Thirteen follow-up CT scans showed abnormalities that led to a change in therapy. CONCLUSIONS: CT-detected abnormalities frequently lead to alterations in therapy, particularly sinus and thoracic CT. Most patients with CT-detected abnormalities have symptoms or signs referable to the site of abnormality. Asymptomatic febrile neutropenic children rarely have CT findings that lead to a change in therapy.

Hematopoietic marrow regeneration in pediatric patients undergoing spinal irradiation: MR depiction.

PURPOSE: To evaluate changes in vertebral marrow signal intensity that occur over time in children undergoing craniospinal radiation therapy, specifically evaluating for the occurrence and timing of marrow regeneration. METHODS: MR images of nine pediatric patients (ages 4 to 12 years) with posterior fossa medulloblastoma who received total spinal irradiation (24 to 40 Gy) and had at least three MR examinations were reviewed. Signal intensity to vertebral body marrow was graded by two pediatric neuroradiologists who were blinded to patient identity and to the timing of the studies. RESULTS: Eight of nine patients demonstrated increasing signal intensity of the vertebral marrow after irradiation, consistent with conversion of hematopoietic to fatty marrow. In each of these patients, this was followed by subsequent decreasing signal intensity in a mottled or peripheral band pattern indicating recovery of hematopoietic marrow. CONCLUSION: Changes in vertebral body signal intensity consistent with marrow reconversion commonly are seen in pediatric patients 11 to 30 months after they undergo total spinal irradiation.

Spinal subdural enhancement after suboccipital craniectomy.

PURPOSE: To characterize transient intraspinal subdural enhancement (potentially mimicking the subarachnoid spread of tumor) seen on MR images in some children after suboccipital craniectomy for posterior fossa tumor resection. METHODS: Radiologic and medical records of 10 consecutive children who had MR imaging for spinal staging after resection of posterior fossa tumor during a 9-month period were reviewed retrospectively. In addition, one case with similar findings of intraspinal enhancement on spinal staging MR images obtained at another institution was included in the review. RESULTS: Intraspinal enhancement thought to be subdural was seen in four of 10 patients undergoing spinal staging MR imaging 6 to 12 days after surgery. In these four patients, MR studies 50 to 18 days later, without intervening treatment, showed resolution of the abnormal enhancement. A fifth patient (from another institution) with similar intraspinal enhancement underwent CT myelography 4 days later, which showed no subarachnoid lesions. No metastases have developed in any of these five patients during the 2.5- to 3.5-year follow-up period. conclusions: From analysis of the MR appearance and on the basis of prior myelographic experience, we suggest an extraarachnoid, probably subdural, location of this enhancement. Awareness of this phenomenon will reduce the rate of false-positive diagnoses of metastatic disease. Preoperative spinal staging should be considered for patients undergoing suboccipital craniectomy.

Neonatal brain tumors: a review.

This article presents two cases of infants with brain tumors and reviews the literature pertinent to congenital and neonatal brain tumors. Information regarding epidemiology, presentation, prognosis, and clinical management are also addressed with specific regard to differences between neonatal and childhood brain tumors. An appeal is made to consider (1) coordination of the care of these children through pediatric multidisciplinary neuro-oncology programs; (2) enrollment, whenever possible, in clinical trials; and (3) submission of available tumor tissue to pediatric tumor banks to assure its availability to interested researchers.

Prognostic significance of type 1 neurofibromatosis (von Recklinghausen Disease) in childhood optic glioma.

Although the association between optic glioma and neurofibromatosis is well recognized, few studies have systematically compared the outcomes of patients with optic gliomas and neurofibromatosis and patients with optic gliomas without neurofibromatosis. In the present study, patients with optic gliomas and Type 1 neurofibromatosis (NF-1) were compared with patients with optic gliomas without NF-1, with respect to survival, time to tumor progression, and tumor location. Forty-four patients with optic gliomas who were evaluated between 1949 and 1991 were studied retrospectively. Sixteen of 44 patients (36%) met the National Institutes of Health criteria for NF-1. The medical records of all patients were examined, and letters of inquiry were sent to every living patient to ascertain current health statuses. Death certificates were obtained to determine causes of death. Follow-up averaged 7.2 years (10.2 yr for patients with NF-1, 5.4 yr for patients without NF-1). The 5- and 10-year survival rates for patients with optic gliomas and NF-1 were 93 and 81%, respectively. For those patients with optic gliomas who did not have NF-1, 5- and 10-year survival rates were 83 and 76%, respectively. Seventeen patients experienced tumor progression (5 with NF-1, 12 without NF-1). A difference was observed in the mean time to tumor progression (first relapse) between the two groups (mean time with NF-1, 8.37 yr; without NF-1, 2.39 yr [P < 0.01]). However, no significant difference in overall survival, as evaluated by a log-rank test of the respective Kaplan-Meier survival curves, was observed between the two groups. A significant difference in distribution of tumor location between the group with NF-1 and the group without NF-1 was also noted (Fisher's exact test, P = 0.0338), although the number of patients evaluated in this series was too small to determine whether this difference in tumor location influenced relapse rate. We conclude that optic gliomas in patients with neurofibromatosis have a different distribution of location as opposed to those in patients without neurofibromatosis, and, for first relapse, the presence of neurofibromatosis is a significant favorable factor.

The risks of metastases from shunting in children with primary central nervous system tumors.

The authors reviewed the hospital charts of 415 pediatric patients treated for benign or malignant primary brain tumors over the past 20 years at the Children's Hospital Medical Center, Seattle. Patients' ages ranged from the neonatal period to 18 years. A shunt was placed in 152 patients (37%), 45 before and 94 after surgery. Confirmation of extraneural metastases was based on clinical and diagnostic examination. Factors analyzed as possibly influencing the occurrence of extraneural metastases were: 1) the shunt: type, valve, location, filter, and revisions; 2) extent of resection; 3) pathology; and 4) treatment regimen. Eight of the 415 patients developed extraneural metastases during life. All eight patients had a medulloblastoma (cerebellar primitive neuroectodermal tumor). These eight patients were separated into Group A (without a shunt) and Group B (with a shunt). In Group A (five patients), the mean interval from primary diagnosis to metastasis was 15 months. Two children had gross total resection of the tumor. The predominant location of metastases in Group A was: bone (two cases); cervical lymph nodes (one); lung/bone (one); and retroperitoneal pelvic mass (one). Three Group A patients had a simultaneous central nervous system (CNS) recurrence. Of the three Group B patients, two had a ventriculoperitoneal (VP) shunt and one a ventriculoatrial (VA) shunt; all were placed postoperatively. One Group B patient had a simultaneous CNS recurrence. No shunt revisions were performed in these three patients. The mean time from primary diagnosis to metastasis was 25 months. One patient had a total tumor resection. The predominant location of metastases was bone (one case), retroperitoneal pelvic mass (one), and abdominal cavity with ascites (one case). Only one patient in the entire series had a filter placed; this resulted in shunt obstruction and was removed 1 month following placement. It is concluded that cerebrospinal fluid shunts, regardless of type, location, revision rate, or filter insertion, do not predispose pediatric patients with brain tumors to develop extraneural metastases. A diagnosis of shunt-related metastases should be based on the development of intra-abdominal (VP shunt) or pulmonary (VA shunt) dissemination primarily with or without additional sites. The diagnosis of medulloblastoma is an important factor related to metastasis occurrence while the extent of resection and postoperative therapy are not influential.

Survival and prognostic factors following radiation therapy and chemotherapy for ependymomas in children: a report of the Children's Cancer Group.

OBJECT: Ependymomas in children continue to generate controversy regarding their histological diagnosis and grading. optimal management, and possible prognostic factors. To increase our knowledge of these tumors the authors addressed these issues in a cohort of children with prospectively staged ependymomas treated with radiotherapy and chemotherapy. METHODS: Children between the ages of 2 and 17.3 years harboring an intracranial ependymoma confirmed by a central review of the tumor's pathological characteristics were treated according to Children's Cancer Group Protocol 921 from 1986 to 1992. Treatment following surgery and postoperative tumor staging (including brain computerized tomography or magnetic resonance [MR] imaging, spinal MR imaging or myelography, and cerebrospinal fluid cytological investigation) included craniospinal irradiation with a local boost to the primary tumor and patient randomization to receive adjuvant chemotherapy with either 1) CCNU, vincristine, and prednisone, or 2) the eight-drugs-in-1-day regimen. Centralized review of the tumor pathological characteristics revealed 20 ependymomas and 12 anaplastic ependymomas in the 32 children included in the study. Diagnoses made at the individual institutions included anaplastic (malignant) ependymoma (15 patients), ependymoma (four patients), ependymoblastoma (nine patients), ependymoastrocytoma (one patient), and primitive neuroectodermal tumor (three patients), which were discordant with the centralized review diagnosis in 22 of 32 cases. Only three of the 32 patients had metastatic disease (two with M and one with M3 stages). At surgery, 47% of tumors were estimated to be totally resected. Among the 14 of 17 patients who suffered a relapse and were evaluated for site of relapse, 10 (71%) had an isolated local relapse, three (21%) had concurrent local and metastatic relapse, and only one (7%) had an isolated metastatic relapse. Kaplan-Meier estimates of 5-year progression-free survival (PFS) and overall survival rates were 50 +/- 10% and 64 +/- 9%, respectively. CONCLUSIONS: Predictors of PFS duration included an estimate of the extent of resection made at surgery (total compared with less than total, p = 0.0001) and the amount of residual tumor on postoperative imaging as verified by centralized radiological review (< or = 1.5 cm2 compared with > 1.5 cm2, p < 0.0001). No other factors, including centrally reviewed tumor histopathological type, location, metastasis and tumor (M and T) stages, patient age, race, gender, or chemotherapy treatment regimen significantly correlated with PFS duration. The pattern of predominantly local relapse and the important influence of residual tumor or the extent of resection on PFS duration confirms a prevailing impression that local disease control is the major factor in the prediction of outcome of ependymoma. Survival rates were comparable with those reported by other investigators who have treated patients with similar doses of radiation and no chemotherapy.

Infant brain tumors.

This article describes the histology and location of brain tumors in young children as well as the presenting features of tumors in this age group and then focuses on three tumor types: medulloblastoma, ependymoma, and chiasmatic optic glioma. The discussion proceeds in terms of prognosis, late effects of treatment, and current and future strategies for treatment.