This phase I trial studies the side effects and best dose of expanded natural killer cells in treating younger patients with brain tumors that have come back or do not respond to treatment. Infusing a particular type of a patient's own white blood cells called natural killer cells that have been through a procedure to expand (increase) their numbers may work in treating patients with recurrent/refractory brain tumors.

PRIMARY OBJECTIVES:

I. To establish the safety, feasibility, efficacy, and maximum tolerated dose (MTD) of administering autologous natural killer (NK) cells that have been propagated ex vivo with artificial antigen-presenting cells (aAPC) and administered directly into the ventricle in recurrent /refractory malignant posterior fossa tumors.

SECONDARY OBJECTIVES:

I. To assess the antitumor activity based on imaging and cytology of autologous NK cell locoregional administration directly into the lateral or fourth ventricle.

II. To determine the persistence of adoptively-transferred expanded NK cells (as performed with excess NK cells, via optional correlative studies).

III. Determine the immunophenotype and function of expanded NK cells. IV. Determine the overall response of medulloblastoma to NK-cell therapy. V. Correlate NK cell persistence, phenotype, and function with overall response.

OUTLINE: This is a dose-escalation study.

Patients receive autologous expanded NK cells intravenously (IV) into the ventricle over 3 minutes once weekly on weeks 1-3. Treatment repeats every 4 weeks for up to 3 courses in the absence of disease progression or unacceptable toxicity. Patients experiencing disease progression may continue treatment at the discretion of the treating physician if pseudo-progression or benefit of slowed progression is suspected.

After completion of study treatment, patients are followed up within 30 days.

Inclusion Criteria:

• Diagnosis: patients with recurrent/refractory medulloblastoma (MB), atypical teratoid (AT)/rhabdoid tumors (RT) or ependymoma involving the brain and/or spine at original diagnosis or relapse; they must have histological verification at diagnosis and/or relapse; patient must have presented with these tumors in the posterior fossa (PF) or relapsed in the PF
• Patient must have either measurable or evaluable tumor
• Presence of or determined by neurosurgery to be a candidate for an implanted catheter in the ventricles to receive NK cell infusion
• Life expectancy of at least 12 weeks in opinion of principal investigator (PI) and/or designee
• Lansky score of 50 or greater if =<16 years of age or a Karnofsky score of 50 or greater if > 16 years of age (NOTE: patients who are unable to walk because of paralysis, but who are in a wheelchair, will be considered ambulatory for the purpose of assessing the performance score)
• Neurologic deficits must have been relatively stable for a minimum of 1 week prior to study enrollment
• Patients must have recovered from the acute toxic effects of all prior anticancer chemotherapy
• Patient must be 4 weeks off any palliative radiation or craniospinal radiation
• Absolute neutrophil count (ANC) of >= 1000/uL
• Platelet count of >= 30,000
• Hemoglobin of >= 9.0 g/dl
• Patients with a seizure disorder may be enrolled if well-controlled and on non-enzyme inducing anticonvulsants
• Patient or patient's legal representative, parent(s), or guardian able to provide written informed consent

Exclusion Criteria:

• Enrolled in another treatment protocol
• Evidence of untreated infection
• Extra-cranial metastasis
• Chronic corticosteroid dependence (except replacement therapy)
• Extensive disease, disease location, and/or co-morbid condition that the PI or designee considers unsafe for surgical intervention of NK cell infusion
• Pregnant or lactating women