Lonafarnib and Temozolomide in Treating Patients with Glioblastoma Multiforme That is Recurrent or Did Not Respond to Previous Treatment with Temozolomide

Inclusion Criteria: * Patients with histologically proven supratentorial glioblastoma multiforme (GBM) or gliosarcoma * Patients must have shown unequivocal evidence for tumor recurrence or progression by magnetic resonance imaging (MRI) scan after radiation therapy; the scan done prior to study entry documenting progression will be reviewed by the treating physician to document tumor volume changes to provide a gross assessment of growth rate * Patients may have had as many as 2 prior chemotherapy regimens for recurrent or progressive tumor; patients must have had prior treatment with Temodar but may not have had prior treatment with farnesyl transferase inhibitors (Sarasar or Zarnestra); patients in phase 1b expansion are required to have received a minimum of two cycles of adjuvant temozolomide (TMZ) * All patients must sign an informed consent indicating that they are aware of the investigational nature of this study in keeping with the policies of this hospital * Patients must have shown unequivocal evidence for tumor progression by MRI or computed tomography (CT) scan; a scan should be performed within 14 days prior to registration and on a steroid dose that has been stable or decreasing for at least 5 days; if the steroid dose is increased between the date of imaging and registration a new baseline magnetic resonance (MR)/CT is required; the same type of scan, i.e., MRI or CT must be used throughout the period of protocol treatment for tumor measurement * Patients (pts) having had recent resection of recurrent or progressive tumor are eligible as long as: * Patients must be status post surgical resection at least 2 weeks prior to study enrollment, have recovered from surgery, have adequate early wound healing and a Karnofsky performance status of \> or = 60 * Residual disease following resection of recurrent tumor is not mandated for eligibility into the study; a CT/MRI should be done within 96 hours (hrs) post-op or at least 4 weeks (wks) post-op (within 14 days of registration); if the steroid dose is increased between the scan date and registration, a new baseline MRI/CT is required on a stable steroid dose for 5 days * Patients must have a Karnofsky performance status of \>= 60 * Patients must have recovered from the toxic effects of prior therapy: 4 weeks from prior cytotoxic therapy and/or at least two weeks from vincristine, 6 weeks from nitrosoureas, 3 weeks from procarbazine administration, and 1 week for non-cytotoxic agents, e.g., interferon, tamoxifen, thalidomide, cis-retinoic acid, etc. (radiosensitizer does not count); any questions related to the definition of non-cytotoxic agents should be directed to the study chair * Absolute neutrophil count (ANC) \>= 1,500/mm\^3 * Platelet count of \>= 100,000/mm\^3 * Serum glutamic pyruvate transaminase (SGPT) \< 2.5 times normal * Alkaline phosphatase \< 2.5 times normal * Bilirubin \< 1.5 mg * Blood urea nitrogen (BUN) \< 1.5 times institutional normal * Creatinine \< 1.5 times institutional normal Exclusion Criteria: * Patients must not be taking primidone, carbamazepine, phenobarbital or phenytoin anticonvulsants; patients changing from these anticonvulsants to other allowable drugs that are not enzyme inducing antiepileptic drugs (EIAEDs) must be off the drugs listed above for at least 72 hours prior the initiation of treatment * Patients with a history of any other cancer (except non-melanoma skin cancer or carcinoma in-situ of the cervix), are ineligible unless in complete remission and off of all therapy for that disease for a minimum of 3 years * Patients must not have: * Uncontrolled active infection * Disease that will obscure toxicity or dangerously alter drug metabolism * Serious intercurrent medical illness * Prior recurrence with a farnesyl transferase inhibitor * Oral contraceptives and other hormonal methods (Depo-Provera) of birth control