The RNAse III endonuclease DICER is a key regulator of microRNA (miRNA) biogenesis and is frequently down-regulated in a variety of malignancies. We characterized the role of Dicer in glioblastoma (GB), specifically demonstrating its effects on the ability of glioma stem-like cells to form tumors in a mouse model of GB. DICER silencing in glioma stem-like cells (GSCs) reduced their stem cell characteristics, while tumors arising from these cells were more aggressive, larger in volume, and displayed a higher proliferation index and lineage differentiation. The resulting tumors, however, were more sensitive to radiation treatment. Our results demonstrate that DICER affects the tumorigenic potential of GSCs, providing a platform for analysis of specific relevant miRNAs and development of potentially novel therapies against GB.

Ionizing radiation (IR) is one of the conventional post-surgical treatments for Glioblastoma Multiforme (GBM). Mesenchymal stem cells (MSCs) constitute a subpopulation of bone marrow derived cells which are actively recruited to the site of radiation and/or tumour microenvironment (TME), both of which have important implications for neovascularization and tumor progression. The goal of this project is to investigate the functional contribution of MSCs in the TME. We postulate that Bone Marrow-MSCs promote radio-resistance in GBM via cell cycle arrest. We tested the effect of MSC on U87 glioblastoma cell line in response to IR. We found that MSC co-culture, MSC-conditioned media (MSCCM) and irradiated MSC-conditioned media (MSCIRCM) did not reduce IR-induced p53 (ser15) phosphorylation, signifying intact p53-dependent DNA damage pathway in all conditions. However, both MSCCM and MSCIRCM temporally increased phospho-Chk2, a kinase involved in ATM-dependent cascade and cell cycle arrest. This increase occurred at 24 hours and reverted to baseline levels by 48 hours. Interestingly, IR (15Gy) caused transiently heightened metabolic rate under MSC and MSC IRCM as opposed to IR-null treatment at 48 hours elevated cell proliferation. MSCCM, but not MSCIRCM, marginally reduced caspase 3/7-dependent apoptotic levels. The combination of IR and MSCCM as well as MSCIRCM first increased protein level of phospho-Chk2 at 24 hours; followed by increased metabolic rate at 48 hours; and lastly, boosted proliferation at 72 hours. This data combined proposes plausible machinery for BM-MSC mediated radio-resistance by initiating cell cycle arrest in tumour cells for DNA damage repair.

Background: The incidence of brain metastases is increasing with the development of improved systemic therapies with limited impact on intracranial disease. The purpose of this study was to determine if there is a threshold tumor size below which local control (LC) rates approach 100% after stereotactic radiosurgery (SRS). Methods: 200 patients with 1237 tumors were identified from a prospective registry of patients having undergone SRS between 2012-2014. Histology consisted predominantly of non-small cell lung cancer (NSCLC), melanoma and breast cancer. Results: The median tumor size was 6 mm in diameter or 70 mm3 and most commonly NSCLC. Thirty-three tumors had local progression at a median time of 8.8 months. The 1- and 2-year actuarial LC for all tumors were 97% and 93%. LC of 100% was seen for intracranial metastases less than 100 mm3 or 6 mm in diameter, independent of histology. Total tumor volume was an independent predictor of overall survival, after adjusting for age, KPS and extracranial disease status. Conclusions: SRS can achieve LC rates approaching 100% for subcentimeter metastases. The earlier detection and prompt treatment of small intracranial metastases may prevent the development of neurological symptoms, the need for surgical resection, and potentially improve overall survival. The results of this study would favour the implementation of routine staging MRIs.

Background: Delayed facial palsy (DFP) after resection of vestibular schwannomas (VS) is described as worsening of facial nerve function after a normal postoperative result. Several mechanisms have been postulated to explain this phenomenon, although none satisfactorily explain all of its features. Furthermore, systematic documentation of recovery rates is lacking. Methods: 403 consecutive cases of VS resection between 2001 and 2015 were reviewed. Patients with preoperative facial palsy were excluded. Patients developing significant facial palsy (HB grade ≥ 3) were categorized into groups based on timing of onset: immediate facial palsy (IFP), “early-onset” DFP (within 48 h), and “late-onset” DFP (after 48 h). IFP patients were subdivided into “minor” (HB grade 3) and “major” (HB grade≥4) groups. These groups were compared with respect to demographics, intraoperative data, and recovery. Results: The late-onset DFP group demonstrated the quickest recovery to HB≤2 (2.9 weeks), followed by the minor IFP group (8.5 weeks), then the early-onset DFP group (53 weeks). Major IFP group exhibited the poorest recovery with only 32% recovering to HB grade≤2 within one year. Conclusions: The bimodal distribution in recovery time in delayed facial palsy patients implies separate underlying phenomena. We propose that a delayed demyelination of the facial nerve occurs in late-onset DFP, and best explains the uniformly rapid recovery observed.

As the central component of heightened vascularization in glioblastoma (GBM), endothelial cells (EC) are arguably the most responsive stromal cells in the tumour microenvironment during conventional treatment using ionizing radiation (IR). Although the inherent tumor tropism and angiogenic property of mesenchymal stem cells (MSC) has been documented in many cancer types including GBM, little is known about the function of MSC that are recruited to the GBM tumor microenvironment. The purpose of this study was to elucidate the effect of MSC on irradiated EC. Here we studied the influence of IR, MSC or MSC condition media (CM) on human umbilical vein endothelial cells (HUVECs). IR was found to up-regulate mRNA levels of CXCL5, CXCL10, ICAM1, VCAM1, VEGF-c and tissue factor (TF) in a dose-dependent manner whereas MSC co-culture boosted the expression of ICAM1, VCAM1, VEGF-c, TF, and MCP3. An additive effect of IR and MSC-culture was most detected with respect to ICAM1, TF and CXCL5 under 2Gy. MSC co-culture decreased IR-induced phospho-p53 in HUVECs at 15Gy. In addition, IR decreased the level of phospho-Akt in HUVECs as well as cell proliferation. Notably, both effects were countered by MSC CM. Interestingly, up-regulation of the same set of IR-driven genes in GBM was positively correlated with poor survival in the TCGA GBM database, a correlation that was lost if using gene list that was obtained from IR and MSC combine. These findings suggest that MSC promotes angiogenesis in GBM by overturning the IR-induced active state of endothelial cells and rendering them radio-resistant.

Background: The Flow diversion in the treatment of Intracranial Aneurysm (FIAT) trial was designed to guide the clinical use of flow diversion. Methods: FIAT proposed randomized allocation flow diversion or standard management (observation, coiling, parent vessel occlusion, or clipping), and a registry of non-randomized patients treated with flow diversion. Primary safety outcome was death or dependency (mRS > 2) at 3 months. Primary efficacy outcome was angiographic occlusion at 3-12 months combined with independent clinical outcome. Results: Of 112 participating patients recruited, 78 were randomized, and 34 received flow diversion within the registry. The study was halted for safety concerns. Twelve of 73 patients (16.4%; CI [9.7% -26.7%]) who were allocated or received flow diversion at any time were dead (n=8) or dependent (n=4) at 3 months or more, crossing a predefined safety boundary. Death or dependency occurred in 5 of 36 patients randomly allocated flow diversion and in 5 of 36 patients allocated standard treatment (13.9%; [6.1%-28.7%]). Efficacy was below hypothesized expectations: 15 of 36 patients (41.7%; [27.1%-57.8%]) randomly allocated flow diversion failed to reach the primary outcome, as compared to 11 of 36 patients allocated standard treatment (30.1%; [18.0%-46.9%]). Conclusions: Flow diversion was not as safe and effective as hypothesized. More randomized trials are needed.

We identified the G protein coupled relaxin receptor RXFP1 in glioblastoma (GB) and discovered C1q-tumor necrosis factor related protein 8 (CTRP8) as a novel ligand of RXFP1 in this common and most aggressive form of brain tumor. The known RXFP1 ligand relaxin isoform 2 (RLN2) is not expressed in the brain indicating that RXFP1 signaling in GB is activated by auto-/paracrine secretion of CTRP8. In GB, RXFP1 signaling included PI3K and PKC activation and resulted in the increased production and secretion of lysosomal protease cathepsin B, a known prognostic marker of GB. In the present study, we have investigated the potential role of RXFP1 in chemoresistance to the commonly used DNA alkylating drug temozolomide (TMZ) in human GB. Upon TMZ treatment, CTRP8/RLN2 mediated activation of RXFP1 was able to mitigate DNA damage in human primary GB cells and enhanced their survival. Activation of RXFP1 resulted in STAT3 pathway activation and the RXFP1- and STAT3-dependent up-regulation of proteins and their activity deemed important for TMZ-induced DNA damage repair. Furthermore, RXFP1 activation resulted in the up-regulation of key anti-apoptotic factors in human GB cells. Our results indicate a novel role for the CTRP8-RXFP1 ligand-receptor system in STAT3-dependent cell invasion, TMZ chemoresistance, and survival and identify RXFP1 as a new protective G protein coupled receptor in GB cells.

The most common brain cancer, glioblastoma (GBM), also has the worst five-year survival rates of all cancers. This is due in part to cancer stem cells, known as brain tumor initiating cells (BTICs), which initiate GBM and are resistant to current therapies. Thus, it is important to study therapeutics targeting BTICs. Innate immune cells, microglia and macrophages (MMs), are also important GBM-associated cells. GBM influences MMs to promote tumor growth. However, my research group discovered that an old drug, amphotericin B, re-stimulates MMs to secrete cytokines that inhibit BTIC self-renewal (Nature Neurosci 17:46-55, 2014). Hence, I hypothesize these cytokines are potential GBM treatments. Methods: Twenty genetically diverse BTIC lines were exposed to interleukin-8 (IL8), monocyte chemoattractant protein-1 (MCP1), and tumor necrosis factor-alpha (TNF), the cytokines secreted most abundantly by amphotericin-stimulated MMs. Assays for proliferation, differentiation, apoptosis, and cell cycle arrest were performed. BTICs were then co-cultured with human GBM-derived MMs stimulated with TNF. Lastly, the relationship between TNF receptors (TNFR1/2) and presumed BTICs was determined in frozen human GBM sections by immunofluorescence. Results: TNF had the most potent inhibitory effect on BTIC proliferation. Also, TNF was able to increase BTIC differentiation, and induce apoptosis and G1 cell cycle arrest. In MM-BTIC co-culture, TNF stimulated MMs to decrease sphere formation. Interestingly, in human GBM tissue TNFR1 co-labeled with OLIG2, a major transcription factor expressed in presumed BTICs, implicating TNF as a BTIC-specific treatment. Conclusion: TNF is a promising candidate for the treatment of GBM.

Background: Brain arteriovenous malformations (AVM’s) are abnormal connections between arteries and veins. Endovascular glue embolization with N-butyl cyanoacrylate (NBCA) is an accepted form of treatment, with most complications related to timing of polymerization. Current literature reports a wide range of polymerization times with large discrepancies between in-vivo and in-vitro results. Methods: Polymerization time was measured for mixtures of lipiodol/NBCA of 50/50, 60/40, 70/30. The influence of pH, temperature and presence of biological catalysts on polymerization rate was investigated in-vivo using submerged droplet tests. PVA-C, silicone and endothelium surfaces were compared and contact angles were measured to assess physical interaction with NBCA. High-speed video of glue injection through a microcatheter was captured to characterize coaxial flow. Results: Polymerization rate increases with pH and temperature. A hydrophilic substrate such as PVA-C provides surface properties that are most similar to endothelium. Endothelium provides a catalytic surface that increases the rate of polymerization. Blood products further increase the polymerization rate with RBC’s providing almost instantaneous polymerization of NBCA upon contact. Characterization of coaxial flow shows dripping to jetting transition with significant wall effect. Conclusions: We have successfully deconstructed and characterized the dynamic behavior of NBCA embolization. A refined understanding of NBCA behavior could help reduce embolization-related complications.

Background: Neurosurgery patients are at higher risk of falls given the morbidity associated with their neurological disease. We present our department’s experience with in-patient falls. Methods: We analyzed our hospital’s database for Neurosurgery in-patient falls from January 1st till December 31st, 2015. Results: Of 1,317 patients admitted under Neurosurgery, 5% (n=63) had in-patient falls. CT head was done in 24% (n=15) of patients who had a fall and 93% (n=14) of the CT head post-fall was reported as no significant interval change. The combined cost of repeat CT imaging reporting no interval changes was approximately $ 7,000. One CT head post-fall showed worsening midline shift but did not impact management. One of the 78% (n=48) post-fall patients who did not get a CT head progressed to coma requiring emergent surgery and another patient suffered an isolated hip fracture requiring operation. 41% (n=26) of falls were from bed and 37% (n=22) were while ambulating. Leading diagnosis of in-patient falls was subdural hematoma (33%, n=21) and tumour (32%, n=20). Conclusions: Identification of risk factors for in-patient falls can reduce hospitalization costs. The highest number of in-patient falls occurs in patients with subdural hematoma and are likely to occur from a patient’s bed.

Tumours of the central nervous system are currently classified based on the 2007 WHO Classification of Tumours, which uses histological features to classify and grade these heterogeneous tumors. With recent advances in the development of clinically relevant molecular signatures, there is an interest to incorporate appropriate molecular markers in to the classification. The views of the neuro-oncology community on such changes would be informative for advising this process. METHODS A survey with 8 questions regarding molecular markers in tumor classification was sent to an email list of Society of Neuro-Oncology members and attendees of prior meetings (n=5065). There were 403 respondents. Analysis was performed using whole group response and based on self-reported sub-specialty. RESULTS Survey results among all respondents show support for incorporating our molecular knowledge of brain tumors into the WHO classification (>80%). As one example, 96% of respondents responded that the integration of 1p/19q co-deletion into the molecular classification of oligodendroglioma was “very” or “critically” important for the management of grade III gliomas. While 30% of all respondents believe that IDH mutation status should affect overall management of GBM. Interestingly, there was some variability among sub-specialties for certain aspects, and as one example neuropathologists were slightly more inclined to disagree that molecular markers should be included in the WHO classification (25% for neuropathologists versus 13% overall). CONCLUSION Based on a survey provided to the neuro-oncology community, we report strong support for the integration of molecular markers into the WHO classification of brain tumors, as well as for using an integrated “layered” diagnostic format.

In the current practice of pathology, the identification of cell markers and their respective distribution represents an indispensable dialogue for diagnostic, predictive, therapeutic, and research purposes. Early immunohistochemical protocols were limited to direct, fluorescent labeled antibodies, yielding quick results but lacking sensitivity. More recently, the use of indirect techniques –utilization of enzyme labels–and various detection systems have continued to advance the complexity of IHC, increasing both its specificity and sensitivity of detecting one or multiple antigen(s) (Ag) simultaneously. As such, IHC has become an affordable, powerful, and readily available means for the identification of candidate biomarkers (mostly lineage markers) in formalin-fixed, paraffin-embedded (FFPE) tissue samples. Pathologists are now asked to “quantify” expression levels of differential prognostic markers–at microscopic level–using this arguably “non-quantitative” technique. Conventionally, histological grading relies mainly on manual counting of positively immunostained cells, a labour intensive protocol that may be associated with subjectivity, intra- and inter- observer variation and reproducibility issues. The subjectivity and lack of reproducibility has prompted the use of computer-assisted or fully automated image analysis technologies. Digital image acquisition systems are becoming commonplace and as such, the demand for complex assessments of digital images of histological slides must be matched with quantitative platforms. In this study, we aim to introduce a computer-assisted image-computing platform that is both accurate and efficient in quantification of isolated and heterogeneous candidate biomarkers in glioblastoma.

Background: Treatment of adult patients with hydrocephalus is often undertaken with a ventriculoperitoneal shunt (VPS). Failure rates have been reported as high as 50% in the first year. Methods: A Quality Improvement (QI) model was used to evaluate and modify VPS-insertion techniques to improve outcome. Malfunction was defined as a change in neurological shunt-related function with correlated diagnostic imaging studies. Prospectively collected data from 2012-2015 was reviewed. Results: 146 patients underwent a new VPS insertion. Diagnoses were: normal pressure hydrocephalus 101 patients, acquired hydrocephalus 28 patients and chronic-congenital hydrocephalus 17 patients. 103 patients had traditional insertion of a ventricular catheter using surface landmarks with 2 catheter misplacements requiring surgery. Image guidance with electromagnetic tracking was instituted with 0 catheter misplacements in 43 consecutive patients. 121 patients had traditional minilaparotomy/trocar placement of the peritoneal catheter with 59/121 (49%) experiencing shunt malfunction and 35/59 (59%) experiencing a second malfunction requiring surgery. Laparoscopic insertion of the peritoneal catheter was instituted in 25 consecutive patients with 3 (12%) distal obstructions. Laparoscopy was also used in 13 patients undergoing VPS revision with 2 (15%) experiencing subsequent malfunction. Conclusions: Changes to standard VPS surgical treatment including the addition of image-guidance and laparoscopic surgical techniques were associated with a significant decrease in shunt malfunction requiring surgery.

Pre-operative functional magnetic resonance imaging (fMRI) has emerged as valuable clinical tool to help surgically manage patients diagnosed with brain tumours. Surgical decision-making may be significantly improved through the provision of fMRI, however its clinical usage is contingent on the level of agreement with direct cortical stimulation (DCS). While previous studies have been undertaken to investigate the spatial agreement between fMRI and DCS, the influence that various factors may have on fMRI sensitivity and specificity is not fully clear. Thus, in a group of eight brain tumour patients who underwent pre-operative fMRI followed intra-operative DCS during an awake craniotomy procedure, we measured the agreement between the two brain mapping techniques looking at the influence of behavioural task, statistical threshold, and task standardization. Results: There were significant differences between motor and language mapping, where agreement was better for the former. Sensitivity and specificity shared an inverse relationship with increasing fMRI threshold, and were significantly reduced in the case where tasks were not standardized. Lastly, false positive occurrences were identified as the dominate source of error in comparison to false negative occurrences. Conclusion: Thus, the results from this work suggest that fMRI can predict intraoperative findings with good accuracy, however, sources of variability may significantly reduce the quality of fMRI data at the single-subject level. Neurosurgeons should carefully evaluate fMRI data with these considerations prior to its inclusion in the surgical-decision making process.

Background: The aim of this study was to assess the outcomes of surgery for multiple sclerosis-related trigeminal neuralgia (MS-TN). Methods: All Manitobans undergoing first surgery for medically refractory MS-TN between 2000 and 2014 were identified. The time interval until additional surgeries were required for recurrent pain, defined as the time to fail (TTF), was determined from a retrospective chart review. Kaplan-Meier analyses were performed and outcomes compared. Results: Twenty-one patients (26 sides) underwent first rhizotomy by GammaKnife (GK, 13), glycerol injection (PGR, 10) or balloon compression (BCR, 3). Second procedures were required in 88% at 15±13 months, including GK (24), PGR (19), BCR (25), microvascular decompression (2) and open surgical partial rhizotomy (Dandy, 4) for an overall total of 99 surgeries (1-12 per side). The additional GK, PGR, and BCR eventually failed and required further surgeries in 40%, 60% and 70% at 1, 2, and 3 years respectively with a trend to longer TTF compared to first surgeries (ns). Follow up of Dandy procedures, however, identified no pain recurrence at 4 to 110 months. Conclusions: The minimally invasive rhizotomies for MS-TN were associated with high rates of recurrence and reoperation. Long term pain relief was best achieved with a Dandy procedure, even after multiple prior rhizotomies.

Recurrent GBM and AA have a dismal prognosis and a high unmet need for effective therapies. Toca 511 (vocimagene amiretrorepvec) is an investigational retroviral replicating vector that encodes the transgene cytosine deaminase (CD). Toca 511 selectively infects, persists and spreads in tumor. Subsequent oral administration of 5-fluorocytosine (Toca FC) produces 5-fluorouracil (5-FU) by CD within infected cells. 5-FU kills cancer cells and myeloid derived suppressor cells, inducing robust antitumor immune responses in animal models. Clinical data from phase 1 trials are consistent with this mechanism of action, and show extended survival compared to historical controls. Toca 5 is a multicenter, randomized, open-label Phase 2/3 trial of Toca 511 and Toca FC versus standard of care administered to patients undergoing resection for first or second recurrence of GBM or AA. Phase 2 will enroll 170 patients. Primary endpoint is overall survival (OS). Key secondary endpoints are safety, objective response rate, clinical benefit rate, progression-free survival, and landmark OS. Key inclusion criteria are age 18-75 years, histologically proven GBM or AA, measurable disease preoperatively of less than 5cm, candidate for equal or greater 80% resection of enhancing tumor based on pre-operative evaluation, and KPS equal or greater to 70. Assays for immune monitoring will be performed and molecular profiling of resected tumor samples will be correlated efficacy.

Background: We retrospectively review benign peripheral nerve sheath tumours (BPNST) managed surgically at the Toronto Western Hospital. The incidence of BPNST is classified by anatomic location and predisposition syndrome. Independent predictors of tumour recurrence and symptom resolution are identified. Methods: 175 patients with 201 tumours were eligible for analysis. Data was collected on patient age, gender, diagnosis of neurofibromatosis (NF), tumour histopathology, tumour location, tumour volume, and extent of resection. Postoperative motor, sensory and pain outcomes were dichotomized as stable/improved or worse than preoperative scores. Relationships between tumour recurrence, or symptom resolution, and predictor variables were assessed with univariate and multiple logistic regression models. Results: Among Schwannomas, subtotal resection, a diagnosis of Schwannomatosis, and larger tumour volume were associated with recurrence (p=0.012, p=0.048, p=0.049, respectively); for neurofibromas, subtotal resection and a diagnosis of NF1 were associated with recurrence (p=0.036, p=0.022, respectively). Multivariate analyses revealed subtotal resection as an independent predictor of recurrence for BPNSTs (p=0.007, OR=13.16, 95%-CI 2.34-52.63). Gross-total resection (p=0.023, OR=4.01, 95%-CI 1.21-13.22) and presence of a preoperative motor deficit (p=0.038, OR=8.06, 95%-CI 4.65-90.91) were independent predictors of stable/improved postoperative motor function for BPNSTs. Conclusions: Gross-total resection is associated with both reduced recurrence and improved postoperative motor function, and should be attempted for all eligible BPNSTs.

Background: Accurate prediction of length of stay (LOS) following elective lumbar fusion may help optimize the utilization of resources and may assist with physician and patient expectations. Methods: Outcomes were collected prospectively among patients undergoing elective lumbar fusion. Prolonged LOS was defined as ≥5 days. The influence of baseline and peri-operative characteristics on the odds of prolonged LOS was assessed by a multivariate logistic regression model. Results: 150 patients underwent elective lumbar fusion surgery. Patient characteristics were as follows: average age was 61.9, average pre-operative back pain measured by the visual analogue scale was 54.3, and 36.5% of patients were identified as having severe disability, defined by an Oswestry disability index over 40. The average LOS was 4.9 days, with 28% having a prolonged LOS. Majority of patients had one level fused (69.7%). The odds of prolonged LOS were increased by severe disability (odds ratio [OR] 3.58, p<0.005), levels fused (OR 2.52, p<0.005), greater than 70 years of age (OR 3.81, p<0.005), and any treatment related adverse event (OR 4.32, p <0.02). There was no significant influence of prolonged surgery (p=0.3) or pre-operative back pain (p=0.23) on LOS. Conclusions: Prolonged length of stay was significantly influenced by severe disability, levels fused, age > 70, and any adverse events.

For decades, debate has persisted regarding the role of surgical resection in newly diagnosed glioblastoma. There is increasing evidence that extent of resection (EoR) is an independent prognostic factor. Previous work has proposed the inclusion of EoR in a risk stratification algorithm but does not incorporate account recent advances in the molecular characterization of tumours. We set out to investigate the effect of EoR on overall survival (OS), and to develop a stratification algorithm incorporating both EoR and modern molecular markers for prognostication. HYPOTHESIS: Greater EoR is independently associated with improved OS. METHODS: We examined 190 consecutive cases of histopathologically confirmed newly-diagnosed glioblastoma who were operated upon between January 1, 2012 and December 31, 2014. Variables including age, sex, postal code, KPS, tumour location, presenting symptoms, treatment history, date of progression, date of reoperation, as well as MGMT, IDH, 1p/19q codeletion, and ATRX status were recorded. Preoperative and postoperative MRIs were reviewed and volumetric tumour burden will be analyzed and EoR will be calculated. RESULTS: Preliminary EoR calculations (n=18) show a positive correlation between EoR and OS. CONCLUSION: A correlation exists between EoR and OS, although multivariable analysis is planned to exclude potential confounders. MRI review, chart review including molecular marker analysis and EoR calculations are ongoing.

Glioblastoma (GBM) is the most common brain cancer. Resistance to front-line systemic therapy with temozolomide (TMZ) is correlated with O6-methylguanine-DNA-methyltransferase (MGMT) expression. Second-line treatment with bevacizumab has not improved overall survival. Dianhydrogalactitol (VAL-083) is a bi-functional alkylating agent that has MGMT-independent cell-kill activity against GBM cell-lines and cancer stem cells in vitro. VAL-083 crosses the blood-brain barrier and showed promise against CNS tumors in prior NCI-sponsored clinical trials. The goal of this clinical trial is to determine appropriate VAL-083 dosing for advancement to Phase III trials as a new treatment for recurrent GBM. METHODS: Patients must have recurrent GBM following surgery, radiation, TMZ and bevacizumab. Phase I: Open-label, single-arm, dose-escalation study. Patients received VAL-083 on days 1,2,3 of a 21-day cycle, until reaching MTD. Phase II: Additional patients enrolled at MTD to further assess safety and outcomes. RESULTS: Phase I: 29 patients were enrolled across 9 dose cohorts (1.5-50 mg/m2/d). 40mg/m2/d was confirmed as MTD. Myelosuppression was mild; no drug-related serious adverse events were reported at doses up to 40mg/m2/d. Dose limiting G4 thrombocytopenia was observed at higher doses. Platelet nadir occurred around day 20 and resolved rapidly and spontaneously. A dose-related survival improvement was observed. Pharmacokinetic analyses show 1-2h plasma terminal half-life; average Cmax 781ng/mL at 40mg/m2/d. Phase II: 14 patients were enrolled at 40mg/m2/d. To date, safety observations in Phase II are consistent with Phase I. CONCLUSIONS: VAL-083 at 40mg/m2/d exhibits a favorable safety profile and dose-related trend toward clinically meaningful improved survival in refractory GBM patient