Search

Gliomas arising in the setting of neurofibromatosis type 1 (NF1) are heterogeneous, occurring from childhood through adulthood, can be histologically low-grade or high-grade, and follow an indolent or aggressive clinical course. Comprehensive profiling of genetic alterations beyond NF1 inactivation and epigenetic classification of these tumors remain limited. Through next-generation sequencing, copy number analysis, and DNA methylation profiling of gliomas from 47 NF1 patients, we identified 2 molecular subgroups of NF1-associated gliomas. The first harbored biallelic NF1 inactivation only, occurred primarily during childhood, followed a more indolent clinical course, and had a unique epigenetic signature for which we propose the terminology “pilocytic astrocytoma, arising in the setting of NF1”. The second subgroup harbored additional oncogenic alterations including CDKN2A homozygous deletion and ATRX mutation, occurred primarily during adulthood, followed a more aggressive clinical course, and was epigenetically diverse, with most tumors aligning with either high-grade astrocytoma with piloid features or various subclasses of IDH-wildtype glioblastoma. Several patients were treated with small molecule MEK inhibitors that resulted in stable disease or tumor regression when used as a single agent, but only in the context of those tumors with NF1 inactivation lacking additional oncogenic alterations. Together, these findings highlight recurrently altered pathways in NF1-associated gliomas and help inform targeted therapeutic strategies for this patient population.

Purpose Prognostically favorable IDH-mutant gliomas are known to produce oncometabolite D-2-hydroxyglutarate (2HG). In this study, we investigated metabolite-based features of patients with grade 2 and 3 glioma using 2HG-specific in vivo MR spectroscopy, to determine their relationship with image-guided tissue pathology and predictive role in progression-free survival (PFS). Methods Forty-five patients received pre-operative MRIs that included 3-D spectroscopy optimized for 2HG detection. Spectral data were reconstructed and quantified to compare metabolite levels according to molecular pathology (IDH1R132H, 1p/19q, and p53); glioma grade; histological subtype; and T2 lesion versus normal-appearing white matter (NAWM) ROIs. Levels of 2HG were correlated with other metabolites and pathological parameters (cellularity, MIB-1) from image-guided tissue samples using Pearson’s correlation test. Metabolites predictive of PFS were evaluated with Cox proportional hazards models. Results Quantifiable levels of 2HG in 39/42 (93%) IDH+ and 1/3 (33%) IDH– patients indicated a 91.1% apparent detection accuracy. Myo-inositol/total choline (tCho) showed reduced values in astrocytic (1p/19q-wildtype), p53-mutant, and grade 3 (vs. 2) IDH-mutant gliomas (p p p p p p = 0.002), total NAA (R = − 0.61; p = 0.002) and cellularity (R = 0.37; p = 0.04) but not MIB-1. Increasing levels of 2HG/tCr (p = 0.0007, HR 5.594) and thresholding (≥ 0.905, median value; p = 0.02) predicted adverse PFS. Conclusion In vivo 2HG detection can reasonably be achieved on clinical scanners and increased levels may signal adverse PFS.

Background IDH-mutant diffuse gliomas are heterogeneous, and improved methods for optimal patient therapeutic stratification are needed. PI3K/AKT/mTOR signaling activity can drive disease progression and potential therapeutic inhibitors of the pathway are available. Yet, the prevalence of PI3K/AKT/mTOR signaling pathway activity in IDH-mutant glioma is unclear and few robust strategies to assess activity in clinical samples exist. Methods PI3K/AKT/mTOR signaling pathway activity was evaluated in a retrospective cohort of 132 IDH-mutant diffuse glioma (91 astrocytoma and 41 oligodendroglioma, 1p/19q-codeleted) through quantitative multiplex immunoprofiling using phospho-specific antibodies for PI3K/AKT/mTOR pathway members, PRAS40, RPS6, and 4EBP1, and tumor-specific anti-IDH1 R132H. Expression levels were correlated with genomic evaluation of pathway intrinsic genes and univariate and multivariate Cox proportional hazard regression models were used to evaluate the relationship with outcome. Results Tumor-specific expression of p-PRAS40, p-RPS6, and p-4EBP1 was common in IDH-mutant diffuse glioma and increased with CNS WHO grade from 2 to 3. Genomic analysis predicted pathway activity in 21.7% (13/60) while protein evaluation identified active PI3K/AKT/mTOR signaling in 56.6% (34/60). Comparison of expression in male versus female patients suggested sexual dimorphism. Of particular interest, when adjusting for clinical prognostic factors, the level of phosphorylation of RPS6 was strongly associated with PFS (P Conclusions Our study emphasizes the value of proteomic assessment of signaling pathway activity in tumors as a means to identify relevant oncogenic pathways and potentially as a biomarker for identifying aggressive disease.

Composite pleomorphic xanthoastrocytoma-ganglioglioma (PXA-GG) is an extremely rare central nervous system neoplasm with 2 distinct but intermingled components. Whether this tumor represents a “collision tumor” of separate neoplasms or a monoclonal neoplasm with divergent evolution is poorly understood. Clinicopathologic studies and capture-based next generation sequencing were performed on extracted DNA from all available PXA-GG at 2 medical centers. Five PXA-GG were diagnosed in 1 male and 4 female patients ranging from 13 to 25 years in age. Four arose within the cerebral hemispheres; 1 presented in the cerebellar vermis. DNA was sufficient for analysis in 4 PXA components and 3 GG components. Four paired PXA and GG components harbored BRAF p.V600E hotspot mutations. The 4 sequenced PXA components demonstrated CDKN2A homozygous deletion by sequencing with loss of p16 (protein product of CDKN2A) expression by immunohistochemistry, which was intact in all assessed GG components. The PXA components also demonstrated more frequent copy number alterations relative to paired GG components. In one PXA-GG, shared chromosomal copy number alterations were identified in both components. Our findings support divergent evolution of the PXA and GG components from a common BRAF p.V600E-mutant precursor lesion, with additional acquisition of CDKN2A homozygous deletion in the PXA component as is typically seen in conventional PXA.

Background Chemotherapy improves overall survival after surgery and radiotherapy for newly diagnosed high-risk IDH-mutant low-grade gliomas (LGGs), but a proportion of patients treated with temozolomide (TMZ) will develop recurrent tumors with TMZ-induced hypermutation. We aimed to determine the prevalence of TMZ-induced hypermutation at recurrence and prognostic implications. Methods We sequenced recurrent tumors from 82 patients with initially low-grade IDH-mutant gliomas who underwent reoperation and correlated hypermutation status with grade at recurrence and subsequent clinical outcomes. Results Hypermutation was associated with high-grade disease at the time of reoperation (OR 12.0 95% CI 2.5-115.5, P = .002) and was identified at transformation in 57% of recurrent LGGs previously exposed to TMZ. After anaplastic (grade III) transformation, hypermutation was associated with shorter survival on univariate and multivariate analysis (HR 3.4, 95% CI 1.2-9.9, P = .024), controlling for tumor grade, subtype, age, and prior radiotherapy. The effect of hypermutation on survival after transformation was validated in an independent, published dataset. Hypermutated (HM) tumors were more likely to develop discontiguous foci of disease in the brain and spine (P = .003). To estimate the overall incidence of high-grade transformation among low-grade IDH-mutant tumors, data from a phase II trial of TMZ for LGG were analyzed. Eight-year transformation-free survival was 53.8% (95% CI 42.8-69.2), and 61% of analyzed transformed cases were HM. Conclusions TMZ-induced hypermutation is a common event in transformed LGG previously treated with TMZ and is associated with worse prognosis and development of discontiguous disease after recurrence. These findings impact tumor classification at recurrence, prognostication, and clinical trial design.

Although physiologic (diffusion-weighted and perfusion-weighted) MRI has shown promise in identifying regions of recurrent tumor (rTumor) in patients with glioblastoma suspected of progression, distinguishing treatment-induced effects (TxE) from rTumor on anatomical MRI remains a challenge. Whereas prior larger-scale machine learning (ML)-based studies mostly utilize anatomical imaging alone and/or perform lesion-level predictions, this study aimed to develop a non-invasive, radiopathomic tool for regional probabilistic mapping of TxE using 208 tissue-samples (55 pathologically-confirmed TxE, 153 recurrent glioblastoma) acquired from 107 patients with known spatial coordinates on pre-surgical MRI. We tested the hypothesis that applying a deep-learning (DL) model that included physiological MRI can: 1) more accurately identify areas of TxE that mimic rTumor on anatomical MRI and 2) better generalize to an independent test set than ML-models or a DL-model that uses anatomical MRI alone. An 80/20 split for training/validation was used after 1/3 of the patients were withheld for testing. Oversampling of TxE samples was employed to address class imbalance and an equal proportion of TxE samples was maintained across all datasets. Three ML-models, their ensemble, and a deep 4D-convolutional-neural-network were trained based on normalized anatomical (post-contrast T1, T2-FLAIR), diffusion-weighted (ADC, FA), and DSC perfusion-weighted (PeakHeight, %recovery) images cropped to 10mm-cubic patches centered on the coordinates from where tissue was obtained. Although Random Forest and voting-ensembled ML-models using all imaging and the anatomical DL-model had the best validation performance (AUC=0.81-0.82), these models did not generalize (test AUC=0.58-0.59). The DL-model including physiologic images had slightly lower validation AUC (0.78) but the best overall test AUC (0.795), indicating superior generalizability. Elevated blood volume (nPeakHeight) was the most important feature. Our DL-model’s interpretability was also demonstrated by disrupting class separation after shuffling voxels within each input patch. These results suggest that using deep-learning with physiologic MRI can improve intratumoral classification of TxE from rTumor.

INTRODUCTION Fractionated radiotherapy is a first-line treatment for glioblastoma, but daily ionizing radiation prevents durable immune infiltration of the tumor microenvironment. Hypofractionated radiotherapy is used to treat glioblastomas at recurrence, but the impact of hypofractionated radiotherapy on the glioblastoma immune microenvironment is incompletely understood. Here we define immune microenvironment changes across multiple immunocompetent intracranial mouse models of glioblastoma after treatment with ionizing radiation mimicking stereotactic radiosurgery (SRS) in humans. METHODS Syngeneic GL261 (3x105 cells/mouse) or SB28 glioblastoma cells (3x104 cells/mouse) were implanted into the frontal lobe of immunocompetent C57BL/6J mice (18 mice/arm x 4 arms). Intracranial bioluminescence was used to assess glioblastoma growth. After tumor engraftment, glioblastomas were treated with conformal SRS (18Gy/1Fx) or sham. Glioblastomas were collected for histologic, single-cell, or molecular analyses 5 days after treatment (6 mice/arm) or at the time of euthanasia after monitoring for survival (12 mice/arm). Glioblastoma immune microenvironment responses were assessed using (1) H&E, (2) single cell mass cytometry (CyTOF) or IHC to define or validate immune cell changes, respectively, or (3) multiplexed cytokine assays to elucidate molecular mechanisms reprograming the glioblastoma immune microenvironment in response to SRS. RESULTS SRS attenuated glioblastoma growth and prolonged survival compared to sham treatment in both immunocompetent intracranial mouse models (GL261: 14 days versus 27 days, p< 0.001, SB28: 19 days versus 22 days, p=0.001). CyTOF showed SRS decreased immunosuppressive macrophage infiltration and increased microglia or CD8+ T cell infiltration of the glioblastoma immune microenvironment. Histologic analyses validated T cell and microglia infiltration after SRS. Glioblastoma cytokine analysis revealed inhibition of pro-tumor/anti-inflammatory cytokines (IL6, LIF) after SRS. CONCLUSIONS Single-fraction SRS durably reprograms glioblastoma macrophage, microglia, and CD8+ T cell populations in preclinical models, suggesting SRS or inhibition of pro-tumor/anti-inflammatory mechanisms underlying the immunosuppressive glioblastoma microenvironment represent immunogenic therapies that may offer a benefit to patients with glioblastoma.

Understanding the molecular landscape and microenvironment of melanoma brain metastases is critical to devise improved treatments. Here, we perform bulk and single cell genomic analysis of melanoma brain metastases to identify molecular correlates of clinical outcome. 84 consecutive patients who underwent surgical resection at a single institution with a histo-pathologically confirmed diagnosis of melanoma brain metastasis were retrospectively identified. In 60 patients (71%) with sufficient brain metastasis tissue for targeted next generation sequencing, DNA mutations were assessed with a CLIA certified sequencing assay. Single nuclear RNA sequencing using the 10x platform was performed on n=6 samples from treatment naïve patients. Overall survival (OS) and CNS progression free survival (CNS PFS) from time of brain metastasis diagnosis were estimated using the Kaplan-Meier method. The median patient age was 62 years old (range: 25-78 years), and the median clinical follow up was 17 months. A total of 33 patients (39%) had BRAFV600E melanoma brain metastases. Multivariate analysis incorporating age, performance status, and extracranial disease revealed BRAF status was an independent prognostic factor for OS (p< 0.05). In patients undergoing targeted next generation sequencing, the most common pathogenic variant was TERT promoter mutation (n=44; 73%). With regard to TCGA molecular melanoma subgroups, NRAS mutant (n=22; 37%) brain metastases were most common followed by BRAF mutant (n=20; 33%), NF1 mutant (n=11; 18%), and triple wildtype (n=7; 12%). Evaluation of clinical outcomes in the context of next generation sequencing results revealed no differences by TERT status but demonstrated worse overall survival in the BRAF mutant molecular group (p< 0.01, log-rank test). Single nuclear sequencing of 36,115 nuclei across 6 samples revealed BRAF wildtype tumors exhibit greater infiltrating immune cell populations including microglia and T cell subtypes. Future work will require integration of these findings with different systemic therapy regimens and across larger, prospective, multi-institutional cohorts.

BACKGROUND The Hedgehog pathway directs gene expression programs that are essential for glioblastoma stem cells in vitro, but a clinical trial suggested Hedgehog pathway inhibition does not improve glioblastoma outcomes (ABTC-0904). In pancreatic and bladder cancer, Hedgehog signaling through the tumor microenvironment induces stromal morphogens to inhibit cancer stem cell proliferation. Thus, we hypothesized Hedgehog signaling through the glioblastoma microenvironment may inhibit glioblastoma stem cell proliferation in vivo. METHODS Hedgehog signaling was studied using patient-derived, fluorescently-labeled IDH wild-type glioblastoma cells (GBM6, SF11956, SF11907, GPMP004, GPMP005) in vitro, in vivo after intracranial implantation in mice, or in 3D co-culture with iPSC-derived organoids comprised of human astrocytes or neurons. Genetic loss-of-function experiments were performed using CRISPR interference. Cell proliferation, stem cell marker expression, Sonic Hedgehog (SHH) expression, or primary cilia architecture were assessed using immunofluorescence, confocal microscopy, or live-cell imaging. Hedgehog signaling or morphogen expression were assessed using QPCR or single-cell RNA sequencing. Pharmacologic experiments were performed using vismodegib to inhibit the Hedgehog pathway, or FK506 to induce stromal morphogen expression. RESULTS Vismodegib increased glioblastoma cell proliferation and stem cell marker expression in vivo or in co-culture with astrocyte organoids. Single-cell RNA sequencing demonstrated glioblastoma cell co-culture with astrocyte organoids induced Hedgehog target genes and stem cell markers in glioblastoma cells, or stromal morphogens in astrocytes, which also expressed primary cilia. CRISPRi suppression of SHH in glioblastoma cells increased glioblastoma cell proliferation and stem cell marker expression in co-culture with astrocyte organoids. Neither vismodegib nor SHH suppression induced glioblastoma cell proliferation or stem cell marker expression in vitro or in co-culture with neuron organoids. FK506 induced astrocyte morphogen expression, inhibiting glioblastoma cell proliferation, stem cell marker expression, and the effects of vismodegib in co-culture with astrocyte organoids. CONCLUSIONS SHH from glioblastoma cells signals through astrocytes to inhibit cancer stem cell proliferation.

INTRODUCTION Pathologically aggressive tumor biology can extend beyond the contrast-enhancing or non-enhancing anatomical lesions in patients with glioma. Identification of malignant regions can help guide diagnosis and subsequent treatment planning. This study leverages a unique multi-parametric MRI dataset with tissue samples of known spatial coordinates to noninvasively predict cellular proliferation (KI-67) and a novel index of tumor aggressiveness (TAI), that combines proliferation, cellularity, and tumor-score. METHODS 420 tissue samples were collected from 162 patients with newly-diagnosed glioma (47% IDH-wildtype). Clinical imaging consisted of T2-weighted, T2-FLAIR, T1-weighted pre- and post-contrast images, and apparent diffusion coefficient (ADC) and fractional anisotropy (FA) from diffusion-weighted imaging. Mean normalized imaging metrics were quantified from 5mm spheres centered at the location of the tissue sample. A single spectrum was reconstructed at the location of each tissue sample from 3D 1H-MR Spectroscopic Imaging (MRSI) before quantifying normalized metabolite peak-heights for choline, creatine, NAA, lactate/lipid, and relative indices. Univariate mixed-effects linear regression models were employed and features with p< 0.2 were selected for subsequent model building. Support vector machine (SVM), random forest, and gradient boosting machine-learning algorithms were trained and tested on a ⅔-⅓ train-test split with 4-fold cross-validation in training to predict a high/low KI-67 and TAI. RESULTS Although none of the individual imaging metrics were significantly associated with KI-67 in the univariate analysis, all diffusion and several MRSI metrics (ncholine, nNAA, CNI, excess choline and creatine) were significantly associated with cellularity. Preliminary multivariate analyses to date suggest that the best radiopathomic model performance is achieved when an SVM was used along with T1-precontrast, nADC, and all metabolite levels (mean cross-validation AUC=0.73 and accuracy=.77). CONCLUSION Our results suggest that multi-parametric physiologic and metabolic MRI are useful for radiopathomic-mapping of tumor aggressiveness and are currently being optimized in a larger cohort.

Background Genomic profiling studies of diffuse gliomas have led to new improved classification schemes that better predict patient outcomes compared to conventional histomorphology alone. One example is the recognition that patients with IDH-wildtype diffuse astrocytic gliomas demonstrating lower-grade histologic features but genomic and/or epigenomic profile characteristic of glioblastoma typically have poor outcomes similar to patients with histologically diagnosed glioblastoma. Here we sought to determine the clinical impact of prospective genomic profiling for these IDH-wildtype diffuse astrocytic gliomas lacking high-grade histologic features but with molecular profile of glioblastoma. Methods Clinical management and outcomes were analyzed for 38 consecutive adult patients with IDH-wildtype diffuse astrocytic gliomas lacking necrosis or microvascular proliferation on histologic examination that were genomically profiled on a prospective clinical basis revealing criteria for an integrated diagnosis of “diffuse astrocytic glioma, IDH-wildtype, with molecular features of glioblastoma, WHO grade IV” per cIMPACT-NOW criteria. Results We identified that this diagnosis consists of two divergent clinical scenarios based on integration of radiologic, histologic, and genomic features that we term “early/evolving” and “undersampled” glioblastoma, IDH-wildtype. We found that prospective genomically guided identification of early/evolving and undersampled IDH-wildtype glioblastoma resulted in more aggressive patient management and improved clinical outcomes compared to a biologically matched historical control patient cohort receiving standard-of-care therapy based on histomorphologic diagnosis alone. Conclusions These results support routine use of genomic and/or epigenomic profiling to accurately classify glial neoplasms, as these assays not only improve diagnostic classification but critically lead to more appropriate patient management that can improve clinical outcomes.

Author(s): Sloan, Emily A; Hilz, Stephanie; Gupta, Rohit; Cadwell, Cathryn; Ramani, Biswarathan; Hofmann, Jeffrey; Kline, Cassie N; Banerjee, Anu; Reddy, Alyssa; Oberheim Bush, Nancy Ann; Chang, Susan; Braunstein, Steve; Chang, Edward F; Raffel, Corey; Gupta, Nalin; Sun, Peter P; Kim, John YH; Moes, Gregory; Alva, Elizabeth; Li, Rong; Bruggers, Carol S; Alashari, Mouied; Wetmore, Cynthia; Garg, Shipra; Dishop, Megan; Van Ziffle, Jessica; Onodera, Courtney; Devine, Patrick; Grenert, James P; Lee, Julieann C; Phillips, Joanna J; Pekmezci, Melike; Tihan, Tarik; Bollen, Andrew W; Berger, Mitchel S; Costello, Joseph F; Perry, Arie; Solomon, David A

10% of all pediatric brain tumors arise in the brainstem. Amongst these gliomas are the most common while other entities are rare and infrequently described in the literature. In this study we investigated the prevalence of non-gliomatous tumors in the brainstem. Amongst the 1323 embryonal tumours received at the RBTC, we identified 17 cases of ETMRs (17/165) that presented as brainstem primaries. Previously grouped within CNS-PNETs, ETMR, is a new WHO diagnostic entity, characterized by C19MC alterations. ETMR is a disease of infancy, the clinical spectrum of which is poorly understood. ETMRs arise at multiple CNS locations including cerebrum being most common (60%,) followed by cerebellum (18%) and midline structures (6%); notably 10% were brainstem primaries, mimicking DIPG radiologically. All patients presented with a short history of progressive neurological symptoms, with most common signs and symptoms of cranial neuropathies, long tract signs and gait disturbance. Median age at diagnosis was 27 months (range 16-75months) with a male to female ratio of 0.9:1. Predominantly localized (M0-94%, M2-3 -6%) majority of patients underwent upfront biopsy or partial resection (15/17:88%), while complete tumor resection was achieved in 2 cases. All patients received heterogenous combination of chemotherapy with and without radiotherapy. Majority of patients progressed rapidly with median time to progression of 4 months and overall survival of < 13 months. The only long-term surviving patient had complete resection dose intensified chemotherapy and radiation (OS 202months). Primary ETMRs in the brainstem are under recognised entities and carry a dismal prognosis. Although rapidly progressive, prompt recognition, maximal resection and management with multimodal adjuvant therapy should be considered in cases with brainstem disease.

The prospect of using immunotherapy for IDH-mutant LGGs has been daunting given the immune-poor microenvironment and low mutational burden. We hypothesized that LGG-targeting T cells might still be present at low frequency and with limited regional infiltration into the tumor. To improve sensitivity, we combined high-density multi-region tumor sampling with high-throughput neoantigen-T cell screening for a patient with WHO Grade II diffuse astrocytoma who eventually progressed, at second recurrence, to anaplastic astrocytoma. We performed maximal-anatomic sampling from 10 distinct regions of the tumor at the initial resection, as well as single sampling at first recurrence, for exome-based prediction of clonal and subclonal expressed neoantigens, RNAseq-based estimation of regional immune cell composition, and T cell receptor (TCR) beta deep sequencing. Based on our predictions, we then generated a barcoded library of patient-specific peptide-HLA multimers loaded with predicted neopeptides. Using this library, neoantigen-specific CD8 T cells were captured and isolated from patient peripheral blood and subjected to single cell TCR sequencing. We screened patient-derived peripheral blood drawn two years after initial resection and identified five T cell clones recognizing three LGG neoepitopes. Two neoepitopes were derived from truncal, tumor-wide mutations, including a truncating splice-site mutation in TP53 and a missense mutation in MRPL46. Each of these neoepitopes were recognized by two distinct TCRs, consistent with TCR convergence. A third neoepitope was derived from a subclonal MRPL46 mutation. Using the TCR beta sequence as a molecular barcode, TCRs specific to the 2 clonal neoepitopes, but not the subclonal neoepitope, were detectable in the glioma and blood at initial resection, as well as in the recurrent glioma. In summary, we demonstrate the existence and persistence of neoantigen-targeting T cells within the blood and tumor of an IDH-mutant LGG patient. These findings suggest a feasible methodology to develop personalized T cell-based immunotherapies for IDH-mutant LGGs.

BackgroundDiagnostic classification of diffuse gliomas now requires an assessment of molecular features, often including IDH-mutation and 1p19q-codeletion status. Because genetic testing requires an invasive process, an alternative noninvasive approach is attractive, particularly if resection is not recommended. The goal of this study was to evaluate the effects of training strategy and incorporation of biologically relevant images on predicting genetic subtypes with deep learning.MethodsOur dataset consisted of 384 patients with newly diagnosed gliomas who underwent preoperative MRI with standard anatomical and diffusion-weighted imaging, and 147 patients from an external cohort with anatomical imaging. Using tissue samples acquired during surgery, each glioma was classified into IDH-wildtype (IDHwt), IDH-mutant/1p19q-noncodeleted (IDHmut-intact), and IDH-mutant/1p19q-codeleted (IDHmut-codel) subgroups. After optimizing training parameters, top performing convolutional neural network (CNN) classifiers were trained, validated, and tested using combinations of anatomical and diffusion MRI with either a 3-class or tiered structure. Generalization to an external cohort was assessed using anatomical imaging models.ResultsThe best model used a 3-class CNN containing diffusion-weighted imaging as an input, achieving 85.7% (95% CI: [77.1, 100]) overall test accuracy and correctly classifying 95.2%, 88.9%, 60.0% of the IDHwt, IDHmut-intact, and IDHmut-codel tumors. In general, 3-class models outperformed tiered approaches by 13.5%-17.5%, and models that included diffusion-weighted imaging were 5%-8.8% more accurate than those that used only anatomical imaging.ConclusionTraining a classifier to predict both IDH-mutation and 1p19q-codeletion status outperformed a tiered structure that first predicted IDH-mutation, then 1p19q-codeletion. Including apparent diffusion coefficient (ADC), a surrogate marker of cellularity, more accurately captured differences between subgroups.

Glioblastoma responses to immune checkpoint inhibition (ICI) are rare, and the molecular mechanisms underlying ICI responses are incompletely understood. Thus, serial glioblastoma samples are valuable resources for identifying biomarkers or therapeutic targets to increase the efficacy of ICI in patients with glioblastoma. We obtained paired glioblastoma samples from 7 patients who underwent sequential surgery, ICI, and eventual salvage surgery for recurrence. Patients were distinguished as ICI responders (n=3) or non-responders (n=4) based on (1) MRI evidence of tumor stability/reduction over 6+ months after ICI, or (2) pathologic evidence of predominant treatment effect at the time of salvage surgery after ICI. FFPE sections from each tumor (n=14) were stained using H&E or IHC/IF for macrophages/microglia (CD68) or T cells (CD3) and analyzed using light or fluorescence microscopy. Six regions-of-interest (ROIs) comprising viable tumor were selected neuropathologist from each sample (n=84 ROIs). ROIs were analyzed using quantitative spatial profiling of 72 proteins on the Nanostring Digital Profiler platform. Glioblastomas responding to ICI were enriched in T-cell proteins (CD3, CD4, CD8) and T-cell activation markers (CD25) at the time of salvage compared to initial surgery. Markers of MAPK signaling were suppressed in pre-ICI samples compared to post-ICI samples in responders. p-ERK was suppressed in post-ICI samples compared to pre-ICI samples in non-responders. Myeloid proteins (CD68, CD163, CD11c) were enriched in post-ICI samples compared to pre-ICI samples in non-responders. Principle components analysis revealed p-ERK and immune proteins (CD3, CD4, CD8, CD20, CD11c, CTLA4, CD68, CD45, CD56, and CD127) accounted for 62% of the variance among pre-ICI and post-ICI samples in responders. In conclusion, temporospatial protein profiling of human glioblastomas reveals molecular mechanisms and biomarkers underlying responses to immune checkpoint inhibition. These data establish a foundation for functional studies to reprogram the immunosuppressive glioblastoma microenvironment and sensitize tumors to immune checkpoint inhibition.

INTRODUCTION Noninvasive, radiopathomic mapping of tumor aggressiveness can benefit patients with glioma by guiding the selection of tissue samples for diagnosis, increasing extent of resection, and non-invasively characterizing residual tumor burden for subsequent treatment. Although prior studies have demonstrated the utility of metabolic metrics quantified from 1H-MR Spectroscopy (MRS) in probing tumor pathology, this study evaluated the benefit of using the entire 1D-spectrum and deep learning for radiopathomic mapping of intratumoral cellularity, proliferation (ki-67), and a new tumor aggressiveness index (TAI) defined as log((n(ki−67)+n(cellularity))*tumor-score). METHODS Multi-voxel 1H-MRS was acquired on 281 patients newly diagnosed with a glioma (47% IDH-wildtype) immediately before surgical resection. After reconstructing individual spectra at the locations where tissue samples were obtained during surgery and normalizing by NAA in contralateral normal-appearing-white-matter, 607 spectra with corresponding histopathology were deemed of sufficient quality for analysis. A 1D convolutional-neural-network with bidirectional long- and short-term memory deep-learning model using the entire spectrum (0.6-3.6ppm) was compared to mixed-effects regression (with choline-to-NAA index[CNI]) and Random Forest (with CNI+normalized peak heights) models for predicting ki-67, cellularity, and TAI. Results & DISCUSSION Using deep-learning on the entire spectrum resulted in 10.3%-22.1% lower mean absolute error (MAE) and 0.32-0.37 higher R2 values compared to using CNI alone or a random forest model with multiple metabolic metrics. MAE values for all 3 deep-learning models were 26-44% < 1 standard deviation of the ground truth, demonstrating reasonable prediction accuracy within the test data set. Although the lowest MAE (0.16) and highest R2 (0.41) was attained when predicting TAI with deep-learning, the prediction of cellularity resulted in the lowest %MAE. Colormaps of predicted pathology identified regions of heightened aggressiveness surrounding tissue samples with most abnormal pathological features that sometimes extended beyond the non-enhancing lesion. Current work is evaluating the clinical utility of our deep-learning model and predicted maps of aggressiveness.

The glioblastoma microenvironment is an immunosuppressive barrier to therapeutic innovation. We hypothesized intratumor convection enhanced delivery (CED) of gene therapy vectors could reprogram the glioblastoma immune microenvironment and elucidate therapeutic vulnerabilities. To test this, SB28 or GL261 glioblastoma allografts were implanted into immunocompetent mice and treated with CED of attenuated adeno-associated virus 9 vectors (AAV9) encoding experimental cytokines (Il1b, Ccl4, or Apoa1) underlying infiltration or activation of anti-tumor immune cells in other intracranial tumors. Serial intracranial bioluminescence was used to assess glioblastoma growth and animals were monitored for survival. The impact of gene therapy perturbations on the glioblastoma immune microenvironment was assessed using histology, immunohistochemistry, single-cell mass cytometry (CyTOF), and multiplexed cytokine assays. Serial body weight and systemic cytokine measurements showed no evidence of treatment toxicity. AAV9-APOA1 or AAV9-IL1B gene therapy CED treatments attenuated SB28 growth and prolonged survival, decreasing immunosuppressive macrophage infiltration and increasing CD8 T cell and microglia infiltration of the glioblastoma microenvironment compared to control AAV9 vectors. Gene therapy CED treatments did not attenuate GL261 growth or prolong survival, but CyTOF of human glioblastomas (n=6) in comparison to preclinical models revealed untreated GL261 glioblastomas were endogenously enriched in CD8 T cells and other lymphoid lineages compared to untreated SB28 or human glioblastomas. Multiplexed cytokine assays demonstrated suppression of intratumor IL6 is a conserved mechanism of action underlying glioblastoma gene therapy responses. Single-cell RNA sequencing analysis of 32,877 cells from human glioblastomas (n=11) showed IL6 is predominantly produced by radial glial like cancer stem cells or endothelial cells in the tumor microenvironment. In support of these findings, survival from intracranial SB28 glioblastomas was prolonged in Il6 knockout C57BL/6J mice compared to wildtype mice. In summary, we report a novel strategy using gene therapy and CED to reprogram the glioblastoma immune microenvironment, revealing IL6 drives glioblastoma immunosuppression and growth.

Mutations in the TERT promoter are the single most common non-coding mutation in cancer and represent the genetic underpinnings of tumor cell immortality. Beyond the two most common point mutations, G228A and G250A, which selectively recruit the ETS factor GABP to activate TERT, the significance of other variants in the TERT promoter are unknown. We identified duplications of wildtype sequence within the core promoter region of TERT in 7 different cancer types that have strikingly similar features including size, insertion position, and inclusion of one of the native ETS motifs. Each duplication activates the TERT promoter to a similar level as G228A and G250A and is critically dependent on the insertion site. The GABP tetramer binds to the TERT duplicated promoter sequence by virtue of the native ETS motif and its duplicated version with precise spacing, and it is necessary for the transcriptional activation by all duplications tested. Spatiotemporal analysis in a multifocal glioblastoma shows the duplication is clonal and its activation of TERT is readily detectable at the single cell level and in bulk tumor tissue. We conclude that recurrent TERT promoter duplications of the native ETS sequence are functionally and mechanistically equivalent to the hotspot mutations that confer tumor cell immortality. The shared mechanism of these divergent somatic genetic alterations suggests a strong selective pressure for recruitment of the GABP tetramer to activate TERT.

There are no approved targeted therapies for meningiomas and the cell types underlying meningeal tumorigenesis are incompletely understood. To address these limitations, we performed single-cell RNA sequencing of 57,114 cells from 8 human meningiomas and 54,607 cells from 3 canine meningiomas. Pseudotime, gene ontology, and copy number variant analyses revealed a population of pericyte-like meningioma cells that were conserved across human and canine tumors and were enriched in expression of Notch3 and other cancer stem cell genes. Deconvolution of cell types from bulk RNA sequencing and DNA methylation profiling of 200 human meningiomas integrated with immunohistochemistry (IHC), immunofluorescence (IF), and RNAScope demonstrated Notch3+ pericytes and Notch3 expression were enriched in high grade or Immune-enriched meningiomas, which were distinguished from other meningioma DNA methylation groups by genes driving vasculature development. IHC and IF of human meninges integrated with lineage tracing approaches using Notch3-CreERT2 ROSAmT/mG alleles in mice demonstrated Notch3 expression was restricted to the perivascular stem cell niche during meningeal development and homeostasis. Mice harboring Notch3-CreERT2 Nf2fl/fl alleles developed meningeal hyperproliferation. Overexpression of constitutively activated Notch3 (Notch3AICD) in Immune-enriched human meningioma cells increased the expression of cancer stem cell genes, driving clonogenic growth in vitro, limiting dilution tumor-initiating capacity in vivo, and resistance to radiotherapy in vivo. A selective Notch3 neutralizing antibody (αNRR3) blocked meningioma cell proliferation and expression of Notch3 target genes, inhibiting meningioma xenograft growth and prolonging overall survival. Single-cell RNA sequencing of 187,366 cells from meningioma xenografts after αNRR3 or radiotherapy treatment ± Notch3AICD overexpression revealed distinct meningioma cell-intrinsic or cell-extrinsic mechanisms driving responses to radiotherapy or αNRR3, respectively. Combined treatment with αNRR3 and radiotherapy additively blocked meningioma xenograft growth and extended survival benefit. In sum, these data shed light on a novel cell type, molecular mechanism, and therapeutic vulnerability in the most common primary intracranial tumor.

BACKGROUND Monitoring lower-grade gliomas (LrGGs) for disease progression is made difficult by the limits of anatomical MRI to distinguish treatment-related tissue changes from tumor progression. MR spectroscopic imaging (MRSI) offers additional metabolic information that can help address these challenges. The goal of this study was to compare longitudinal changes in multiparametric MRI, including diffusion-weighted imaging, perfusion imaging, and 3D MRSI, for LrGG patients who progressed at the final time point and those who remained clinically stable. METHODS Forty-one patients with LrGG who were clinically stable were longitudinally assessed for progression. Changes in anatomical, diffusion, perfusion, and MRSI data were acquired and compared between patients who remained clinically stable and those who progressed. RESULTS Thirty-one patients remained stable, and 10 patients progressed. Over the study period, progressed patients had a significantly greater increase in normalized choline, choline-to-N-acetylaspartic acid index (CNI), normalized creatine, and creatine-to-N-acetylaspartic acid index (CRNI), than stable patients. CRNI was significantly associated with progression status and WHO type. Progressed astrocytoma patients had greater increases in CRNI than stable astrocytoma patients. CONCLUSIONS LrGG patients in surveillance with tumors that progressed had significantly increasing choline and creatine metabolite signals on MRSI, with a trend of increasing T2 FLAIR volumes, compared to LrGG patients who remained stable. These data show that MRSI can be used in conjunction with anatomical imaging studies to gain a clearer picture of LrGG progression, especially in the setting of clinical ambiguity.

Author(s): Sloan, Emily A; Cooney, Tabitha; Oberheim Bush, Nancy Ann; Buerki, Robin; Taylor, Jennie; Clarke, Jennifer L; Torkildson, Joseph; Kline, Cassie; Reddy, Alyssa; Mueller, Sabine; Banerjee, Anu; Butowski, Nicholas; Chang, Susan; Mummaneni, Praveen V; Chou, Dean; Tan, Lee; Theodosopoulos, Philip; McDermott, Michael; Berger, Mitchel; Raffel, Corey; Gupta, Nalin; Sun, Peter P; Li, Yi; Shah, Vinil; Cha, Soonmee; Braunstein, Steve; Raleigh, David R; Samuel, David; Scharnhorst, David; Fata, Cynthia; Guo, Hua; Moes, Gregory; Kim, John YH; Koschmann, Carl; Van Ziffle, Jessica; Onodera, Courtney; Devine, Patrick; Grenert, James P; Lee, Julieann C; Pekmezci, Melike; Phillips, Joanna J; Tihan, Tarik; Bollen, Andrew W; Perry, Arie; Solomon, David A

Increases in the extent of resection of both contrast-enhanced (CE) and non–contrast-enhanced (NCE) tissue are associated with substantial survival benefits in patients with isocitrate dehydrogenase wild-type glioblastoma. The fact, however, remains that these lesions exist within the framework of complex neural circuitry subserving cognition, movement, and behavior, all of which affect the ultimate survival outcome. The prognostic significance of the interplay between CE and NCE cytoreduction and neurological morbidity is poorly understood. OBJECTIVE: To identify a clinically homogenous population of 228 patients with newly diagnosed isocitrate dehydrogenase wild-type glioblastoma, all of whom underwent maximal safe resection of CE and NCE tissue and adjuvant chemoradiation. We then set out to delineate the competing interactions between resection of CE and NCE tissue and postoperative neurological impairment with respect to overall survival. METHODS: Nonparametric multivariate models of survival were generated via recursive partitioning to provide a clinically intuitive framework for the prognostication and surgical management of such patients. RESULTS: We demonstrated that the presence of a new postoperative neurological impairment was the key factor in predicting survival outcomes across the entire cohort. Patients older than 60 yr who suffered from at least one new impairment had the worst survival outcome regardless of extent of resection (median of 11.6 mo), whereas those who did not develop a new impairment had the best outcome (median of 28.4 mo) so long as all CE tissue was resected. CONCLUSION: Our data provide novel evidence for management strategies that prioritize safe and complete resection of CE tissue.

BACKGROUND Meningiomas are the most common primary intracranial tumors in humans and dogs, but biologic drivers and cell types underlying meningeal tumorigenesis are incompletely understood. Here we integrate meningioma single-cell RNA sequencing with stem cell approaches to define a perivascular stem cell underlying vertebrate meningeal tumorigenesis. METHODS Single-cell RNA sequencing was performed on 57,114 cells from 8 human meningiomas, 54,607 cells from 3 dog meningiomas, and human meningioma xenografts in mice. Results were validated using immunofluorescence (IF), immunohistochemistry (IHC), and deconvolution of bulk RNA sequencing of 200 human meningiomas. Mechanistic and functional studies were performed using clonogenic and limiting dilution assays, xenografts, and genetically engineered mouse models. RESULTS Copy number variant identification from human meningioma single cells distinguished tumor cells with loss of chr22q from non-tumor cells with intact chr22q. A single cluster distinguished by expression of Notch3 and other cancer stem cell genes had an intermediate level of loss of chr22q, suggesting this cluster may represent meningioma stem cells. In support of this hypothesis, pseudotime trajectory analysis demonstrated transcriptomic progression starting from Notch3+ cells and encompassing all other meningioma cell types. Notch3+ meningioma cells had transcriptomic concordance to mural pericytes, and IF/IHC of prenatal and adult human meninges, as well as lineage tracing using a Notch3-CreERT2 allele in mice, confirmed Notch3+ cells were restricted to the perivascular stem cell niche in mammalian meningeal development and homeostasis. Integrating human and dog meningioma single cells revealed Notch3+ cells in tumor and non-tumor clusters in dog meningiomas. Notch3 IF/IHC and cell-type deconvolution of bulk RNA sequencing showed Notch3+ cells were enriched in high-grade human meningiomas. Notch3 overexpression in human meningioma cells increased clonogenic growth in vitro, and increased tumorigenesis and tumor growth in vivo, decreasing overall survival. CONCLUSIONS Notch3+ stem cells in the perivascular niche underlie vertebrate meningeal tumorigenesis.

Low-grade gliomas (LGGs) are the most common childhood brain tumor in the general population and in individuals with the Neurofibromatosis type 1 (NF1) cancer predisposition syndrome. Surgical biopsy is rarely performed prior to treatment in the setting of NF1, resulting in a paucity of tumor genomic information. To define the molecular landscape of NF1-associated LGGs (NF1-LGG), we integrated clinical data, histological diagnoses, and multi-level genetic/genomic analyses on 70 individuals from 25 centers worldwide. Whereas, most tumors harbored bi-allelic NF1 inactivation as the only genetic abnormality, 11% had additional mutations. Moreover, tumors classified as non-pilocytic astrocytoma based on DNA methylation analysis were significantly more likely to harbor these additional mutations. The most common secondary alteration was FGFR1 mutation, which conferred an additional growth advantage in multiple complementary experimental murine Nf1 models. Taken together, this comprehensive characterization has important implications for the management of children with NF1-LGG, distinct from their sporadic counterparts.

BackgroundPrognostic markers for meningioma are needed to risk-stratify patients and guide postoperative surveillance and adjuvant therapy.ObjectiveTo identify a prognostic gene signature for meningioma recurrence and mortality after resection using targeted gene-expression analysis.MethodsTargeted gene-expression analysis was used to interrogate a discovery cohort of 96 meningiomas and an independent validation cohort of 56 meningiomas with comprehensive clinical follow-up data from separate institutions. Bioinformatic analysis was used to identify prognostic genes and generate a gene-signature risk score between 0 and 1 for local recurrence.ResultsWe identified a 36-gene signature of meningioma recurrence after resection that achieved an area under the curve of 0.86 in identifying tumors at risk for adverse clinical outcomes. The gene-signature risk score compared favorably to World Health Organization (WHO) grade in stratifying cases by local freedom from recurrence (LFFR, P

Treatment failure in glioblastoma is often attributed to intratumoral heterogeneity (ITH), which fosters tumor evolution and generation of therapy-resistant clones. While ITH in glioblastoma has been well-characterized at the genomic and transcriptomic levels, the extent of ITH at the epigenomic level and its biological and clinical significance are not well understood. In collaboration with neurosurgeons, neuropathologists, and biomedical imaging experts, we have established a novel topographical approach towards characterizing epigenomic ITH in three-dimensional (3-D) space. We utilize pre-operative MRI scans to define tumor volume and then utilize 3-D surgical neuro-navigation to intra-operatively acquire 10+ samples representing maximal anatomical diversity. The precise spatial location of each sample is mapped by 3-D coordinates, enabling tumors to be visualized in 360-degrees and providing unprecedented insight into their spatial organization and patterning. For each sample, we conduct assay for transposase-accessible chromatin using sequencing (ATAC-Seq), which provides information on the genomic locations of open chromatin, DNA-binding proteins, and individual nucleosomes at nucleotide resolution. We additionally conduct whole-exome sequencing and RNA sequencing for each spatially mapped sample. Integrative analysis of these datasets reveals distinct patterns of chromatin accessibility within glioblastoma tumors, as well as their associations with genetically defined clonal expansions. Our analysis further reveals how differences in chromatin accessibility within tumors reflect underlying transcription factor activity at gene regulatory elements, including both promoters and enhancers, and drive expression of particular gene expression sets, including neuronal and immune programs. Collectively, this work provides the most comprehensive characterization of epigenomic ITH to date, establishing its importance for driving tumor evolution and therapy resistance in glioblastoma. As a resource for further investigation, we have provided our datasets on an interactive data sharing platform – The 3D Glioma Atlas – that enables 360-degree visualization of both genomic and epigenomic ITH.

The FGFR1 gene encoding fibroblast growth factor receptor 1 has emerged as a frequently altered oncogene in the pathogenesis of multiple low-grade neuroepithelial tumor (LGNET) subtypes including pilocytic astrocytoma (PA), dysembryoplastic neuroepithelial tumor (DNT), rosette-forming glioneuronal tumor (RGNT), and extraventricular neurocytoma (EVN). These activating FGFR1 alterations in LGNET can include tandem duplication of the exons encoding the intracellular tyrosine kinase domain, in-frame gene fusions most often with TACC1 as the partner, or hotspot missense mutations within the tyrosine kinase domain (either p.N546 or p.K656). However, the specificity of these different FGFR1 events for the various LGNET subtypes and accompanying genetic alterations are not well defined, nor are the histopathologic features of pilocytic astrocytomas with FGFR1 alterations versus those harboring the more common BRAF mutations or fusions. Here we performed comprehensive genomic and epigenomic characterization on a diverse cohort of 30 LGNET with FGFR1 alterations. We identified that RGNT harbors a distinct epigenetic signature compared to other LGNET with FGFR1 alterations, and is uniquely characterized by FGFR1 kinase domain hotspot missense mutations in combination with either PIK3CA or PIK3R1 mutation, often with accompanying NF1 or PTPN11 mutation. In contrast, EVN harbors its own distinct epigenetic signature and is characterized by FGFR1-TACC1 fusion as the solitary pathogenic alteration. Additionally, DNT and PA are characterized by either kinase domain tandem duplication or hotspot missense mutations, occasionally with accompanying NF1 or PTPN11 mutation, but lacking the accompanying PIK3CA or PIK3R1 mutation that characterizes RGNT. The glial component of LGNET with FGFR1 alterations typically has a predominantly oligodendroglial morphology, and many of the pilocytic astrocytomas with FGFR1 alterations lack the biphasic pattern, piloid processes, and Rosenthal fibers that characterize pilocytic astrocytomas with BRAF mutation or fusion. Together, this analysis refines the classification and histopathologic spectrum of LGNET with FGFR1 alterations.

BACKGROUND The overall prognostic significance of maximal surgical resection in patients with diffuse low-grade glioma has been well established. Nonetheless, prior studies omit the combined importance of molecular subclass, patient characteristics, and chemoradiation. Similar to findings recently published in newly diagnosed glioblastoma, incorporation of these interactive factors may redefine the relative benefit of cytoreductive surgery. METHODS We examine the interactive effects of volumetric extent of resection with molecular and clinical factors to develop a new roadmap for cytoreductive surgery. Based on a 20-year retrospective cohort of 556 patients with WHO II diffuse low-grade glioma treated with surgery at UCSF 444 had complete records for survival modeling and recursive partitioning (RPA) to investigate multivariate relationships of overall and progression free survival. RESULTS Regardless of molecular subtype, patients with tumor volume under 55cm3 and postoperative volume of residual under 1.9cm3 experience the longest OS (median OS: not reached). Patients with volume of residual over 1.9cm3 experience a OS similar to that of patients with large (over 55cm3) oligodendrogliomas (median OS: not reached). Patients faring worst have large (over 55cm3) astrocytic gliomas (median OS: 84.8 months). Patients not treated with chemotherapy and either ATRX wild-type tumors or ATRX-mutant tumors with small (under 1cm3) volume of residual have the longest PFS together with chemotherapy treated patients who receive either no radiation or radiation for p53-mutant tumors under 30cm3 (median PFS 119 months). Patients with the shortest PFS are under 32-years with larger volume of residual (>1cm3), who receive no chemotherapy for ATRX-mutant tumors together with patients who receive both chemoradiation for larger (>30cm3) p53 mutant tumors (median PFS 30.8 months). CONCLUSION This is the first study to combine extent of resection with molecular and clinical information which paves the way for rethinking surgical strategies for individual patients with newly diagnosed low-grade gliomas.

TheFGFR1gene encoding fibroblast growth factor receptor 1 has emerged as a frequently altered oncogene in the pathogenesis of multiple low-grade neuroepithelial tumor (LGNET) subtypes including pilocytic astrocytoma, dysembryoplastic neuroepithelial tumor (DNT), rosette-forming glioneuronal tumor (RGNT), and extraventricular neurocytoma (EVN). These activatingFGFR1alterations in LGNET can include tandem duplication of the exons encoding the intracellular tyrosine kinase domain, in-frame gene fusions most often withTACC1as the partner, or hotspot missense mutations within the tyrosine kinase domain (either at p.N546 or p.K656). However, the specificity of these differentFGFR1events for the various LGNET subtypes and accompanying genetic alterations are not well defined. Here we performed comprehensive genomic and epigenomic characterization on a diverse cohort of 30 LGNET withFGFR1alterations. We identified that RGNT harbors a distinct epigenetic signature compared to other LGNET withFGFR1alterations, and is uniquely characterized byFGFR1kinase domain hotspot missense mutations in combination with eitherPIK3CAorPIK3R1mutation, often with accompanyingNF1orPTPN11mutation. In contrast, EVN harbors its own distinct epigenetic signature and is characterized byFGFR1-TACC1fusion as the solitary pathogenic alteration. Additionally, DNT and pilocytic astrocytoma are characterized by either kinase domain tandem duplication or hotspot missense mutations, occasionally with accompanyingNF1orPTPN11mutation, but lacking the accompanyingPIK3CAorPIK3R1mutation that characterizes RGNT. The glial component of LGNET withFGFR1alterations typically has a predominantly oligodendroglial morphology, and many of the pilocytic astrocytomas withFGFR1alterations lack the biphasic pattern, piloid processes, and Rosenthal fibers that characterize pilocytic astrocytomas withBRAFmutation or fusion. Together, this analysis improves the classification and histopathologic stratification of LGNET withFGFR1alterations.

BackgroundDifferentiating treatment-induced injury from recurrent high-grade glioma is an ongoing challenge in neuro-oncology, in part due to lesion heterogeneity. This study aimed to determine whether different MR features were relevant for distinguishing recurrent tumor from the effects of treatment in contrast-enhancing lesions (CEL) and non-enhancing lesions (NEL).MethodsThis prospective study analyzed 291 tissue samples (222 recurrent tumor, 69 treatment-effect) with known coordinates on imaging from 139 patients who underwent preoperative 3T MRI and surgery for a suspected recurrence. 8 MR parameter values were tested from perfusion-weighted, diffusion-weighted, and MR spectroscopic imaging at each tissue sample location for association with histopathological outcome using generalized estimating equation models for CEL and NEL tissue samples. Individual cutoff values were evaluated using receiver operating characteristic curve analysis with 5-fold cross-validation.ResultsIn tissue samples obtained from CEL, elevated relative cerebral blood volume (rCBV) was associated with the presence of recurrent tumor pathology (P < 0.03), while increases in normalized choline (nCho) and choline-to-NAA index (CNI) were associated with the presence of recurrent tumor pathology in NEL tissue samples (P < 0.008). A mean CNI cutoff value of 2.7 had the highest performance, resulting in mean sensitivity and specificity of 0.61 and 0.81 for distinguishing treatment-effect from recurrent tumor within the NEL.ConclusionAlthough our results support prior work that underscores the utility of rCBV in distinguishing the effects of treatment from recurrent tumor within the contrast enhancing lesion, we found that metabolic parameters may be better at differentiating recurrent tumor from treatment-related changes in the NEL of high-grade gliomas.

Background Emerging data suggest that a subset of patients with diffuse isocitrate dehydrogenase (IDH)-mutant low-grade glioma (LGG) who receive adjuvant temozolomide (TMZ) recur with hypermutation in association with malignant progression to higher-grade tumors. It is currently unclear why some TMZ-treated LGG patients recur with hypermutation while others do not. MGMT encodes O6-methylguanine-DNA methyltransferase, a DNA repair protein that removes cytotoxic and potentially mutagenic lesions induced by TMZ. Here, we hypothesize that epigenetic silencing of MGMT by promoter methylation facilitates TMZ-induced mutagenesis in LGG patients and contributes to development of hypermutation at recurrence. Methods We utilize a quantitative deep sequencing assay to characterize MGMT promoter methylation in 109 surgical tissue specimens from initial tumors and post-treatment recurrences of 37 TMZ-treated LGG patients. We utilize methylation arrays to validate our sequencing assay, RNA sequencing to assess the relationship between methylation and gene expression, and exome sequencing to determine hypermutation status. Results Methylation level at the MGMT promoter is significantly higher in initial tumors of patients that develop hypermutation at recurrence relative to initial tumors of patients that do not (45.7% vs 34.8%, P = 0.027). Methylation level in initial tumors can predict hypermutation at recurrence in univariate models and multivariate models that incorporate patient age and molecular subtype. Conclusions These findings reveal a mechanistic basis for observed differences in patient susceptibility to TMZ-driven hypermutation. Furthermore, they establish MGMT promoter methylation level as a potential biomarker to inform clinical management of LGG patients, including monitoring and treatment decisions, by predicting risk of hypermutation at recurrence.

Brain tumors are the most common solid tumors of childhood, and the genetic drivers and optimal therapeutic strategies for many of the different subtypes remain unknown. Here, we identify that bithalamic gliomas harbor frequent mutations in the EGFR oncogene, only rare histone H3 mutation (in contrast to their unilateral counterparts), and a distinct genome-wide DNA methylation profile compared to all other glioma subtypes studied to date. These EGFR mutations are either small in-frame insertions within exon 20 (intracellular tyrosine kinase domain) or missense mutations within exon 7 (extracellular ligand-binding domain) that occur in the absence of accompanying gene amplification. We find these EGFR mutations are oncogenic in primary astrocyte models and confer sensitivity to specific tyrosine kinase inhibitors dependent on location within the kinase domain or extracellular domain. We initiated treatment with targeted kinase inhibitors in four children whose tumors harbor EGFR mutations with encouraging results. This study identifies a promising genomically-tailored therapeutic strategy for bithalamic gliomas, a lethal and genetically distinct brain tumor of childhood.

Background Oncolytic measles virus (MV) is effective in xenograft models of many tumor types in immune-compromised mice. However, no murine cell line exists that is tumorigenic, grows in immune-competent mice, and is killed by MV. The lack of such a model prevents an examination of the effect of the immune system on MV oncotherapy. Methods Cerebellar stem cells from human CD46-transgenic immunocompetent mice were transduced to express Sendai virus C-protein, murine C-Myc, and Gfi1b proteins. The resultant cells were injected into the brain of NSG mice, and a cell line, called CSCG, was prepared from the resulting tumor. Results CSCG cells are highly proliferative, and express stem cell markers. These cells are permissive for replication of MV and are killed by the virus in a dose- and time-dependent manner. CSCG cells form aggressive tumors that morphologically resemble medulloblastoma when injected into the brains of immune-competent mice. On the molecular level, CSCG tumors overexpress natriuretic peptide receptor 3 and gamma-aminobutyric acid type A receptor alpha 5, markers of Group 3 medulloblastoma. A single intratumoral injection of MV‒green fluorescent protein resulted in complete tumor regression and prolonged survival of animals compared with treatments with phosphate buffered saline (P = 0.0018) or heat-inactivated MV (P = 0.0027). Conclusions This immune-competent model provides the first platform to test therapeutic regimens of oncolytic MV for Group 3 medulloblastoma in the presence of anti-measles immunity. The strategy presented here can be used to make MV-sensitive murine models of any human tumor for which the driving mutations are known.

Author(s): Pekmezci, Melike; Stevers, Meredith; Phillips, Joanna J; Van Ziffle, Jessica; Bastian, Boris C; Tsankova, Nadejda M; Kleinschmidt-DeMasters, Bette K; Rosenblum, Marc K; Tihan, Tarik; Perry, Arie; Solomon, David A

TERT promoter mutation (TPM), found in over 80% of IDH-wildtype glioblastomas (GBMs) and oligodendrogliomas, leads to reactivation of telomerase and consequently tumor cell immortalization, which is potentially reversible. TERT could therefore serve as an effective target in treating tumors with TPM, if TPM is present throughout the tumor. Previous studies using a single sample or minimal sampling per tumor have shown potentially conflicting results, suggesting TPM is clonal in some cases and subclonal in others. Here we use spatially mapped tumor samples representing maximal tumor sampling to address this critical issue. Sanger sequencing was performed on 311 newly diagnosed and recurrent tumor samples from 19 IDH-wildtype GBMs and 10 oligodendrogliomas. To validate Sanger sequencing and resolve potentially ambiguous samples, deep amplicon sequencing was performed on 164 samples. To determine tumor purity and TERT expression levels, whole exome sequencing (164 samples) and RNA-Seq (129 samples) data sets were analyzed computationally. Sanger and amplicon sequencing showed that TPM was present in 305 of 311 samples (98.1%). TPM was not detected in 6 samples which had tumor purity estimates too low to be accurately determined by FACETS and lacked evidence of any driver mutation. Variant allele frequencies (VAFs) of TPM showed high positive correlation with those of clonal alterations in GBMs (r(90) = .93, p < .0001) and oligodendrogliomas (r(48) = .96, p < .0001). TPM VAFs also showed high positive correlation with tumor purity in both GBMs (r(112) = .92, p < .0001) and oligodendrogliomas (r(48) = .89, p < .0001). TPM VAF showed a moderate positive correlation with TERT expression in GBMs (r(78) = .40, p < .001) and oligodendrogliomas (r(47) = .49, p < .001). Therefore, TPM is a tumor-wide, clonal mutation in both newly diagnosed and recurrent GBMs and oligodendrogliomas. TPM VAF is moderately correlated with TERT expression.

BACKGROUND The prognosis of WHO grade II low-grade gliomas (LGG) is varied with potential for long survival.Given their relatively intact immune system and slow growth rate, vaccines are an attractive treatment strategy for LGG in an attempt to defer more toxic treatments. The goals of this pilot study were to evaluate safety and immunological effects of vaccination with GBM6-AD, an allogeneic glioblastoma stem cell line lysate, with poly-ICLC in LGG. METHODS Eligible patients were ≥ 18 years old, ≥ 70 KPS, with recurrent LGG or imaging consistent with LGG, and amenable to resection. Patients were randomized to vaccine prior to surgery (Arm 1) or not (Arm 2) and all received adjuvant vaccine. Co-primary outcomes were safety and immune response in the tumor, with exploratory outcomes of survival and immunologic effects in peripheral blood. RESULTS A total of 17 eligible patients were evaluable – nine into Arm 1 and eight into Arm 2. Median age was 33 years, with median time from initial diagnosis of 4.7 years (0 – 20). Two patients (11.8%) previously received radiotherapy and seven (41.2%) prior systemic therapy. No dose limiting toxicities or grade 3 AEs were observed. Neoadjuvant vaccination induced up regulation of type-1 cytokines and chemokines in peripheral blood, and CD8+ T cell clones that reacted to the vaccine were also detected in the tumor. Median follow-up time from first post-operative vaccine was 20.8 months with median PFS of 11.0 months and time to change in therapy of 23.7 months. Of the six patients to receive additional treatment, three had second surgery only one confirming malignant progression to anaplastic oligodendroglioma. CONCLUSION Treatment was well-tolerated with no regimen-limiting toxicity. GBM6-AD plus poly-ICLC induced effector CD8+ T cell response in peripheral blood and enables some vaccine-reactive CD8+ T cells to migrate into the TME. Further investigation is warranted.

Familial melanoma-astrocytoma syndrome is a tumor predisposition syndrome caused by inactivating germline alteration of theitalicCDKN2A/italictumor suppressor gene on chromosome 9p21. While some families with germlineitalicCDKN2A/italicmutations are prone to development of just melanomas, other families develop both melanomas, astrocytomas, and occasionally other nervous-system neoplasms including peripheral nerve sheath tumors and meningiomas. The histologic spectrum of the astrocytomas that arise as part of this syndrome is not well described, nor are the additional genetic alterations that drive these astrocytomas apart from the germlineitalicCDKN2A/italicinactivation. Herein, we report the case of a young man with synchronous development of a pleomorphic xanthoastrocytoma, diffuse astrocytoma, and paraspinal mass radiographically consistent with a peripheral nerve sheath tumor. His paternal family history is significant for melanoma, glioblastoma, and oral squamous cell carcinoma. Genomic profiling revealed that he harbors a heterozygous deletion in the germline of chromosome 9p21.3 encompassing theitalicCDKN2A/italicanditalicCDKN2B/italictumor suppressor genes. Both the pleomorphic xanthoastrocytoma and diffuse astrocytoma were found to have homozygous deletion ofitalicCDKN2A/B/italicdue to somatic loss of the other copy of chromosome 9p containing the remaining intact alleles. Additional somatic alterations includeditalicBRAF/italicp.V600E mutation in the pleomorphic xanthoastrocytoma anditalicPTPN11/italic,italicATRX/italic, anditalicNF1/italicmutations in the diffuse astrocytoma. The presence of germlineitalicCDKN2A/B/italicinactivation together with the presence of multiple anatomically, histologically, and genetically distinct astrocytic neoplasms, both with accompanying somatic loss of heterozygosity for theitalicCDKN2A/B/italicdeletion, led to a diagnosis of familial melanoma-astrocytoma syndrome. This remarkable case illustrates the histologic and genetic diversity that astrocytomas arising as part of this rare glioma predisposition syndrome can demonstrate. .

BACKGROUND Unlike cancers affecting many solid organs, gliomas exist within the context of complex neural circuitry. It remains unknown whether glioma-neuron interactions play a role in maintaining functional circuits underlying cognition. We test the hypothesis that malignant gliomas remodel functional circuits through glioma-neuron interactions. METHODS Using language processing as a model for functional circuit dynamics, we enrolled 53 patients with dominant hemisphere IDH-wild-type glioblastoma. Task related circuit dynamics were measured using electrocorticography. Magnetoencephalography measures of functional connectivity identified intratumoral connectivity (HFC) and suppressed connectivity (LFC) regions. Primary patient samples and cultures from HFC and LFC-sites were assessed by single-cell RNA sequencing, pre/post-synaptic marker expression, cocultured with murine hippocampal neurons, and induced neuron organoids. Hippocampal tumor xenografts were created. Language/survival statistics were performed to correlate with functional connectivity measures. RESULTS Speech production evokes neuronal population spikes within the entire area of tumor-infiltrated cortex, far beyond the cortical territory normally involved in expressive language. Primary patient samples from HFC-regions are enriched for glioblastoma cells with a synaptogenic profile as characterized by pre-and post-synaptic marker expression at both tissue and cellular levels. RNA-sequencing and proteomic analyses from HFC samples revealed a neurogenic signature including thrombospondin-1 originating from glioma cells in HFC-regions and non-tumor astrocytes in LFC-regions. HFC xenografts demonstrated increased total number of synapses. Importantly, when compared with gliomas without intratumoral functional connectivity, connected gliomas have worse language task performance (r= -0.54,p=0.03) and shorter OS (medianOS-64 weeks compared with 107-weeks,p=0.04). CONCLUSION Glioma infiltrated regions generate task-relevant neural responses, with speech production evoking neuronal activity throughout tumor-involved cortex in the dominant hemisphere. An enriched population of synaptogenic glioma cells are organized within functionally connected intratumoral regions and this confers negative functional and survival outcomes. Together, these findings indicate that malignant gliomas can functionally remodel neural circuitry, thereby impairing neurological function and promoting tumor progression.

Author(s): Calixto-Hope, Lucas; Lee, Julieann; Sloan, Emily; Hofmann, Jeffrey; Van Ziffle, Jessica; Onodera, Courtney; Grenert, James; Devine, Patrick; Kline, Cassie; Banerjee, Anu; Clarke, Jennifer; Taylor, Jennie; Ann Oberheim-Bush, Nancy; Buerki, Robin; Butowski, Nicholas; Chang, Susan; McDermott, Mike; Aghi, Manish; Theodosopoulos, Philip; Hervey-Jumper, Shawn; Berger, Mitchel; Raffel, Corey; Gupta, Nalin; Kleinschmidt-DeMasters, Bette; Wood, Matthew; Grafe, Marjorie; Guo, Hua; Sun, Peter; Torkildson, Joseph; Cooney, Tabitha; Fata, Cynthia; Scharnhorst, David; Samuel, David; Bannykh, Serguei; Khatib, Ziad; Maher, Ossama; Chamyan, Gabriel; Pelaez, Liset; Brathwaite, Carole; Jin, Lee-way; Lechpammer, Mirna; Born, Donald; Vogel, Hannes; Lee, Han; Phillips, Joanna; Pekmezci, Melike; Bollen, Andrew; Tihan, Tarik; Perry, Arie; Solomon, David | Abstract: Abstract BACKGROUND Rosette-forming glioneuronal tumor (RGNT) is an uncommon CNS tumor originally described in the fourth ventricle characterized by a low-grade glial neoplasm admixed with a rosette-forming neurocytic component. METHODS We reviewed clinicopathologic features of 42 patients with RGNT. Targeted next-generation sequencing was performed, and genome-wide methylation profiling is underway. RESULTS The 20 male and 22 female patients had a mean age of 25 years (range 3–47) at time of diagnosis. Tumors were located within or adjacent to the lateral ventricle (n=16), fourth ventricle (15), third ventricle (9), and spinal cord (2). All 31 tumors assessed to date contained FGFR1 activating alterations, either in-frame gene fusion, kinase domain tandem duplication, or hotspot missense mutation in the kinase domain (p.N546 or p.K656). While 7 of these 31 tumors harbored FGFR1 alterations as the solitary pathogenic event, 24 contained additional pathogenic alterations within PI3-kinase or MAP kinase pathway genes: 5 with additional PIK3CA and NF1 mutations, 4 with PIK3CA mutation, 3 with PIK3R1 mutation (one of which also contained focal RAF1 amplification), 5 with PTPN11 mutation (one with additional PIK3R1 mutation), and 2 with NF1 deletion. The other 5 cases demonstrated anaplastic features including hypercellularity and increased mitotic activity. Among these anaplastic cases, 3 harbored inactivating ATRX mutations and two harbored CDKN2A homozygous deletion, in addition to the FGFR1 alterations plus other PI3-kinase and MAP kinase gene mutations seen in those RGNT without anaplasia. CONCLUSION Independent of ventricular location, RGNT is defined by FGFR1 activating mutations or rearrangements, which are frequently accompanied by mutations involving PIK3CA, PIK3R1, PTPN11, NF1, and KRAS. Whereas pilocytic astrocytoma and ganglioglioma are characterized by solitary activating MAP kinase pathway alterations (e.g. BRAF fusion or mutation), RGNT are genetically more complex with dual PI3K-Akt-mTOR and Ras-Raf-MAPK pathway activation. Rare anaplastic examples may show additional ATRX and/or CDKN2A inactivation.

The design and evaluation of immunotherapies in IDH-mutant lower grade gliomas (LGG) is hindered by a poor understanding of the LGG T-cell repertoire. We present data on the temporal evolution, intratumoral spatial distribution, and prognostic value of the T-cell repertoire in IDH-mutant LGGs. We performed immunogenomic profiling using T-cell receptor beta-chain sequencing of 163 glioma and peripheral blood samples from 33 immunotherapy-naive glioma patients (22 astrocytomas, 11 oligodendrogliomas). T-cell repertoire evolution was analyzed in a subset of 26 patients (69 samples) with matched primary (WHO grade II) and recurrent (WHO grade II-IV) glioma samples. T-cell repertoire diversity was defined as the number of unique T-cell clonotypes by V-gene, J-gene, and CDR3 nucleotide sequences. Malignant transformed (Grade III or IV) recurrent gliomas demonstrated increased T-cell repertoire diversity compared to their patient-matched primary tumors (p=0.0023), but grade II recurrences did not show the same increased diversity (p=0.26). This increase in T-cell repertoire diversity was greater in patients who underwent transformation in the context of TMZ-associated hypermutation compared to spontaneously transformed counterparts (p=0.035). In grade II primary astrocytomas (n=17), T-cell repertoire diversity above the median (186 unique T-cell clonotypes per sample) was associated with worse transformation-free (HR=4.2, p=0.045) and overall survival (HR=6.4, p=0.025). Next, we evaluated intratumoral immune heterogeneity in 7 patients by sampling from up to 10 distinct and maximally-separated intratumoral sites per LGG (64 samples). Eighty-two to 96% of unique clonotypes within a given tumor were present only within a single sampled site. Despite this heterogeneity, six LGG patients harbored T-cell clonotypes present tumor-wide across all sampled sites within a given tumor. Ten of 24 (42%) tumor-wide T-cell clonotypes were enriched in the glioma compared to matched peripheral blood, suggesting glioma-specificity. Taken together, T-cell receptor profiling in LGGs may have utility both as a prognostic biomarker and to identify glioma-specific T-cells.

BackgroundChromosomal instability is associated with earlier progression in isocitrate dehydrogenase (IDH)–mutated astrocytomas. Here we evaluated the prognostic significance of polysomy in gliomas tested for 1p/19q status.MethodsWe analyzed 412 histologic oligodendroglial tumors with use of 1p/19q testing at 8 institutions from 1996 to 2013; fluorescence in situ hybridization (FISH) for 1p/19q was performed. Polysomy was defined as more than two 1q and 19p signals in cells. Tumors were divided into groups on the basis of their 1p/19q status and polysomy and were compared for progression-free survival (PFS) and overall survival (OS).ResultsIn our cohort, 333 tumors (81%) had 1p/19q loss; of these, 195 (59%) had concurrent polysomy and 138 (41%) lacked polysomy, 79 (19%) had 1p/19q maintenance; of these, 30 (38%) had concurrent polysomy and 49 (62%) lacked polysomy. In agreement with prior studies, the group with 1p/19q loss had significantly better PFS and OS than did the group with 1p/19q maintenance (P < 0.0001 each). Patients with 1p/19q loss and polysomy showed significantly shorter PFS survival than patients with 1p/19q codeletion only (P < 0.0001), but longer PFS and OS than patients with 1p/19q maintenance (P < 0.01 and P < 0.0001). There was no difference in survival between tumors with >30% polysomic cells and those with ConclusionsThe presence of polysomy in oligodendroglial tumors with codeletion of 1p/19q predicts early recurrence and short survival in patients with 1p/19q codeleted tumors.

Author(s): Lee, Julieann C; Mazor, Tali; Lao, Richard; Wan, Eunice; Diallo, Alpha B; Hill, Nicholas S; Thangaraj, Naina; Wendelsdorf, Katherine; Samuel, David; Kline, Cassie N; Banerjee, Anuradha; Auguste, Kurtis; Raffel, Corey; Gupta, Nalin; Berger, Mitchel; Raleigh, David R; Shai, Anny; Phillips, Joanna J; Bollen, Andrew W; Tihan, Tarik; Perry, Arie; Costello, Joseph; Solomon, David A

High-grade neuroepithelial tumor with BCOR exon 15 internal tandem duplication (HGNET BCOR ex15 ITD) is a recently proposed tumor entity of the central nervous system (CNS) with a distinct methylation profile and characteristic genetic alteration. The complete spectrum of histologic features, accompanying genetic alterations, clinical outcomes, and optimal treatment for this new tumor entity are largely unknown. Here, we performed a comprehensive assessment of ten new cases of HGNET BCOR ex15 ITD. The tumors mostly occurred in young children and were located in the cerebral or cerebellar hemispheres. On imaging all tumors were large, well-circumscribed, heterogeneous masses with variable enhancement and reduced diffusion. They were histologically characterized by predominantly solid growth, glioma-like fibrillarity, perivascular pseudorosettes, and palisading necrosis, but absence of microvascular proliferation. They demonstrated sparse to absent GFAP expression, no synaptophysin expression, variable OLIG2 and NeuN positivity, and diffuse strong BCOR nuclear positivity. While BCOR exon 15 internal tandem duplication was the solitary pathogenic alteration identified in six cases, four cases contained additional alterations including CDKN2A/B homozygous deletion, TERT amplification or promoter hotspot mutation, and damaging mutations in TP53, BCORL1, EP300, SMARCA2, and STAG2. While the limited clinical follow-up in prior reports had indicated a uniformly dismal prognosis for this tumor entity, this cohort includes multiple long-term survivors. Our study further supports inclusion of HGNET BCOR ex15 ITD as a distinct CNS tumor entity and expands the known clinicopathologic, radiographic, and genetic features.

Pleomorphic xanthoastrocytoma (PXA) is an astrocytic neoplasm that is typically well circumscribed and can have a relatively favorable prognosis. Tumor progression to anaplastic PXA (WHO grade III), however, is associated with a more aggressive biologic behavior and worse prognosis. The factors that drive anaplastic progression are largely unknown. We performed comprehensive genomic profiling on a set of twenty-three PXAs from 19 patients, including 15 with anaplastic PXA. Four patients had tumor tissue from multiple recurrences, including two with anaplastic progression. We find that PXAs are genetically defined by the combination of CDKN2A biallelic inactivation and RAF alterations that were present in all 19 cases, most commonly as CDKN2A homozygous deletion and BRAF p.V600E mutation but also occasionally BRAF or RAF1 fusions or other rearrangements. The third most commonly altered gene in anaplastic PXA was TERT, with 47% (7/15) harboring TERT alterations, either gene amplification (n=2) or promoter hotspot mutation (n=5). In tumor pairs analyzed before and after anaplastic progression, two had increased copy number alterations and one had TERT promoter mutation at recurrence. Less commonly altered genes included TP53, BCOR, BCORL1, ARID1A, ATRX, PTEN, and BCL6. All PXA in this cohort were IDH and histone H3 wildtype, and did not contain alterations in EGFR. Genetic profiling performed on six regions from the same tumor identified intratumoral genomic heterogeneity, likely reflecting clonal evolution during tumor progression. Overall, anaplastic PXA is characterized by the combination of CDKN2A biallelic inactivation and oncogenic RAF kinase signaling as well as a relatively small number of additional genetic alterations, with the most common being TERT amplification or promoter mutation. These data define a distinct molecular profile for PXA and suggest additional genetic alterations, including TERT, may be associated with anaplastic progression. This article is protected by copyright. All rights reserved.

PBTC-042 was a phase I trial of palbociclib to determine the maximum tolerated dose (MTD) and describe toxicities in children. Palbociclib is an oral, selective cyclin dependent kinase 4/6 inhibitor. METHODS: A rolling-6 design was utilized. Eligible patients were children ≥4 and ≤21 years-old with a progressive/refractory CNS tumor with intact retinoblastoma protein, measurable disease, and ability to swallow capsules. Pharmacokinetic studies were performed during the first course. Here, we report on the heavily pretreated stratum, which included patients who received >4 prior treatment regimens (either chemotherapy or biologic agent), and/or craniospinal irradiation, and/or myeloablative chemotherapy plus stem cell rescue. Palbociclib was initiated at 50 mg/m2/day for 21 consecutive days of a 28-day course. This was one dosage level below the MTD for the less heavily pretreated stratum (75 mg/m2). RESULTS: Fourteen eligible patients were enrolled (median age 12.8 years; male 79%). Eleven patients (79%) had either ependymoma or medulloblastoma. Four eligible and evaluable patients were enrolled at 50 mg/m2 with no DLTs. This prompted a dosage increase to 75 mg/m2. Ten eligible subjects were enrolled and 7 were evaluable for DLT assessment. One of 7 evaluable patients experienced a DLT (grade 3 thrombocytopenia). This established 75 mg/m2 as the MTD for more heavily pretreated patients. Mean ± SD palbociclib apparent oral clearance was 34.6 ± 18.4 L/h/m2. CONCLUSION: The MTD for palbociclib on a 3 week on/1 week off schedule in children with brain tumors is 75 mg/m2 and does not appear to be influenced by the degree of prior therapy.