Journal of Neuro-Oncology

Background: Extent of resection remains central to meningioma management, yet Simpson grading is subjective and may not reflect measurable postoperative residual disease. We compared surgeon-reported Simpson grade, report-derived radiological grading, and residual tumour volumetry across a multicentre cohort. Methods: We performed a retrospective study across two tertiary neurosciences centres comprising four hospitals, including patients undergoing primary cranial meningioma resection from 2006 to 2025. Postoperative magnetic resonance imaging (MRI) reports were harmonised using weakly supervised natural language processing based on term frequency-inverse document frequency (TF-IDF) and a linear support vector machine classifier. Residual tumour volume was segmented from contrast-enhanced postoperative MRI and log-transformed. Concordance between Simpson and radiological gross-total/subtotal resection classification was assessed using absolute agreement and prevalence-adjusted bias-adjusted kappa (PABAK). Cox models assessed recurrence-free survival, with bootstrap validation and anatomical and scan-timing sensitivity analyses. Results: Among 912 patients, recurrence or residual progression occurred in 281. Surgical-radiological agreement was substantial but imperfect (absolute agreement 74%; PABAK 0.61), with lower agreement in skull-base and parafalcine-parasagittal tumours. In adjusted models, recurrence hazard increased with Simpson grade (hazard ratio 1.54, 95% confidence interval 1.37-1.72), radiological grade (1.92, 1.68-2.20), and log-transformed residual volume (1.20, 1.16-1.24; all p<0.0005). Optimism corrected concordance increased from Simpson grade to radiological grade and log-volumetry (0.692, 0.733, and 0.748), with this ranking preserved across sensitivity analyses. Conclusions: Imaging-based postoperative residual disease measures outperformed Simpson grade. TF-IDF-assisted report-derived grading provides a scalable bridge to volumetry, while quantitative residual volume offers the strongest prognostic representation.

PurposeMaas et al. recently showed that a microenvironment-determined risk continuum, driven by the shift from microglia-like to myeloid-derived macrophage-like tumor-associated macrophages (TAMs), independently predicts meningioma progression beyond WHO grade. Whether this gradient is recoverable from bulk RNA-seq has not been tested. MethodsWe computed a microglia-to-macrophage ssGSEA ratio using expanded gene sets (15 microglia, 17 macrophage) anchored to Maas core markers across 968 meningiomas from 5 GEO datasets, validated it against pseudo-bulk profiles from the Maas snRNA-seq cohort (n=25), and tested recurrence-free survival (RFS) association by Cox regression in a 101-patient subset (73 events, median follow-up 110.2 months). ResultsThe ratio correlated with single-cell microglia proportion (overlap-controlled r=0.70, 95% CI 0.42-0.86) and discriminated WHO grades and transcriptomic clusters, confirming biological recoverability. The ratio did not predict RFS (HR 0.92, 95% CI 0.72-1.16, p=0.46). A quantitative attenuation analysis predicts the Maas IHC HR of 2.00 attenuates to HR 1.24-1.40 after proxy measurement error (r2=0.22-0.49) and NF2-wildtype dilution (30-45%), yielding only 15-40% power at 73 events. An exploratory NF2-expression proxy subgroup (uncorrected p=0.056) showed a trend in NF2-low tumors (HR 0.68, 95% CI 0.46-1.01) absent in NF2-high tumors (HR 0.98, p=0.89). The Chen 34-gene tumor-intrinsic panel also reached near-chance discrimination (C-index 0.552). ConclusionSingle-cell-anchored ssGSEA recovers the Maas gradient in bulk RNA-seq but attenuates it below detectability in moderate-sized, NF2-unselected cohorts. The prognostic component is bounded by power and NF2 stratification, not an intrinsic modality failure; NF2-annotated cohorts with approximately 480 events are required for definitive testing.

Background/ObjectivesIncreasing evidence indicates that gliomas co-opt mechanisms of excitatory synaptic transmission and plasticity to support tumor progression, yet these processes remain poorly characterized in lower-grade gliomas (LGGs). Here, we investigated whether genes associated with excitatory synaptic function are linked to patient prognosis in LGG. MethodsA curated panel of 36 synaptic genes was analyzed in LGG using RNA-sequencing and clinical data from The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) datasets. Correlations among gene expression levels were analyzed using the Evergene platform. ResultsAmong the genes investigated, DLG2, DLG3, and DLG4, which encode the postsynaptic scaffolding proteins PSD-93, SAP-102, and PSD-95, respectively, showed strong associations with patient overall survival (OS). Higher expression of each gene was consistently associated with longer OS across both datasets. Expression of DLG2-DLG4 was higher in oligodendroglioma and IDH-mutant, 1p/19q co-deleted tumors, and lower in astrocytoma and IDH-wild-type tumors. Furthermore, expression of all three genes positively correlated with a broad gene signature related to excitatory synaptic transmission and synaptic plasticity, including multiple components of glutamatergic signaling and postsynaptic organization. ConclusionsThese findings suggest that elevated expression of DLG2-DLG4 is associated with a transcriptional program resembling differentiated neuronal-like features and favorable clinical outcome in LGG. Simple SummaryLower-grade gliomas are brain tumors with highly variable outcomes, and better markers are needed to predict how patients will fare. Recent research suggests that these tumors may use mechanisms normally involved in communication between brain cells, but this is not well understood in these cancer types. In this study, we analyzed large patient datasets to examine genes related to synaptic function. We found that higher expression of three genes involved in synaptic membrane organization, DLG2, DLG3, and DLG4 was consistently associated with longer patient survival. These genes were also linked to a broader pattern of gene expression suggestive of neural transmission and plasticity. Our findings suggest that some lower-grade gliomas may adopt characteristics of normal brain cells that are associated with less aggressive behavior. This work may help guide future research on prognostic markers and improve understanding of brain tumor biology.

Glioblastomas integrate into the brain globally, where they provoke neuronal hyperactivity to enhance tumor growth and invasion. Communication of glioblastomas with neurons is not only present locally, but has preclinically been shown to extend towards the contralateral hemisphere through white matter tracts. However, it remains unknown how the distant hyperactivity that is often found in patients relates to structural embedding of the tumor into the larger brain network. 29 newly diagnosed IDH-wildtype glioblastoma patients and 25 age and sex matched healthy controls were included. To define structural tumor embedding, we overlayed each patient-specific tumor mask with a normative structural connectome obtained from diffusion MRI. We identified the average number of streamlines intersecting the tumor, extracting the tumors average tract density ( Lesion-Tract Density Index, L-TDI). For a subgroup of patients (n = 17), we determined structural embedding directly from diffusion scans and subsequent tractography. To identify regions connecting to the tumor, we seeded from each patients tumor rim outside FLAIR hyperintensities in the white matter to the 210 cortical regions of the Brainnetome atlas. We then counted the number of tumor-connecting regions, termed PATNET hereafter. Finally, participants underwent eyes-closed resting-state magnetoencephalography. We used broadband power as a proxy for neuronal spiking activity of each cortical region. To capture deviant brain activity in tumor and non-tumor regions, we regionally standardized broadband values using controls. We then sought to establish an association of deviant peritumor activity with both L-TDI and PATNET. Subsequently, we investigated whether tumor-connected regions showed more deviant activity than unconnected regions. Finally, we explored the clinical relevance of L-TDI and PATNET. Greater structural tumor embedding significantly related to more deviant peritumor activity (rhoLTDI= 0.47, PLTDI = .010; rhoPATNET = 0.54, PPATNET = .024), with larger tumors showing greater embedding and more hyperactivity than smaller tumors. Furthermore, distant tumor-connected regions showed more hyperactivity than unconnected regions, but only in patients with peritumor hyperactivity (F(1,15) = 11.02, P =.005). Finally, higher PATNET associated with lower KPS (U = 61.5, P = .015). Glioblastomas structural embedding explains hyperactivity around the tumor and in distant cortical regions, such that distant hyperactivity occurs primarily when there is tumor hyperactivity and the region is structurally connected to the tumor. Moreover, patient-specific tumor embedding relates to functional status.

IntroductionCraniopharyngioma (CP) comprises two distinct histological subtypes, adamantinomatous craniopharyngioma (ACP) and papillary craniopharyngioma (PCP), which are often challenging to distinguish preoperatively. Approximately 95% of PCP harbor the BRAF V600E mutation, whereas ACP lacks this alteration, making PCP uniquely sensitive to BRAF and MEK inhibition. However, in the absence of a reliable preoperative classification strategy, targeted therapy has been limited to recurrent disease or to cases with histological confirmation. This study aims to describe and prospectively evaluate a pragmatic preoperative classification strategy and short-course neoadjuvant BRAF and MEK inhibition followed by surgery in newly diagnosed, preoperatively classified PCP. Methods and analysisThis is a prospective, single-arm, open-label study. Patients with newly diagnosed craniopharyngioma will be screened using an integrated preoperative strategy combining imaging-based prediction and selective cerebrospinal fluid (CSF) cell-free DNA testing for BRAF V600E in indeterminate cases. Twelve participants preoperatively predicted as PCP and BRAF V600E positive will receive dabrafenib 150 mg twice daily plus trametinib 2 mg once daily for up to three 28-day cycles, followed by transnasal endoscopic surgery. Assessments are scheduled at days 7, 14, 28, 56, and 84 until surgery. The primary endpoint is objective response rate, assessed by contrast-enhanced MRI using RANO 2.0 criteria. Secondary outcomes include progression-free survival, local disease control, endocrine outcomes of the hypothalamic-pituitary-adrenal and hypothalamic-pituitary-thyroid axes, visual and cognitive outcomes, postoperative diabetes insipidus, surgical complexity, and concordance between the preoperative classification strategy and postoperative pathology and BRAF V600E status. Exploratory analyses will evaluate treatment-related changes in tumor vascularity, tissue characteristics, and post-treatment molecular alterations in tumor tissue. Ethics and disseminationThis protocol has been approved by the Ethics Committee of Huashan Hospital, Fudan University (KY2024-028). Written informed consent will be obtained from all participants. Results will be disseminated through peer-reviewed publications and scientific conferences. Trial registration numberChiCTR2400081636 STRENGTHS AND LIMITATIONS OF THIS STUDYO_ST_ABSStrengthC_ST_ABS[tpltrtarr] This study proposes an integrated, clinically applicable preoperative strategy that combines imaging-based prediction with selective cerebrospinal fluid cell-free DNA analysis to identify papillary craniopharyngioma (PCP) prior to surgery. [tpltrtarr]It prospectively evaluates short-course neoadjuvant BRAF and MEK inhibition in newly diagnosed PCP, addressing a clinically relevant gap in current management. [tpltrtarr]Standardized, multidimensional assessments are performed across the neoadjuvant, perioperative, and early postoperative periods, capturing radiographic, surgical, endocrine, visual, and cognitive outcomes. Limitation[tpltrtarr] The single-arm, open-label design without a surgical control group limits direct comparison with upfront surgery. [tpltrtarr]Despite the integrated prediction strategy, preoperative misclassification cannot be excluded entirely.

BackgroundGliomas are heterogeneous tumors with poor outcomes following current therapies, including immunotherapy. The tumor microenvironment (TME) is a critical determinant of therapeutic response in gliomas. We have classified the immune TME of gliomas by multiplex immunofluorescence (mIF). MethodsTissue taken at initial resection from 354 patients with newly-diagnosed glioma grades 2-4 were analyzed using three mIF panels of markers for T, B, and myeloid cells. Tumor cores were characterized by the relative abundances of: (i) 15 primary immune phenotypes, (ii) 96 secondary immune phenotypes, and, (iii) 29 biologically meaningful multi-marker immune phenotypes. ResultsUsing unsupervised cluster analysis of WHO grade 4 gliomas we identified four subtypes , {beta}, {gamma}, and {delta} that were internally reproducible. Immune subtype was characterized by high abundance of antigen-presenting cells (APCs) and low levels of MHC II- monocytes. Subtype {beta} was high in regulatory T cells and myeloid cells, but low in lymphocytes with effector functions. Subtype {gamma} displayed high abundance of immune cell phenotypes, particularly lymphocytes with effector or helper functions. Subtype {delta} was low in lymphoid and myeloid immune phenotypes and APCs, with poorer outcomes. Grade 3 tumors could also be classified into , {beta}, {gamma}, and {delta} subtypes, indicating generalizability of these immune TME subtypes across high grade gliomas. ConclusionsWe have identified internally reproducible criteria for classifying gliomas according to the immune microenvironment, findings that could aid our understanding of the natural progression of low- and high-grade gliomas and inform the rational application of immune-oncologic therapeutic interventions.

Background Despite WHO grade and IDH status, significant survival differences remain in diffuse gliomas. We hypothesized that a brain-aging transcriptomic signature, reflecting neuroinflammation, myeloid infiltration, and synaptic loss, would independently predict survival and allow for molecular reclassification. Methods A neurodegeneration score was derived via PCA of brain MRI volumes from 1,057 OASIS-3 subjects and projected onto 888 TCGA-LGG/GBM (discovery) and 693 CGGA gliomas (validation). A 14-gene signature of glial/myeloid (GFAP, AQP4, TYROBP, TREM2, C1QA, CD68, ITGAM) and neuronal (SYP, DLG4, GRIN1, GRIA1, SNAP25, SYN1, RBFOX3) genes were computed. Elastic-net Cox regression identified a 3-gene panel (C1QA, CD68, GRIA1). Kaplan-Meier, multivariate Cox, decision curve, and single-cell RNA-seq analyses were performed. Results High brain-aging scores predicted poorer overall survival (p < 0.0001) and remained an independent prognostic factor after adjusting for WHO grade and IDH status (z = 4.72, p < 0.001); chronological age was non-significant (p = 0.231). In IDH-mutant gliomas, significance was confirmed in both cohorts (TCGA p = 0.027; CGGA p < 0.0001). Bidirectional reclassification showed high-risk Grade 2 tumors with Grade 3-like survival (p = 0.00089), and indolent Grade 3 tumors resembling Grade 2 by Ki-67. Single-cell RNA-seq confirmed macrophage localization of signature genes; DCA demonstrated net benefit over grade alone at 5-30% probability thresholds. Conclusions A brain-aging transcriptomic signature independently predicts glioma survival beyond WHO grade and IDH status, validated in an independent Chinese cohort, with clinical utility for identifying high-risk Grade 2 and sparing over-treatment of indolent Grade 3 tumors.

Glioblastoma is an aggressive and highly heterogeneous brain tumor with poor prognosis despite multimodal treatment strategies. Understanding the molecular diversity of the disease is essential for improving tumor stratification and identifying potential therapeutic targets. In this study, we investigate whether network-based analysis can reveal biologically meaningful subgroups of glioblastoma tumors. Using RNA sequencing and mutation data from the TCGA-GBM cohort, we constructed patient-specific protein-protein interaction networks based on genes that are differentially expressed or harbor somatic mutations. These networks capture the molecular alterations associated with individual tumors within the context of the human interactome. We then derived similarities between tumors using a binary representation of network nodes and the Jaccard similarity metric, enabling the construction of a patient similarity graph. Community detection algorithms (Louvain and Leiden) were applied to this graph to identify clusters of tumors with similar molecular network profiles. Our analysis revealed six tumor communities characterized by distinct gene compositions and enriched biological processes. For each community, we identified candidate biomarkers and network hubs that may represent potential therapeutic targets. Several of the identified genes correspond to known drug targets, while others represent potential candidates for further investigation. These results illustrate how integrating molecular alterations with network-based modeling can help stratify glioblastoma tumors and uncover molecular mechanisms that may guide the development of more personalized therapeutic strategies.

Background/ObjectivesGlioblastoma is an aggressive primary brain tumor that develops within a chronically low-oxygen microenvironment, yet most preclinical studies are performed under atmospheric oxygen conditions that poorly reflect in vivo physiology. This study investigated how sustained culture under physiological oxygen tension (physioxia; 5% O{square}) influences glioblastoma cell behavior, signaling, and therapeutic response. MethodsMultiple patient-derived glioblastoma models were cultured under normoxia (21% O{square}) or sustained physioxia (5% O{square}) for at least seven days before experimentation. Cell migration, proliferation, cell cycle distribution, expression of the epithelial-to-mesenchymal transition-associated transcription factor Slug (SNAI2), PDGFR{beta}-associated signaling, and sensitivity to 5-fluorouracil were evaluated using transwell migration assays, cell counting, flow cytometry, RT-qPCR, immunoblotting, and BrdU incorporation assays. Additional patient-derived cultures established and maintained continuously under physioxia were used to examine the effects of oxygen history. ResultsSustained physioxia consistently increased migration across all glioblastoma models while reducing proliferation in normoxia-adapted cell lines through increased G0/G1 cell cycle arrest. Physioxia significantly increased Slug expression in all models and enhanced PDGFR{beta}, AKT, and ERK phosphorylation in a cell line-dependent manner. Therapeutic sensitivity to 5-fluorouracil was also altered, with physioxia conferring increased resistance in selected glioblastoma models but not universally. Patient-derived cultures maintained continuously under physioxia retained enhanced migratory capacity and exhibited increased proliferation compared with normoxia, indicating that prior oxygen exposure influences proliferative responses while the pro-migratory phenotype remains conserved. ConclusionsPhysiological oxygen tension is a major regulator of glioblastoma cell behavior, influencing migration, proliferation, signaling, and therapeutic response. These findings demonstrate that conventional normoxic culture conditions can obscure biologically relevant phenotypes and support incorporating physioxia into experimental design to improve the physiological and translational relevance of preclinical glioblastoma research.

Background: Non-invasive diagnosis, reliable recurrence surveillance remain critical unmet needs in gliomas. Glioma induces profound systemic immune alterations despite its anatomical confinement to the central nervous system. Circulating immune cells, particularly monocytes, are key mediators of tumor-host crosstalk and may retain tumor-induced transcriptional imprints. However, their potential clinical utility as blood-based biomarkers for detection and monitoring, remain largely unexplored. Methods and findings: In this study, we performed integrated single-cell RNA sequencing of blood immune cells and demonstrated that circulating CD14+ monocytes are significantly expanded in glioma patients, exhibiting features of differentiation arrest and increased transcriptional plasticity. These cells harbor glioma-specific molecular signatures distinct from those observed in healthy controls and patients with other tumors. Leveraging these findings, we developed an ensemble machine learning diagnostic model based on transcriptomic profiles of circulating CD14+ monocytes (training cohort, n=107), which achieved a mean area under the receiver operating characteristic curve (AUC) of 0.971 during cross-validation. In an independent cohort of 567 participants, the model maintained high diagnostic accuracy, yielding an AUC of 0.877 for distinguishing glioma from controls and other tumors. And it achieved a recurrence detection AUC of 0.969 in 51 postoperative samples. Moreover, in a prospective follow-up study involving 30 glioma patients, lower model-derived scores of postoperation were significantly associated with prolonged progression-free survival (log-rank test, P=0.043), supporting its prognostic utility. Conclusion: We demonstrate circulating CD14+ monocytes undergo glioma-specific transcriptional reprogramming, generating systemic tumor-associated signal captured via transcriptomic profiling. This blood-based diagnostic model provides non-invasive, scalable approach for glioma detection, recurrence surveillance, outcome prediction.

BackgroundVenous thromboembolism (VTE) remains a major cause of perioperative morbidity in cranial neurosurgery, yet clinical practice varies widely, and formal guidelines are inconsistent. Understanding internationally sampled neurosurgical practice is essential for informing consensus and future trials. MethodsAn international, 2-stage cross-sectional, internet-based survey was conducted. Practising neurosurgeons performing elective adult cranial surgery were eligible. Descriptive statistics were used to summarise practice. Responses covered patterns of pre-operative haemostasis decision making, use and timing of mechanical and/or chemical prophylaxis, use of perioperative imaging prior to anticoagulation, and frequency of clinical assessment for VTE. Associations with geographical income status, subspecialty, and years post-certification were statistically tested. Practice heterogeneity was quantified and contextual influence was summarised using mean effect sizes across stratifying variables in order to determine domains of true equipoise. ResultsOf 585 responses, 456 (78%) met criteria for inclusion: representing 322 units across 78 countries (71% high-income). Thirteen per cent reported no departmental VTE plan; 23% followed no guidelines and 12% used multiple. Routine pre-operative testing almost universally included haemoglobin/platelets/haematocrit, with fibrinogen more common in high-income settings. Compared with high-income country respondents, low- and middle-income respondents reported higher haemoglobin transfusion thresholds (>90 g/dL; p<0.001) and shorter antiplatelet interruption (p[&le;]0.03), and less frequent outpatient VTE assessment (p<0.001). Mechanical prophylaxis was common (TEDs 81%, IPC 62%), typically started pre-or intra-operatively. Among those completing the chemoprophylaxis section (n=310), 57% required a CT or MRI scan before LMWH which was then initiated on average 31.4 hours after surgery. 1% of respondents did not routinely use LMWH. Many clinical decisions demonstrated statistical equipoise ie. high heterogeneity with low contextual influence. ConclusionPeri-operative haemostasis and VTE prophylaxis practices in adult elective cranial neurosurgery vary substantially worldwide, with some decisions reflecting geographical or socioeconomic differences and many others reflecting true clinical equipoise rather than contextual determinants. By mapping contemporary real-world practice across diverse health-system contexts, this study provides a necessary empirical foundation for rational trial design and future guideline development.

BackgroundThe Netrin-1 dependence receptor pathway plays critical roles in neural development, but its expression landscape and prognostic significance in glioblastoma (GBM) remain poorly characterized. MethodsSingle-cell RNA-seq data from 148,019 cells across 34 tumors (Neftel et al., 2019) were analyzed to map Netrin-1 pathway gene expression across GBM cellular states. Differential gene expression and pathway enrichment analyses were performed on NEO1-defined subpopulations. Bulk RNA-seq survival analysis was conducted across three independent GBM cohorts TCGA (n=106), CGGA mRNAseq_325 (n=137), and CGGA mRNAseq_693 (n=237), totaling 480 patients. Primary analysis used continuous Cox regression (per-SD hazard ratios); meta-analysis employed fixed-effects inverse-variance weighting. ResultsIn GBM single-cell data, Netrin-1 pathway genes showed state-specific enrichment --NEO1, DCC, NTN1, and RGMB were predominantly expressed in oligodendrocyte-precursor (OPC) and neural-progenitor (NPC) states. Cells positive for NEO1 were enriched for neural differentiation programs (nervous system development, p=9.6x10-; Axon Guidance, p=2.8x10-), whereas NEO1-negative cells were dominated by ribosomal/translational and immune activation programs. In the 3-cohort survival meta-analysis, NTN1 (Netrin-1 ligand) emerged as the sole gene reaching meta-analytic significance as a risk factor (Meta HR=1.163 per SD, 95% CI 1.056-1.281, p=0.0021, I{superscript 2}=0%, 3/3 cohorts concordant), while DCC and RGMB showed directionally consistent protective trends (DCC: Meta HR=0.938, 95% CI 0.858-1.025, p=0.156; RGMB: Meta HR=0.979, 95% CI 0.881-1.087, p=0.686; both 3/3 cohorts concordant). NEO1 itself did not independently predict survival (Meta HR=1.008, 95% CI 0.885-1.147, p=0.910). After Bonferroni correction for 10 genes tested (threshold p<0.005), only NTN1 met strict significance. In exploratory sex-stratified analysis of a single cohort (CGGA 693, n=237), NEO1 and NTN1 exhibited female-specific risk enhancement (NEO1: HR=1.417, p=0.014; NTN1: HR=1.249, p=0.019), with minimal effects in males. UNC5B showed context-dependent risk in MGMT-unmethylated tumors (HR=1.331, p=0.037). These sex-dimorphic findings require independent validation. ConclusionsThe Netrin-1 pathway exhibits divergent prognostic trends in GBM, with NTN1 as a risk factor and DCC trending toward protection--consistent with the dependence receptor model. These findings, which should be interpreted as hypothesis-generating, nominate NTN1 as a candidate therapeutic target and highlight the potential importance of sex-stratified evaluation in future Netrin-1-directed trials. Independent replication in larger cohorts is warranted.

BackgroundGlioblastoma is an aggressive and incurable brain tumor. Clinical trials of immune checkpoint inhibitors showed no clinical benefit in glioblastoma when given after surgery. However, a clinical trial in which PD1 inhibition was given prior to second surgery did show pharmacodynamic evidence for activity. This suggests the possibility that immune checkpoint inhibitors may be more effective in a setting where large tumors are present. Here we have studied immune responses to large tumors in an autochthonous mouse model of glioblastoma. MethodsGlioblastoma was induced by transfection with oncogenic plasmids injected directly into the lateral ventricle of neonatal mice. Immune responses were assessed using a combination of spectral flow cytometry and immunohistochemistry. ResultsThere was a marked immune response to large tumors, with significant increases in CD4 T cells and dendritic cells. T cell changes occurred primarily at leptomeningeal/perivascular border sites. A large proportion of CD4 T cells expressed PD1 and half of these were regulatory T cells. NK cells were also increased in mice with large tumors, but were predominantly in immature states. The mouse model accurately recapitulates the formation of palisading necroses. These contain apoptotic cells and avidly recruit myeloid cells that are induced to express large amounts of TGF{beta}. ConclusionsLarge glioblastoma tumors generate a border site population of PD1 positive T cells that may explain the pharmacodynamic response in neoadjuvant trials, and a palisading necrosis-driven immunosuppressive mechanism that may explain why responses are insufficient to provide a significant clinical benefit. KEY POINTSThe SB mouse model accurately recapitulates immune features of human glioblastoma Large tumors induce a significant border site immune response Palisading necroses in large tumors counter this with a strong immunosuppressive response IMPORTANCE OF STUDYImmune checkpoint inhibitors have not shown efficacy in glioblastoma when used post-surgery, but do show pharmacodynamic activity when used in patients prior to second surgery (i.e. neoadjuvant). This suggest the possibility that immune checkpoint inhibition is more effective when large tumors are present. Using a clinically-relevant autochthonous mouse model, we show here that large tumors induce an immune response that is evident in leptomeningeal border sites. Large tumors in this mouse model also generate palisading necroses, a well-known diagnostic feature in glioblastoma tumors. These palisading necroses generate large amounts of TGF{beta}, providing a mechanism by which large tumors can suppress border site immune responses. This further supports the concept that palisading necroses are drivers of glioblastoma malignancy and suggests novel strategies to enhance responses to immune checkpoint inhibition in this cancer.

Background: GBM is one of the most common and most aggressive brain tumors in adults, and upfront standard of care treatment has limited efficacy. Immune checkpoint inhibitor strategies have significantly improved outcomes in various solid tumors but have not proven effective in GBM, suggesting other strategies may be needed to realize their full potential. Methods: GBM patients were treated with upfront standard of care chemoradiation with temozolomide and pembrolizumab, followed by adjuvant temozolomide and pembrolizumab for six nine-week cycles. Depending on production of sufficient vaccine, patients were randomized into HSPPC-96 vaccine or placebo group (q4 weeks) while those with failed vaccine production continued on study unblinded as an ancillary group. The primary objective was overall survival at one year, and secondary endpoints were progression-free survival at six months, overall and progression-free survival, radiographic response, and tolerability by patient-reported outcomes and adverse event documentation. Results: 90 patients were screened, 32 were treated (8 vaccine, 9 placebo, 15 ancillary), and 26 were evaluable for radiographic responses prior to accrual termination. The study did not meet its primary endpoint of overall survival at one year (65.5% in vaccine group, 75% in placebo). Progression-free endpoints were mildly improved in the vaccine group but were not significant, and response rates were not significantly different. The regimen was well-tolerated and safe. Conclusions: Though limited by early discontinuation, these findings do not support the combination of pembrolizumab and HSPPC-96 vaccine with standard of care therapy. Trials Registration: ClinicalTrials.gov identifier: NCT03018288

Background. Meningiomas exhibit well-established hormonal biology, yet no study has examined whether myeloid immune infiltration interacts with estrogen-responsive transcription in this tumor type. Methods. We applied three-method consensus immune deconvolution (EPIC, MCPcounter, CIBERSORTx) to 968 harmonized meningioma RNA-seq transcriptomes from five public datasets, stratified by Thirimanne et al. (2024) transcriptomic subtypes. Competitive gene set enrichment compared macrophage-high versus macrophage-low tertiles with sex-adjusted, purity-adjusted, and method-independent sensitivity analyses. Survival modeling tested both total macrophage burden and a decomposed microglia-to-macrophage ratio validated against single-cell ground truth (pseudo-bulk r = 0.77). Results. Macrophage-high tumors showed significant suppression of estrogen response gene sets (FDR = 4.9 x 10-5) despite paradoxical ESR1 upregulation (log2FC = +0.40, FDR = 2.5 x 10-26) and PGR downregulation (log2FC = -0.34, FDR = 2.7 x 10-3), indicating post-receptor transcriptional disruption. This signal strengthened after sex adjustment (FDR = 1.9 x 10-6) and was confirmed across a multi-layer sensitivity battery (eleven analyses including reference-matrix-independent, purity-adjusted, rotation-based self-contained, and empirical-null tests; all FDR < 3 x 10-4 in the relevant convergent tests). Myeloid infiltration was strongly subtype-dependent (Kruskal-Wallis p = 7.4 x 10-16) but grade-independent (p = 0.399), with CSF1R enriched in the macrophage-dominant Cluster B. Neither total macrophage score (HR = 0.90, p = 0.53; N = 102) nor a decomposed microglia/macrophage ratio (HR = 0.92, p = 0.46; N = 101) predicted recurrence-free survival. Conclusions. The pre-registered primary endpoint - macrophage infiltration score predicting recurrence-free survival - was not supported; the estrogen-immune dissociation emerged from secondary exploratory gene-set analysis and requires independent validation. Macrophage-infiltrated meningiomas exhibit a previously unreported dissociation between maintained ESR1 expression and suppressed estrogen-responsive transcription, with implications for hormonal therapy stratification.

Oligodendroglioma is a primary central nervous system tumor classified by the presence of isocitrate dehydrogenase (IDH) mutations and codeletion of 1p/19q. Here we describe the generation of an IDH-mutant 1p/19q-codeleted oligodendroglioma mouse model using in utero electroporation. We identified IDH1R132H, PIK3CAE545K, CicKO, Fubp1KO and Cdkn2aKO as the optimal combination (termed OligoCdkn2a) to drive fully penetrant tumors that histologically resemble human grade II/III IDH-mutant, 1p/19q-codeleted oligodendroglioma. Replacing Cdkn2a with Trp53 loss in this mouse model shifted tumor histology towards high grade astrocytoma. OligoCdkn2a tumors displayed metabolic and transcriptional changes associated with IDH and CIC mutations, and single cell sequencing identified a bias towards oligodendrocyte differentiation compared to an IDH wild-type glioblastoma mouse model. OligoCdkn2a tumors represent the first mouse model system to recapitulate the genetic, histological and transcriptional features of human IDH-mutant 1p/19q-codeleted oligodendrogliomas, offering a platform to further dissect tumor biology and test new therapeutic strategies.

Background: Cognitive impairment is a prevalent long-term sequela in glioma patients, yet its cerebrovascular correlates remain poorly characterized. Arterial spin labeling (ASL) perfusion MRI offers a non-invasive means to quantify cerebral blood flow (CBF) and may serve as a sensitive correlate of radiotherapy (RT)-induced neurovascular injury. Methods: Fifty WHO Grade 2/3 glioma patients and 50 matched healthy controls underwent pseudo-continuous ASL (pCASL) MRI and a standardized cognitive test battery. Regional CBF was compared between patients (n=44, after quality control) and controls (n=50) using ANCOVA with age, sex, and deep white matter CBF as covariates. In irradiated patients (~5 years post-RT), RT dose-CBF associations were assessed using region-wise regression, and regional CBF was compared among controls and low-dose ([&le;]15 Gy) versus high-dose ([&ge;]40 Gy) regional RT exposure groups. Cognition-CBF associations were evaluated in a priori domain-specific regions of interest. Results: Compared with controls, patients showed frontoparietal cortical hypoperfusion, with significantly lower CBF in middle frontal and superior/inferior parietal cortices (all q<0.01; partial -squared=0.128-0.147). Region-wise regression showed no significant linear RT dose-CBF associations after correction. However, subgroup analyses identified RT dose-sensitive regions with [&ge;]40 Gy exposure that showed lower adjusted CBF than controls, most prominently in the left precentral and caudal middle frontal cortices (q<0.01; adjusted-{Delta}CBF{approx}-27.2--28.8 mL/100g/min). Perfusion in the left precentral and postcentral gyri of irradiated patients correlated positively with motor performance. Conclusions: pCASL reveals persistent cortical hypoperfusion in glioma patients that spatially corresponds with RT dose exposure and associates with cognitive performance, positioning ASL as a promising non-invasive biomarker of RT-related neurovascular injury.

Introduction: Precise anatomical navigation is fundamental to safe endoscopic pituitary surgery, a high-stakes procedure characterised by a challenging learning curve. While traditional navigation systems often rely on workflow-disrupting probes or static preoperative imaging, advancements in computer vision AI (CVAI) now enable dynamic, real-time anatomical segmentation directly from live surgical video1-3. Our group has previously conducted a series of preclinical human-computer interaction studies to refine the system's design, alongside digital and high-fidelity physical simulations demonstrating the benefit of AI assistance in improving overall performance, training, and safety4-8. Building on this foundation, the current study represents a first-in-human application of real-time CVAI assistance in the neurosurgical operating room, serving to assess feasibility and safety, and to iteratively improve the system. Method: Guided by DECIDE-AI and IDEAL frameworks, this single-centre evaluation comprises an initial proof-of-concept phase (n=6) for endoscopic transsphenoidal pituitary surgeries. The AI model utilised a DINOv3-derived vision transformer architecture, deployed via a high-performance edge computing unit to achieve low-latency, real-time inference without reliance on cloud infrastructure2. Given the high-risk nature of the procedure and the early stage of clinical AI integration, the system was initially deployed as an educational adjunct on a secondary monitor, ensuring the primary surgical feed remains uncompromised. Functionality and safety were assessed via structured questionnaire, prospective observation, and blinded retrospective review of the recordings of the endoscopic surgical video feed and wider operating room environment. Continuous multi-stakeholder feedback through validated human factors surveys drove iterative technical refinements between cases. Results: Six patients with pituitary adenomas were enrolled. The CVAI system was successfully deployed in four cases, demonstrating acceptable real-time sella segmentation accuracy. Deployment failed pre-operatively in two cases owing to a single recurring system reboot bug. Iterative refinement between cases were driven by our experience and surgical team feedback. This resulted in the integration of additional anatomical structure segmentations (e.g., carotid arteries), enhanced model accuracy via training dataset expansion, and hardware firmware upgrades. Multi-stakeholder surveys demonstrated satisfactory system feasibility, usability, and acceptability among the surgical team. Both prospective observation and retrospective video review confirmed the absence of adverse events, including no significant distraction to the primary surgeon, and there were no AI-related clinical complications. Conclusion: This first-in-human early clinical evaluation demonstrates the feasibility, safety and iterative development of real-time, CVAI-based anatomical navigation during high-stakes neurosurgery. Future work will include a larger single-centre case series (IDEAL Stage 2a) with more surgical teams to further iterate the system and explore its impact on training and workflow. As the underpinning technology improves, deployment will transition to direct intra-operative decision support and integration with other intra-operative navigational technologies.

BackgroundPrevious machine learning models to intraoperatively predict the molecular status of gliomas using stimulated Raman histology (SRH), such as DeepGlioma, have achieved high performance (91.5% accuracy) on curated datasets. However, when used intraoperatively, DeepGlioma (162M parameters) runs slowly on current SRH hardware and underperforms due to its lack of an image rejection mechanism and its validation on curated images. Here, we introduce SRH-Informed Glioma classificatioN with Attention Learning (SIGNAL) (27M parameters), a lighter model with a built-in attention-based rejection mechanism that outperforms DeepGlioma on uncurated clinical datasets. MethodsSIGNAL was developed using 1.56 million SRH fields-of-view from 967 adult diffuse glioma patients collected between December 2017 and July 2025. We used 412 patients from NYU for training and internal validation and a multi-institutional, international cohort of 555 patients for testing. SIGNAL uses a ResNet50 backbone pretrained using a hierarchical contrastive loss function followed by a multi-head multi-layer perceptron (MLP). Using a patch-based attention threshold of 0.6, a final MLP was trained to predict glioma subtypes: glioblastoma, oligodendroglioma, or astrocytoma. ResultsSIGNAL outperformed DeepGlioma, achieving greater overall accuracy (90.10% vs. 72.59%) while running faster (16.0 vs. 6.7 patches/s). SIGNAL also outperformed DeepGlioma on all three molecular classification tasks, including IDH mutation (accuracy: 93.51% vs. 79.22%), 1p19q codeletion (93.51% vs. 88.31%), and ATRX loss (89.61% vs. 83.98%). SIGNALs attention mechanism had a strong positive linear correlation with mean patch cellularity (r=0.96, p<0.001) and a strong negative correlation with patch blood coverage (r=-0.99,p<0.001). Finally, subtype and molecular accuracy between tumor core and margin samples were equivalent despite significantly lower patch retention in tumor margins (44.5% vs 60.2%, p<0.0001). ConclusionSIGNAL is a lightweight model for intraoperative molecular classification of gliomas using SRH imaging. Its attention-based image quality filter allows for excellent performance, quick processing, and highly interpretable outputs critical for reliable use in intraoperative workflows. Brief 1-2 Sentence DescriptionWe present SIGNAL, a lightweight machine learning model for intraoperative molecular classification of diffuse gliomas using stimulated Raman histology, whose core innovation is a learned attention mechanism that filters diagnostically uninformative tissue, such as blood and acellular regions, before classification, enabling robust real-world generalizability. Validated on 555 patients across four international centers, SIGNAL outperforms the previous state-of-the-art model DeepGlioma on glioma subtype classification (90.10% vs. 72.59% accuracy) while running 2.4 times faster on intraoperative hardware.

Glioblastoma (GBM) develops within a complex tumor ecosystem whose temporal dynamics remain poorly understood. Here, we performed longitudinal single-cell RNA sequencing and spatial transcriptomics across multiple timepoints in two widely used murine GBM models - CT2A and GL261 - which differ markedly in aggressiveness and response to immune checkpoint blockade. Tumor cell transcriptomes revealed model-specific programs: CT2A cells progressively upregulated epithelial-mesenchymal transition (EMT), non-classical MHC Class I, and progressively, hypoxia response pathways, resembling the human mesenchymal GBM cell state, while GL261 cells exhibited MHC Class II expression and developmental signatures resembling oligodendrocyte progenitor and astrocytic states. Ligand-receptor interaction analyses identified thrombospondins (Thbs1, Thbs2) and osteopontin (Spp1) as CT2A-specific tumor ligands mediating tumorigenic interactions with immune cells, with downstream targets enriched for EMT and TGF-{beta} pathways. Conversely, the GL261 model presented a differential potential to engage neuronal and perivascular guidance networks, with Glutamate and L1 cell adhesion molecule (L1cam) as lead signaling partners. The CT2A immune compartment exhibited progressive microglia-to-macrophage phenotypic conversion, enhanced macrophage infiltration driven by Spp1, and elevated T cell exhaustion, while GL261 maintained a distinct adaptive immune communication hub via MHC class II-CD4 signaling. Elevated THBS1, THBS2, and SPP1 expression correlated with poor survival in human GBM datasets. Together, these findings reveal divergent tumor-immune ecosystems in CT2A and GL261 that recapitulate distinct aspects of human GBM, with implications for therapeutic targeting.