Deciphering the causal relationship between plasma and cerebrospinal fluid metabolites and glioblastoma multiforme: a Mendelian Randomization study.

BACKGROUND: Glioblastoma Multiforme (GBM) is one of the most aggressive and fatal brain cancers. The study of metabolites could be crucial for understanding GBM's biology and reveal new treatment strategies. METHODS: The GWAS data for GBM were sourced from the FinnGen database. A total of 1400 plasma metabolites were collected from the GWAS Catalog dataset. The cerebrospinal fluid (CSF) metabolites data were collected from subsets of participants in the WADRC and WRAP studies. We utilized the inverse variance weighting (IVW) method as the primary tool to explore the causal relationship between metabolites in plasma and CSF and glioblastoma, ensuring the exclusion of instances with horizontal pleiotropy. Additionally, four supplementary analytical methods were applied to reinforce our findings. Aberrant results were identified and omitted based on the outcomes of the leave-one-out sensitivity analysis. Conclusively, a reverse Mendelian Randomization analysis was also conducted to further substantiate our results. RESULTS: The study identified 69 plasma metabolites associated with GBM. Of these, 40 metabolites demonstrated a significant positive causal relationship with GBM, while 29 exhibited a significant negative causal association. Notably, Trimethylamine N-oxide (TMAO) levels in plasma, not CSF, were found to be a significant exposure factor for GBM (OR = 3.1627, 95% CI = (1.6347, 6.1189), P = 0.0006). The study did not find a reverse causal relationship between GBM and plasma TMAO levels. CONCLUSIONS: This research has identified 69 plasma metabolites potentially associated with the incidence of GBM, among which TMAO stands out as a promising candidate for an early detectable biomarker for GBM.

Development and validation of a nomogram to predict leptomeningeal metastases in lung adenocarcinoma: Cervical lymph node metastasis is an important association factor.

BACKGROUND: The goal of this study was to create a nomogram using routine parameters to predict leptomeningeal metastases (LMs) in advanced lung adenocarcinoma (LAC) patients to prevent needless exams or lumbar punctures and to assist in accurately diagnosing LMs. METHODS: Two hundred and seventy-three patients with LMs and brain metastases were retrospectively reviewed and divided into derivation (n = 191) and validation (n = 82) cohorts using a 3:7 random allocation. All LAC patients with LMs had positive cerebrospinal fluid cytology results and brain metastases confirmed by magnetic resonance imaging. Binary logistic regression with backward stepwise selection was used to identify significant characteristics. A predictive nomogram based on the logistic model was assessed through receiver operating characteristic curves. The validation cohort and Hosmer-Lemeshow test were used for internal validation of the nomogram. RESULTS: Five clinicopathological parameters, namely, gene mutations, surgery at the primary lung cancer site, clinical symptoms of the head, N stage, and therapeutic strategy, were used as predictors of LMs. The area under the curve was 0.946 (95% CI 0.912-0.979) for the training cohort and 0.861 (95% CI 0.761-0.961) for the internal validation cohort. There was no significant difference in performance between the two cohorts (p = 0.116). In the internal validation, calibration plots revealed that the nomogram predictions were well suited to the actual outcomes. CONCLUSIONS: We created a user-friendly nomogram to predict LMs in advanced lung cancer patients, which could help guide treatment decisions and reduce unnecessary lumbar punctures.

Single-cell atlas reveals the immunosuppressive microenvironment and Treg cells landscapes in recurrent Glioblastoma.

Patients diagnosed with glioblastoma (GBM) have the most aggressive tumor progression and lethal recurrence. Research on the immune microenvironment landscape of tumor and cerebrospinal fluid (CSF) is limited. At the single-cell level, we aim to reveal the recurrent immune microenvironment of GBM and the potential CSF biomarkers and compare tumor locations. We collected four clinical samples from two patients: malignant samples from one recurrent GBM patient and non-malignant samples from a patient with brain tumor. We performed single-cell RNA sequencing (scRNA-seq) to reveal the immune landscape of recurrent GBM and CSF. T cells were enriched in the malignant tumors, while Treg cells were predominately found in malignant CSF, which indicated an inhibitory microenvironment in recurrent GBM. Moreover, macrophages and neutrophils were significantly enriched in malignant CSF. This indicates that they an important role in GBM progression. S100A9, extensively expressed in malignant CSF, is a promising biomarker for GBM diagnosis and recurrence. Our study reveals GBM's recurrent immune microenvironment after chemoradiotherapy and compares malignant and non-malignant CSF samples. We provide novel targets and confirm the promise of liquid CSF biopsy for patients with GBM.

CSF Biopsy in Glioma: A Brief Review.

Glioma is the most common intracranial malignant tumor. Over the past several years, liquid biopsy in diagnosis and treatment of solid tumors have made many progressions, but there is still a gap from a large clinical application of liquid biopsy in glioma due to many limitations. However, in recent years, researchers have made many explorations into liquid biopsy in glioma. In the future, the liquid biopsy of glioma, especially cerebrospinal fluid, will have a broad prospect. In this review, we will discuss the current research progressions of CSF biopsy in glioma in recent years.

A review on genetic alterations in CNS metastases related to breast cancer treatment. Is there a role for liquid biopsies in CSF?

Acquired mutations or altered gene expression patterns in brain metastases (BM) and/or leptomeningeal metastases (LM) of breast cancer may play a role in therapy-resistance and offer new molecular targets and treatment options. Despite expanding knowledge of genetic alterations in breast cancer and their metastases, clinical applications for patients with central nervous system (CNS) metastases are currently limited. An emerging tool are DNA-techniques that may detect genetic alterations of the CNS metastases in the cerebrospinal fluid (CSF). In this review we discuss genetic studies in breast cancer and CNS metastases and the role of liquid biopsies in CSF.

IDH1 p.R132H ctDNA and D-2-hydroxyglutarate as CSF biomarkers in patients with IDH-mutant gliomas.

INTRODUCTION: We aimed to evaluate IDH1 p.R132H mutation and 2-hydroxyglutarate (2HG) in cerebrospinal fluid (CSF) as biomarkers for patients with IDH-mutant gliomas. METHODS: CSF was collected from patients with infiltrating glioma, and 2HG levels were measured by liquid chromatography-mass spectrometry. IDH1 p.R132H mutant allele frequency (MAF) in CSF-ctDNA was measured by digital droplet PCR (ddPCR). Tumor volume was measured from standard-of-care magnetic resonance images. RESULTS: The study included 48 patients, 6 with IDH-mutant and 42 with IDH-wildtype gliomas, and 57 samples, 9 from the patients with IDH-mutant and 48 from the patients with IDH-wildtype gliomas. ctDNA was detected in 7 of the 9 samples from patients with IDH-mutant glioma, and IDH1 p.R132H mutation was detected in 5 of the 7 samples. The MAF ranged from 0.3 to 39.95%. Total 2HG level, D-2HG level, and D/L-2HG ratio in CSF were significantly higher in patients with IDH-mutant gliomas than in patients with IDH-wildtype gliomas. D-2HG level and D/L-2HG ratio correlated with total tumor volume in patients with IDH-mutant gliomas but not in patients with IDH-wildtype gliomas. CONCLUSION: Our results suggest that detection of IDH1 p.R132H mutation by ddPCR and increased D-2HG level in CSF may help identify IDH-mutant gliomas. Our results also suggest that D-2HG level and D/L-2HG ratio correlate with tumor volume in patients with IDH-mutant gliomas. Further prospective studies with larger cohorts are needed to validate these findings.

Low-coverage whole-genome sequencing of cerebrospinal-fluid-derived cell-free DNA in brain tumor patients.

This protocol summarizes the pipeline for analysis of tumor-derived cell-free DNA (cfDNA) from cerebrospinal fluid (CSF) using low-coverage whole-genome sequencing (lcWGS). This approach enables resolution of chromosomal and focal copy-number variations (CNVs) as oncologic signatures, particularly for patients with central nervous system tumors. Our strategy tolerates sub-nanogram cfDNA input and is thus optimized for CSF samples where cfDNA yields are typically low. Overall, the detection of tumor-specific signatures in CSF-derived cfDNA is a promising biomarker for personalization of brain-tumor therapy. For complete details on the use and execution of this protocol, please refer to Liu et al. (2021).

Cerebrospinal fluid cytokine levels are associated with macrophage infiltration into tumor tissues of glioma patients.

BACKGROUND: Diffuse gliomas are the most common malignant tumors of the central nervous system with poor treatment efficacy. Infiltration of immune cells into tumors during immunosurveillance is observed in multiple tumor entities and often associated with a favorable outcome. The aim of this study was to evaluate the infiltration of immune cells in gliomas and their association with cerebrospinal fluid (CSF) cytokine concentrations. METHODS: We applied immunohistochemistry in tumor tissue sections of 18 high-grade glioma (HGG) patients (4 anaplastic astrocytoma, IDH-wildtype WHO-III; 14 glioblastomas (GBM), IDH-wildtype WHO-IV) in order to assess and quantify leucocytes (CD45) and macrophages (CD68, CD163) within the tumor core, infiltration zone and perivascular spaces. In addition, we quantified the concentrations of 30 cytokines in the same patients' CSF and in 14 non-inflammatory controls. RESULTS: We observed a significantly higher percentage of CD68+ macrophages (21-27%) in all examined tumor areas when compared to CD45+ leucocytes (ca. 3-7%); CD163+ cell infiltration was between 5 and 15%. Compared to the tumor core, significantly more macrophages and leucocytes were detectable within the perivascular area. The brain parenchyma showing a lower tumor cell density seems to be less infiltrated by macrophages. Interleukin (IL)-7 was significantly downregulated in CSF of GBM patients compared to controls. Additionally, CD68+ macrophage infiltrates showed significant correlations with the expression of eotaxin, interferon-γ, IL-1ß, IL-2, IL-10, IL-13, IL-16 and vascular endothelial growth factor. CONCLUSIONS: Our findings suggest that the infiltration of lymphocytes is generally low in HGG, and does not correlate with cytokine concentrations in the CSF. In contrast, macrophage infiltrates in HGG are associated with CSF cytokine changes that possibly shape the tumor microenvironment. Although results point towards an escape from immunosurveillance or even exploitation of immune cells by HGG, further studies are necessary to decipher the exact role of the immune system in these tumors.

Longitudinal Bottom-Up Proteomics of Serum, Serum Extracellular Vesicles, and Cerebrospinal Fluid Reveals Candidate Biomarkers for Early Detection of Glioblastoma in a Murine Model.

Glioblastoma Multiforme (GBM) is a brain tumor with a poor prognosis and low survival rates. GBM is diagnosed at an advanced stage, so little information is available on the early stage of the disease and few improvements have been made for earlier diagnosis. Longitudinal murine models are a promising platform for biomarker discovery as they allow access to the early stages of the disease. Nevertheless, their use in proteomics has been limited owing to the low sample amount that can be collected at each longitudinal time point. Here we used optimized microproteomics workflows to investigate longitudinal changes in the protein profile of serum, serum small extracellular vesicles (sEVs), and cerebrospinal fluid (CSF) in a GBM murine model. Baseline, pre-symptomatic, and symptomatic tumor stages were determined using non-invasive motor tests. Forty-four proteins displayed significant differences in signal intensities during GBM progression. Dysregulated proteins are involved in cell motility, cell growth, and angiogenesis. Most of the dysregulated proteins already exhibited a difference from baseline at the pre-symptomatic stage of the disease, suggesting that early effects of GBM might be detectable before symptom onset.

Genomic comparison between cerebrospinal fluid and primary tumor revealed the genetic events associated with brain metastasis in lung adenocarcinoma.

Lung adenocarcinoma (LUAD) is most common pathological type of lung cancer. LUAD with brain metastases (BMs) usually have poor prognosis. To identify the potential genetic factors associated with BM, a genomic comparison for BM cerebrospinal fluid (CSF) and primary lung tumor samples obtained from 1082 early- and late-stage LUAD patients was performed. We found that single nucleotide variation (SNV) of EGFR was highly enriched in CSF (87% of samples). Compared with the other primary lung tissues, copy number gain of EGFR (27%), CDK4 (11%), PMS2 (11%), MET (10%), IL7R (8%), RICTOR (7%), FLT4 (5%), and FGFR4 (4%), and copy number loss of CDKN2A (28%) and CDKN2B (18%) were remarkably more frequent in CSF samples. CSF had significantly lower tumor mutation burden (TMB) level but more abundant copy number variant. It was also found that the relationships among co-occurrent and mutually exclusive genes were dynamically changing with LUAD development. Additionally, CSF (97% of samples) harbored more abundant targeted drugs related driver and fusion genes. The signature 15 associated with defective DNA mismatch repair (dMMR) was only identified in the CSF group. Cancer associated pathway analysis further revealed that ErbB (95%) and cell cycle (84%) were unique pathways in CSF samples. The tumor evolution analysis showed that CSF carried significantly fewer clusters, but subclonal proportion of EGFR was remarkably increased with tumor progression. Collectively, CSF sequencing showed unique genomic characteristics and the intense copy number instability associated with cell cycle disorder and dMMR might be the crucial genetic factors in BM of LUAD.

A Set of 17 microRNAs Common for Brain and Cerebrospinal Fluid Differentiates Primary Central Nervous System Lymphoma from Non-Malignant Brain Tumors.

The diagnosis of primary central nervous system (CNS) lymphoma, which is predominantly of the diffuse large B-cell lymphoma type (CNS DLBCL), is challenging. MicroRNAs (miRs) are gene expression-regulating non-coding RNAs that are potential biomarkers. We aimed to distinguish miR expression patterns differentiating CNS DLBCL and non-malignant CNS diseases with tumor presentation (n-ML). Next generation sequencing-based miR profiling of cerebrospinal fluids (CSFs) and brain tumors was performed. Sample source-specific (CSF vs. brain tumor) miR patterns were revealed. Even so, a set of 17 miRs differentiating CNS DLBCL from n-ML, no matter if assessed in CSF or in a tumor, was identified. Along with the results of pathway analyses, this suggests their pathogenic role in CNS DLBCL. A combination of just four of those miRs (miR-16-5p, miR-21-5p, miR-92a-3p, and miR-423-5p), assessed in CSFs, discriminated CNS DLBCL from n-ML samples with 100% specificity and 67.0% sensitivity. Analyses of paired CSF-tumor samples from patients with CNS DLBCL showed significantly lower CSF levels of miR-26a, and higher CSF levels of miR-15a-5p, miR-15b-5p, miR-19a-3p, miR-106b-3p, miR-221-3p, and miR-423-5p. Noteworthy, the same miRs belonged to the abovementioned set differentiating CNS DLBCL from non-malignant CNS diseases. Our results not only add to the basic knowledge, but also hold significant translational potential.

Tumor DNA Mutations From Intraparenchymal Brain Metastases Are Detectable in CSF.

Discordant responses between brain metastases and extracranial tumors can arise from branched tumor evolution, underscoring the importance of profiling mutations to optimize therapy. However, the morbidity of brain biopsies limits their use. We investigated whether cell-free DNA (cfDNA) in CSF could serve as an effective surrogate marker for genomic profiling of intraparenchymal (IP) brain metastases. METHODS: CSF and blood were collected simultaneously from patients with progressive brain metastases undergoing a craniotomy or lumbar puncture. Mutations in both biofluids were measured using an error-suppressed deep sequencing method previously published by our group. Forty-three regions of 24 cancer-associated genes were assayed. RESULTS: This study enrolled 14 patients with either IP brain metastases (n = 12) or cytology-positive leptomeningeal disease (LMD, n = 2) and two controls with normal pressure hydrocephalus. Primary cancer types were lung, melanoma, renal cell, and colorectal. cfDNA was measurable in all sixteen samples of CSF. Cancer-associated mutations were found in the CSF of ten patients (eight with IP [67%] and two with LMD [100%]) and plasma of five patients (five with IP [42%] and none with LMD). All patients with plasma cfDNA had extracranial tumors. Among the five patients in the cohort who also had mutation data from time-matched brain metastasis tissue, four patients (80%) had matching mutations detected in CSF and brain, whereas only one patient (20%) had matching mutations detected in plasma and brain. CONCLUSION: The detection of mutational DNA in CSF is not restricted to LMD and was found in two thirds of patients with IP brain metastases in our cohort. Analysis of CSF can be a viable alternative to biopsy for detection of somatic mutations in brain metastases.

Diagnostic biomarkers from proteomic characterization of cerebrospinal fluid in patients with brain malignancies.

Recent technological advances in molecular diagnostics through liquid biopsies hold the promise to repetitively monitor tumor evolution and treatment response of brain malignancies without the need of invasive surgical tissue accrual. Here, we implemented a mass spectrometry-based protein analysis pipeline which identified hundreds of proteins in 251 cerebrospinal fluid (CSF) samples from patients with four types of brain malignancies (glioblastoma, lymphoma, brain metastasis, and leptomeningeal disease [LMD]) and from healthy individuals with a focus on glioblastoma in a retrospective and confirmatory prospective observational study. CSF proteome deregulation via disruption of the blood brain barrier appeared to be largely conserved across brain tumor entities. CSF analysis of glioblastoma patients identified two proteomic clusters that correlated with tumor size and patient survival. By integrating CSF data with proteomic analyses of matching glioblastoma tumor tissue and primary glioblastoma cells, we identified potential CSF biomarkers for glioblastoma, in particular chitinase-3-like protein 1 (CHI3L1) and glial fibrillary acidic protein (GFAP). Key findings were validated in a prospective cohort consisting of 35 glioma patients. Finally, in LMD patients who frequently undergo repeated CSF work-up, we explored our proteomic pipeline as a mean to profile consecutive CSF samples. Therefore, proteomic analysis of CSF in brain malignancies has the potential to reveal biomarkers for diagnosis and therapy monitoring.

Intranasal administration of the chemotherapeutic perillyl alcohol results in selective delivery to the cerebrospinal fluid in rats.

Perillyl alcohol (POH) has been extensively studied for the treatment of peripheral and primary brain tumors. The intranasal route of administration has been preferred for dosing POH in early-stage clinical trials associated with promising outcomes in primary brain cancer. However, it is unclear how intranasal POH targets brain tumors in these patients. Multiple studies indicate that intranasally applied large molecules may enter the brain and cerebrospinal fluid (CSF) through direct olfactory and trigeminal nerve-associated pathways originating in the nasal mucosa that bypass the blood-brain barrier. It is unknown whether POH, a small molecule subject to extensive nasal metabolism and systemic absorption, may also undergo direct transport to brain or CSF from the nasal mucosa. Here, we compared CSF and plasma concentrations of POH and its metabolite, perillic acid (PA), following intranasal or intravascular POH application. Samples were collected over 70 min and assayed by high-performance liquid chromatography. Intranasal administration resulted in tenfold higher CSF-to-plasma ratios for POH and tenfold higher CSF levels for PA compared to equal dose intravascular administration. Our preclinical results demonstrate POH undergoes direct transport from the nasal mucosa to the CSF, a finding with potential significance for its efficacy as an intranasal chemotherapeutic for brain cancer.

Precise detection of the germinomatous component of intracranial germ cell tumors of the basal ganglia and thalamus using placental alkaline phosphatase in cerebrospinal fluid.

PURPOSE: The disadvantages of biopsy for lesions in the basal ganglia and thalamus include a risk of various complications, difficulty in selecting the target tissue in some cases due to indistinct neuroimaging findings and limited availability of sample tissue. Placental alkaline phosphatase (PLAP) plays a decisive role in the diagnosis and management of intracranial germ cell tumors (IGCTs) in the basal ganglia and thalamus. The present study aimed to demonstrate the ability, specificity, and optimal use of PLAP values obtained from cerebrospinal fluid (CSF). METHODS: Twenty patients with lesions in the basal ganglia and thalamus were enrolled in this study: 11 had IGCTs and 9 had non-IGCTs. The values of PLAP and other established tumor markers in the CSF were measured in all patients before treatment. RESULTS: The mean follow-up period was 76.0 months (range, 3-168) for all lesions. PLAP was elevated in all 11 patients with IGCTs in the basal ganglia or thalamus, whereas none of the patients with non-IGCT exhibited elevated PLAP. Thus, the sensitivity and specificity of PLAP were both 100%. CONCLUSION: Our data demonstrated that the PLAP value can specifically identify the germinomatous component even in cases of IGCTs in the basal ganglia or thalamus with high sensitivity and specificity. PLAP is undoubtedly beneficial for the safe and timely detection of the germinomatous component of IGCTs in the basal ganglia and thalamus, because reliance on PLAP measurement enables us to avoid invasive surgical procedures and facilitates the prompt initiation of chemoradiation therapy.

Combining targeted sequencing and ultra-low-pass whole-genome sequencing for accurate somatic copy number alteration detection.

This study investigated the feasibility of combining targeted sequencing and ultra-low-pass whole-genome sequencing (ULP-WGS) for improved somatic copy number alteration (SCNA) detection, due to its role in tumorigenesis and prognosis. Cerebrospinal fluid and matched blood samples were obtained from 29 patients with brain metastasis derived from lung cancer. Samples were subjected to targeted sequencing (genomic coverage: 300 kb) and 2×ULP-WGS. The SCNA was detected by the CTLW_CNV, Control-FreeC, and CNVkit methods and their accuracy was analyzed. Eighteen tumor samples showed consistent SCNA results between the three methods, while a small fraction of samples resulted in different SCNA estimations. Further analysis indicated that consistency of SCNA highly correlated with the difference of baseline depth (normalized depth of regions without SCNA events) estimation between methods. Conflict Index showed that CTLW_CNV significantly improved the accuracy of SCNA detection through precise baseline depth estimation. CTLW_CNV combines targeted sequencing and ULP-WGS for improved SCNA detection. The improvement in detection accuracy is mainly due to a refined baseline depth estimation, guided by single-nucleotide polymorphism allele frequencies within the deeply sequenced region (targeted sequencing). This method is especially suitable for tumor samples with biased aneuploidy, a previously under-estimated genomic characteristic across different cancer types.

Molecular diagnosis of diffuse glioma using a chip-based digital PCR system to analyze IDH, TERT, and H3 mutations in the cerebrospinal fluid.

PURPOSE: Conventional genetic analyzers require surgically obtained tumor tissues to confirm the molecular diagnosis of diffuse glioma. Recent technical breakthroughs have enabled increased utilization of cell-free tumor DNA (ctDNA) in body fluids as a reliable resource for molecular diagnosis in various cancers. Here, we tested the application of a chip-based digital PCR system for the less invasive diagnosis (i.e., liquid biopsy) of diffuse glioma using the cerebrospinal fluid (CSF). METHODS: CSF samples from 34 patients with diffuse glioma were collected from the surgical field during craniotomy. Preoperative lumbar CSF collection was also performed in 11 patients. Extracted ctDNA was used to analyze diagnostic point mutations in IDH1 R132H, TERT promoter (C228T and C250T), and H3F3A (K27M) on the QuantStudio® 3D Digital PCR System. These results were compared with their corresponding tumor DNA samples. RESULTS: We detected either of the diagnostic mutations in tumor DNA samples from 28 of 34 patients. Among them, we achieved precise molecular diagnoses using intracranial CSF in 20 (71%). Univariate analyses revealed that the World Health Organization (WHO) grade (p = 0.0034), radiographic enhancement (p = 0.0006), and Mib1 index (p = 0.01) were significant predictors of precise CSF-based molecular diagnosis. We precisely diagnosed WHO grade III or IV diffuse gliomas using lumbar CSF obtained from 6 (87%) of 7 patients with tumors harboring any mutation. CONCLUSION: We established a novel, non-invasive molecular diagnostic method using a chip-based digital PCR system targeting ctDNA derived from CSF with high sensitivity and specificity, especially for high-grade gliomas.

Prognostic relevance of CSF and peri-tumoral edema volumes in glioblastoma.

BACKGROUND: We conducted a segmental volumetric analysis of pre-operative brain magnetic resonance images (MRIs) of glioblastoma patients to identify brain- and tumor-related features that are prognostic of survival. METHODS: Using a dataset of 210 single-institutional adult glioblastoma patients, total volumes of the following tumor- and brain-related features were quantified on pre-operative MRIs using a fully automated segmentation tool: tumor enhancement, tumor non-enhancement, tumor necrosis, peri-tumoral edema, grey matter, white matter, and cerebrospinal fluid (CSF). Their association with survival using Cox regression models, adjusting for the well-known predictors of glioblastoma survival. The findings were verified in a second dataset consisting of 96 glioblastoma patients from The Cancer Imaging Archive and The Cancer Genome Atlas (TCIA/TCGA). RESULTS: CSF volume and edema were independently and consistently associated with overall survival of glioblastoma patients in both datasets. Greater edema was associated with increased hazard or decreased survival [adjusted hazard ratio (aHR) with 95% confidence interval (CI): 1.34 [1.08-1.67], p = 0.008 (institutional dataset); and, 1.44 [1.08-1.93], p = 0.013 (TCIA/TCGA dataset)]. Greater CSF volume also correlated with increased hazard or decreased survival [aHR 1.27 [1.02-1.59], p = 0.035 (institutional dataset), and 1.42 [1.03-1.95], p = 0.032 (TCIA/TCGA dataset)]. CONCLUSIONS: Higher brain CSF volume and higher edema levels at diagnosis are independently associated with decreased survival in glioblastoma patients. These results highlight the importance of a broader, quantitative brain-wide radiological analyses and invite investigations to understand tumor-related causes of increased edema and possibly increased CSF volume.