Role of Machine Learning in Liquid Biopsy of Brain Tumours.

Liquid biopsy has multiple benefits and is used extensively in other fields of oncology, but its role in neuro-oncology has been limited so far. Multiple tumour-derived materials like circulating tumour cells (CTCs), tumour-educated platelets (TEPs), cell-free DNA (cfDNA), circulating tumour DNA (ctDNA), and miRNA are studied in CSF, blood (plasma, serum) or urine. Large and complex amounts of data from liquid biopsy can be simplified by machine learning using various algorithms. By using this technique, we can diagnose brain tumours and differentiate low versus highgrade glioma and true progression from pseudo-progression. The potential of liquid biopsy in brain tumours has not been extensively studied, but it has a bright future in the coming years. Here, we present a literature review on the role of machine learning in liquid biopsy of brain tumours.

Comprehensive application of AI algorithms with TCR NGS data for glioma diagnosis.

T-cell receptor (TCR) detection can examine the extent of T-cell immune responses. Therefore, the article analyzed characteristic data of glioma obtained by DNA-based TCR high-throughput sequencing, to predict the disease with fewer biomarkers and higher accuracy. We downloaded data online and obtained six TCR-related diversity indices to establish a multidimensional classification system. By comparing actual presence of the 602 correlated sequences, we obtained two-dimensional and multidimensional datasets. Multiple classification methods were utilized for both datasets with the classification accuracy of multidimensional data slightly less to two-dimensional datasets. This study reduced the TCR ß sequences through feature selection methods like RFECV (Recursive Feature Elimination with Cross-Validation). Consequently, using only the presence of these three sequences, the classification AUC value of 96.67% can be achieved. The combination of the three correlated TCR clones obtained at a source data threshold of 0.1 is: CASSLGGNTEAFF_TRBV12_TRBJ1-1, CASSYSDTGELFF_TRBV6_TRBJ2-2, and CASSLTGNTEAFF_TRBV12_TRBJ1-1. At 0.001, the combination is: CASSLGETQYF_TRBV12_TRBJ2-5, CASSLGGNQPQHF_TRBV12_TRBJ1-5, and CASSLSGNTIYF_TRBV12_TRBJ1-3. This method can serve as a potential diagnostic and therapeutic tool, facilitating diagnosis and treatment of glioma and other cancers.

Novel insights toward diagnosis and treatment of glioneuronal and neuronal tumors in young adults.

Aim: Glioneuronal and neuronal tumors are rare primary central nervous system malignancies with heterogeneous features. Due to the rarity of these malignancies diagnosis and treatment remains a clinical challenge. Methods: Here we performed a narrative review aimed to investigate the principal issues concerning the diagnosis, pathology, and clinical management of glioneuronal tumors. Results: Diagnostic criteria have been recently overturned thanks to a better characterization on a histological and molecular biology level. The study of genomic alterations occurring within these tumors has allowed us to identify potential therapeutic targets including BRAF, FGFR, and PDGFRA. Conclusion: Techniques allowing molecular sequencing DNA methylation assessment of the disease are essential diagnostic tools. Targeting agents should be included in the therapeutic armamentarium after loco-regional treatment failure.

[Box: see text].

Breaking new ground: can artificial intelligence and machine learning transform papillary glioneuronal tumor diagnosis?

Papillary glioneuronal tumors (PGNTs), classified as Grade I by the WHO in 2016, present diagnostic challenges due to their rarity and potential for malignancy. Xiaodan Du et al.'s recent study of 36 confirmed PGNT cases provides critical insights into their imaging characteristics, revealing frequent presentation with headaches, seizures, and mass effect symptoms, predominantly located in the supratentorial region near the lateral ventricles. Lesions often appeared as mixed cystic and solid masses with septations or as cystic masses with mural nodules. Given these complexities, artificial intelligence (AI) and machine learning (ML) offer promising advancements for PGNT diagnosis. Previous studies have demonstrated AI's efficacy in diagnosing various brain tumors, utilizing deep learning and advanced imaging techniques for rapid and accurate identification. Implementing AI in PGNT diagnosis involves assembling comprehensive datasets, preprocessing data, extracting relevant features, and iteratively training models for optimal performance. Despite AI's potential, medical professionals must validate AI predictions, ensuring they complement rather than replace clinical expertise. This integration of AI and ML into PGNT diagnostics could significantly enhance preoperative accuracy, ultimately improving patient outcomes through more precise and timely interventions.

Serum beta2-microglobulin acts as a biomarker for severity and prognosis in glioma patients: a preliminary clinical study.

BACKGROUND: Gliomas are the deadliest malignant tumors of the adult central nervous system. We previously discovered that beta2-microglobulin (B2M) is abnormally upregulated in glioma tissues and that it exerts a range of oncogenic effects. Besides its tissue presence, serum B2M levels serve as biomarkers for various diseases. This study aimed to explore whether serum B2M levels can be used in the diagnosis and prognosis of gliomas. METHODS: Medical records from 246 glioma patients were retrospectively analyzed. The relationship between preoperative serum B2M levels and clinicopathological features was examined. Kaplan-Meier analysis, alongside uni- and multivariate Cox regression, assessed the association between B2M levels, systemic inflammatory markers, and glioma patient prognosis. Receiver operating characteristic (ROC) curve analysis evaluated the diagnostic significance of these biomarkers specifically for glioblastoma (GBM). RESULTS: Patients with malignant gliomas exhibited elevated preoperative serum B2M levels. Glioma patients with high serum B2M levels experienced shorter survival times. Multivariate Cox analysis determined the relationship between B2M levels (hazard ratio = 1.92, 95% confidence interval: 1.05-3.50, P = 0.034) and the overall survival of glioma patients. B2M demonstrated superior discriminatory power in distinguishing between GBM and non-GBM compared to inflammation indicators. Moreover, postoperative serum B2M levels were lower than preoperative levels in the majority of glioma patients. CONCLUSIONS: High preoperative serum B2M levels correlated with malignant glioma and a poor prognosis. Serum B2M shows promise as a novel biomarker for predicting patient prognosis and reflecting the therapeutic response.

Analysis of Metabolic Changes in Endogenous Metabolites and Diagnostic Biomarkers for Various Diseases Using Liquid Chromatography and Mass Spectrometry.

Analysis of endogenous metabolites in various diseases is useful for searching diagnostic biomarkers and elucidating the molecular mechanisms of pathophysiology. The author and collaborators have developed some LC/tandem mass spectrometry (LC/MS/MS) methods for metabolites and applied them to disease-related samples. First, we identified urinary conjugated cholesterol metabolites and serum N-palmitoyl-O-phosphocholine serine as useful biomarkers for Niemann-Pick disease type C (NPC). For the purpose of intraoperative diagnosis of glioma patients, we developed the LC/MS/MS analysis methods for 2-hydroxyglutaric acid or cystine and found that they could be good differential biomarkers. For renal cell carcinoma, we searched for various biomarkers for early diagnosis, malignancy evaluation and recurrence prediction by global metabolome analysis and targeted LC/MS/MS analysis. In pathological analysis, we developed a simultaneous LC/MS/MS analysis method for 13 steroid hormones and applied it to NPC cells, we found 6 types of reductions in NPC model cells. For non-alcoholic steatohepatitis (NASH), model mice were prepared with special diet and plasma bile acids were measured, and as a result, hydrophilic bile acids were significantly increased. In addition, we developed an LC/MS/MS method for 17 sterols and analyzed liver cholesterol metabolites and found a decrease in phytosterols and cholesterol synthetic markers and an increase in non-enzymatic oxidative sterols in the pre-onset stage of NASH. We will continue to challenge themselves to add value to clinical practice based on cutting-edge analytical chemistry methodology.

Pediatric high grade gliomas: A comprehensive histopathological, immunohistochemical and molecular integrated approach in routine practice.

Pediatric high grade gliomas have undergone remarkable changes in recent time with discovery of new molecular pathways. They have been added separately in current WHO 2021 blue book. All the entities show characteristic morphology and immunohistochemistry. Methylation data correctly identifies these entities into particular group of clusters. The pediatric group high grade glioma comprises- Diffuse midline glioma, H3K27-altered; Diffuse hemispheric glioma, H3G34-mutant; Diffuse pediatric-type high-grade glioma, H3-wild type & IDH-wild type; Infant hemispheric glioma and Epithelioid glioblastoma/Grade 3 pleomorphic xanthoastrocytoma and very rare IDH-mutant astrocytoma. However it is not always feasible to perform these molecular tests where cost-effective diagnosis is a major concern. Here we discuss the major entities with their characteristic histopathology, immunohistochemistry and molecular findings that may help to reach to suggest the diagnosis and help the clinician for appropriate treatment strategies. We have also made a simple algorithmic flow chart integrated with histopathology, immunohistochemistry and molecular characteristics for better understanding.

Glial and glioneuronal tumors: Navigating the complexity of evolving concepts and new classification.

The World Health Organization (WHO) published the 5th edition classification of tumors of central nervous system in 2021, commonly abbreviated as WHO CNS5, which became the new standard for brain tumor diagnosis and therapy. This edition dramatically impacted tumor diagnostics. In short it introduced new tumors, changed the names of previously recognized tumors, and modified the working definition of previously known tumors. The new system appears complex due to the integration of morphological and multiple molecular criteria. The most radical changes occurred in the field of glial and glioneuronal tumors, which constitutes the lengthy first chapter of this new edition. Herein we present an illustrative outline of the evolving concepts of glial and glioneuronal tumors. We also attempt to explain the rationales behind this substantial change in tumor classification and the challenges to update and integrate it into clinical practice. We aim to present a concise and precise roadmap to aid navigation through the intricate conceptual framework of glial and glioneuronal tumors in the context of WHO CNS5.

Identification and clinical validation of diverse cell-death patterns-associated prognostic features among low-grade gliomas.

Low-grade glioma (LGG) is heterogeneous at biological and transcriptomic levels, and it is still controversial for the definition and typing of LGG. Therefore, there is an urgent need for specific and practical molecular signatures for accurate diagnosis, individualized therapy, and prognostic evaluation of LGG. Cell death is essential for maintaining homeostasis, developing and preventing hyperproliferative malignancies. Based on diverse programmed cell death (PCD) related genes and prognostic characteristics of LGG, this study constructed a model to explore the mechanism and treatment strategies for LGG cell metastasis and invasion. We screened 1161 genes associated with PCD and divided 512 LGG samples into C1 and C2 subtypes by consistent cluster analysis. We analyzed the two subtypes' differentially expressed genes (DEGs) and performed functional enrichment analysis. Using R packages such as ESTIMATE, CIBERSOTR, and MCPcounter, we assessed immune cell scores for both subtypes. Compared with C1, the C2 subtype has a poor prognosis and a higher immune score, and patients in the C2 subtype are more strongly associated with tumor progression. LASSO and COX regression analysis screened four characteristic genes (CLU, FHL3, GIMAP2, and HVCN1). Using data sets from different platforms to validate the four-gene feature, we found that the expression and prognostic correlation of the four-gene feature had a high degree of stability, showing stable predictive effects. Besides, we found downregulation of CLU, FHL3, and GIMAP2 significantly impairs the growth, migration, and invasive potential of LGG cells. Take together, the four-gene feature constructed based on PCD-related genes provides valuable information for further study of the pathogenesis and clinical treatment of LGG.

The diagnostic challenge of lack of choline level elevation on 1H-MR spectroscopy in grade II-III gliomas.

The accurate diagnosis of brain tumour is very important in modern neuro-oncology medicine. Magnetic resonance spectroscopy (MRS) is supposed to be a promising tool for detecting cancerous lesions. However, the interpretation of MRS data is complicated by the fact that not all cancerous lesions exhibit elevated choline (Cho) levels. The main goal of our study was to investigate the lack of Cho lesion /Cho ref elevation in the population of grade II-III gliomas. 89 cases of gliomas grade II and III were used for the retrospective analysis - glioma (astrocytoma or oligodendroglioma) grade II (74 out of 89 cases [83%]) and III (15 out of 89 cases [17%]) underwent conventional MRI extended by MRS before treatment. Histopathological diagnosis was obtained either by biopsy or surgical resection. Gliomas were classified to the group of no-choline elevation when the ratio of choline measured within the tumour (Cho lesion ) to choline from NABT (Cho ref ) were equal to or lower than 1. Significant differences were observed between ratios of Cho lesion /Cr lesion calculated for no-choline elevation and glial tumour groups as well as in the NAA lesion /Cr lesion ratio between the no-choline elevation group and glial tumour group. With consistent data concerning choline level elevation and slightly lower NAA value, the Cho lesion /NAA lesion ratio is significantly higher in the WHO II glial tumour group compared to the no-choline elevation cases ( p < 0.000). In the current study the results demonstrated possibility of lack of choline elevation in patients with grade II-III gliomas, so it is important to remember that the lack of elevated choline levels does not exclude neoplastic lesion.

The Clinical Utility of a Tiered Approach to Pediatric Glioma Molecular Characterization for Resource-Limited Settings.

PURPOSE: Molecular characterization is key to optimally diagnose and manage cancer. The complexity and cost of routine genomic analysis have unfortunately limited its use and denied many patients access to precision medicine. A possible solution is to rationalize use-creating a tiered approach to testing which uses inexpensive techniques for most patients and limits expensive testing to patients with the highest needs. Here, we tested the utility of this approach to molecularly characterize pediatric glioma in a cost- and time-sensitive manner. METHODS: We used a tiered testing pipeline of immunohistochemistry (IHC), customized fusion panels or fluorescence in situ hybridization (FISH), and targeted RNA sequencing in pediatric gliomas. Two distinct diagnostic algorithms were used for low- and high-grade gliomas (LGGs and HGGs). The percentage of driver alterations identified, associated testing costs, and turnaround time (TAT) are reported. RESULTS: The tiered approach successfully characterized 96% (95 of 99) of gliomas. For 82 LGGs, IHC, targeted fusion panel or FISH, and targeted RNA sequencing solved 35% (29 of 82), 29% (24 of 82), and 30% (25 of 82) of cases, respectively. A total of 64% (53 of 82) of samples were characterized without targeted RNA sequencing. Of 17 HGG samples, 13 were characterized by IHC and four were characterized by targeted RNA sequencing. The average cost per sample was more affordable when using the tiered approach as compared with up-front targeted RNA sequencing in LGG ($405 US dollars [USD] v $745 USD) and HGGs ($282 USD v $745 USD). The average TAT per sample was also shorter using the tiered approach (10 days for LGG, 5 days for HGG v 14 days for targeted RNA sequencing). CONCLUSION: Our tiered approach molecularly characterized 96% of samples in a cost- and time-sensitive manner. Such an approach may be feasible in neuro-oncology centers worldwide, particularly in resource-limited settings.

Clinical characteristics and prognostic factors of adult brainstem gliomas: A retrospective analysis of histologically-proven 40 cases.

To illustrate the clinical characteristics and prognostic factors of adult patients pathologically confirmed with brainstem gliomas (BSGs). Clinical data of 40 adult patients pathologically diagnosed with BSGs admitted to Beijing Shijitan Hospital from 2009 to 2022 were recorded and retrospectively analyzed. The primary parameters included relevant symptoms, duration of symptoms, Karnofsky performance status (KPS), tumor location, type of surgical resection, diagnosis, treatment, and survival. Univariate and multivariate analyses were evaluated by Cox regression models. The gliomas were located in the midbrain of 9 patients, in the pons of 14 cases, in the medulla of 5 cases, in the midbrain and pons of 6 cases and invading the medulla and pons of 6 cases, respectively. The proportion of patients with low-grade BSGs was 42.5%. Relevant symptoms consisted of visual disturbance, facial paralysis, dizziness, extremity weakness, ataxia, paresthesia, headache, bucking, dysphagia, dysacousia, nausea, dysphasia, dysosmia, hypomnesia and nystagmus. 23 (57.5%) patients accepted stereotactic biopsy, 17 (42.5%) patients underwent surgical resection. 39 patients received radiotherapy and 34 cases were treated with temozolomide. The median overall survival (OS) of all patients was 26.2 months and 21.5 months for the median progression-free survival (PFS). Both duration of symptoms (P = .007) and tumor grading (P = .002) were the influencing factors for OS, and tumor grading was significantly associated with PFS (P = .001). Duration of symptoms for more than 2 months and low-grade are favorable prognostic factors for adult patients with BSGs.

Ensemble learning prediction framework for EGFR amplification status of glioma based on terahertz spectral features.

Epidermal growth factor receptor (EGFR) plays a pivotal role in the initiation and progression of gliomas. In particular, in glioblastoma, EGFR amplification emerges as a catalyst for invasion, proliferation, and resistance to radiotherapy and chemotherapy. Current approaches are not capable of providing rapid diagnostic results of molecular pathology. In this study, we propose a terahertz spectroscopic approach for predicting the EGFR amplification status of gliomas for the first time. A machine learning model was constructed using the terahertz response of the measured glioma tissues, including the absorption coefficient, refractive index, and dielectric loss tangent. The novelty of our model is the integration of three classical base classifiers, i.e., support vector machine, random forest, and extreme gradient boosting. The ensemble learning method combines the advantages of various base classifiers, this model has more generalization ability. The effectiveness of the proposed method was validated by applying an individual test set. The optimal performance of the integrated algorithm was verified with an area under the curve (AUC) maximum of 85.8 %. This signifies a significant stride toward more effective and rapid diagnostic tools for guiding postoperative therapy in gliomas.

Revisiting prognostic factors of gliomatosis cerebri in adult-type diffuse gliomas.

PURPOSE: There is lack of comprehensive analysis evaluating the impact of clinical, molecular, imaging, and surgical data on survival of patients with gliomatosis cerebri (GC). This study aimed to investigate prognostic factors of GC in adult-type diffuse glioma patients. METHODS: Retrospective chart and imaging review was performed in 99 GC patients from adult-type diffuse glioma (among 1,211 patients; 6 oligodendroglioma, 16 IDH-mutant astrocytoma, and 77 IDH-wildtype glioblastoma) from a single institution between 2005 and 2021. Predictors of overall survival (OS) of entire patients and IDH-wildtype glioblastoma patients were determined. RESULTS: The median OS was 16.7 months (95% confidence interval [CI] 14.2-22.2) in entire patients and 14.3 months (95% CI 12.2-61.9) in IDH-wildtype glioblastoma patients. In entire patients, KPS (hazard ratio [HR] = 0.98, P = 0.004), no 1p/19q codeletion (HR = 10.75, P = 0.019), MGMTp methylation (HR = 0.54, P = 0.028), and hemorrhage (HR = 3.45, P = 0.001) were independent prognostic factors on multivariable analysis. In IDH-wildtype glioblastoma patients, KPS (HR = 2.24, P = 0.075) was the only independent prognostic factor on multivariable analysis. In subgroup of IDH-wildtype glioblastoma with CE tumors, total resection of CE tumor did not remain as a significant prognostic factor (HR = 1.13, P = 0.685). CONCLUSIONS: The prognosis of GC patients is determined by its underlying molecular type and patient performance status. Compared with diffuse glioma without GC, aggressive surgery of CE tumor in GC patients does not improve survival.

Molecular Pathology of Gliomas.

Gliomas are the most common adult and pediatric primary brain tumors. Molecular studies have identified features that can enhance diagnosis and provide biomarkers. IDH1/2 mutation with ATRX and TP53 mutations defines diffuse astrocytomas, whereas IDH1/2 mutations with 1p19q loss defines oligodendroglioma. Focal amplifications of receptor tyrosine kinase genes, TERT promoter mutation, and loss of chromosomes 10 and 13 with trisomy of chromosome 7 are characteristic features of glioblastoma and can be used for diagnosis. BRAF gene fusions and mutations in low-grade gliomas and histone H3 mutations in high-grade gliomas also can be used for diagnostics.

Demographic bias in misdiagnosis by computational pathology models.

Despite increasing numbers of regulatory approvals, deep learning-based computational pathology systems often overlook the impact of demographic factors on performance, potentially leading to biases. This concern is all the more important as computational pathology has leveraged large public datasets that underrepresent certain demographic groups. Using publicly available data from The Cancer Genome Atlas and the EBRAINS brain tumor atlas, as well as internal patient data, we show that whole-slide image classification models display marked performance disparities across different demographic groups when used to subtype breast and lung carcinomas and to predict IDH1 mutations in gliomas. For example, when using common modeling approaches, we observed performance gaps (in area under the receiver operating characteristic curve) between white and Black patients of 3.0% for breast cancer subtyping, 10.9% for lung cancer subtyping and 16.0% for IDH1 mutation prediction in gliomas. We found that richer feature representations obtained from self-supervised vision foundation models reduce performance variations between groups. These representations provide improvements upon weaker models even when those weaker models are combined with state-of-the-art bias mitigation strategies and modeling choices. Nevertheless, self-supervised vision foundation models do not fully eliminate these discrepancies, highlighting the continuing need for bias mitigation efforts in computational pathology. Finally, we demonstrate that our results extend to other demographic factors beyond patient race. Given these findings, we encourage regulatory and policy agencies to integrate demographic-stratified evaluation into their assessment guidelines.

Dual-space disentangled-multimodal network (DDM-net) for glioma diagnosis and prognosis with incomplete pathology and genomic data.

Objective. Effective fusion of histology slides and molecular profiles from genomic data has shown great potential in the diagnosis and prognosis of gliomas. However, it remains challenging to explicitly utilize the consistent-complementary information among different modalities and create comprehensive representations of patients. Additionally, existing researches mainly focus on complete multi-modality data and usually fail to construct robust models for incomplete samples.Approach. In this paper, we propose adual-space disentangled-multimodal network (DDM-net)for glioma diagnosis and prognosis. DDM-net disentangles the latent features generated by two separate variational autoencoders (VAEs) into common and specific components through a dual-space disentangled approach, facilitating the construction of comprehensive representations of patients. More importantly, DDM-net imputes the unavailable modality in the latent feature space, making it robust to incomplete samples.Main results. We evaluated our approach on the TCGA-GBMLGG dataset for glioma grading and survival analysis tasks. Experimental results demonstrate that the proposed method achieves superior performance compared to state-of-the-art methods, with a competitive AUC of 0.952 and a C-index of 0.768.Significance. The proposed model may help the clinical understanding of gliomas and can serve as an effective fusion model with multimodal data. Additionally, it is capable of handling incomplete samples, making it less constrained by clinical limitations.

TGIF2 is a potential biomarker for diagnosis and prognosis of glioma.

Background: TGFB-induced factor homeobox 2 (TGIF2), a member of the Three-Amino-acid-Loop-Extension (TALE) superfamily, has been implicated in various malignant tumors. However, its prognostic significance in glioma, impact on tumor immune infiltration, and underlying mechanisms in glioma development remain elusive. Methods: The expression of TGIF2 in various human normal tissues, normal brain tissues, and gliomas was investigated using HPA, TCGA, GTEx, and GEO databases. The study employed several approaches, including Kaplan-Meier analysis, ROC analysis, logistic regression, Cox regression, GO analysis, KEGG analysis, and GSEA, to explore the relationship between TGIF2 expression and clinicopathologic features, prognostic value, and potential biological functions in glioma patients. The impact of TGIF2 on tumor immune infiltration was assessed through Estimate, ssGSEA, and Spearman analysis. Genes coexpressed with TGIF2 were identified, and the protein-protein interaction (PPI) network of these coexpressed genes were constructed using the STRING database and Cytoscape software. Hub genes were identified using CytoHubba plugin, and their clinical predictive value was explored. Furthermore, in vitro experiments were performed by knocking down and knocking out TGIF2 using siRNA and CRISPR/Cas9 gene editing, and the role of TGIF2 in glioma cell invasion and migration was analyzed using transwell assay, scratch wound-healing assay, RT-qPCR, and Western blot. Results: TGIF2 mRNA was found to be upregulated in 21 cancers, including glioma. High expression of TGIF2 was associated with malignant phenotypes and poor prognosis in glioma patients, indicating its potential as an independent prognostic factor. Furthermore, elevated TGIF2 expression positively correlated with cell cycle regulation, DNA synthesis and repair, extracellular matrix (ECM) components, immune response, and several signaling pathways that promote tumor progression. TGIF2 showed correlations with Th2 cells, macrophages, and various immunoregulatory genes. The hub genes coexpressed with TGIF2 demonstrated significant predictive value. Additionally, in vitro experiments revealed that knockdown and knockout of TGIF2 inhibited glioma cell invasion, migration and suppressed the epithelial-mesenchymal transition (EMT) phenotype. Conclusion: TGIF2 emerges as a potential biomarker for glioma, possibly linked to tumor immune infiltration and EMT.

Pathway-based stratification of gliomas uncovers four subtypes with different TME characteristics and prognosis.

Increasing evidences have found that the interactions between hypoxia, immune response and metabolism status in tumour microenvironment (TME) have clinical importance of predicting clinical outcomes and therapeutic efficacy. This study aimed to develop a reliable molecular stratification based on these key components of TME. The TCGA data set (training cohort) and two independent cohorts from CGGA database (validation cohort) were enrolled in this study. First, the enrichment score of 277 TME-related signalling pathways was calculated by gene set variation analysis (GSVA). Then, consensus clustering identified four stable and reproducible subtypes (AFM, CSS, HIS and GLU) based on TME-related signalling pathways, which were characterized by differences in hypoxia and immune responses, metabolism status, somatic alterations and clinical outcomes. Among the four subtypes, HIS subtype had features of immunosuppression, oxygen deprivation and active energy metabolism, resulting in a worst prognosis. Thus, for better clinical application of this acquired stratification, we constructed a risk signature by using the LASSO regression model to identify patients in HIS subtype accurately. We found that the risk signature could accurately screen out the patients in HIS subtype and had important reference value for individualized treatment of glioma patients. In brief, the definition of the TME-related subtypes was a valuable tool for risk stratification in gliomas. It might serve as a reliable prognostic classifier and provide rational design of individualized treatment, and follow-up scheduling for patients with gliomas.