Glioblastoma—A Contemporary Overview of Epidemiology, Classification, Pathogenesis, Diagnosis, and Treatment: A Review Article

Gliomas are the most common primary tumors of the central nervous system, among which glioblastoma (GBM) is the most aggressive and deadly form, accounting for 60% of all brain tumors in adults [1]. The current classification of the World Health Organization (WHO, 2021) has significantly increased the role of molecular markers in defining tumor subtypes, leading to a redefinition of the term “glioblastoma”, which refers to high-grade tumors that meet IDH-wildtype criteria and specific molecular and histological characteristics [2]. This change has important implications for diagnosis, prognosis, and clinical trial design [3].

In epidemiological terms, glioblastoma is the most common primary malignant brain tumor in adults, accounting for approximately 45–50% of all gliomas [4]. The global incidence of this disease varies significantly between countries and continents. The highest incidence rates are found in highly developed countries—in the United States (US), an average of 3.19 cases per 100,000 individuals per year, and in Australia, approximately 3.4/100,000 [5]. In Europe, these figures are similar to those in the US, although there are slight regional differences between northern and southern countries [6]. Recent population-based analyses indicate that overall incidence trends have remained relatively stable over the past decade, with a slight upward tendency in aging populations [5]. Moreover, demographic disparities are evident: higher incidence is reported among individuals of White ethnicity compared with Black or Asian populations [7], and socioeconomic differences continue to influence both diagnosis rates and outcomes [8]. The high incidence in developed countries may be partly related to better admittance to diagnostic imaging, higher life expectancy, and more complete case reporting, while in lower-income regions, the actual incidence may be underestimated due to underdiagnosis and limited access to the neurooncological care [1].

In terms of demographics, GBM most commonly occurs in individuals aged 55–75, with the peak incidence in the sixth decade of life [9]. The disease also shows gender differences, men are affected about 50% more often than women, which may be related to hormonal alterations, differences in the expression of X-linked genes, and environmental factors [10]. In children, glioblastoma multiforme is rare and has a different genetic profile, suggesting a different pathogenesis mechanism [11]. It is also important to emphasize that, according to the WHO CNS5 classification, the term “glioblastoma” applies exclusively to adult-type diffuse gliomas that are IDH-wildtype, while pediatric high-grade gliomas constitute a separate group defined by distinct molecular alterations [12]. Consequently, tumors formerly diagnosed as pediatric GBM are now classified under pediatric-type diffuse high-grade gliomas, which differ in epidemiology, biology, and prognosis from adult GBM [2].

Clinically, GBM remains a cancer with an extremely poor prognosis. Despite advances in imaging and molecular diagnostics and the introduction of integrated treatment involving maximum surgical resection, radiotherapy, and chemotherapy with temozolomide, the median survival of patients is approximately 14–18 months, and the five year survival rate does not exceed 5% [9]. Recurrence of the disease is almost inevitable and poses a significant clinical problem, resulting, among other causes, from the presence of cancer stem cells, resistance to treatment, and limited penetration of drugs through the blood–brain barrier [13].

The etiology of GBM remains largely unclear. Potential risk factors include previous exposure to ionizing radiation, immunosuppression, genetic predisposition (including Li-Fraumeni syndrome, Turcot syndrome, type 1 neurofibromatosis), and environmental factors [14,15]. Based on the above findings, glioblastoma multiforme remains one of the greatest challenges in modern neurooncology. The aim of this review is to present the current state of knowledge on the etiopathogenesis, molecular classification, epidemiology, diagnosis, and therapeutic strategies (both available and developing) for GBM. In addition, the article identify key directions for future research that could help improve patient prognosis and quality of life in this exceptionally malignant cancer.

In the last few years, the combination of molecular data and morphology has resulted in significant improvements in the classification of central nervous system (CNS) malignancies. Epigenetics plays a key role in shaping the complex biology of gliomas, and its impact extends far beyond the traditionally discussed MGMT promoter methylation. Recent studies have shown that global patterns of DNA methylation, histone modifications, and disruptions of chromatin architecture constitute a distinct layer of tumor regulation that significantly influences intratumoral heterogeneity, glioma cell plasticity, and clonal evolution [16]. The chemical classification of DNA methylation has made it possible to identify unique molecular subtypes of glioblastoma—including classical, proneural, mesenchymal, and neural—each with its own epigenetic landscape and gene expression profile. This approach has highlighted that epigenetics not only passively reflects the tumor state but actively shapes it by driving, among other processes, the phenotypic switching of tumor cells between proneural and mesenchymal states—one of the mechanisms underlying clinical aggressiveness and therapeutic resistance [17].

Alterations in DNA methylation and histone modifications modulate the activity of key pathways such as EGFR, PI3K/AKT, Notch, and Wnt, thereby regulating the behavior of tumor cells and glioblastoma stem cells (GSCs). Moreover, epigenetic dysregulation affects interactions between tumor cells and the microenvironment, for example by controlling the transcription of cytokines, chemokines, and immune checkpoint ligands, contributing to the immunosuppressive nature of the GBM TME [18,19]. Methylation patterns are stable, heritable, and at the same time responsive to environmental cues, making them a central mechanism driving dynamic phenotypic plasticity and, consequently, the development of treatment resistance.

In the context of glioblastoma heterogeneity, single-cell sequencing data have been particularly transformative, reshaping our understanding of the tumor’s internal architecture over the last decade. Single-cell RNA-seq studies demonstrated that GBM consists of multiple coexisting subpopulations, differing in transcriptional profiles, metabolic states, degrees of differentiation, and vulnerabilities to therapy [20,21]. Importantly, these populations respond to therapeutic pressure in markedly different ways, and treatment frequently leads to selective expansion of resistant clones. Further studies have confirmed the presence of a continuous spectrum of cellular states, rather than sharply defined subtypes, suggesting that GBM heterogeneity is dynamic and fluid, regulated by an intertwined epigenetic–transcriptional network [22,23]. Chen et al. additionally observed that macrophage subpopulations within the tumor may undergo epigenetically driven EZH2 activation, promoting pro-tumor phenotypes and further destabilizing the tumor’s cellular landscape [24].

When integrated with methylation analyses and multi-regional tumor sampling, single-cell data describe GBM clonal evolution as a nonlinear process involving the simultaneous branching of subclones, their cooperation, and occasionally, phenotypic convergence [25,26]. This gives rise to distinct ecological niches within a single tumor mass, including hypoxic zones, perivascular regions, and areas enriched in GSCs, all of which critically influence tumor progression and therapeutic resistance [27,28,29].

The contemporary model of glioblastoma pathogenesis therefore assumes a tight correlation between genetic mutations (EGFR, PTEN, TP53, TERT), epigenetics, and microenvironmental niches, which synergistically define tumor biology, aggressiveness, recurrence potential, and treatment resistance. Integrating epigenetics into practical tumor classification (as proposed in recent work on DNA-methylation-based CNS tumor classes) aligns with the WHO CNS5 framework, which emphasizes the need for multidimensional data integration to improve diagnostic and prognostic precision [30,31] (Figure 1).

The clinical picture of glioblastoma multiforme (GBM) is varied and depends on the location of the tumor, its growth rate, and its massive effects. Patients most often present with: focal symptoms such as limb weakness, speech impairment (aphasia), visual disturbances, and seizures occur depending on the part of the brain affected by the tumor; damage to motor areas leads to muscle weakness, while involvement of speech centers leads to communication difficulties; symptoms of increased intracranial pressure, as headaches, nausea, vomiting, and impaired consciousness are the result of increased pressure within the skull caused by tumor growth and brain swelling; or cognitive and behavioral changes in which many patients experience memory impairment, personality changes, disorientation, and psychomotor retardation, reflecting the tumor’s impact on higher brain functions [32,33]. The course of the disease is characterized by a high tendency for local infiltration, rapid growth, and resistance to treatment, which means that despite aggressive therapy, relapses are almost inevitable [34]. Table 1 summarizes the most common clinical manifestations and the typical course of glioblastoma (GBM).

Glioblastoma multiforme manifests a wide spectrum of neurological disorders, the nature of which depends on the location and extent of the tumor. The disease progresses very aggressively, and despite advances in diagnosis and treatment, the prognosis remains poor. Rapid diagnosis and an interdisciplinary approach are crucial to improving patients’ quality of life [32,35,37].

The diagnosis of glioblastoma multiforme (GBM) is one of the most important stages in clinical management, determining both the therapeutic strategy and the patient’s prognosis. Due to its high biological heterogeneity and aggressive course, the diagnosis of GBM requires an integrated approach combining classical imaging methods, histopathological analyses, and advanced molecular techniques [2].

In recent years, neurooncological diagnostics have undergone a significant transformation, from traditional morphological criteria to a complex molecular-genetic profile. According to the 2021 WHO classification, changes have been introduced, emphasizing the need to include genetic markers such as TERT mutations, IDH status, and MGMT promoter methylation [38]. Today, the diagnosis of GBM is therefore based on a so-called integrated diagnostic model, combining histological, immunohistochemical, molecular, and imaging data [13]. At the same time, the development of non-invasive technologies such as liquid biopsy, radiogenomics, and next-generation sequencing (NGS) allows for assessment of minimal residual disease (MRD), increasingly accurate monitoring of disease progression, and importantly earlier detection of recurrence [32].

Glioblastoma multiforme remains a highly malignant tumor characterized by a short median survival. Multimodal treatment combining surgery, radiotherapy, and chemotherapy offers hope. In recent years, immunological and viral strategies, molecularly targeted therapies, and modern drug delivery technologies have been rapidly developing. All aimed at overcoming tumor heterogeneity and its highly immunosuppressive microenvironment. The first step in treatment remains a maximally safe resection, which reduces tumor mass and improves the volumetric effects of adjuvant therapies. However, the diffuse, infiltrative nature of GBM usually prevents cure with surgery alone. To date, the mainstay of first-line therapy after tumor resection is fractionated radiotherapy (usually 60 Gy/30 fractions) with concurrent and adjuvant temozolomide (Stupp protocol). In clinical practice, modifications (e.g., different doses, shorter treatment schedules in elderly patients) and combinations with new immunological methods are also being investigated. However, temozolomide remains the main systemic drug, usually providing greater benefit in patients with MGMT promoter methylation. Intensified chemotherapy has not yet replaced standard temozolomide in first-line treatment [68]. However, new therapies are known. Personalized neoantigen vaccines (peptide, mRNA, dendritic) have the ability to induce T-cell responses and local lymphocytic infiltration in GBM. Early clinical trials show promising signs of immunogenicity, although the impact on survival needs to be confirmed in randomized trials and depends on factors such as the use of glucocorticosteroids [81]. CAR-T cells targeting antigens such as IL13Rα2, EGFRvIII, or EGFR (in multitarget constructs) and administered intrathecally/directly into the CNS cavity show the ability to elicit rapid responses in some patients. The latest reports (2025) describe significant, though often transient, tumor regressions with acceptable toxicity, and intensive work is underway to add further targets and improve the durability of the effect [82,83]. BiTE (bispecific T cell activating) molecules, which bring T cells closer to tumor cells, may be a promising adjunct to targeted immunotherapy. Preliminary studies with EGFRvIII/EGFR demonstrate their activity in preclinical models and early-phase clinical trials [84]. Oncolytic adenoviruses (e.g., DNX-2401/Delta-24-RGD) and other vectors (PVSRIPO—oncolytic poliovirus-like virus) have passed phase I/II and, in combination with checkpoint inhibitors, show clinical signs of improved response in some patients; DNX-2401 + pembrolizumab showed increased response and improved 12-month survival in early phase studies. In short, oncolytics stimulate both direct lytic antitumor activity and secondary immune response [85]. Device-based therapies are also known—Tumour Treating Fields (TTFields). TTFields (low-intensity alternating electric field, clinical device “Optune”) added to the standard showed improved survival in initial studies, and real-world data and reviews from 2023 to 2025 confirm clinical utility in selected patients; issues related to treatment availability and compliance remain relevant [86]. Nanotechnologies and drug delivery systems, as well as stem cell-based therapies, offer new perspectives in the treatment of glioblastoma multiforme. Nanoparticles, liposomes, and other carrier systems are designed to cross the blood–brain barrier, selectively deliver cytostatics/oligonucleotides, and reduce systemic toxicity. Despite encouraging preclinical results, clinical outcomes remain limited by issues with the distribution, elimination, and immunogenicity of nanomaterials [85]. Neural and mesenchymal stem cells are being considered as carriers of prodrugs, gene vectors, or regenerative agents due to their tropism for tumor sites; safety aspects (transformation potential) and function control remain under investigation [87].

Recent studies have demonstrated that the heparanase inhibitor RDS 3337 significantly modulates the balance between apoptosis and autophagy in U87 glioma cells, providing novel insights into the molecular mechanisms governing cell death and potential therapeutic targets. Pharmacological inhibition of heparanase with RDS 3337 resulted in the accumulation of the lipidated form of LC3-II and elevated p62/SQSTM1 levels, indicative of impaired autophagic-lysosomal flux. Concomitantly, autophagy inhibition was associated with activation of apoptotic pathways, as evidenced by increased levels of cleaved caspase-3, the appearance of cleaved PARP1, and DNA fragmentation. These findings suggest that heparanase supports autophagic processes, and its inhibition can shift the cellular response toward apoptosis, potentially enhancing the susceptibility of cancer cells to anticancer therapies. Accordingly, heparanase inhibitors such as RDS 3337 represent a promising strategy to modulate cell death pathways and control tumor progression by regulating the interplay between autophagy and apoptosis [88].

Immunotherapy for glioblastoma (GBM) has generated considerable interest but, to date, has produced limited clinical success. Multiple factors intrinsic to GBM biology help explain these disappointing outcomes. First, the GBM tumor microenvironment (TME) is potently immunosuppressive: tumor and stromal cells secrete TGF-β and other inhibitory mediators, recruit immunosuppressive myeloid populations, and upregulate checkpoint ligands, collectively creating barriers to effective antitumor T-cell responses [27,28]. Second, profound intratumoral heterogeneity and plasticity—including the presence of glioblastoma stem cells (GSCs) occupying protective niches such as perivascular and hypoxic zones—generate diverse cellular states that can evade immune recognition and repopulate the tumor after selective pressure [26,29,30,80]. Third, specific genetic alterations commonly found in GBM can shape immune resistance: for example, loss of PTEN has been linked to a more immunosuppressive microenvironment and reduced sensitivity to immune-based therapies. These features reduce neoantigen burden, impair antigen presentation, and limit productive T-cell infiltration, all of which blunt the efficacy of immune checkpoint inhibitors [20,27].

Mechanisms of resistance to checkpoint blockade in GBM are multifactorial and often overlapping. Antigenic heterogeneity and antigen loss lead to immune escape after targeted immune activation; immunosuppressive myeloid cells (TAMs) and regulatory populations can actively suppress effector T cells; impaired trafficking across the blood–brain barrier (and intratumoral stromal barriers) limits immune cell access; and adaptive resistance mechanisms upregulate alternative inhibitory pathways beyond PD-1/PD-L1, necessitating combinatorial blockade [26,27,28]. In addition, the low-to-moderate tumor mutational burden typical of many GBMs limits the abundance of neoantigens recognizable by the adaptive immune system, reducing the probability of durable responses to monotherapy checkpoint inhibitors [27,28].

Given these obstacles, contemporary efforts have shifted toward rational combination approaches designed to remodel the TME, increase antigen exposure, and produce more durable antitumor immunity. Oncolytic virotherapy represents a particularly attractive partner: oncolytic viruses can selectively lyse tumor cells, release tumor antigens in an inflammatory context, and reprogram the local immune milieu—thereby converting immunologically “cold” regions into “hot” ones and enhancing responsiveness to checkpoint blockade or other immunomodulators [82,85]. Early translational and preclinical work suggests that combining oncolytic platforms with immune stimulatory agents (e.g., cytokines, TLR agonists) or immune checkpoint inhibitors can potentiate antitumor T-cell activity and overcome some mechanisms of resistance, although these approaches still face delivery, safety, and regulatory challenges [82,85,87].

Other combinatorial strategies under active investigation include vaccines (personalized peptide or neoantigen vaccines) used to broaden the T-cell repertoire, and adoptive cellular therapies (e.g., CAR T cells) engineered to recognize GBM-associated antigens. Vaccination approaches have shown promise in inducing immune responses in selected patients, but clinical benefits are often transient, likely due to antigen heterogeneity and immune suppression within the TME [81]. CAR T-cell therapy in GBM has demonstrated feasibility in preclinical and early clinical studies, yet efficacy has been limited by antigen heterogeneity, on-target off-tumor toxicity concerns, and the hostile intratumoral environment; combination strategies that include cytokine support, epigenetic modulators, or local oncolytic vectors may be required to improve persistence and function of transferred cells [82,83].

To maximize the potential of immunotherapy in GBM, future trials should emphasize rational patient selection and biomarker-driven designs: identifying tumors with more permissive immune microenvironments, specific genetic alterations that predict benefit (or resistance), or particular methylation/epigenetic signatures could prioritize patients most likely to respond. Integration of single-cell and spatial profiling will be essential to understand the cellular and molecular determinants of response and resistance and to guide combinatorial regimens that target tumor cells, GSC niches, and immunosuppressive myeloid compartments simultaneously [26,79,80]. In summary, while monotherapy immune checkpoint inhibition has largely failed to deliver transformative benefit in unselected GBM cohorts, carefully designed combination strategies—particularly those that convert the TME, broaden antigen presentation, and counteract myeloid-mediated suppression—remain the most promising path forward [27,28,81,82,83,85,87].

Future therapeutic strategies for glioblastoma (GBM) are increasingly moving toward highly personalized, molecularly guided approaches supported by real-time adaptive clinical trial designs. Advances in genomic, epigenomic, and transcriptomic characterization—including single-cell and spatial profiling—have revealed the profound heterogeneity and dynamic plasticity of GBM, demonstrating that each tumor comprises multiple coexisting cellular states that evolve under therapeutic pressure [26,79,80]. This complexity underscores the need for individualized treatment concepts that incorporate a patient’s specific mutational profile, transcriptional program, epigenetic landscape, and immune microenvironment. Emerging personalized strategies include molecular subtype–based classification consistent with contemporary WHO CNS5 recommendations, which use integrated genomic and methylation signatures to guide prognosis and therapeutic choice [2,38]. Additionally, neoantigen-targeted vaccines and engineered cellular therapies, such as CAR T cells, are increasingly being designed to match patient-specific antigenic repertoires, aiming to overcome the challenges posed by antigenic heterogeneity and immunosuppression [81,82,83]. Insights into chromatin state, methylation patterns, and epigenetically driven signaling may further support the selection of synergistic combination therapies capable of reshaping tumor behavior and immune responsiveness [26,79].

As these individualized approaches become more sophisticated, progress in GBM therapy will increasingly depend on the implementation of adaptive clinical trial frameworks that allow investigators to modify therapeutic arms based on continuously updated biological and clinical data. Because GBM evolves rapidly and recurrent tumors often differ markedly from their initial presentation, adaptive designs provide a mechanism to align treatment decisions with the tumor’s current molecular profile rather than relying on static baseline assessments [25,80]. Such trial models can integrate serial multiomic monitoring, including repeat biopsies, liquid biopsy assays, and longitudinal single-cell or methylation analyses, enabling dynamic adjustment of treatment in response to emerging resistance mechanisms. This flexibility facilitates rational testing of combination therapies, such as pairing oncolytic viruses with checkpoint inhibitors or immune-stimulatory agents to remodel the tumor microenvironment, or selecting patients for vaccine-based or cellular immunotherapies on the basis of real-time antigen-expression patterns [81,82,83,85,87]. By uniting precision oncology with continuously adaptive therapeutic evaluation, these future directions aim to overcome the historical limitations of fixed-design trials and to better address the biological variability and evolutionary dynamics that define GBM. As multiomic technologies become increasingly accessible in clinical practice, such adaptive, personalized frameworks may ultimately offer a path toward more durable and meaningful treatment responses in a disease that has long resisted conventional therapeutic strategies. A summary of described therapies have been presented in Table 2.

Glioblastoma multiforme (GBM) is one of the most aggressive tumors of the central nervous system, characterized by high genetic heterogeneity, invasiveness, and resistance to treatment. Despite advances in molecular diagnostics, the prognosis remains poor, with a median survival of approximately 15 months. The introduction of the integrated WHO 2021 classification, combining histological and genetic assessment (including IDH, MGMT, TERT, EGFR, CDKN2A/B status) has enabled more precise diagnosis and prognosis. The development of modern methods such as NGS, radiogenomics, and liquid biopsy contributes to more accurate disease monitoring and represents a step towards precision medicine. The standard of care is still based on maximal resection, radiotherapy, and hepatemozolomide (Stupp regimen). However, frequent recurrences result from the presence of glioblastoma stem cells and the limited efficacy of drugs that cross the blood–brain barrier. In response, new therapies are being developed: immunotherapy (CAR-T, dendritic vaccines), virotherapy, Tumour Treating Fields, and drug nanocarriers. The future of GBM treatment involves further personalization of therapy, integration of molecular, imaging, and clinical data, and the use of artificial intelligence. Such an interdisciplinary, adaptive approach offers an actual chance to improve the prognosis and quality of life of patients with this exceptionally malignant tumor.

Conceptualization, K.K., M.Z. and M.G.-S.; investigation, K.K. and K.B.; writing—original draft preparation, K.K. and K.B.; writing—review and editing, S.Ł., M.Z. and M.G.-S.; visualization, K.K. and M.G.-S.; supervision, S.Ł., M.Z. and M.G.-S.; funding acquisition, M.G.-S. All authors have read and agreed to the published version of the manuscript.

Not applicable.

Not applicable.

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

The authors declare no conflicts of interest.

The APC have been covered by Medical University of Bialystok (B.SUB.25.318).