Population-based cancer registries indicate that glioblastoma has an incidence of ~3.2 cases per 100,000 person-years (1). The disease predominantly affects adults and older individuals. By contrast, glioblastoma is relatively uncommon in children and adolescents aged 0-19 years, accounting for only ~2.9% of all brain and other CNS tumors reported in this age group. The incidence increases steadily with an advancing age and peaks in older adults, with a median age at diagnosis of ~64 years in large epidemiological series. In the majority of cases, no predisposing hereditary cancer syndrome or identifiable genetic susceptibility is present (1). The etiology of glioblastoma remains largely unknown, with the majority of cases arising sporadically and without identifiable environmental or inherited predisposing factors. Aside from prior exposure to therapeutic cranial irradiation and rare cancer predisposition syndromes, no definitive causal factors have been established (1).

Under the 2021 WHO classification, glioblastoma is defined as an IDH-wildtype diffuse astrocytic tumor that fulfills CNS WHO grade 4 criteria (7). Diffuse astrocytic tumors harboring IDH mutations are classified separately as astrocytoma, IDH-mutant, CNS WHO grade 2-4, reflecting their distinct molecular biology and clinical behavior (7,9). Historically, glioblastomas were subdivided into so-called primary (de novo) tumors, which arise without a known precursor lesion, and secondary tumors, which develop through malignant progression from lower-grade diffuse astrocytomas (9). These historical designations describe two major biological pathways, characterized by rapidly presenting IDH-wildtype tumors in older adults and more slowly evolving IDH-mutant tumors in younger patients. However, these terms are no longer formal diagnostic entities within the current WHO framework, having been superseded by molecularly defined classifications (7,9).

The clinical presentation of glioblastoma is largely determined by tumor location, size, growth kinetics and the extent of associated peritumoral edema. Patients commonly present with a progressive headache, focal neurological deficits, such as hemiparesis, aphasia, or visual field disturbances, seizures and cognitive or personality changes resulting from tumor infiltration or mass effect (15). Lesions involving the frontal or temporal lobes frequently manifest with behavioral changes, executive dysfunction, or new-onset epilepsy, whereas tumors affecting the motor cortex or internal capsule more often present with focal motor weakness. In a minority of cases, patients may present acutely with stroke-like symptoms caused by tumor-associated hemorrhage or sudden worsening of cerebral edema leading to increased intracranial pressure (15).

Neuroimaging, most commonly magnetic resonance imaging, typically demonstrates a heterogeneously enhancing mass with central necrosis, surrounding vasogenic edema, and associated mass effect. Despite these characteristic features, imaging alone cannot reliably distinguish glioblastoma from other high-grade gliomas or metastatic brain lesions. Consequently, a histopathological examination combined with the molecular analysis of tumor tissue obtained through stereotactic biopsy or surgical resection remains essential for a definitive diagnosis and classification (7,15).

The 2021 WHO classification formalized an integrated diagnostic framework in which histopathological features are combined with key molecular alterations to define diffuse gliomas. Within this system, glioblastoma is defined as an IDH-wildtype diffuse astrocytic tumor that demonstrates at least one criterion of CNS WHO grade 4 disease, including microvascular proliferation, necrosis, telomerase reverse transcriptase (TERT) promoter mutation, epidermal growth factor receptor (EGFR) amplification, or the combined presence of whole-chromosome 7 gain and chromosome 10 loss (7). By contrast, diffuse astrocytic tumors harboring IDH mutations are classified separately as astrocytoma, IDH-mutant, CNS WHO grade 2-4, regardless of histological grade. These tumors are biologically distinct from glioblastoma and are associated with a substantially different clinical behavior and prognosis (7,9).

Mutations in the IDH1 and IDH2 genes represent a fundamental molecular determinant in diffuse gliomas. IDH-wildtype tumors constitute the vast majority of glioblastomas and correspond to the canonical aggressive phenotype that typically presents in older adults with rapid clinical progression. By contrast, IDH-mutant diffuse astrocytomas more commonly arise in younger patients, often evolve over a period of several years from lower-grade precursor lesions, and are associated with a significantly prolonged overall survival (9).

At the molecular level, IDH mutations result in the accumulation of the oncometabolite 2-hydroxyglutarate, which drives widespread epigenetic reprogramming and contributes to the development of the glioma CpG island methylator phenotype. Across multiple clinical series, the presence of an IDH mutation has been consistently associated with an improved response to alkylating chemotherapy and a longer survival, independent of MGMT promoter methylation status and treatment modality (9).

MGMT promoter methylation is the most clinically relevant predictive biomarker for response to temozolomide in IDH-wildtype glioblastoma. The MGMT gene encodes a DNA repair enzyme that removes alkyl adducts from the O⁶ position of guanine, thereby directly reversing the cytotoxic DNA lesions induced by temozolomide. The hypermethylation of the MGMT promoter leads to the transcriptional silencing of the gene, reduced DNA repair capacity and an increased susceptibility of tumor cells to temozolomide-induced cell death (8).

The pivotal EORTC-NCIC trial demonstrated a marked survival advantage for patients with MGMT-methylated glioblastoma treated with combined chemoradiation, with a median overall survival of 23.4 months vs. 12.6 months in those with unmethylated tumors (6). These findings established MGMT promoter methylation as both a potent prognostic marker and a critical determinant of therapeutic benefit from temozolomide in clinical practice (8).

The G-CIMP describes a distinct epigenetic state characterized by widespread promoter hypermethylation and is strongly associated with IDH-mutant tumors. G-CIMP-positive gliomas are mainly observed in younger patient populations, demonstrate a lower proliferative activity, and are associated with significantly improved clinical outcomes compared with G-CIMP-negative tumors (9). The recognition of this phenotype further refines prognostic stratification within IDH-mutant astrocytomas and may have implications for responsiveness to epigenetic or molecularly targeted therapies (9).

In addition to molecular biomarkers, several clinical and treatment-related variables provide independent prognostic information in glioblastoma. A younger age at diagnosis and good functional performance status are consistently associated with longer survival across multiple studies (16,17). The extent of surgical resection is particularly influential, with volumetric analyses demonstrating that resection of at least 95 percent of contrast-enhancing tumor is associated with a significant survival advantage compared with subtotal resection or biopsy alone (16,17). Conversely, tumors involving deep or eloquent brain regions, such as the motor cortex, language areas, or basal ganglia, are often less amenable to aggressive resection and are associated with poorer outcomes (16,17).

Collectively, age, performance status, IDH mutation status, MGMT promoter methylation and the extent of resection constitute the core elements of contemporary prognostic models used to guide risk stratification and therapeutic decision-making in glioblastoma.

Maximal safe surgical resection is the cornerstone of initial glioblastoma management when anatomically and clinically feasible (6,17). The primary surgical objective is to achieve the greatest possible cytoreduction of contrast-enhancing tumor, while preserving neurological function and maintaining quality of life. Advances in neurosurgical techniques, including neuronavigation, intraoperative neuromonitoring, awake language mapping, intraoperative ultrasound, intraoperative magnetic resonance imaging and 5-aminolevulinic acid-guided fluorescence surgery, have significantly improved the extent of resection, while minimizing operative morbidity (6,17).

Multiple studies have demonstrated that complete or near-complete resection of enhancing tumor, commonly defined as at least 95% volumetric reduction, is associated with significantly prolonged survival compared with subtotal resection or biopsy alone (6,15,17,18). In large clinical series, median overall survival increases from ~9-10 months following biopsy alone to 15-18 months following gross-total resection when combined with standard post-operative chemoradiation (17). Nevertheless, only ~50-70% of patients are candidates for near-complete resection. Tumors located in deep or eloquent brain regions, poor baseline performance status, advanced age, or significant medical comorbidities frequently limit the feasibility of aggressive surgical intervention. In such cases, patients typically undergo limited debulking or stereotactic biopsy and experience substantially poorer clinical outcomes (6,17).

Post-operative external beam radiotherapy remains a critical component of standard glioblastoma treatment. The established regimen delivers a total dose of 60 Gy in daily fractions of 1.8-2.0 Gy over a 6-week period (6). Randomized clinical trials evaluating radiation dose escalation beyond 60 Gy have failed to demonstrate a survival benefit and have instead shown increased rates of late neurotoxicity, including radiation necrosis and cognitive impairment (6). Contemporary three-dimensional conformal and intensity-modulated radiotherapy techniques allow more precise target coverage, while reducing radiation exposure to surrounding normal brain tissue and critical structures (6).

For fit, non-elderly patients, conventional fractionation to 60 gray remains the standard approach (16). In elderly or medically frail patients, hypofractionated radiotherapy schedules, such as 40-50 Gy delivered in 15-20 fractions, have demonstrated comparable survival outcomes with improved tolerability and shorter overall treatment duration (15,19). The selection of fractionation strategy is generally guided by patient age, functional status, comorbidities, and individual treatment goals (15,19).

Temozolomide is the standard systemic therapy for patients newly diagnosed with glioblastoma (6,20). It is an oral alkylating agent with favorable CNS penetration that exerts cytotoxic effects primarily through methylation of the O⁶ position of guanine and the N7 position of adenine (20). In the pivotal EORTC-NCIC trial, the addition of concurrent and adjuvant temozolomide to radiotherapy improved median overall survival from 12.1 to 14.6 months and increased the 2-year survival rate from 10 to 26% (6). These findings established the so-called Stupp protocol as the global standard of care for newly diagnosed glioblastoma (6).

Prior research evaluating alternative alkylating chemotherapy regimens, including the combination of procarbazine, lomustine and vincristine (PCV), has not demonstrated a clear survival benefit in patients newly diagnosed with glioblastoma (18). In a phase III randomized trial, Cairncross et al (18) reported that the addition of PCV chemotherapy failed to produce a meaningful improvement in clinical outcomes and was associated with greater toxicity compared with standard treatment, underscoring the limited role of this regimen in this disease setting. These results further reinforced the adoption of temozolomide-based chemoradiation as the preferred therapeutic approach

During the concurrent chemoradiation phase, temozolomide is administered at a daily dose of mg/m² throughout the 6-week course of radiotherapy. Following a 4-week treatment break, adjuvant temozolomide is typically administered at 150-200 mg/m² on days 1-5 of each 28-day cycle for six to twelve cycles, depending on patient tolerance and institutional practice (6,20). The most common adverse effects include hematologic toxicity, particularly thrombocytopenia and lymphopenia, as well as fatigue and nausea. Although dose reductions or early discontinuation are required in a substantial proportion of patients, cumulative temozolomide exposure appears to be associated with improved clinical outcomes, particularly in tumors with MGMT promoter methylation (20).

Glioblastoma is characterized by a high propensity for early recurrence and therapeutic failure, reflecting the convergence of multiple resistance mechanisms operating at different biological levels. These include limitations in drug delivery, enhanced DNA repair capacity, hierarchical cellular organization, an immunosuppressive tumor microenvironment and extensive intratumoral heterogeneity with dynamic clonal evolution under therapeutic pressure (21-27).

The blood-brain barrier represents a major anatomic and functional obstacle to effective systemic therapy in glioblastoma. It is composed of specialized endothelial cells connected by tight junctions, along with a basement membrane, pericytes and astrocytic end-feet, together forming a highly selective interface between the systemic circulation and CNS parenchyma (3). This structure restricts the passage of most macromolecules and many small-molecule agents, such that an estimated 98% of systemically administered drugs exhibit limited penetration into the brain (3).

Although glioblastoma disrupts blood-brain barrier integrity in regions of contrast enhancement, barrier function is frequently preserved in infiltrative tumor margins. This spatial heterogeneity results in regions of subtherapeutic drug exposure within the tumor, often referred to as drug-penetration deserts, which contribute to treatment failure (3). In addition, efflux transporters such as P-glycoprotein and multidrug resistance-associated proteins are overexpressed in both brain endothelium and glioblastoma cells, actively exporting chemotherapeutic agents in an ATP-dependent manner and further reducing effective intratumoral drug concentrations (3).

Strategies under investigation to overcome blood-brain barrier-related resistance include osmotic or pharmacologic barrier disruption, nanoparticle-based drug delivery systems, convection-enhanced delivery via intraparenchymal catheters, and the development of engineered biologics with improved CNS penetration (3,22).

Temozolomide exerts its primary cytotoxic effect through the formation of O⁶-methylguanine DNA adducts, which mispair with thymine during DNA replication and initiate mismatch repair-mediated cell death when unrepaired (20,23). MGMT directly counteracts this mechanism by removing methyl groups from the O⁶ position of guanine, thereby restoring normal base pairing and preventing activation of the mismatch repair pathway (8,23). A high expression of MGMT, typically driven by an unmethylated MGMT promoter or transcriptional upregulation, therefore confers intrinsic resistance to temozolomide (8,23).

In addition to MGMT activity, defects in mismatch repair genes, such as MSH6 or MLH1 can paradoxically promote resistance by preventing recognition of O⁶-methylguanine-thymine mismatches. This allows tumor cells to evade lethal mismatch repair cycles and continue replicating despite ongoing DNA damage (23). Together, these mechanisms explain early treatment failure in a subset of patients and underscore why MGMT promoter methylation, although highly informative, remains an incomplete predictor of sensitivity to alkylating chemotherapy (8,23).

TTFields are low-intensity, intermediate-frequency alternating electric fields in the range of 100-300 kilohertz, delivered non-invasively through transducer arrays applied to the scalp. These fields selectively disrupt mitosis in rapidly dividing tumor cells by interfering with mitotic spindle formation, chromosome segregation and cytokinesis, ultimately leading to mitotic catastrophe and apoptosis (10). Additional biological effects include the impairment of DNA damage repair, induction of autophagy, reduced cellular invasiveness and the potential modulation of blood-brain barrier permeability (10,28).

In the phase III EF-14 trial, the addition of tumor-treating fields to maintenance temozolomide following chemoradiation significantly prolonged median progression-free survival from 4.0-6.7 months and median overall survival from 16.0-20.9 months, corresponding to a hazard ratio of 0.63(10). The principal adverse effect was low-grade scalp irritation at array application sites, while systemic toxicity was minimal (28). Real-world registry data from the PRiDE study have confirmed the feasibility and safety of tumor-treating fields in broader clinical populations, with survival outcomes comparable to or slightly exceeding those observed in clinical trials (28). However, prolonged daily usage requirements, typically >18 h per day, as well as cosmetic and quality-of-life considerations, have limited adoption in some clinical settings (28).

LITT is a minimally invasive neurosurgical technique that employs stereotactically placed laser fibers to deliver focal thermal energy, resulting in the coagulative necrosis of tumor tissue under real-time magnetic resonance thermometry guidance (11). This approach is particularly attractive for recurrent glioblastoma located in deep or eloquent brain regions where open surgical resection carries substantial morbidity (11).

Clinical reports have suggested that LITT performed at the first recurrence is associated with a median overall survival of ~11-12 months from the time of treatment, compared with roughly 9-10 months in patients managed with biopsy and systemic therapy alone. The procedure is typically associated with relatively short hospital stays and acceptable complication rates (11). Although the available evidence is largely derived from retrospective analyses and single-arm series, LITT is increasingly incorporated into individualized salvage treatment strategies for carefully selected patients (11).

Dendritic cell vaccines are designed to enhance antitumor immunity by priming or amplifying tumor-specific T-cell responses through antigen presentation in an immunostimulatory context (12,21). Autologous dendritic cells are generated ex vivo from peripheral blood monocytes, loaded with tumor lysate or defined antigens, matured, and subsequently administered via intradermal, subcutaneous, or intranodal routes (12,21). Among these platforms, DCVax-L, an autologous tumor lysate-loaded dendritic cell vaccine, has been the most extensively studied in glioblastoma (12,29).

In a large phase III trial and subsequent pooled analyses, the addition of DCVax-L to standard therapy was associated with encouraging survival outcomes in patients both newly diagnosed with and recurrent glioblastoma when compared with external control cohorts, including the presence of a subset of long-term survivors (12,29). A recent meta-analysis of dendritic cell vaccine studies reported a significant survival benefit in patients with glioblastoma treated with dendritic cell vaccines in addition to standard therapy. The pooled analysis demonstrated an overall survival hazard ratio of approximately 0.71, indicating a meaningful reduction in mortality compared with standard therapy alone (30). Treatment was generally well tolerated, with predominantly low-grade, transient flu-like symptoms. Nevertheless, variability in trial design, logistical complexity of vaccine production, prolonged manufacturing timelines, and cost have limited widespread implementation outside specialized centers (12,21,29).

CAR T-cell therapy involves the genetic modification of autologous T-cells to express synthetic receptors that recognize tumor-associated antigens and trigger T-cell activation upon antigen binding (13). In glioblastoma, targets under active investigation include interleukin-13 receptor α2, EGFR variant III, human epidermal growth factor receptor 2 and GD2(13). Early-phase clinical trials of interleukin-13 receptor α2-directed CAR T-cells delivered intratumorally or intraventricularly have demonstrated objective radiographic responses, including partial and complete responses, in heavily pretreated patients with recurrent glioblastoma (13). However, these responses have generally been transient, with disease progression often occurring within months (13).

Barriers to durable efficacy include antigen heterogeneity and antigen loss, limited persistence of CAR T-cells within the CNS, profound local immunosuppression and logistical challenges related to manufacturing timelines in the setting of rapidly progressive disease (13,25,31). Current research efforts are focused on the development of multi-antigen CAR constructs, armored CAR T-cells with enhanced resistance to exhaustion, combination strategies incorporating checkpoint inhibition or myeloid-targeted therapies, and optimization of delivery routes and dosing schedules (13,21,31).

Immune checkpoint inhibitors targeting the programmed death 1 and programmed death ligand 1 pathway have revolutionized the treatment of several systemic malignancies, but clinical outcomes in glioblastoma have been disappointing (14,26,31). In the randomized CheckMate 143 trial, nivolumab failed to improve overall survival compared with bevacizumab in patients with recurrent glioblastoma, despite an acceptable safety profile (14,26). Similarly, phase II studies evaluating pembrolizumab or nivolumab, either alone or in combination with bevacizumab, have demonstrated limited efficacy in unselected patient populations (26,31).

Several factors likely contribute to these results, including low tumor mutational burden, the limited infiltration of effector T-cells, dominant myeloid-mediated immunosuppression and the regulatory effects of the blood-brain barrier (25,31). Consequently, current strategies emphasize rational combination approaches that integrate checkpoint blockade with vaccines, cellular therapies, oncolytic viruses, or agents targeting regulatory T-cells and tumor-associated macrophages, as well as biomarker-driven patient selection (21,25,31).

Bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor A, received accelerated approval for recurrent glioblastoma based on high radiographic response rates and improvements in progression-free survival (32). However, subsequent randomized research has consistently failed to demonstrate a corresponding improvement in overall survival in either recurrent or newly diagnosed disease settings (32,33). As a result, bevacizumab is best regarded as a palliative therapy that can reduce peritumoral edema, improve or stabilize neurological symptoms and facilitate corticosteroid tapering, without fundamentally altering the natural history of the disease (32,33).

Reirradiation, typically delivered using hypofractionated schedules and occasionally combined with bevacizumab, represents a treatment option for selected patients with localized recurrence and preserved performance status (34). The NRG/RTOG 1205 trial demonstrated an improved progression-free survival with the combination of reirradiation and bevacizumab compared with bevacizumab alone, although no significant overall survival benefit was observed (34). These findings reinforce the role of anti-angiogenic therapy and reirradiation as disease-modifying rather than curative interventions in the recurrent setting.