Implementation of novel approaches for management and treatment of brain tumours

Primary brain tumours remain among the most difficult cancers to treat because of their molecular heterogeneity, infiltrative growth and the blood–brain barrier (BBB), which restricts drug delivery. For almost two decades, management of glioblastoma, the most common malignant primary brain tumour in adults, has rested on maximal safe resection followed by radiotherapy with temozolomide. Recent years have brought a more differentiated picture. Integrated molecular classification (WHO CNS5) and DNA methylation profiling have sharpened diagnosis; fluorescence-guided surgery and tumour treating fields have extended the established armamentarium; the IDH inhibitor vorasidenib (hazard ratio for progression or death 0.39; 95% CI 0.27–0.56 versus placebo) and BRAF/MEK inhibition with dabrafenib plus trametinib have delivered the first clearly biomarker-directed successes in adult and paediatric glioma respectively; and immunotherapy, including checkpoint blockade, dendritic cell vaccination and intraventricular CAR T cells, is under active investigation with mixed results. Regulatory approvals of vorasidenib (2024) and dordaviprone (2025) have since extended the biomarker-directed options, and ultrasound-mediated BBB opening illustrates a delivery-focused strategy. In parallel, machine-learning tools now support tumour classification, segmentation and response assessment, and are well placed to inform patient selection, end-point definition and trial design. This review summarises the clinical evidence for each approach, highlights pharmacological issues (CNS penetration, response end-points, safety and exposure–response relationships) and discusses barriers to implementation, including trial design, external validation of AI tools and equitable access.