Evolution of Preclinical Models for Glioblastoma Modelling and Drug Screening.

purpose of reviewIsocitrate dehydrogenase wild-type glioblastoma is an extremely aggressive and fatal primary brain tumour, characterised by extensive heterogeneity and diffuse infiltration of brain parenchyma. Despite multimodal treatment and diverse research efforts to develop novel therapies, there has been limited success in improving patient outcomes. Constructing physiologically relevant preclinical models is essential to optimising drug screening processes and identifying more effective treatments. RECENT

findingsTraditional in-vitro models have provided critical insights into glioblastoma pathophysiology; however, they are limited in their ability to recapitulate the complex tumour microenvironment and its interactions with surrounding cells. In-vivo models offer a more physiologically relevant context, but often do not fully represent human pathology, are expensive, and time-consuming. These limitations have contributed to the low translational success of therapies from trials to clinic. Organoid and glioblastoma-on-a-chip technology represent significant advances in glioblastoma modelling and enable the replication of key features of the human tumour microenvironment, including its structural, mechanical, and biochemical properties. Organoids provide a 3D system that captures cellular heterogeneity and tumour architecture, while microfluidic chips offer dynamic systems capable of mimicking vascularisation and nutrient exchange. Together, these technologies hold tremendous potential for high throughput drug screening and personalised, precision medicine. This review explores the evolution of preclinical models in glioblastoma modelling and drug screening, emphasising the transition from traditional systems to more advanced organoid and microfluidic platforms. Furthermore, it aims to evaluate the advantages and limitations of both traditional and next-generation models, investigating their combined potential to address current challenges by integrating complementary aspects of specific models and techniques.