Modeling Glioblastoma for Translation: Strengths and Pitfalls of Preclinical Studies.

Glioblastoma (GB) is an extremely aggressive tumor for which effective therapy is still in its infancy. Although several candidate therapeutics have been identified in functional preclinical assays, clinical trials have not supported their effectiveness in GB patients. The poor clinical efficacy of the treatments can be attributed to the insufficient mimicry of GB in patients by the preclinical models used. In this review article, we provide a comprehensive overview of the available GB preclinical models, which are classified according to their origin (animal or human), species, type and modeling strategy (two- or three-dimensional cell culture, in vivo grafting or in silico modeling). Moreover, the article compares developing cutting-edge technologies, including GB-derived organoids, bioprinting, microfluidic devices, and their multimodal integration in GB-on-chip systems, which aim to replicate the GB microenvironment with high precision. In silico and in vivo approaches are also reviewed, including zebrafish transplantation models. The costs, benefits, applications and clinical relevance of each model system and/or modeling strategy are discussed in detail and compared. We highlight that the most appropriate, or combination of, GB preclinical models must be selected (or even customized) based on the specific aims and constraints of each study. Finally, to improve the reliability and translational relevance of GB research, we propose a practical roadmap that addresses critical challenges in preclinical assay development, ranging from short-term adjustments to long-term strategic planning.