Patient‐Derived Glioblastoma Organoids for AI‐Guided Prediction of Therapeutic Response

Glioblastoma (GBM) remains one of the most treatment-refractory brain tumors, with poor survival driven by intratumoral heterogeneity, adaptive therapy resistance, and an immunosuppressive microenvironment that conventional preclinical models have failed to capture. Although these systems offer partial biological insights, none simultaneously preserve patient-specific tumor architecture, microenvironmental complexity, and scalability needed for translational research. Patient-derived GBM organoids address this gap by recapitulating 3D tumor architecture, glioma stem cell (GSC) hierarchies, extracellular matrix (ECM) interactions, and clinically relevant drug response phenotypes in a human-specific system. This review examines GBM organoid biology and engineering, from GSC isolation and ECM scaffold design to maturation, vascular integration, and immune co-culture. We discuss established platforms including NeoCOR and GLICO, and the rationale for incorporating functional vasculature, tumor-immune microenvironment components, blood-brain-barrier (BBB), and blood-cerebrospinal fluid barrier (BCSFB) interfaces. Critically, we propose that integrating artificial-intelligence (AI) and machine learning (ML) with GBM organoid platforms represents a major opportunity in the field. Convolutional neural networks, ML-based multi-omics integration, drug-genome correlation analysis, and predictive modelling of therapy resistance can transform organoid readouts into clinically actionable decisions. This AI-organoid framework offers a path toward real-time, patient-specific therapy selection, supporting precision-oncology workflows, informing clinical trial design, and reducing GBM therapy failure.