Glioblastoma's near-universal lethality has long been attributed not just to its aggressiveness but to the fact that different regions of the same tumor behave like entirely different diseases. A spatially resolved drug-sensitivity framework now offers a concrete explanation for why uniform systemic treatments so consistently fail — and points toward a more architecturally aware treatment strategy.
Using MRI-guided multi-region sampling, investigators collected paired enhancing cores and non-enhancing margins from GBM patients and subjected each compartment to whole-exome sequencing, RNA-seq transcriptomics, and a 15-drug targeted panel tested in patient-derived 3D organoid models. A striking spatial asymmetry emerged: oncogene mutations were more spatially variable than tumor suppressor alterations regardless of MRI contrast status. At the transcriptional level, mesenchymal subtypes — characterized by extracellular matrix remodeling and immune activation programs — dominated the contrast-enhancing core, while proneural subtypes with neurological process signatures were enriched at the infiltrative, non-enhancing margin. This regional identity translated directly into drug response: ERK inhibition via ulixertinib showed preferential efficacy in cores, whereas PI3K pathway blockade with paxalisib or CC-115 was more potent at margins. Across both compartments, the pan-Bcl2 inhibitor navitoclax and the epigenetic agent trotabresib emerged as the most consistently effective monotherapies.
This work fits into an accelerating movement in neuro-oncology toward treating GBM as a spatially heterogeneous ecosystem rather than a monolithic target. Prior single-biopsy studies almost certainly underestimated the tumor's molecular diversity, and the introduction of supramaximal resection as a survival-extending procedure now creates a practical window for collecting multi-region tissue. The key limitation here is the small cohort size inherent to such intensive sampling protocols, and 3D organoid drug testing, while increasingly validated, does not fully recapitulate the immunosuppressive tumor microenvironment in vivo. Still, the finding that margin-targeted PI3K inhibition and core-targeted ERK inhibition could be rationally combined represents an incremental but clinically meaningful advance toward spatially informed combination regimens.