Understanding why glioblastoma remains almost universally fatal despite treatment requires better laboratory models — and a subtle but consequential variable may have been systematically overlooked. The physical size of tumor cell clusters used in lab experiments appears to substantially change how those tumors behave and how well standard chemotherapy works against them, raising questions about the reproducibility of pre-clinical drug testing data.
Researchers embedded glioblastoma spheroids spanning 1,000 to 10,000 cells into methacrylamide-functionalized gelatin hydrogels — a three-dimensional biomaterial designed to approximate brain tissue mechanics — and systematically tracked invasion behavior and temozolomide (TMZ) response. Both wild-type and TMZ-resistant cell lines were tested. A key finding was an inverse relationship between spheroid size and apoptotic cell count, meaning larger clusters harbored proportionally fewer dying cells. Critically, TMZ efficacy was significantly size-dependent: high single doses outperformed metronomic (low, frequent) dosing in suppressing migration from smaller spheroids, whereas both dosing strategies performed equivalently against large TMZ-resistant spheroids.
These results carry two distinct implications. First, for cancer researchers, spheroid size is not a trivial methodological detail — it is an experimental variable that shapes outcomes, yet it is inconsistently reported across the literature, potentially obscuring which findings are truly comparable. Second, for translational relevance, the heterogeneous size of residual tumor clusters at surgical margins in actual GBM patients means no single spheroid model fully captures clinical reality. This study is preclinical and in-vitro, so it cannot directly prescribe dosing strategies for patients. However, it meaningfully challenges the assumption that TMZ dosing regimens can be optimized without accounting for tumor burden geometry. The finding that metronomic dosing — favored in some recurrence protocols — may underperform against smaller residual clusters is particularly noteworthy and warrants follow-up in vivo. Overall, this is an incremental but methodologically important contribution to glioblastoma tissue engineering.