A serum-derived 3D tumor model platform for personalized prediction and monitoring of chemotherapeutic response in pancreatic ductal adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) remains a highly lethal cancer, largely due to late diagnosis, tumor heterogeneity, and a dense, immunosuppressive stroma that limits therapeutic efficacy. While regimens like FOLFIRINOX and gemcitabine-based therapies offer some benefit, treatment selection remains empirical, with no reliable predictive models to guide personalized decisions. We adapted our previously validated serum-derived educated spheroid technology creating 3D spheroids using PDAC patient serum. These spheroids maintained structural integrity, viability, and consistent size over eight days, avoiding overgrowth. They also exhibited extracellular matrix deposition such as type I collagen, and expressed key genes involved in drug resistance and tumor progression including COL1A1, FN1, MMP2, CXCL1, and CXCL2. Using the Target-Independent Cell Killing (TICK) strategy, we established individualized chemograms to assess true therapeutic response helping clinicians in refining the optimal treatment protocol. In a 16-case study, our model achieved high concordance with clinical responses across gemcitabine, Gem-Pac, and FOLFIRINOX treatments supporting its utility in personalized care. Finally, we demonstrated that predictive accuracy was highest when patient serum was collected within a short window prior to treatment initiation. These findings support PDAC patient serum-educated spheroids as a rapid, non-invasive, and physiologically relevant tool for guiding personalized chemotherapy and monitoring treatment response in real time.