Abstract
Pancreatic cancer remains one of the most lethal malignancies worldwide, with pancreatic ductal adenocarcinoma (PDAC) accounting for the majority of cases and exhibiting a persistently poor prognosis. Although chimeric antigen receptor T (CAR-T) cell therapy has achieved remarkable success in hematological malignancies, its therapeutic efficacy in pancreatic cancer remains limited. This review summarizes recent advances in CAR-T cell therapy for pancreatic cancer, with a focus on representative tumor-associated targets, including mesothelin (MSLN), claudin 18.2 (CLDN18.2), prostate stem cell antigen (PSCA), CD155, CD276, growth arrest-specific protein 6 (GAS6), and glypican-1 (GPC1), while also highlighting emerging next-generation CAR-T engineering strategies, including nanobody-based antigen recognition, cytokine-armored CAR-T cells, allogeneic CAR-NKT platforms, dual-target and logic-gated CAR systems, and innovative delivery approaches. Current preclinical and early clinical evidence suggests that several targets, particularly MSLN and CLDN18.2, show promising antitumor activity; however, durable clinical responses remain difficult to achieve. The major barriers include the dense desmoplastic stroma, highly immunosuppressive tumor microenvironment (TME), antigen heterogeneity, antigen loss, limited CAR-T persistence and expansion, T-cell exhaustion, and on-target, off-tumor toxicity. To overcome these obstacles, emerging strategies have focused on remodeling the TME, engineering armored or dual-target CAR-T cells, developing logic-gated CAR systems, improving CAR-T persistence, and optimizing delivery approaches through nanomaterials, oncolytic viruses, in situ CAR-T generation, alternative immune-cell carriers, and locoregional administration. Overall, CAR-T therapy for pancreatic cancer is progressing from preclinical exploration toward clinical translation. Future success will likely depend on rational target selection, multi-dimensional TME modulation, advanced CAR engineering, precision delivery, and biomarker-guided patient stratification.