Pancreatic ductal adenocarcinoma (PDAC) is among the most lethal malignancies, with a five-year survival rate below 10%. Its poor prognosis stems from late diagnosis, rapid metastasis, and resistance to conventional therapies. A central contributor is the tumor microenvironment (TME), characterized by dense stroma and pancreatic stellate cells (PSCs), which promote therapy resistance by remodeling the extracellular matrix, inducing hypoxia, and limiting drug penetration. These barriers highlight the need for novel therapeutic strategies. This thesis investigated the potential of cold atmospheric plasma (CAP), a non-thermal partially-ionized gas rich in reactive oxygen and nitrogen species (RONS), as a treatment modality for PDAC. Advanced 3D triple co-culture spheroid models (PDAC cells, PSCs and endothelial cells) and the in ovo chorioallantoic membrane (CAM) assay with intravascular chemotherapy injection were developed and employed to evaluate CAP’s effects on angiogenesis and in combination with standard chemotherapeutics (gemcitabine, paclitaxel). Contrary to expectations to improve drug delivery and efficacy, low-dose CAP did not induce angiogenesis. Instead, it maintained vascular quiescence, reduced VEGF-A secretion, and slightly inhibited endothelial tube formation. CAM assays confirmed unchanged vascularization, underscoring CAP’s context-dependent vascular effects. CAP monotherapy reduced cancer cell viability, with selectivity varying by model. In BxPC-3:hPSC21:HMEC-1 spheroids, CAP damaged both cancer and stellate cells, and enhanced low-dose chemotherapy efficacy, suggesting dose-sparing potential. In MiaPaCa-2:RLT-PSC:HMEC-1 spheroids, CAP preserved stellate cells and slightly enhanced low-dose chemotherapy efficacy, however, in some conditions, combined high-dose chemotherapy and CAP reduced treatment efficacy. CAM experiments confirmed chemotherapy cytotoxicity but showed no consistent benefit from CAP addition. CAP represents a potent but context-dependent therapeutic tool in PDAC. Its effects vary with dose, CAP parameters, tumor type, and TME components, highlighting the potential to reduce required chemotherapy doses at low-doses and the risk of inefficient tumor killing at higher doses. The co-culture models developed herein provide valuable platforms for testing multimodal therapies in a realistic TME context. Future research should refine and standardize CAP treatment parameters and explore its integration with existing therapies, potentially enabling CAP to complement multimodal PDAC treatment.