3D bioprinted ascites-derived models for gastric cancer patients with peritoneal metastasis

Gastric cancer with peritoneal metastasis is associated with poor prognosis and marked heterogeneity in treatment response. Reliable patient-derived models for functional drug testing remain limited. This study aimed to develop a malignant ascites-derived three-dimensional bioprinted model and evaluate its utility for drug sensitivity assessment. Ascites-derived cells were incorporated into gelatin methacryloyl and hyaluronic acid methacryloyl hydrogels to generate bioink for three-dimensional bioprinting. Cell viability, morphology, histopathology, and genomic features were assessed. Drug sensitivity testing was performed using standard chemotherapeutic agents, and dose–response curves were analyzed to derive area under the curve values. An exploratory sensitivity threshold was determined using the Jenks natural breaks method, and in vitro drug responses were compared with clinical outcomes. Three-dimensional bioprinted models were successfully established in 13 of 14 cases (92.9%). Constructs were fabricated within 6 hours, developed organized glandular-like or solid architectures within 7–10 days, and maintained viability for at least two weeks. Immunohistochemistry showed preservation of gastric cancer-related markers, including CDX2 and CEA, and generally consistent expression patterns of HER2, PD-L1, and Claudin 18.2 compared with parental tumors. Whole-exome sequencing demonstrated preservation of key genomic features, with consistent correlations in variant allele frequencies of shared mutations between bioprinted models and parental ascites (median Pearson r = 0.817, range 0.679–0.921). Drug sensitivity testing revealed marked inter-patient heterogeneity across oxaliplatin, fluorouracil, and paclitaxel. An exploratory normalized area under the curve threshold of 0.675 was derived to classify in vitro drug responses as sensitive or resistant. The resulting classifications were generally consistent with clinical outcomes. This study demonstrates the feasibility of establishing a malignant ascites-derived three-dimensional bioprinted model. The model retains key features of the parental tumors and shows potential for individualized in vitro drug sensitivity testing. These findings support its potential as a platform for translational research in gastric cancer with peritoneal metastasis. Not applicable.