Genomic-driven targeted therapy evaluation using patient-derived organoids and xenografts in gastric cancer

Background: Gastric cancer (GC) exhibits significant molecular heterogeneity that hampers the effectiveness of standard therapies. Reliable preclinical models that accurately reflect patient-specific biology are urgently needed to inform precision treatments. Patient-derived organoids (PDOs) and xenografts (PDXs) provide complementary platforms for assessing targeted therapies. Methods: We developed gastric cancer organoids (GCOs) representing diverse molecular subtypes, along with matched PDX and organoid-derived xenograft (ODX) models. Histology, IHC, whole-exome sequencing, and copy-number analysis were used to verify their fidelity to patient tumors. Targeted drug screening was conducted based on actionable genomic alterations (TP53 mutation, PIK3CA mutation/amplification, CDK4/6 amplification). In vivo therapeutic efficacy and resistance mechanisms were evaluated using PDX models, including a Palbociclib-resistant line (3CG-278R) generated through prolonged exposure. Results: GCOs retained the morphological, immunophenotypic, and genomic features of their parental tumors and PDX counterparts, including recurrent amplifications in PIK3CA, AKT1-3, MTOR, CCND1, CCNE1, and CDK4/6. Genomic-guided drug screening revealed distinct vulnerabilities: CDK4/6-amplified GCOs showed strong sensitivity to Palbociclib (IC₅₀ ~0.01 μM); PIK3CA-mutant GCOs exhibited potent responses to PI3K inhibitors (IC₅₀ = 0.004-0.013 μM); TP53-mutant GCOs responded robustly to MDM2 inhibitors with clear p53/p21 pathway reactivation. In vivo, Palbociclib and the AKT inhibitor Ipatasertib significantly delayed tumor progression (>40 days to endpoint) and improved survival in CDK4/6-amplified PDXs with minimal toxicity. The Palbociclib-resistant PDX model (3CG-278R) showed CDK6 overexpression and transcriptional reprogramming marked by NF-κB and AP-1 activation, mucin upregulation, and decreased FOX-family tumor suppressor activity. Combination therapy with Palbociclib plus Ipatasertib restored apoptotic signaling and effectively overcame resistance. Conclusions: This integrated PDO/PDX/ODX platform accurately models gastric cancer biology, supports genomic-guided therapeutic stratification, and uncovers actionable mechanisms behind targeted therapy resistance. These patient-derived systems provide a strong translational framework for discovering effective precision treatments and rational drug combinations in gastric cancer.