Comprehensive landscape of tumor angiogenesis via integrating RNA sequencing and three-dimensional microphysiological system.

e16058 Background: The prognosis of gastric cancer (GC) patients with peritoneal metastasis with malignant ascites is very poor. We developed an in vitro tumor angiogenesis platform “TumorAngioChip” to reproduce tumor angiogenesis in 45 gastric cancer patients and mapping their angiogenesis characteristics. This platform, which can predict the features of tumor angiogenesis and drug responsiveness in patient’s samples, opens up potential for a variety of combination therapies, including ramucirumab. Methods: We built a PDC library for the study based on samples of 45 GC patients receiving treatment at Samsung Medical Center. For all patient samples, whole exome and transcriptome sequencing was performed for subtype clustering. TumorAngioChip was fabricated by design prototyping through 3D printing and mass production through injection molding. Patient-derived tumor angiogenesis reconstructed on TumorAngioChip was analyzed using an algorithm for morphological image processing of 3D images obtained through confocal microscopy. Results: Of 45 GC patients, we investigated TumorAngioChip analysis and transcriptome sequencing to classify them as KDR-High (N = 24) and KDR-Low (N = 21) ( P = 4.481e-08, Wilcoxon signed-rank test). In the TumorAngioChip–patient mapping analysis, we clearly demonstrated that GC patients with high KDR level regardless to their GC TCGA subtype had high index of tumor-induced angiogenesis and cancer invasiveness measured by TumorAngioChip ( KDR-High vs. KDR-low group; mean sprouting length: 1.455 × 103 vs. 0.946 × 103, respectively, P < 0.0001; vessel density: 3.470 vs. 2.637, respectively, P < 0.0001; sprouts number: 219 vs. 111, respectively, P < 0.0001; vessel total length: 4.931 × 104 vs. 3.919 × 104, respectively, P = 0.0011). Conclusions: This study is the first attempt to elucidate cancer angiogenesis in large-scale GC patients via a microfluidic-based in vitro system. By integrating transcriptome sequencing and TumorAngioChip-based 3D morphological image processing, we demonstrated tumor angiogenesis mapping for GC patients and further revealed the benefits of ramucirumab in individual patients. The application of TumorAngioChip shows its potential as a novel preclinical drug screening platform to identify effective angiogenesis inhibitors reflecting TME for the first time.