Abstract 4179: Three-dimensional spatial multi-omics of gastric tumor sections reveal hidden features beyond 2D analyses

Key Points

• The goal is to explore the limitations of 2D analyses in capturing the complex structure of gastric tumors using 3D spatial multi-omics.
• Utilized the G4X platform for in-situ RNA and protein readouts
• Profiled serial FFPE sections to quantify targeted RNAs and proteins
• Registered sections into a common 3D frame to reconstruct a volumetric atlas
• Acquired label-free 3D holotomography for structural visualization
• Applied AI-based single-cell segmentation for detailed cell type mapping.
• Identified strong variability in cell-type compositions in 2D slices
• Revealed a coherent exhausted T-cell domain only in 3D analyses
• Confirmed continuity between T cells, extracellular matrix, and tumor cells using holotomography
• Highlighted the potential for clinically relevant immune niches to be missed in single-section analyses
• Demonstrated a practical workflow for volumetric tumor-immune profiling.

Abstract