Spatial omics at the forefront: emerging technologies, analytical innovations, and clinical applications
Yunhe Liu, Yibo Dai, Linghua Wang · 2025 · 122 citationsRead the paper
Spatial omics transforms our understanding of cancer by revealing how tumor cells and the microenvironment are organized, interact, and evolve within tissues. Here, we synthesize advances in spatial technologies that map tumor ecosystems with unprecedented fidelity. We highlighted analytical breakthroughs-including multimodal integration and emerging spatial foundation models-that resolve functional niches and spatial communities, converting spatial patterns into mechanistic insights. We summarize how spatially organized features, from immune hubs to microbiota and neural interfaces, shape tumor evolution and clinical outcomes. We then outline how spatial approaches illuminate precancer biology, metastatic adaptation, and therapy response. Bridging discovery and translation, we provide a practical roadmap for incorporating spatial readouts into clinically oriented study design. We conclude by discussing persistent challenges in standardization and scalability and how high-plex spatial discoveries may be distilled into scalable, AI-enabled, clinically deployable assays, positioning spatial omics as a cornerstone of next-generation predictive and precision oncology.
What the full text shows
Working prototype- Commercial imaging-based spatial transcriptomics platforms can detect up to 18,000 genes. — 18,000 genes (Using CosMx whole transcriptome (WTX) panels)
- Commercial spatial proteomics platforms typically detect a limited number of proteins, with most in a specific range. — 20–60 proteins (Most commercialized platforms)
- High-resolution spatial genomics can map a vast number of genomic loci and transcripts. — >100,000 genomic loci and ~18,000 genes (Using two-layer seqFISH+)
- Spatial metabolomics via MALDI-IMS achieves specific spatial resolution for small molecule mapping. — 10–50 μm (MALDI-IMS on intact tissue sections)
- Same-section multi-omics can simultaneously detect whole transcriptomes and a high number of proteins. — >100 proteins (Using Stereo-CITE Proteo-Transcriptomics Set (STOmics))
- Setup
- Review of commercial platforms (e.g., Xenium, CosMx, PhenoCycler, G4X) and research-grade methods (e.g., seqFISH+, DBiT-seq) applied to human cancer tissues and FFPE samples.
- How
- The paper reviews a variety of spatial omics technologies including imaging-based and sequencing-based transcriptomics, mass spectrometry-based metabolomics, and multi-modal integration. It also evaluates computational frameworks for cell segmentation, noise correction, and 3D tissue reconstruction.
- Limitations
- High-plex workflows face hurdles in cost and throughput · Short-read constraints (~100 bases) in some in situ sequencing · Diffraction-limited resolution of microscopes hampers densely packed transcript resolution · Challenges in model interpretability and reproducibility for AI frameworks
Summarized by Gemma 4 from the open-access full text · 11 Oct 2026. Read it