Four Leading Spatial Transcriptomics Platforms Benchmarked Across Human Tumors
Comprehensive comparison of Stereo-seq, Visium HD, CosMx, and Xenium platforms reveals key performance differences for cancer research applications.
Summary
Researchers systematically compared four cutting-edge spatial transcriptomics platforms using identical human tumor samples from colon, liver, and ovarian cancers. The study evaluated Stereo-seq v1.3, Visium HD FFPE, CosMx 6K, and Xenium 5K across multiple performance metrics including gene detection sensitivity, spatial resolution, and cell identification accuracy. Each platform showed distinct strengths: Stereo-seq and Visium HD detected the most genes genome-wide, while CosMx and Xenium provided superior single-cell resolution and transcript localization precision. The researchers created a comprehensive dataset with over 8 million cells and developed SPATCH, a web portal for data access and visualization.
Detailed Summary
This landmark benchmarking study addresses a critical gap in spatial transcriptomics by systematically comparing four leading commercial platforms using identical human tumor samples. As spatial transcriptomics revolutionizes cancer research by revealing gene expression patterns within tissue architecture, researchers need reliable performance data to choose optimal platforms.
The team collected fresh tumor samples from colon adenocarcinoma, hepatocellular carcinoma, and ovarian cancer patients, creating serial tissue sections processed identically across all platforms. They evaluated Stereo-seq v1.3, Visium HD FFPE, CosMx 6K, and Xenium 5K using comprehensive metrics including gene detection sensitivity, spatial resolution, transcript diffusion control, and cell segmentation accuracy. Adjacent tissue sections were analyzed with CODEX protein imaging and single-cell RNA sequencing to establish ground truth comparisons.
Key findings revealed platform-specific advantages: Stereo-seq v1.3 and Visium HD FFPE detected 17,000+ genes genome-wide with excellent capture sensitivity, while CosMx 6K and Xenium 5K provided superior subcellular resolution and transcript localization precision despite smaller gene panels (6,175 and 5,001 genes respectively). CosMx showed the highest transcript detection per cell, while Xenium demonstrated superior spatial clustering performance. All platforms successfully identified major cell types, though with varying accuracy levels.
The study produced an unprecedented multi-omics dataset comprising 8.13 million cells with manual annotations and ground truth protein data. The researchers developed SPATCH, a user-friendly web portal enabling data visualization and download. This comprehensive resource will accelerate computational method development and inform platform selection for cancer research, ultimately advancing precision medicine approaches through better understanding of tumor microenvironments.
Key Findings
- Stereo-seq and Visium HD detected 17,000+ genes while CosMx and Xenium covered 6,175 and 5,001 genes respectively
- CosMx 6K showed highest transcript detection per cell with superior single-molecule sensitivity
- Xenium 5K demonstrated best spatial clustering performance for identifying tissue regions
- All platforms successfully identified major cancer cell types with varying accuracy levels
- Study created 8.13 million cell dataset with SPATCH web portal for public access
Methodology
Researchers used identical serial tissue sections from three human tumor types (colon, liver, ovarian cancers) processed uniformly across four commercial spatial transcriptomics platforms. Adjacent sections were analyzed with CODEX protein imaging and single-cell RNA sequencing to establish ground truth comparisons with manual cell type annotations.
Study Limitations
The study examined only three cancer types from single patients each, potentially limiting generalizability across tumor heterogeneity. Platform performance may vary with different tissue types, fixation methods, or sample processing conditions not tested here.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.