THiRST: Computer Vision Empowered Three-dimensional High-Resolution Spatial Transcriptomics Analytics

THiRST reconstructs multi-resolution 3D spatial transcriptomic maps from serial kidney sections, boosting matched-unit same-cluster proportions from 63.8%/65.6% to 70.7%/70.0% at adjacent offsets.

• Why it matters: Spatial transcriptomics is often confined to single 2D slices, limiting analysis of continuous 3D tissue structure and consistency across cross-sections.
• What they did: The framework performs morphology-guided coarse-to-fine registration, spatial matching of ST units, cross-section association of segmented structures, and integrates Visium V2 with high-definition ST.
• The result: In serial kidney sections, improved registration increased expression concordance and enabled 3D glomerular/object-level analysis, with 54.1% of differential genes detected only in 3D.

AI summary of the abstract · may contain errors · not medical advice · read the paper.