科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Cancer research and treatment2026-09-18

Integrating AI-Based Digital Pathology with Spatial Transcriptomics for Deeper Insight.

Mohammad Amin Rezania, Haechan Yi, Choi Jaeil, Dakeun Lee, Myung-Giun Noh

原始摘要(英文原文)· Original abstract
Histopathology preserves tissue architecture, whereas spatial transcriptomics measures gene expression within its native spatial context. Integrating these modalities with artificial intelligence (AI) can link morphology to spatial molecular states and scale spatial discoveries to routine pathology cohorts. Early hematoxylin and eosin (H&E)-to-spatial models demonstrated the feasibility of local gene-expression prediction. However, transcriptome-wide reconstruction remains constrained by gene-specific predictability, molecular sparsity, image quality, platform heterogeneity, and limited generalization. Xenium, CosMx, and related technologies are moving molecular supervision from multicellular spots toward cellular and subcellular units. In parallel, OmiCLIP exemplifies the shift from task-specific regression toward reusable visual-omics representations. We review these experimental and computational transitions and their applications in tumor microenvironment (TME) analysis, biomarker discovery, prognosis, and treatment-response prediction. We argue that the most credible translational goal is not to infer every transcript from H&E, but to identify biologically coherent spatial states with reproducible morphological correlates. Predictive uncertainty should be quantified, and these states should be validated across independent platforms and cohorts before being mapped across large archival collections. Future progress should extend this framework beyond static two-dimensional transcriptomic maps toward spatial multi-omics, longitudinal morphomolecular trajectory prediction, and three-dimensional tissue modeling.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Integrating AI-Based Digital Pathology with Spatial Transcriptomics for Deeper Insight. — 科研速览 Science Skim