Research
The deep learning framework Hist2Prot infers single-cell spatial proteomics from histology to predict biomarkers and clinical outcomes
Nature · By Jiansong Fan ·
The Brief
- Hist2Prot is a deep learning framework described by Fan et al. in Nature Cancer, published online 13 November 2026.
- It aims to infer single-cell spatial proteomic patterns from routine hematoxylin and eosin (H&E) histology slides.
- The model is proposed to predict molecular biomarkers, prognosis and clinical outcomes.
Why it matters
If validated, it suggests that routinely collected H&E slides may contain information about spatial protein patterns that can be used for biomarker and outcome prediction.
Who’s affected
- cancer researchers
- pathologists
- computational pathology researchers
- #spatial proteomics
- #histology
- #deep learning
- #biomarkers
- #cancer
- #pathology
Nature Cancer, Published online: 13 November 2026; doi:10.1038/s43018-026-01270-8 Fan et al. propose an artificial intelligence model to infer spatial proteomic patterns at the single-cell level from routine hematoxylin and eosin slides to predict molecular biomarkers, prognosis…