Answers are grounded in this paper and cite their sources.
Quantitative cellular morphology provides a rich but underutilized window into cell identity and state. We introduce sc‑HistoMorphomics, a scalable framework for profiling nuclear morphology in whole‑slide H&E images and correlating these measurements with spatial transcriptomics data. Our pipeline combines deep‑learning based segmentation with 194 interpretable morphometric features capturing nuclear shape, eosin and hematoxylin texture, and neighbourhood architecture, enabling high‑throughput, multimodal characterization of tissue organization. Applied to endometrium, breast cancer tissue, and pancreas, sc-HistoMorphomics revealed conserved high‑dimensional feature structure across tissues and reproducible association between morphology and gene expression. Eosin‑based texture features were associated with senescence signatures in endometrium, and morphology-based clustering recapitulated broad stromal, lymphocyte, and epithelial tumor cell classes in breast cancer. In pancreas, negative correlations between CDKN1A and multiple morphological features were consistent across cell types and donors. Together, these results establish sc‑HistoMorphomics as a generalizable and interpretable tool for linking morphology with spatially resolved molecular state across diverse tissues.
No takes yet. Share an insight, caveat, or question.
Alizadeh et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: