Back

A Single-Cell Atlas of Uterine Carcinosarcoma from Diverse Ancestries

Subhash, S.; Bammert, M.-T.; Yueh, B.; Chu, T.; Barbi, M.; Alici, A.; Eskiocak, O.; Ozler, K. A.; Nizam, A.; Katcher, A.; Chung, C.; Shah, V.; Ozcelik, E.; Robine, N.; Frimer, M.; Goldberg, G. L.; Beyaz, S.

2026-04-24 cancer biology
10.64898/2026.04.21.720013 bioRxiv
Show abstract

Uterine carcinosarcoma (UCS) is an aggressive endometrial malignancy characterized by epithelial-mesenchymal plasticity, early metastasis, and poor therapeutic response. However, its single-cell organization and microenvironmental interactions remain incompletely defined, particularly in patients of African ancestry who are underrepresented in genomic datasets despite a disproportionate disease burden. Here, we generate a single-cell transcriptomic atlas of UCS from a diverse cohort enriched for patients of African ancestry, integrated with whole-genome sequencing. We identify pronounced inter-patient heterogeneity and resolve epithelial-like, mesenchymal-like, transitional, and stem-like malignant states. Copy-number and trajectory analyses reveal multiple subclones linked along epithelial, progenitor-like, and mesenchymal programs, consistent with metaplastic transitions. Compared to normal endometrium, primary tumors exhibit epithelial-mesenchymal-transition (EMT), mTORC1, and glycolytic programs, whereas metastases show enhanced TNF-NF{kappa}B signaling linked to invasion. The tumor microenvironment comprises diverse immunosuppressive myeloid states and heterogeneous cancer-associated fibroblast populations, including pericyte-like and matrix-remodeling subsets that act as communication hubs via chemokine and immune-checkpoint signaling. These data define the UCS cellular ecosystem in which malignant plasticity is coupled to stromal-immune cell remodeling in a patient cohort of diverse ancestries.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.