Authors
Ge Zhao, Zhijie Hu, Baicheng Yang, Wei Shi
Published in
Neoplasia (New York, N.Y.). Volume 83. Pages 101392. Nov 06, 2026. Epub Nov 06, 2026.
Abstract
Single-cell atlases increasingly organize tumors according to recurrent transcriptional states, but state-based integration can weaken or obscure the molecular evidence that distinguishes malignant from non-malignant provenance. Ewing sarcoma provides a mechanistically tractable example because FET::ETS fusions anchor malignant provenance while EWSR1::FLI1 regulatory output spans proliferative, migratory, mesenchymal and matrix-producing states, including cancer-associated fibroblast-like malignant populations. We propose malignant-to-stromal annotation leakage as a testable post-identification failure mode in which provenance-defining evidence is weakened or discarded during reference mapping, annotation or atlas reuse, allowing transcriptional similarity to become a surrogate for cellular identity. Such misassignment can distort estimates of stromal abundance, cell-cell communication, prognostic signatures and therapeutic target attribution. We therefore propose a provenance-aware architecture that keeps malignant provenance, oncogenic regulatory activity, phenotypic state and spatial or clinical context separately traceable while retaining uncertainty. We further outline an evidence hierarchy and a falsifiable benchmarking strategy based on orthogonal molecular evidence, supported where possible by spatial and pathological validation. Although Ewing sarcoma provides the principal exemplar, the framework is designed for testing in other tumors in which plastic malignant states converge transcriptionally with non-malignant compartments.
PMID:
43008689
Bibliographic data and abstract were imported from PubMed on 11 Nov 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 6
- Comments 0