KRT15 identified by scRNA-Seq and machine learning as stemness regulator and prognostic biomarker in ESCC

Postoperative recurrence and metastasis in esophageal squamous cell carcinoma (ESCC) are closely associated with cancer stem cells (CSCs), though the heterogeneity and key molecular mechanisms underlying CSC-driven progression remain incompletely understood. In this study, we identified a malignant, stem-like subpopulation in ESCC using single-cell sequencing data and screened for subpopulation-specific markers via machine learning algorithms, identifying KRT15 as a candidate. Functional experiments in vitro and in vivo demonstrated that the overexpression of KRT15 promoted proliferation, migration, invasion, and stemness in ESCC cells, while its knockdown suppressed these phenotypes. Clinically, high KRT15 expression was significantly associated with poorer overall survival and progression-free survival and served as an independent prognostic risk factor. Collectively, our findings indicate that KRT15 acts as a functional regulator of stemness and invasiveness in ESCC, highlighting its potential as a therapeutic target and a prognostic biomarker for postoperative risk stratification.