Single-cell transcriptomic insights into ccRCC: a stemness gene signature for prognosis and treatment response prediction.

Abstract

Clear cell renal cell carcinoma (ccRCC), the most prevalent renal malignancy, exhibits remarkable intratumoral heterogeneity that challenges precise prognostication and treatment stratification. Leveraging single-cell RNA sequencing (scRNA-seq) and an advanced artificial intelligence (AI)-driven analytical framework, we comprehensively dissected the cellular complexity of ccRCC and developed a robust prognostic model. Analyzing scRNA-seq data from 44 ccRCC samples, we identified a distinct proliferative epithelial cell subtype strongly correlated with adverse clinical outcomes. Through a sophisticated LASSO-XGBoost machine learning approach, we constructed a novel stemness-related gene signature (SGS) that demonstrated exceptional predictive capabilities across multiple independent cohorts. The SGS effectively stratified patients into high-risk and low-risk groups, with high-risk individuals experiencing significantly reduced overall survival. Notably, our model outperformed conventional clinicopathological parameters and existing prognostic signatures. Critically, patients with elevated SGS scores exhibited diminished responsiveness to targeted therapies and immune checkpoint inhibitors, suggesting its potential as a predictive biomarker for treatment efficacy. Our findings not only illuminate the pivotal role of proliferative epithelial cells in ccRCC progression but also underscore the transformative potential of AI-driven approaches in precision oncology. This study provides a foundational framework for enhanced patient stratification and personalized therapeutic interventions in ccRCC.