Background Clear cell renal cell carcinoma (ccRCC) is the most common histological subtype of kidney cancer and shows marked heterogeneity in progression, metastasis, and therapeutic response. Sphingolipid metabolism has emerged as an important regulator of tumor progression and the tumor microenvironment, but its cell‐state‐specific role in ccRCC remains unclear. Methods Public datasets from multiple cohorts were integrated, including single‐cell RNA sequencing data from GEO and bulk transcriptomic and clinical data from Xena, ArrayExpress, and ICGC. Nonnegative matrix factorization was used to resolve cellular heterogeneity and identify sphingolipid metabolism–related characteristic genes across distinct cell subsets. A prognostic model was constructed by screening 101 machine learning combinations and validated in independent cohorts. Additional analyses included immune characterization, pathway enrichment, drug sensitivity prediction, protein–protein interaction network analysis, and mutation profiling. Results Single‐cell analysis identified 23 cell clusters representing nine major cell types in ccRCC and further resolved multiple sphingolipid metabolism–related metagene‐defined subclusters within immune and stromal compartments. Based on these features, 101 machine learning combinations were evaluated, and a final 12‐gene prognostic signature was established using the Lasso+SuperPC model. The model showed stable prognostic performance across independent cohorts and outperformed conventional clinical variables. It was also associated with distinct immune features, predicted therapeutic vulnerabilities, and mutation‐related genomic characteristics. Conclusion This study provides a single‐cell‐guided framework for understanding sphingolipid metabolism–related heterogeneity in the ccRCC microenvironment and establishes a 12‐gene prognostic signature with potential value for risk stratification and therapeutic prioritization.
Huang et al. (Thu,) studied this question.