A radiogenomics biomarker based on tumor angiogenesis for non-invasive prognosis of clear cell renal cell carcinoma

• Yuanchao Li ^[1] ; Youting Huang ^[1] ; Haiyun Xu ^[1] ; Shuohui Yang ^[1]
  1. [1] Shanghai University of Traditional Chinese Medicine

     Shanghai University of Traditional Chinese Medicine

     China

     
• Localización: Clinical & translational oncology, ISSN 1699-048X, Vol. 28, Nº. 5, 2026, págs. 1825-1838
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Purpose:

    Tumor angiogenesis drives prognostic heterogeneity in clear cell renal cell carcinoma (ccRCC), but macroscopic imaging cannot predict angiogenesis-related gene dysregulation. We aimed to develop a noninvasive radiogenomic model for assessing angiogenesis-associated gene signatures.

    Methods:

    Transcriptomic profiles from TCGA-KIRC were analyzed via “ConsensusClusterPlus” to identify angiogenesis subtypes. Univariate Cox, LASSO and Multivariate Cox regression selected prognostic angiogenesis-related genes, constructing a signature-based risk model. Prognostic nomograms integrated genomic markers with clinical variables. Radiomic features from TCIA CT images identified biomarkers stratifying angiogenesis expression, forming a radiogenomic prognostic nomogram. Performance was validated using receiver operating characteristic curves, calibration plots, and decision curve analysis.

    Results:

    The ccRCC patients were stratified into two angiogenesis-based molecular subtypes. An eight-gene angiogenesis signature predicted overall survival in TCGA, categorizing patients into low-/high-risk groups. Six radiomic features predicting signature expression were identified (Training AUC = 0.753; Testing AUC = 0.814). The combined radiogenomic-clinical nomogram achieved time-dependent survival AUCs of 0.870 (1-year), 0.811 (3-year), and 0.784 (5-year).

    Conclusion:

    The radiogenomics model correlates significantly with angiogenesis-related gene expression and enables prognostic stratification in ccRCC, supporting precision treatment selection and advancing personalized theranostics.