A CT-based radiomics model for noninvasive prediction of MMP9 expression and prognostic evaluation in renal cell carcinoma

Renal cell carcinoma (RCC) represents one of the most common malignant tumors of the urinary system, characterized by high heterogeneity and aggressiveness. Studies indicate that matrix metalloproteinase-9 (MMP9) plays a critical role in the invasion, metastasis, and prognosis of RCC, and is recognized as a potential molecular biomarker and therapeutic target. This study aims to develop a radiomics-based prediction model for noninvasive identification of MMP9 expression levels in RCC patients and to assess its clinical relevance in patient prognosis. This retrospective study utilized enhanced computed tomography images and clinical data from The Cancer Genome Atlas database. We manually delineated volumes of interest, extracted radiomic features, and performed standardization procedures. Key features were selected using minimum redundancy maximum relevance and recursive feature elimination algorithms, followed by the construction of a radiomics model for predicting MMP9 expression through the gradient boosting machine algorithm. Survival analysis was performed using Kaplan–Meier curves and Cox proportional hazards regression models. Furthermore, immune cell infiltration analysis and enrichment analysis were employed to explore the underlying biological mechanisms associated with MMP9. MMP9 demonstrated independent prognostic value in RCC patients (hazard ratio = 1.447). The radiomics model incorporating 3 features, constructed using multiple machine learning algorithms, exhibited favorable predictive performance for MMP9 expression levels (area under the curve = 0.865). The radiomics score mapping MMP9 expression levels served as an independent risk predictor for RCC patients (hazard ratio = 2.287, 95% confidence interval = 1.025–5.103, P = .043). Furthermore, MMP9 correlates with pathways such as immune response regulation signals, carbon metabolism, and PI3K-Akt, as well as immune cell infiltration, including M2-type macrophages and regulatory T cells. MMP9 serves as an independent prognostic factor in RCC. The enhanced computed tomography radiomics model composed of 3 features can noninvasively predict MMP9 expression levels in RCC patients.