MRI-based Habitat Radiomics for Evaluating Lymph Node Metastasis in Renal Cell Carcinoma

Motivation: The prognostic property of regional lymph node metastasis (RLNM) has been widely recognized, but the diagnostic workup has stagnated for renal cell carcinoma (RCC). Goal(s): This study aimed to develop a machine learning model using MRI-based habitat radiomics to enhance the preoperative assessment of RLNM in RCC. Approach: A multi-center retrospective study. Results: Using 25 optimal habitat radiomics features combined with the diameter of lymph node, the support vector machine model achieved areas under the receiver operating characteristic curves of 0.87 and 0.89 in the internal and external test cohorts, respectively, both exceeding those of the node diameter alone. Impact: The MRI-based habitat radiomics combined model demonstrates a robust non-invasive capability for assessing regional lymph node metastasis in renal cell carcinoma, providing significant insights for clinical staging, surgical decision-making, and prognostic evaluation.