Clinical validation of Node-RADS for nodal staging in colon cancer: a retrospective study of 360 patients

Objective This study aimed to assess the effectiveness of the Node Reporting and Data System version 1.0 (Node-RADS 1.0) for preoperative evaluation of lymph node metastasis (LNM) in colon cancer and to benchmark its diagnostic performance against conventional CT-based size criteria. Materials and methods This retrospective study enrolled patients who underwent radical resection for colon cancer at Shantou Central Hospital between January 2015 and January 2019. All patients received radical resection followed by histopathological evaluation of regional lymph nodes. Preoperative computed tomography (CT) images were independently reviewed by two experienced radiologists using the Node-RADS scoring system. Inter-observer agreement was assessed, and diagnostic performance was evaluated using receiver operating characteristic (ROC) analysis. Results A total of 360 patients were included. Lymph node positivity was associated with several CT features, including nodal size, internal heterogeneity, necrosis, margin characteristics, fatty hilum status, and the Node-RADS score (all p < 0.05). The Node-RADS score demonstrated the highest diagnostic performance (area under the curve (AUC) = 0.835, 95% CI: 0.792–0.871) with an optimal cutoff of ≥3, good inter-observer reliability (intraclass correlation coefficient (ICC) = 0.962), and higher discriminative ability than the conventional 10-mm size criterion (AUC 0.546; p < 0.05). Conclusion In this large retrospective cohort, contrast-enhanced CT–based Node-RADS demonstrated reliable diagnostic performance for preoperative assessment of LNM in colon cancer and outperformed the conventional size-based criterion. The good inter-observer agreement further supports the reproducibility and clinical feasibility of Node-RADS in routine clinical practice. Key Points Question Can CT-based Node-RADS overcome the low sensitivity and underestimation of nodal metastasis associated with the conventional 10-mm short-axis criterion in preoperative colon cancer staging? Findings In 360 patients, Node-RADS at ≥3 achieved an AUC of 0.835, 96.6% sensitivity, and good interobserver agreement, outperforming the 10-mm criterion with fewer missed metastases. Critical relevance statement In 360 patients, Node-RADS at ≥3 achieved an AUC of 0.835, 96.6% sensitivity, and good interobserver agreement, outperforming the 10-mm criterion with fewer missed metastases