Diagnostic Performance of Node Reporting and Data System Magnetic Resonance Imaging Score in Detecting Metastatic Cervical Lymph Nodes of Nasopharyngeal Carcinoma

Xinggang Yang; Jiaqing Yang; Jia Li; Junyan Leng; Yu Qiu; Xuelei Ma

doi:10.1177/11795549241231564

Diagnostic Performance of Node Reporting and Data System Magnetic Resonance Imaging Score in Detecting Metastatic Cervical Lymph Nodes of Nasopharyngeal Carcinoma

Clin Med Insights Oncol. 2024 Apr 1:18:11795549241231564. doi: 10.1177/11795549241231564. eCollection 2024.

Authors

Xinggang Yang¹, Jiaqing Yang^{1

2}, Jia Li¹, Junyan Leng³, Yu Qiu³, Xuelei Ma¹

Affiliations

¹ Division of Biotherapy, Cancer Center, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China.
² West China School of Medicine, West China Hospital, Sichuan University, Chengdu, China.
³ West China Hospital of Stomatology, Sichuan University, Chengdu, China.

Abstract

Background: The Node Reporting and Data System (Node-RADS) is a recently proposed classification system for the categorization of lymph nodes in radiological images. This study was conducted to retrospectively evaluate the diagnostic accuracy of the Node-RADS score for metastatic cervical lymph nodes on magnetic resonance imaging (MRI) of patients with nasopharyngeal carcinoma (NPC).

Methods: We retrospectively analyzed cervical lymph nodes of NPC cases. Two radiologists independently evaluated each lymph node on the MRI scans using Node-RADS. Interobserver agreement between 2 radiologists for Node-RADS score assessment was evaluated by linear weighted kappa statistics. The correlation between metastasis and the Node-RADS score of each lymph node was analyzed using multivariate regression analysis. To investigate the diagnostic performance of the Node-RADS score, we further conducted receiver operating characteristic curve analysis. Correspondently, the sensitivity, specificity, positive predictive value, and negative predictive value of each different cutoff (>1, >2, >3, and >4) were computed.

Results: In all, 119 patients with NPC were assessed, including 203 cervical lymph nodes consisting of 140 (69%) of 203 metastatic and 63 (31%) of 203 benign. The kappa agreement between the 2 readers for the Node-RADS score was 0.863 (95% CI = 0.830-0.897, P < .001). Node-RADS score on MRI scan was shown to be an independent predictive factor of lymph node metastasis after multivariate regression analysis (odds ratio [OR] = 6.745, 95% CI = 3.964-11.474, P < .001). Node-RADS achieved an area under the curve (AUC) of 0.950 (95% CI = 0.921-0.979) in diagnosing metastatic lymph nodes. When Node-RADS >2 was identified as the best cutoff based on balanced values, the sensitivity and positive predictive value were 0.92 and 0.94, respectively.

Conclusions: Our study suggests that the Node-RADS score has high accuracy in predicting NPC cervical lymph node metastasis. Nevertheless, this conclusion requires confirmation in a larger cohort of patients with NPC.

Keywords: Node-RADS; cervical lymph node; chemoradiation; induction therapy; lymph node metastasis; nasopharyngeal carcinoma.