Head-to-head pathologic validation of Node-RADS vs size-based criteria for cervical lymph node metastasis in nasopharyngeal carcinoma

Background

Accurate diagnosis of cervical lymph node (CLN) metastasis is essential for staging nasopharyngeal carcinoma (NPC). Conventional size-based MRI criteria exhibit high specificity but low sensitivity, leading to underdiagnosis. This study pathologically validated the Node Reporting and Data System (Node-RADS) vs size criteria for diagnosing CLN metastasis in NPC.

Materials and methods

Patients with histologically confirmed NPC who underwent ultrasound-guided fine-needle aspiration cytology (FNAC) of CLNs between March 2014 and April 2023 were included. All patients had a pretreatment head and neck MRI within 7 days before FNAC. Two blinded radiologists independently assigned Node-RADS scores (1–5) to each biopsied node according to the standardized workflow. Diagnostic performance was compared against conventional size-based criteria using receiver operating characteristic (ROC) analysis.

Results

A total of 273 patients (202 men; median age, 51.0 years [IQR: 37–65]) with 335 CLNs were included, of which 150 (44.8%) were malignant on FNAC. Using a cutoff of Node-RADS ≥ 4, sensitivity and specificity were 96.0% and 98.9%, respectively, with an AUC of 0.97 (95% CI: 0.96–0.99), significantly higher than that of size-based criteria (AUC, 0.82; 95% CI: 0.79–0.86, p < 0.05). Among 227 size-negative nodes, Node-RADS achieved a sensitivity of 87.8% and specificity of 99.4%, correctly identifying 43 of 49 metastatic nodes missed by size criteria (AUC, 0.94; 95% CI: 0.89–0.98).

Conclusion

The Node-RADS system significantly outperforms conventional size-based criteria for detecting CLN metastasis in NPC, particularly in size-negative nodes. By offering high sensitivity while maintaining excellent specificity, Node-RADS represents a robust imaging tool that could enhance nodal staging accuracy and precise treatment planning.

Key Points

Question Can a standardized MRI-based nodal scoring system improve the detection of metastatic CLNs in NPC, particularly when nodes remain below conventional size thresholds?

Findings In this pathologically validated cohort, Node-RADS outperformed conventional size criteria and correctly identified most metastatic CLNs missed by size-based assessment.

Clinical relevance The Node-RADS improves MRI-based detection of metastatic CLNs in NPC while preserving high specificity, supporting more accurate nodal staging and more precise radiotherapy target delineation.

Graphical Abstract