6078 Background: Imaging extranodal extension (iENE) is a critical prognostic factor in nasopharyngeal carcinoma (NPC). While international consensus has standardized iENE grading, its prognostic value in NPC remains to be validated. Furthermore, evidence is lacking regarding the utility of iENE-based risk stratification in guiding individualized treatment, particularly for identifying candidates benefiting from adjuvant chemotherapy (AC). Methods: A total of 694 non-metastatic NPC patients with cervical lymph node metastasis who received intensity-modulated radiotherapy (IMRT) at Sun Yat-sen University Cancer Center from 2016 to 2017 were included. iENE was graded based on international group consensus by two radiologists through a consensus reading protocol: G0 (no extranodal extension), G1 (infiltration of perinodal fat), G2 (matted nodes), and G3 (invasion of adjacent structures including muscles and neurovascular bundles). Progression-free survival and overall survival were analyzed using the Kaplan–Meier method and log-rank test. Multivariable Cox proportional hazards models were applied to adjust for confounders. Time-dependent receiver operating characteristic (TD-ROC) curve analysis was used to evaluate the predictive performance of clinical variables. Stratified analysis for prognosis was performed based on treatment strategies including concurrent chemoradiotherapy(CCRT), induction chemotherapy plus CCRT(IC+CCRT), and IC+CCRT followed by adjuvant chemotherapy (IC+CCRT+AC). Results: The cohort included 257 (37.0%) G0, 117 (16.9%) G1, 216 (31.1%) G2, and 104 (15.0%) G3 patients. Survival analysis for OS, PFS, DMFS, and LRRFS revealed that G1 patients showed a comparable prognosis to G0 patients(5-year PFS: 82.1% vs. 89.1%, P > 0.05), whereas G2 and G3 patients exhibited significantly worse outcomes; therefore, G0 and G1 were combined into a low-risk subgroup. Multivariable analysis adjusting for T stage, N stage, volume of maximal lymph node, clinical stage, and treatment regimens confirmed that the refined iENE grading was an independent prognostic factor for PFS (G2 vs. G0/1: HR 1.77, 95% CI 1.25–2.51, P = 0.001; G3 vs. G0/1: HR 1.90, 95% CI 1.08–3.36, P = 0.027). TD-ROC curve demonstrated that the predictive performance of iENE for 1-, 3-, and 5-year PFS and OS was superior to that of T stage and pre-treatment EBV DNA. Most importantly, subgroup analysis indicated that the addition of AC significantly improved PFS outcomes in the G2/G3 subgroup (P = 0.006 and P = 0.009, respectively), whereas no survival benefit was observed in the low-risk G0/G1 subgroup (P = 0.133). Conclusions: The refined iENE grading system provides robust risk stratification for NPC. This system outperforms traditional biomarkers like EBV DNA, better distinguishing prognostic outcomes in patients and effectively guiding the application of intensified treatment.
Zhang et al. (Wed,) studied this question.