Establishment of a prognostic scoring model for regional recurrent nasopharyngeal carcinoma after neck dissection

Objective: The main aim of this study was to establish a scoring model to predict risk of progression and survival in patients with regionally recurrent nasopharyngeal carcinoma (NPC). Methods: Three hundred and forty-eight patients subjected to neck dissection from 2003 to 2017 were included for study. Clinicopathologic information for each patient was analyzed. Independent prognostic factors were selected using the Cox proportional hazards model and incorporated into the scoring model. Concordance index (C-index) and calibration curves were used to verify discrimination and calibration, respectively and the results validated using bootstrap resampling. Results: Microscopic positive lymph node > 2 [hazard ratio (HR), 2.19; 95% confidence interval (CI), 1.30-3.68; P = 0.003], extranodal extension (HR, 2.75; 95% CI, 1.69-4.47; P < 0.001), and lower neck involvement (HR, 1.78; 95% CI, 1.04-3.04; P = 0.034) were identified from multivariate analysis as independent factors for overall survival (OS). A qualitative 4-point scale was generated to stratify patients into 4 risk groups for predicting OS and progression-free survival (PFS). The novel scoring model demonstrated enhanced discrimination (C-index = 0.69; 95% CI, 0.62-0.76) relative to the original recurrent tumor-node-metastasis (rTNM) staging system (C-index = 0.56; 95% CI, 0.50-0.62), and was internally validated with a bootstrap-adjusted C-index of 0.70. The calibration curve showed good agreement between predicted probabilities and actual observations. Conclusions: The scoring system established in this study based on a large regionally recurrent NPC cohort fills a gap regarding assessment of risk and prediction of survival outcomes after neck dissection in this population and could be further applied to identify high-risk patients who may benefit from more aggressive intervention.