Preoperative enhanced CT lymph node parameters and postoperative pathology for predicting regional control failure after neck dissection for oral cancer

Abstract

Background Regional control failure (RCF) after neck dissection (ND) remains a major determinant of poor survival in oral squamous cell carcinoma (OSCC). The objective of this research was to devise and confirm a combined predictive model incorporating preoperative contrast-enhanced computed tomography (CT) lymph node parameters and postoperative histopathological findings for RCF in OSCC patients. Methods This single-center retrospective investigation encompassed 797 OSCC patients who had undergone ND within the timeframe between January 2015 and December 2020. Postoperative RCF was defined as pathologically confirmed cervical lymph node event occurring within a 5-year follow-up period. Patients who remained event-free for at least 5 years were classified as the success group. Patients were allocated chronologically by surgery date: those treated from January 2015 to December 2018 formed the training set (n = 523; success n = 340, failure n = 183), and those treated from January 2019 to December 2020 formed the testing set (n = 274; success n = 178, failure n = 96). Preoperative CT parameters—longest diameter (LL), shortest diameter (LS), area (LA), and relative CT value (LCT) of target cervical lymph nodes—were measured. Postoperative pathological variables included cytokeratin 13 (CK13), cytokeratin 17 (CK17), perineural invasion (PNI), lymphovascular space invasion (LVSI), surgical margin status, and tumor differentiation grade. Results Of the 797 patients, 279 experienced RCF and 518 did not. Multivariable analysis identified LL, LS, LA, CK17 positivity, PNI, LVSI, and poor differentiation as independent predictors of RCF, whereas negative surgical margin (vs. positive/close) was significantly protective (all P < 0.05). The combined nomogram integrating CT parameters and pathological features achieved an area under the curve of 0.964 in the training cohort and 0.961 in the testing cohort. Conclusion The combined model incorporating preoperative CT lymph node dimensions and postoperative pathological factors accurately predicts RCF after ND for OSCC, offering a practical tool for individualized risk stratification.