Abstract
INTRODUCTION: Traditional Cox regression may yield biased estimates when competing events are present, limiting the accuracy of prognostic analyses in head and neck adenoid cystic carcinoma (HNACC). This study applied the Fine-Gray competing risks model to identify independent prognostic factors associated with HNACC-related mortality and develop a predictive nomogram using data from the Surveillance, Epidemiology, and End Results (SEER) database. METHODS: Patients diagnosed with HNACC between 2004 and 2015 were identified from the SEER database. Univariable analyses were performed using Gray's test and the cumulative incidence function, while multivariable analyses employed Cox regression and Fine-Gray proportional subdistribution hazards models. A nomogram was developed to predict 3-, 5-, and 10-year cancer-specific survival (CSS) and validated in an independent cohort. RESULTS: A total of 2,688 eligible patients were included. During follow-up, 1,046 deaths occurred, of which 673 were attributable to HNACC. The Fine-Gray model identified age, T-stage, N-stage, POCRT status, PORT status, and perineural invasion (PNI) as independent prognostic factors for CSS. These variables were incorporated into a nomogram that demonstrated excellent discrimination, with concordance indices of 0.818, 0.806, and 0.822 for 3-, 5-, and 10-year predictions in the training cohort, and 0.909, 0.931, and 0.965, respectively, in the validation cohort. CONCLUSIONS: The competing risks model identified key prognostic factors influencing CSS in HNACC. The derived nomogram provides individualized survival estimates, offering a practical tool to support clinical decision-making.