Nomogram for predicting cancer-specific survival of patients with buccal squamous cell carcinoma: a population-based study

Abstract

Background The use of the traditional American Joint Committee on Cancer (AJCC) staging system alone has limitations in predicting the survival of buccal squamous cell carcinoma (BSCC) patients. We aimed to establish a comprehensive prognostic nomogram whose prognostic value compared with AJCC system. Methods Patients were obtained from the SEER (Surveillance, Epidemiology, and End Result) database. The Cox regression model was used to select variables using a backward stepwise selection method. Based on the predictive model for determining prognostic factors, a nomogram was developed to predict the cancer-specific survival rates of BSCC patients at 3, 5 and 8 years. We used several standard model validation methods to evaluate the performance of the survival model: concordance index (C-index), area under the time-dependent receiver operating characteristic curve (AUC), net reclassification improvement (NRI), calibration plots, decision curve analyses (DCAs), and integrated discrimination improvement (IDI). Results Multivariate analysis showed that age, AJCC stage, N stage, M stage, pathological grade, and surgical status were risk factors for survival. The c-index chart, the AUC chart, and the calibration chart show that the performance of the nomogram is better, and the NRI and IDI values show that the performance of the nomogram is better than that of the AJCC staging system. The 3-year, 5-year and 8-year DCA curves of nomogram show that it has more net benefit than traditional AJCC staging system, both in terms of training and validation queues. Conclusion We developed and validated the first BSCC prognosis nomogram, which has a better prognostic value than the separate AJCC staging system. Our BSCC prognosis nomogram is a valuable tool for explaining the risks of the next 3, 5 and 8 years in clinical practice.