Novel Prediction Models for Patients With Oral Squamous Cell Carcinoma at Different Anatomical Sites

Abstract

The individualized prediction of oral cavity squamous cell carcinoma (OC-SCC) is essential and should be as comprehensive as possible. The aim of this study was to identify new risk factors and develop nomograms comparing all anatomic sites of the oral cavity. We performed a retrospective cohort study using the Surveillance, Epidemiology, and End Results (SEER) database. All patients with OC-SCC diagnosed from 2004 to 2015 were selected and divided into the training cohort and the validation cohort. Age, gender, race, marital status, primary site, tumor grade, American Joint Committee on Cancer (AJCC) stage, TNM stage, surgical treatment, radiotherapy and chemotherapy were identified as predictor variables. The overall survival (OS) and disease specific survival (DSS) were identified as outcome variables. Kaplan-Meier method with log-rank test, univariate and multivariate cox regression analysis were performed. Independent prognostic factors were used to develop 3- and 5-year nomograms. Hazard ratio (HR) and corresponding 95% confidence interval (CI) showed the influence of each factor on OS or DSS. Concordance indexes (C-indexes) and calibration curves verified the nomograms internally and externally. A total of 12,346 patients were included. Marital status and chemotherapy were independent prognostic factors (P < .05). Tumors occurring on the cheek mucosa had the highest risk in OS (HR, 2.0, 95% CI, 1.7-2.3) and DSS (HR, 4.7, 95% CI, 3.6-6.0), while tumors occurring on the lip had the lowest risk in OS (HR, 1.0) and DSS (HR,1.0). The C-indexes for OS in the training and validation sets were 0.767 and 0.770, respectively, and for DSS were 0.800 and 0.799, respectively. Marital status and chemotherapy independently affect OC-SCC patients' survival. The prognosis is least favorable for tumors occurring on the cheek mucosa and most favorable for tumors occurring on the lip.