The prognosis of patients with advanced tongue squamous cell carcinoma (ATSCC) is poor, and their overall survival (OS) is relatively short. Currently, the TNM stage system is often used clinically to assess the prognosis of patients, but the evaluation index of the TNM stage system is relatively single and does not specifically demonstrate relevant prognostic data. Therefore, the purpose of this study was to construct a dynamic online nomogram for predicting the prognosis of patients with ATSCC and to provide some reference for personalized clinical treatment of patients.

Clinical and prognostic information on patients with pathologically confirmed ATSCC from 2000 to 2018 was extracted from the SEER database and randomly divided into a training cohort and a validation cohort in a 7:3 ratio. Multifactorial and univariate Cox regression analyses were used to identify prognostic risk factors. Dynamic online nomogram were constructed using R software. Area under the curve (AUC), C-index, calibration curve, and decision curve analysis (DCA) with time-dependent ROC curves were used to assess the clinical utility of the nomogram. Kaplan-Meier survival curves were used to compare the prognosis of different patient categories.

A total of 3828 patients with ATSCC were screened in the SEER database.Age,race, primary site, AJCC T,N and M stage, lymph nodes surgery, radiotherapy, chemotherapy and marital status were independent influences on OS(P < 0.05). In the training cohort, the C-index of the OS-related line plot was 0.733 and the AUC for predicting 3-year OS was 0.867. In the validation cohort, the C-index was 0.738 and the AUC for 3-year OS was 0.899. Calibration plots and DCA curves showed good predictive performance of the model in both the training and validation cohorts. Kaplan-Meier survival curves showed that chemotherapy, lymph nodes surgery,married,primary site(tongue base) and radiotherapy had better OS than the non-chemotherapy, non-surgery, single, primary site(tongue anterior), and non-radiotherapy groups, respectively (all P < 0.05).

The established dynamic online nomogram has good predictive performance, which helps to personalize and combine the actual clinical patients to comprehensively predict the prognosis of ATSCC patients and may have better clinical application than the TNM stage system.