Prognostic exploration of all-cause death in gingival squamous cell carcinoma: A retrospective analysis of 2076 patients

Abstract

Background. We aimed to establish a prognostic model for gingival squamous cell carcinoma (GSCC) that was superior to traditional AJCC staging and to perform a comprehensive comparison of the newly established nomogram with the AJCC staging system. Methods. We extracted 2,076 patients with gingival squamous cell carcinoma who had been entered into the SEER (Surveillance, Epidemiology, and End Results) database between 2004 and 2015, and randomly divided 70% of them into the training cohort and the other 30% into the validation cohort. Cox regression analysis was performed in combination with clinical experience and age, race, sex, marital status, tumor location, histological subtype, tumor grade, AJCC stage, chemotherapy status, radiotherapy status, and surgery status as possible prognostic factors. We evaluated and compared the two cohorts using the consistency index (C-index), area under the receiver operating characteristic curves, calibration curves, discriminant improvement index, and decision-curve analysis. Results. The Cox retrospective analysis showed that age, AJCC stage, tumor grade, histological subtype, radiotherapy status, and surgery status were significant factors to include in the new model of gingival squamous cell carcinoma. The other indicators were also better for the new model than for the AJCC staging system. Conclusion. We have developed and validated a nomogram for performing reliable gingival squamous cell carcinoma prognoses. The prognostic value of the nomogram is higher than that of the AJCC staging system. We expect that the inclusion of more-comprehensive and authoritative data (i.e., not just limited to residents of the United States) would also allow the construction of reliable nomograms for other populations.