Competing risk models versus traditional Cox models for prognostic factors' prediction and care recommendation in patients with advanced laryngeal squamous carcinoma: a population-based study.

Abstract

To explore the prognostic factors and the optimal treatment modalities for patients with stage IVA laryngeal squamous cell carcinoma (LSCC), so as to improve the survival rate of patients. Patients with stage IVA LSCC between 2004 and 2019 were selected from the Surveillance, Epidemiology, and End Results (SEER) database. We used competing risk models to build nomograms for predicting cancer-specific survival (CSS). The effectiveness of the model was assessed using the calibration curves and the concordance index (C-index). The above results were compared with the nomogram established by Cox regression analysis. The patients were grouped into low-risk and high-risk groups by competing risk nomogram formula. And the Kaplan-Meier (K-M) method and log-rank test were used to make sure that these groups had a survival difference. Overall, 3612 patients were included. Older age, black race, a higher N stage, a higher pathological grade, and a larger tumor size were independent risk factors for CSS; married marital status, total/radical laryngectomy, and radiotherapy were protective factors. The C-index was 0.663, 0.633, and 0.628 in the train set and 0.674, 0.639, and 0.629 in the test set of the competing risk model, and 0.672, 0.640, and 0.634 in the traditional Cox nomogram for 1, 3, and 5years. In overall survival and CSS, the prognosis of the high-risk group was poorer than that of the low-risk group. For patients with stage IVA LSCC, a competing risk nomogram was created to help screen risk population and guide clinical decision-making.