Development and validation of nomograms to predict survival of primary cervical lymphoma: A surveillance, epidemiology, and end results (SEER) database analysis

Abstract

Due to the rarity of primary cervical lymphoma (PCL), the long-term survival of patients with cervical lymphoma and factors influencing survival are unknown. This study aimed to compare the survivals of patients with PCL and those with other cervical tumors and construct a clinical prediction model to assess the prognosis of patients with PCL. Patients with PCL from the Surveillance, Epidemiology, and End Results database were allocated randomly in a 7:3 ratio to the training and validation sets. Cox proportional hazard and Fine-Gray models were used to verify independent factors influencing overall survival (OS) and disease-specific survival (DSS), and nomograms were constructed. Receiver operating characteristic curve analysis and decision curve analysis (DCA) were used to test the performance and clinical utility of the models, respectively. We included 206 patients with PCL. The areas under the curves (AUCs) and DCA showed that all models had clinical benefits; The models constructed in this study had a predictive performance for patients with PCL. It can guide clinicians to rationalize the treatment plan for patients.