Nomograms for predicting overall survival and cancer-specific survival in patients with head and neck non-Hodgkin lymphoma: A population-based study

Abstract

This study aimed to develop comprehensive nomograms for patients with head and neck non-Hodgkin lymphoma (H&NNHL) to determine their overall survival (OS) and cancer-specific survival (CSS). In this study, 602 H&NNHL patients were analyzed from the Surveillance, Epidemiology, and End Results database. The R software was used to randomly divide the patients into the training cohort (n = 421) and the validation cohort (n = 181) in a 7-to-3 ratio. To develop nomograms for projecting OS and CSS, multivariable Cox regression was used to acquire independent predictive factors. We have constructed nomograms to predict the 3-, 5-, and 8-year OS and CSS probabilities of H&NNHL patients. The consistency index of the nomograms for OS (CSS) was 0.74 (0.778) and 0.734 (0.775), in the training and validation cohort respectively, and was higher than that of the Ann Arbor staging system. Calibration plotting showed that our models have good calibration ability. Moreover, assessments of the area under the time-dependent receiver operating characteristics curve, net reclassification improvement, integrated discrimination improvement and decision curve analysis demonstrated that our nomograms performed better and were more clinically useful than the Ann Arbor staging system. This is the first research to establish comprehensive nomograms for predicting OS and CSS in patients with H&NNHL at 3-, 5-, and 8-year. The validation of the models demonstrated good performance. It can provide clinicians with reference information for determining customized clinical treatment options and providing personalized prognoses. Indexes such as the concordance index, the area under the time-dependent receiver operating characteristics curve, calibration curves, the net reclassification improvement, the integrated discrimination improvement, and decision-curve analysis were used to compare new survival models to the classical Ann Arbor staging system.