Construction and application value of a survival prediction model for patients with extramedullary plasmacytoma based on SEER database
Objective To construct a survival prediction model for the patients with extramedullary plasmacytoma (EMP),and to evaluate its application value.Methods The data of the patients diagnosed as EMP from 2000 to 2019 were collected from the National Cancer Institute's Surveillance,Epidemiology,and End Results (SEER) database.Random forest algorithm was used to screen variables,multivariate Cox pro-portional hazards regression model was used to determine the independent risk factors of EMP prognosis,and a nomogram model was constructed based on these factors.The total score of the patients was calculated ac-cording to the nomogram model,and the total samples were divided into the low,medium and high risk groups according to the optimal cutoff value by using X-tile software,and the Kaplan-Meier survival curves of the three groups were analyzed.The predictive performance of the model was evaluated using area under the curve (AUC) of time-dependent receiver operating characteristic (t-ROC) curve,AUC after X years N repeated K-fold cross-validation,calibration curve,and decision curve analysis (DCA).Results A total of 1458 patients with EMP were included,including 796 cases of deletions and 662 cases of death.In the deletion and deaths groups,the majority were aged 60-<75 years old (35.4％,41.2％),male (63.1％,66.3％),white race peo-ple (81.3％,80.4％),and married (68.2％,59.4％).Finally,age (45-<60 years old/60-<75 years old/≥ 75 years old),marital status (single/others),non-primary malignant tumor,and without surgery and radio-therapy were determined to be the independent risk factors for the survival of EPM patients (P<0.05).Based on the above independent risk factors,a survival prediction model was constructed,and a nomogram was drawn.According to the optimal cutoff value,the total sample was divided into the score of 0-96 group (low risk group),score of 98-135 group (medium risk group) and score of 139-191 group (high risk group).The Kaplan-Meier survival analysis revealed that there was statistically significant difference in the survival time a-mong the three groups (P<0.0001).AUC of t-ROC curves for 1,3,5 years and AUC after cross-validation all were>0.7,indicating that the model had a good degree of differentiation.The calibration curves suggested a good consistency between the prediction and practical over survival rate,and DCA indicated that the model could improve the clinical benefits.Conclusion The constructed survival prediction model for EMP patients based on the SEER database could help the clinicians to identify the prognostic risk factors and predict the o-verall survival rate of the patients.
extramedullary plasmacytomaSEER databaseprognosisnomogram predictive model
万方数据
维普
