Prognostic Factors Of Non-metastatic Renal Cell Carcinoma And Establishment Of Cancer Clinical Prediction Model

Objective: To explore the prognostic risk factors of non-metastatic primary renal cancer based on the Surveillance,Epidemiology,and End Results Database(SEER),and construct the nomogram survival prediction model in combination with TNM staging system and other factors,in order to provide help for clinical diagnosis and treatment.Methods: The clinical data of all patients with non-metastatic primary renal cell carcinoma from 2010 to 2015 were downloaded from SEER Database.And set exclusion criteria to determine the number of finalists.The categorical variables were processed for continuous variables,all included variables were coded,descriptive statistical analysis of patient variables,Kaplan-Merier univariate survival analysis was performed on the patient variables,and the differences between groups were tested using the Log-Rank method.The test level was set to 0.05(? = 0.05),and identify the variables that affect the prognosis of renal cancer,incorporate meaningful variables into the Cox multivariate proportional hazard regression model,and determine the independent risk factors that affect the prognosis of renal cancer;add the independent risk factors that affect the prognosis of renal cancer into the nomogram model.The nomogram model used the Bootstrap self-sampling and internal validation to calculate the C-index?AUC and plot the calibration curve.Results: Based on the SEER database,a total of 51,561 patients with non-metastatic primary renal cell cancer were finally included.32,706 patients(63.4%)were male patients,and 29,749(57.7%)were older than 60 years old.Kaplan-Merier single factor survival analysis shows gender,age,histological type,laterality,tumor size,single tumor,grade,T stage,N stage,and treatment were correlated with the prognosis of non-metastatic primary renal cancer patients(P<0.05);Cox multifactor proportional hazard regression risk model shows gender,age,histological type,tumor size,single tumor,grade,T stage,N stage,and treatment is independent risk factors of the prognosis;based on the regression model,construct a nomogram model,the internally adjusted consistency index is 0.736,the area under the ROC curve is 0.736,and the model shows good discrimination;the calibration curve fits the straight line with the slope of 1,indicating that the predicted risk of the constructed nomogram model is close to the actual risks,and the conformity of the model is good.Conclusions: The Cox multivariate proportional hazard regression model shows gender,age,histological type,tumor size,single tumor,grade,T stage,N stage,and treatment is independent risk factors of the patients' prognosis.The nomogram model which based on the TNM staging system combined with the clinical characteristics of the patient shows good stability,suggesting that the model can provide potential value for clinical diagnosis and treatment.