| Objective:To analyze the factors affecting the prognosis of Ewing sarcoma and construct a survival prediction model to provide a clinical reference.Methods:(1)Select the clinical case data of Ewing sarcoma between 2010 and 2015 from the US SEER database;analyze the relevant literature data and select the characteristics of the case data to establish relevant prognostic factors;clean the data according to the inclusion and exclusion criteria;continuous variables by using the X-tile program to get the best cut-off points and converts it into a categorical variable;uses the Kaplan-Meier method to analyze the impact of prognostic factors on the survival rate of Ewing sarcoma patients;uses the computer to use random sampling to divide into the training set and validation set by the ratio of 80% and 20%.(2)Univariate Cox analysis and multivariate Cox analysis were used for the training set to screen the prognostic factors;the independent prognostic factors obtained were incorporated into the Cox proportional hazard model,and a nomogram was constructed to visualize the model.(3)The area under the ROC curve,the C-index and the calibration curve are used to evaluate the reliability and accuracy of the prediction model,the validation set is included in the prediction model for internal verification,and the prediction performance of the model is further evaluated;the decision curve analysis(DCA)method is used to evaluate the prediction model’s performance Clinical applicability,comparing the net benefit rate between the predictive model and the AJCC TNM staging.(4)Develop the dynamic nomogram on the web version to realize the use of the prediction model by mobile phones and computers.Results:(1)This project included 480 patients with Ewing sarcoma,11 prognostic factors(age,sex,race,primary tumor site,tumor size,TNM staging,surgery,radiotherapy and chemotherapy);the best cut-off points of age for continuous variables were 18 years old and 27 years old,tumor size was 58 mm and 115mm;Kaplan-Meier survival analysis results: age,race,tumor primary site,tumor size,T stage,N stage,M stage,surgery,radiotherapy and chemotherapy ten prognostic factors were meaningful(P<0.05);random sampling method was used to group into a training set(n=387)and validation set(n=93).(2)Univariate Cox analysis and multivariate Cox analysis results: age,tumor size,M staging and chemotherapy are statistically significant(P<0.05),and can be included in the survival prediction model.(3)The ROC curve shows that the 1-,3-,and 5-year AUC of the training set are 0.84,0.76,and 0.73,and the 1-,3-,and 5-year AUC of the validation set are 0.85,0.83,and 0.71.The C-index of the training set was 0.77,and the C-index of the validation set was 0.79.The calibration curve shows that the fitted prediction curve in the training set and the validation set is close to the standard curve,that is,the constructed survival prediction model has a high prediction fit;DCA shows that the net benefit rate of the survival prediction model is higher than that of the traditional AJCC TNM staging.(4)The developed dynamic nomogram can normally run on the website.Conclusions:(1)Age,tumor size,M stage,and chemotherapy are independent prognostic factors for Ewing sarcoma.(2)The survival prediction model has good predictive power and clinical practicability and can be used to evaluate the prognostic survival of patients with Ewing sarcoma. |
PDF Full Text Request