| Background: Ovarian carcinosarcoma(OCS)is an extremely rare histologic type of ovarian cancer,which is characterized by the high degree of relapse,high rates of aggression and poor prognosis.Due to its low incidence,the reports at home and abroad are mostly small-sample retrospective studies.At present,there is no unified conclusion about the relevant factors affecting the prognosis of OCS,and lacking an effective prognostic assessment tools.Objective: The aim of this study was to explore the independent prognostic factors of patients with OCS,and to establish a nomogram survival prediction model based on these factors,in order to achieve the individualized prediction of the survival of OCS patients and provide a new reference and direction for clinical work.Methods:1.The clinical data of OCS patients were collected from Surveillance,Epidemiology and End Results database(SEER)in United States between 1988 and 2015,and were randomly divided these patients into training cohort and validation cohort in a ratio of 7:3;2.Possible influencing variables were screened by univariate regression analysis,then the variables with p < 0.10 were included in the multivariable Cox regression analysis,and identified the independent prognostic factors(p<0.05);3.R software was used to establish the nomogram survival prediction model based on the result of independent prognostic factors;4.The prediction ability of the new survival model was compared with the traditional AJCC staging system,and using the area under the time-dependent receiver operating characteristics curve(AUC),concordance index(C-index),the net reclassification improvement(NRI),the integrated discrimination improvement(IDI)to evaluate the differentiation of the new model,and also using calibration plotting and decision-curve analysis(DCA)to evaluate the calibration ability and clinical effectiveness of the model respectively.Results:1.A total of 820 OCS patients were included in this study,including 574 in the training cohort and 246 in the validation cohort.2.Univariate Cox regression analysis suggested that age,grade,size,AJCC stage,surgery and chemotherapy were possible influencing variables of OCS patients(p<0.10).3.Multivariable Cox regression analysis suggested that age,grade,size,AJCC stage,surgery and chemotherapy were the independent prognostic factors of OCS patients(p<0.05).4.Based on the above 6 independent prognostic factors,the nomogram survival prediction model was successfully established.5.The discrimination of nomogram was better than AJCC stage.In the training cohort the 1-,2-,3-and 5-year AUC of the nomogram values were 0.704,0.729,0.739 and 0.732 respectively,while the AUC of AJCC stage predicted the 1-,2-,3-and 5-year OS were 0.600,0.637,0.661,and 0.700;In the validation cohort,the 1-,2-,3-and 5-year AUC of the nomogram values were0.761,0.719,0.721 and 0.734 respectively,while the AUC of AJCC stage predicted the1-,2-,3-and 5-year OS were 0.598,0.600,0.644 and 0.679.Their C-index of the training cohort were0.680 and 0.591 respectively,and their C-index of the validation cohort were 0.703 and0.590.All of the NRI and IDI values being greater than zero for both the training and validation cohorts.6.The calibration ability of nomogram was good,and the predicted 1-,2-,3-and 5-year OS had a higher coincidence rate with the actual survival rate of OCS patients.7.The clinical validity of nomogram was better than AJCC stage,and the 1-,2-,3-and 5-year DCA curves were found to higher than AJCC stage.Conclusion:1.Age,grade,size,AJCC stage,surgery,chemotherapy were the independent prognostic factors of OCS patients.2.Compared with the traditional AJCC staging system,the nomogram we had established in this study had better discrimination,calibration ability and clinical validity,it could predict the 1-,2-,3-and 5-year overall survival probabilities of OCS patients more accurately,and could also provide a new reference and direction for clinical staff to assess the individual survival prognosis of OCS patients. |
PDF Full Text Request