Analysis Of Risk Factors For Recurrence Of Transurethral Resection Of Bladder Tumor And Establishment Of Predictive Model

Objective:Retrospective analysis of recurrence after transurethral resection of bladder tumor,screening for risk factors affecting postoperative recurrence of bladder cancer and establishing a predictive model of recurrence,providing an effective reference for clinicians to objectively and accurately formulate postoperative individual treatment options.Methods:A retrospective analysis was performed on 530 patients with non-muscle invasive bladder cancer at the Second Affiliated Hospital of Suzhou University from January 2009 to January 2016.Compared the relationship of clinical data(age,gender,postoperative smoking history,hypertension with diabetes),tumor characteristics(tumor size,tissue grade,tumor stage,tumor source,tumor number,tumor growth site,tissue type),and related treatment(surgical methods,submucosal perfusion drug injection,and bladder infusion)with tumor recurrence.The chi-square test was used for comparison between groups,the Kaplan-Meier method was used for univariate analysis,and the time curve recurrence-free survival was drawn.Log-Rank was used to test the related factors that affect tumor recurrence.The 530 patients were randomly divided into a modeling group(353,2/3)and a validation group(177,1/3).Univariate analysis of significant variables in the modeling group for COX regression analysis,the relevant independent risk factors,functional equations and prognostic index(PI)affecting tumor recurrence.According to the value of the PI,the low-risk group,the intermediate-risk group and the high-risk group were analyzed by Kaplan-Meier method as a covariate to prove whether there was a difference between the three.The validation group data was brought into the model for ROC analysis,and the accuracy of the model was verified.The difference was statistically significant at P<0.05.Bonferroni correction was performed for multiple comparisons of variables,and the difference of P<0.0167(0.05/3)was statistically significant.Results:Among 530 patients,there were 147 patients had recurrence during follow-up,with the recurrence rate of 27.7%and the average recurrence-free survival time of 30.98 months.The results of chi-square test and univariate analysis showed that Postoperative smoking history(no VS yes,P=0.031<0.05),tumor diameter(<3.0cm VS?3.0cm,P=0.009<0.05),tumor source(primary VS recurrence,P=0.015<0.05),tissue Type(pure transitional epithelial carcinoma vs non-pure transitional epithelial carcinoma,P=0.000<0.05),number of tumors(single VS 2-5 vs>6,P=0.009<0.017),tumor stage(Tis VS Ta VS T1,P=0.000<0.017),tissue classification(low malignant VS low-grade VS High-grade,P=0.000<0.017)and intraoperative submucosal perfusion drug injection(no VS yes,P<0.05)was associated with recurrence.The univariate analysis of meaningful variables was brought into the modeling group for COX regression for multivariate analysis.Multivariate analysis include the univariate analysis of meaningful variables was brought into the modeling group for COX regression.The results revealed that tissue grading,tumor number,tumor stage,and pathological conditions were independent risk factors,derive the COX function equation and prognostic index(PI)and establish a PI value prediction risk model.Finally,the ROC analysis of the PI value model was performed by the validation group,and the results show that the area under the ROC curve(AUC)was 0.82,HR=0.02,and the 95%CI(0.78,0.86),and the results is valuable.Conclusions:The 3-year recurrence rate of non-muscle invasive bladder cancer was 27.7%.The recurrence rate is still high.Studies show that postoperative smoking,tumor stage,histological grade,number of tumors,tumor size,tissue type,tumor source,submucosal drug intravesical injection are associated with postoperative tumor recurrence,Among those risk factors,we found that tumor stage,tissue grade,number of tumors,and tissue type are independent risk factors for postoperative bladder tumor recurrence,and submucosal perfusion drug injection is a protective factor.The PI prediction model is obtained by COX regression.After verification,the prediction is valuable,which can provide effective reference for clinicians to make objectively and accurately formulate postoperative individual treatment plans.