Meta Analysis Of Risk Factors Of Endometrial Carcinoma Complicated With Venous Thromboembolism And Establishment Of Predictive Model

Objective:In order to provide a scientific theoretical foundation for the clinical i dentification and prevention of postoperative venous thromboembolism in endometria l cancer,a meta-analysis method was used to identify risk factors for postoperative ve nous thromboembolism in endometrial cancer,and a risk prediction model was create d based on a high risk factor.Methods:Part Ⅰ:Search databases in both Chinese and English systematically for risk factors in patients with endometrial cancer complicated by venous thromboembol ism,extract data,estimate bias risk,analyze data using Revman 5.3 software,and poo l data using fixed or random models depending on the study’s degree of heterogeneity.Part Ⅱ:1275 endometrial cancer patients who received care at the Department of Gynaecology,Lanzhou University First Hospital,between January 2010 and October2022 were gathered using an electronic medical record system.Among them,128 indi viduals without thrombosis throughout the same period were randomly assigned to the1:2 control group,whereas 64 patients with venous thrombosis were included in the s tudy group,the thrombosis group.Gather patient medical information for statistical an alysis,develop a nomographic model with R software,and assess the model’s perform ance in foretelling postoperative venous thromboembolism.Results:1.A total of 24 papers and 90,037 patients were included in this Meta-an alysis,which showed that age>60 years(OR=2.33,95%CI:1.64～3.32,P<0.001),dia betes((OR=1.99,95%CI:1.12～3.51,P<0.05),time to surgery(MD=14.80,95%CI:3.43～26.17,P<0.05),chemotherapy(OR=2.58,95%CI:1.72～3.86,P<0.001),FIGO sta ge III-IV(OR=4.62,95%CI:3.51～6.08,P<0.001),clear cell carcinoma(OR=3.64,95%CI:2.88 4.60,P<0.001),hypofractionated(OR=3.64,95%CI:2.88～4.60,P<0.001),platelet count>300×10~9/L(OR=3.98,95%CI:2.94～5.40,P<0.001),D-dimer(MD=1.01,95%CI:0.55～1.47,P<0.001),and CA125>35 U/m L(OR=6.54,95%CI:3.49～12.26,P<0.001)were risk factors for venous thrombosis.2.Based on the results of the multifactorial logistic regression analysis,age(OR=1.295,95%CI:1.168～1.435,P<0.001),Charlson comorbidity index(OR=4.945,95%CI:1.602～15.265,P=0.005),platelet count(OR=1.022,95%CI:1.010～1.034,P<0.001),D-dimer(OR=4.508,95%CI:1.678～12.110,P=0.003),systemic immune inflam mation index(OR=1.313,95%CI:1.125～1.534,P=0.046)as predictors and the regre ssion coefficients of each variable were used as weights to create a column line graph model to predict the risk of postoperative VTE in patients with endometrial cancer.3.The model’s internal Bootstrap check(1000 self-samplings)shows that it has g reat resolution,and the model’s C-index is 0.915.The calibration curve reveals a mean absolute error of 0.026between the model’s predicted probability and actual probabili ty,and the prediction curve is near the standard curve,demonstrating the nomogram’s strong consistency.According to the results of the clinical decision curve(DCA),the t hreshold likelihood of venous thromboembolism in EC patients ranged from 1%to 98%.Conclusions:1.Meta-analysis results showed that age(>60 years),diabetes mellit us,time of surgery,chemotherapy,FIGO stage III-IV,clear cell carcinoma,hypofracti onation,platelet count(>300×10~9/L),D-dimer,CA125(>35U/m L)were risk factors f or venous thrombosis in patients with endometrial cancer.They should be paid attenti on to in clinical work,and clinicians should take effective measures to address the risk factors and guide patients to perform functional exercise early to reduce the occurren ce of VTE.2.The model for predicting the risk of complicating VTE after endometrial cance r includes age,Charlson comorbidity index,platelet count,D-dimer,and systemic im mune inflammation index.The model can achieve individualized prediction of the pro bability of VTE in patients.Internal validation has shown that the model performs wel l and can provide a scientific and theoretical basis for VTE prevention.