Risk Factors For Lymph Node Metastasis Of Endometrial Carcinoma And Its Predictive Value

Objective:To study the relationship between clinicopathological indexes and lymph node metastasis of endometrial carcinoma,to explore the risk factors of lymph node metastasis of endometrial carcinoma(EC),and to construct a risk prediction model of lymph node metastasis.Methods:the basic clinical information,pathological data and preoperative laboratory examination results of 247 patients with EC who underwent surgery in the Department of Obstetrics and Gynecology of our Hospital January 2018 to August 2021 were collected.Mann-WhitneyUtest was used for continuous variables and grade data of single factor analysis,and Pearson χ~2 test or continuity corrected χ~2 test was used for classified data.Receiver operating characteristic curve was used to determine the critical value of meaningful continuous variables in univariate analysis to predict lymph node metastasis.Multivariate analysis used binary logistic regression stepwise forward method to explore the risk factors of lymph node metastasis in patients with EC;On this basis,the Z-test was used to compare the predictive value between single predictive index and combined index.Finally,a logistic regression model for predicting lymph node metastasis in EC patients was constructed based on the best combination of predictive indexes.The ROC curve was used to determine the risk threshold of the model,based on which EC patients were stratified.Results:1.247 patients with EC included in this study underwent total hysterectomy+double appendectomy + pelvic lymphadenectomy ±paraaortic lymphadenectomy The postoperative pathological results showed that a total of 27 patients had lymph node metastasis,and the lymph node metastasis rate was 10.9%.2.Univariate analysis of lymph node metastasis:age of EC patients(P=0.013),menopause(P=0.046),preoperative serum carbohydrate antigen 125(CA125)(P<0.001),Ki67(P=0.011),pathological type of non endometrioid carcinoma(P<0.001),FIGO stage(P<0.001),depth of myometrial invasion≥ 1/2(P<0.001),parauterine or vaginal involvement(P<0.001),Adnexal involvement(P=0.003),lymphovascular space invasion(P<0.001),estrogen receptor negative(P<0.001),progesterone receptor negative(P<0.001),p53(P=0.006)and p16(P=0.027)were related to lymph node metastasis in patients with EC.Taking EC patients without lymph node metastasis as the control group and patients with pelvic lymph node metastasis as the comparison group,according to the meaningful continuous variable indexes in the univariate analysis of ROC curve analysis,the cut-off value of lymph node metastasis in EC patients is calculated as follows:age 61(years),CA125 118.86U/ml,Ki67 45(%).3.Results of multivariate analysis of lymph node metastasis:age,menopause,CA125,pathological type and estrogen receptor were included to construct multivariate logistic regression equation.The results showed that CA125≥118.26 U/ml(OR=7.899,95%CI 2.599-24.547,P<0.05)increased the risk of lymph node metastasis;the depth of myometrial invasion≥1/2 of myometrium(OR=3.255,95%CI 1.069-9.09 7,P<0.05)increased the risk of lymph node metastasis;positive lymphovascular space invasion(OR=4.224,95%CI 1.391-12.822,P<0.05)increased the risk of lymph node metastasis;negative estrogen receptor(OR=12.878,95%CI4.375-37.9,P<0.05)increased the risk of lymph node metastasis.The differences of the above four indexes were statistically significant.Logistic regression equation for predicting lymph node metastasis in patients with EC:In[P/(1-P)]=7.286+2.078X1+1.18 X2+1.441 X3+2.555 X4.Note:P represents the risk probability of lymph node metastasis,ranging from 0 to 1;X1 represents the binary expression defined according to the positive threshold of CA125(118.26),and the value is 0(CA125<118.26 U/ml)or 1(CA125≥118.26 U/ml);X2 indicates the infiltration of muscle layer,and the value is 0(infiltration depth<1/2 muscle layer)or 1(infiltration depth≥ 1/2 muscle layer);X3 indicates positive lymphovascular space invasion,with a value of 0(negative)or 1(positive);X4 represents the expression of estrogen,with a value of 0(estrogen positive)or 1(estrogen negative)].4.The results of testing the effectiveness of logistic regression model suggest that the model is generally meaningful(P<0.05),Hosmer lemesho test(H-Ltest)was used to qualitatively evaluate the goodness of fit effect of the model,and the prediction ability of the model was strong P>0.05).5.The area under the curve(AUC)of lymph node metastasis was 0.892(95%CI 0.846-0.927).The risk threshold of logistic regression model is 0.25 calculated by ROC curve and Jordan index.Z test showed that the predictive ability of the prediction model was significantly better than that of other single predictors CA125(P<0.001),ER(P=0.003),depth of myometrial infiltration(P<0.001),lymphovascular space invasion(P<0.001)and the combination of two high-risk factors(depth of myometrial infiltration and Lymphovascular Space Invasion)(P<0.001).Prediction model AUC prediction model sensitivity(92.3%),specificity(77.8%).Conclusion:1.This study showed that high preoperative serum CA125 level(CA125≥118.26 U/ml),infiltration depth≥1/2 muscle layer,positive lymphatic vessel space invasion and negative estrogen receptor were independent risk factors for lymph node metastasis in patients with EC.2.The critical value of CA125 for predicting lymph node metastasis is 118.26 U/ml.on this basis,a logistic regression model can be constructed by combining preoperative CA125≥118.26 U/ml,infiltration depth≥1/2 muscle layer,positive lymphovascular space invasion and negative estrogen receptor to predict lymph node metastasis of low-risk endometrial cancer.The risk stratification of EC patients using the model can provide a reference basis for doctors’ choice of surgical procedures,patients’ follow-up treatment and prognosis prediction.