Establishment And Verification Of Nomogram Model For Predicting The Risk Of Malignant Transformation Of Hydatidiform Mole

Objective:To analysis the risk factors of malignant development of gestational trophoblastic tumor after hydatidiform mole pregnancy.An individualized line chart model for predicting malignant development of hydatidiform mole was established.In order to guide the clinical screening of high-risk groups of hydatidiform mole malignant transformation.In order to carry out intervention measures as soon as possible to reduce the occurrence and development of gestational trophoblastic tumor after hydatidiform mole.Methods:The patients with hydatidiform mole who underwent standardized diagnosis and treatment in the Department of Obstetrics and Gynecology of the Affiliated Hospital of Yan’an University from March 2011 to March 2021 were retrospectively analyzed.At first,276 patients with hydatidiform mole were included,of which 66 were excluded according to the exclusion criteria,and finally 210 patients with hydatidiform mole were included.Follow up patients who meet the requirements in accordance with the follow-up requirements of the 2018 FIGO guidelines.Based on the results,the patients were divided into two groups:non-malignant group(n=175)and malignant group(n=35).The general data(include age,reproductive history,gestational weeks,diameter of uterine mass,etc.)and blood routine results within 5 days were collected in all cases before the first uterine clearance.The Blood routine results include white blood cell count(WBC),neutrophil absolute value(N),lymphocyte absolute value(L),monocyte absolute value(M),hemoglobin(Hb),mean corpuscular volume(MCV),red blood cell distribution width(RDW),blood platelet count(PLT)and transformation parameters such as neutrophil to lymphocyte ratio(NLR),neutrophil to monocyte ratio(NMR),Lymphocyte to monocyte ratio(LMR),platelet to lymphocyte ratio(PLR)and systemic immune inflammatory index(SII),etc.And the results of bloodβ-h CG before curettage.The above-mentioned indexes were analyzed by single factor and multi-factor analysis.The risk factors of malignant transformation of hydatidiform mole were screened by logistic regression model.According to the results of logistic regression,the corresponding nomogram model is established in R software.The predictive value of ROC curve analysis model for malignant transformation of hydatidiform mole was drawn.The Bootstrap method is used to verify the corresponding nomogram model internally,and then the corresponding distinguishing ability and calibration of the model are further evaluated.Result:1.Among the 210 patients with hydatidiform mole,35 patients turned malignant into gestational trophoblastic tumor,with a malignant transformation rate of 16.67%.2.There was no significant difference in pregnancy,delivery,abortion and cesarean section history between malignant transformation group and the non-malignant transformation group(P>0.05).3.The age before the first uterine clearance,the gestational age and the diameter of uterine mass in the malignant transformation group were significantly higher than those in the non-malignant transformation group(P<0.05).In the malignant transformation group,bloodβ-h CG>100000 was found in 25 cases,accounting for 71.43%,while in the non-malignant transformation group,bloodβ-h CG>100000 was found in 76 cases,accounting for 43.43%.There was statistical significance between them(P<0.05).4.There was no significant difference in WBC,N,M,MCV,PLR,NMR and SII between malignant transformation group and non-malignant transformation group(P>0.05).5.The RDW and NLR in malignant transformation group were higher than those in the non-malignant transformation group.L,Hb,PLT and LMR in malignant transformation group were lower than those in the non-malignant transformation group.All of them were statistically significant(P<0.05).6.The AUC of age,gestational week,diameter of uterine mass,area under the curve of bloodβ-h CG,L,HB,RDW,PLT,NLR and LMR were 0.638,0.615,0.637,0.640,0.703,0.608,0.664,0.642,0.699 and 0.723,respectively,the differences were statistically significant(P<0.05).7.Multivariate logistic regression analysis showed that age(OR=1.057,95%CI:1.004～1.113)and bloodβ-HCG(OR=3.276,95%CI:1.304～8.231),RDW(OR=1.944,95%CI:1.249～3.025),NLR(OR=1.419,95%CI:1.944～2.133)and LMR(OR=0.511,95%CI:0.318～0.823),both were independent risk factors for malignant transformation of hydatidiform mole(P<0.05).8.The sensitivity of the established nomogram model is 80.0%,and the specificity is 78.9%.The AUC was 0.845(95%CI:0.783～0.906),and the C-index of validation set model was 0.844(95%CI:0.783～0.905),the difference was statistically significant.The Hosmer-Lemeshow test was carried out and the results were compared(χ~2=5.116,P=0.745).According to the corresponding calibration plot,it can be seen that the calibration curve drawn in the verification set fits well with the reference line,indicating that the nomogram prediction model has good discrimination and consistency.Conclusion:The nomogram model based on age,pre-clearance bloodβ-h CG,RDW,NLR and LMR to predict the risk of malignant transformation of hydatidiform mole has good differentiation and calibration,high classification performance and certain practical value.It is helpful to improve the screening and early identification of high-risk patients with malignant transformation,and has guiding significance for the formulation of intervention measures.