Development and Validation of A Nomogram Model for Predicting Vaginal Invasion in Patients with Stage ⅠB-ⅡA Cervical Cancer

JI Hui; LIU Xiaoli; QU Jialong

doi:10.3639/j.issn.2095-3097.2026.03.004

Abstract：Objective To explore the risk factors for vaginal invasion in patients with stage ⅠB-ⅡA cervical cancer,construct a nomogram model,and validate the predictive performance of the model for vaginal invasion.Methods A total of 380 cervical cancer patients admitted to the Affiliated Zhangjiagang Hospital of Soochow University from September 2021 to September 2024 were retrospectively collected.The patients were randomly divided into a modeling cohort(n=266)and an internal validation cohort(n=114)in a 7:3 ratio.Additionally,150 cervical cancer patients admitted to Zhangjiagang Hospital of Traditional Chinese Medicine were retrospectively collected as an external validation cohort.Clinical data of the patients were collected.LASSO regression was used to screen key variables for inclusion in multivariate analysis.A nomogram model was constructed based on the independent factors identified by multivariate analysis.The predictive performance of the model was validated using receiver operating characteristic(ROC)curve and decision curve analysis(DCA).Results The incidence of vaginal invasion in patients with stage ⅠB-ⅡA cervical cancer in this study was 17.37％(66/380).LASSO regression analysis screened 7 key variables.Multivariate analysis showed that older age(OR=1.173,95％CI:1.097-1.254),positive MRI diagnosis of vaginal invasion(OR=3.004,95％CI:1.185-7.616),and a high systemic inflammatory response index(OR=1.327,95％CI:1.002-1.757)were independent risk factors for vaginal invasion in stage ⅠB-ⅡA cervical cancer patients,while a high prognostic nutritional index(OR=0.912,95％CI:0.859-0.968)was an independent protective factor(P<0.05).ROC curve analysis showed that the area under the curve was 0.866(95％CI:0.803-0.930)for the modeling cohort,0.828(95％CI:0.761-0.896)for the internal validation cohort,and 0.834(95％CI:0.774-0.897)for the external validation cohort.DCA results showed that within the high-risk threshold range of 0-0.6,intervening based on the nomogram model provided a higher standardized net benefit compared to intervening in all patients or in none.Conclusion The nomogram model constructed in this study can relatively accurately predict the risk of vaginal invasion in patients with stage ⅠB-ⅡA cervical cancer,providing a reference for tumor staging assessment and treatment planning for such patients.