A XGBoost predictive model of reproductive outcomes in patients following hysteroscopic adhesiolysis

Abstract

Abstract Background Infertility is the primary clinical symptom and reason for visiting patients with intrauterine adhesions (IUAs). Exploring the factors influencing the prognosis of IUAs and establishing a predictive model for reproductive outcomes after hysteroscopic adhesiolysis (HA) are extremely important for the selection of clinical treatment and prognostic assessment. Methods The clinical informations of 369 individuals diagnosed with and treated for IUAs were obtained from the Intrauterine Adhesion Multicenter Prospective Clinical Database (IUADB, NCT05381376) and randomly divided into the training and validation cohorts. A univariate analysis was performed to identify relevant clinical indicators, followed by a least absolute shrinkage and selection operator (LASSO) regression for regularization and SHapley Additive exPlanation (SHAP) for extreme gradient boosting (XGBoost) predictive model visualization. Finally, receiver operating characteristic (ROC) curves were constructed to assess the model’s efficiency. Results Univariate analysis and LASSO regression demonstrated that 12 clinical indicators were significantly associated with postoperative reproductive outcomes in IUAs patients. SHAP visualization indicated that postoperative fallopian tube ostia, blood supply, uterine cavity shape and age had the highest significance. The area under the ROC curve (AUC) of the XGBoost model in the training and validation cohorts was 0.987 (95% CI 0.9787–0.996) and 0.9851 (95% CI 0.9668-1), respectively. These values were significantly higher than those of the American Fertility Society (AFS) classification, the Chinese Society for Gynaecological Endoscopy (CSGE) classification and endometrial thickness (all P < 0.01). Conclusions The XGBoost model had higher accuracy in predicting postoperative reproductive outcomes in IUAs patients. Clinically, our model may be useful for managing and categorizing IUAs and determining optimal action to aid in pregnancy. Trial registration: The study was an observational cohort study, and the data were obtained from the Chinese Uterine Adhesion Database (ClinicalTrials.gov; NCT05381376; 19/05/2022).