Abstract
OBJECTIVE: Infertility caused by endometriosis (EM) directly affects the possibility of pregnancy in women of gestational age. This study aims to establish a prediction model to accurately predict the natural pregnancy outcome of patients with EM, providing valuable information for clinical decision-making. METHODS: We retrospectively selected a total of 496 patients who underwent their first laparoscopic surgery for infertility at the Obstetrics and Gynecology Department of Jingzhou Central Hospital from January 2016 to June 2023. An improved endometriosis fertility index (EFI) predictive model was created based on ultrasound radiomics and urinary proteomics gathered during the patient's initial admission, using two machine learning algorithms. The predictive model was evaluated for C-index, calibration, and clinical applicability through receiver working characteristic curve, decision curve analysis. RESULTS: The improved EFI prediction model nomogram, based on five ultrasound radiomics parameters and three urine proteomics, had AUC values of 0.921 (95% CI: 0.864-0.978) and 0.909 (95% CI: 0.852-0.966) in the training and validation sets, respectively, while the traditional EFI prediction model had AUC values of 0.889 (95% CI: 0.832-0.946) and 0.873 (95% CI: 0.816-0.930) in the training and validation sets, respectively. Additionally, the nomogram exhibited better discrimination ability and achieved an overall better benefit against threshold probability than the EFI model and decision tree in the decision curve analysis (DCA). CONCLUSION: The combined ultrasound radiomics-urine proteomics model was better able to predict natural pregnancy-associated patients with EM compared to the classical EFI score. This can help clinicians better predict an individual patient's risk of natural pregnancy following a first-ever laparoscopic surgery and facilitate earlier diagnosis and treatment.